Protection of Materials and Structures from the Space Environment (Space Technology Proceedings)

PROTECTION OF MATERIALS AND STRUCTURES FROM THE SPACE ENVIRONMENT Space Technology Proceedings VOLUME 6 PROTECTION ...

Author: Jacob I. Kleiman

79 downloads 942 Views 10MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

PROTECTION OF MATERIALS AND STRUCTURES FROM THE SPACE ENVIRONMENT

Space Technology Proceedings VOLUME 6

PROTECTION OF MATERIALS AND STRUCTURES FROM THE SPACE ENVIRONMENT ICPMSE-7

Edited by Jacob I. Kleiman Integrity Testing Laboratory Inc. Markham, Toronto, Canada

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN-10 ISBN-13 ISBN-10 ISBN-13

1-4020-4281-7 (HB) 978-1-4020-4281-2 (HB) 1-4020-4319-8 (e-book) 978-1-4020-4319-2 (e-book)

Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. www.springer.com

Printed on acid-free paper

All Rights Reserved © 2006 Springer No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands

CONTENTS

Introduction

xi

Acknowledgements

xiii

Organization

xv

Radiation Effects of Protons and Electrons on Backfield Silicon Solar Cells Z. Hu, S. He, and D. Yang Solar Array Arcing in LEO: How Much Charge is Discharged? D. C. Ferguson, B. V. Vayner, and J. T. Galofaro

1

9

Self-Restoration as SEU Protection Mechanism for Reconfigurable On-Board Computing Platform L. Kirischian, V. Geurkov, I. Terterian, and J. Kleiman

21

Synergistic Effect of Protons and Electrons on Radiation Damage of Methyl Silicone Rubber L. Zhang, S. He, D. Yang, and Q. Wei

35

Influence of Electron Radiation on Outgassing of Spacecraft Materials R. H. Khassanchine, A. N. Timofeev, A. N. Galygin, V. I. Kostiuk, and V. M. Tsvelev Effect of Surface Charging on the Erosion Rate of Polyimide Under 5 eV Atomic Oxygen Beam Exposure M. Tagawa, S. Seikyu, K.-I. Maeda, K. Yokota, and N. Ohmae Influence of Space Environment on Spectral Optical Properties of Thermal Control Coatings V. M. Prosvirikov, A. V. Grigorevskiy, L. V. Kiseleva, A. P. Zelenkevich, and V. M. Tsvelev Mitigation of Thruster Plume-Induced Erosion of ISS Sensitive Hardware C. Pankop, J. Alred, and P. Boeder v

43

51

61

71

vi

CONTENTS

Degradation of Thermal Control Coatings Under Influence of Proton Irradiation L. S. Novikov, G. G. Solovyev, V. N. Vasil’ev, A. V. Grigorevskiy, and L. V. Kiseleva Mitigation of Damage to the International Space Station (ISS) from Water Dumps W. Schmidl, J. Visentine, and R. Mikatarian Investigation of Synergistic Effects of Proton and Electron Radiation on the Dyeing of Optical Quartz Glass H. Liu, S. He, H. Geng, D. Yang, and V. V. Abraimov The Role of “Abnormal” Electron Fluxes with Energy <1 MeV in the Surface Charging Dose of Spacecraft O. R. Grigoryan, L. S. Novikov, V. N. Shevelva, K. Kudela, V. L. Petrov, and I. V. Tchurilo Vacuum Ultraviolet Radiation Effects on DC93-500 Silicone Film J. A. Dever, B. A. Banks, and L. Yan Enhancement of Atomic Oxygen-Induced Erosion of Spacecraft Polymeric Materials by Simultaneous Ultraviolet Exposure K. Yokota, N. Ohmae, and M. Tagawa Ground Simulation of Hypervelocity Space Debris Impacts on Polymers R. Verker, E. Grossman, N. Eliaz, I. Gouzman, S. Eliezer, M. Fraenkel, and S. Maman Testing of Spacecraft Materials for Long Flights in Low Earth Orbit L. S. Novikov, V. N. Chernik, S. F. Naumov, S. P. Sokolova, T. I. Gerasimova, A. O. Kurilyonok, and T. N. Smirnova

87

93

107

115

123

141

153

167

M/OD Impacts on the Multipurpose Logistics Module: Post Flight Inspection Results J. L. Hyde, R. P. Bernhard, and E. L. Christinsen

175

Fuel Oxidizer Reaction Products (FORP) Contamination of Service Module and Release of N-Nitrosodimethylamine in a Humid Environment from Crew EVA Suits Contaminated with FORP W. Schmidl, R. Mikatarian, C.-W. Lam, B. West, V. Buchanan, L. Dee, D. Baker, and S. Koontz

193

CONTENTS

Effect of Vacuum Thermocycling on Properties of Unidirectional M40J/AG-80 Composites Y. Gao, D. Yang, S. He, and Z. Li Damage Characteristics of Zr41 Ti14 Cu12.5 Ni10 Be22.5 Bulk Metallic Glass Impacted by Hypervelocity Projectiles C. Yang, C. Z. Fan, Y. Z. Jia, X. Y. Wang, X. Y. Zhang, H. Y. Wang, Q. Jing, G. Li, R. P. Liu, L. L. Sun, J. Zhang, and W. K. Wang Effect of VUV Radiation on Properties and Chemical Structure of Polyethylene Terephthalate Film G. Peng, D. Yang, and S. He Status of Solar Sail Material Characterization at NASA’S Marshall Space Flight Center D. L. Edwards, C. Semmel, M. Hovater, M. Nehls, P. Gray, W. Hubbs, and G. Wertz Atomic Oxygen Durability Evaluation of a UV Curable Ceramer Protective Coating B. A. Banks, C. A. Karniotis, D. Dworak, and M. Soucek

vii

209

217

225

233

247

Cermet Thermal Conversion Coatings for Space Applications B. W. Woods, D. W. Thompson, and J. A. Woollam

265

Multifunction Smart Coatings for Space Applications R. V. Kruzelecky, E. Haddad, B. Wong, W. Jamroz, M. Soltani, M. Chaker, D. Nikanpour and X. X. Jiang

277

Effects of Space Environment Exposure on the Blocking Force of Silicone Adhesive P. Boeder, R. Mikatarian, M. J. Lorenz, S. Koontz, K. Albyn, and M. Finckenor

295

Dry Sliding Wear of Ti-6Al-4V Alloy at Low Temperature in Vacuum Y. Liu, D. Yang, S. He, and Z. Ye

309

Erosion of Kapton H by Hyperthermal Atomic Oxygen: Dependence on O-Atom Fluence and Surface Temperature D. M. Buczala and T. K. Minton

317

viii

CONTENTS

Transparent Arcproof Protective Coatings: Performance and Manufacturability Issues J. Griffin, N. Uppala, J. Vemulapalli, and P. D. Hambourger

331

The Study of the Effects of Atomic Oxygen Erosion on the Microstructure and Property of VO2 Thermochromic Coating Using CSA’S Space Simulation Apparatus X. X. Jiang, D. Nikanpour, M. Soltani, M. Chaker, R. V. Kruzelecky, and E. Haddad

341

Damage Kinetics of Quartz Glass by Proton Radiation Q. Wei, S. Y. He, and D. Z. Yang

351

Microscopic Mechanisms and Dynamics Simulations of O+ (4 S3/2 ) Reacting with Methane L. Sun and G. C. Schatz

359

Theoretical Study of Reactions of Hyperthermal O(3 P) with Perfluorinated Hydrocarbons D. Troya and G. C. Schatz

365

Simulation of UV Influence on Outgassing of Polymer Composites R. H. Khassanchine, A. N. Galygin, A. V. Grigorevskiy, and A. N. Timofeev

377

The Impact of High-Velocity Particles on Thermal Pipelines in Spacecraft N. D. Semkin, K. E. Voronov, and L. S. Novikov

385

Physical Mechanism of Solar Cell Shunting Under High Velocity Impact of Solid Particles V. A. Letin, A. B. Nadiradze, and L. S. Novikov

393

Determination of Ground-Laboratory to in-Space Effective Atomic Oxygen Fluence for DC 93-500 Silicone K. K. De Groh, B. A. Banks, and D. Ma

401

Atomic Oxygen Concentration Using Reflecting Mirrors M. Tagawa, K. Matsumoto, H. Doi, K. Yokota, and N. Ohmae

417

Atomic Oxygen Source Calibration Issues: A Universal Approach C. White, J. C. Valer, A. Chambers, and G. Roberts

431

CONTENTS

Low-Cost Space Missions for Scientific and Technological Investigations D. Rankin, R. E. Zee, F. M. Pranajaya, D G. Foisy, and A. M. Beattie

ix

443

Subject Index

455

Author Index

461

INTRODUCTION

This publication presents the proceedings of ICPMSE-7, the Seventh International Conference on Protection of Materials and Structures from Space Environment, held in Toronto May 10–13, 2004. The ICPMSE series of meetings became an important part of the LEO space community since it was started in 1991. Since then, the meeting has grown steadily, establishing itself as the only North American event covering the various aspects of materials protection in LEO and attracting a large number of engineers, researchers, managers, and scientists from industrial companies, scientific institutions and government agencies in Canada, USA, Asia, and Europe, thus becoming a true international event. The ICPMSE-7 meeting continued the tradition of the previous meetings including in the program the topics on protection of materials in GEO and Deep Space. The conference was organized by Integrity Testing Laboratory Inc. (ITL), and hosted by the University of Toronto’s Institute for Aerospace Studies (UTIAS). The meeting was sponsored by: a) The Materials and Manufacturing Ontario (MMO) and the CRESTech, two Ontario Centres of Excellence that from April 1, 2004 joined under the Ontario Centres of Excellence Inc. (OCE Inc) a not-for-profit, member-based corporation dedicated to establishing Ontario as the place to be for innovation; b) MD Robotics; c) The Integrity Testing Laboratory (ITL) and d) The University of Toronto Institute for Aerospace Studies (UTIAS). Over 80 people from countries covering the American, European and Asian continents registered for the conference representing the major space agencies and the major companies, institutions and government organizations involved in space activities, indicating a further increase in international co-operation in this critical area of protection of materials in space. The papers in the proceedings were organized into six major sections as follows: Session O: Opening Session Session A: Space Environmental Effects: Radiation and Charging Session B: Space Environmental Effects: Synergism of AO/VUV/TC Session C: Space Environmental Effects: Synergism of AO/VUV/TC Session D: Space Environmental Effects: Instrumentation and Calibration

xi

xii

INTRODUCTION

Session E: New Materials and Processes Session F: Modeling and Computer Simulations In addition, poster sessions were organized that covered the same subjects. Jacob Kleiman Chairman/Organizing Committee/ICMSE-7 Integrity Testing Laboratory Inc. 20 January, 2005

ACKNOWLEDGEMENTS

We would like to acknowledge the following for their generous support of ICPMSE-7, the Seventh International Conference on Protection of Materials and Structures from Space Environment;

r r r r r

CRESTech Materials and Manufacturing Ontario (MMO) MD Robotics The Integrity Testing Laboratory (ITL), The University of Toronto Institute for Aerospace Studies (UTIAS)

As well, we would like to acknowledge all the people from ITL and UTIAS that contributed their time and effort and especially Janina Zuchlinski a bright York University co-op student for their help in preparation of the materials for publication. Jacob Kleiman Integrity Testing Laboratory Inc. Conference Chairman

xiii

ORGANIZATION

7th International Conference on “Protection of Materials and Structures from Space Environment” ICPMSE-7 May 10–13, 2004 Toronto, Canada Chairperson: Dr. Jacob Kleiman, ITL Inc./UTIAS, Canada

Organizing Committee B. A. Banks, NASA, Cleveland, USA M. Dinguirard, ONERA/DESP, France D. L. Edwards, NASA, Huntsville, USA T. Minton, Montana State University, USA D. Nikanpour, Canadian Space Agency, Canada L. S. Novikov, Moscow State University, Russia V. G. Sitalo, Design Bureau “Yuzhnoe,” Ukraine M. Tagawa, Kobe University, Japan R. C. Tennyson, UTIAS, Toronto, Canada P. C. Trulove, Air Force, USA M. Van Eesbeek, ESA, Noordwijk, the Netherlands E. Werling, CNES, France D. Yang, Harbin Institute of Technology, China

Program Committee J. Golden, Boeing, Houston, USA M. Finkenor, NASA, Huntsville, USA S. Koontz, NASA, Houston, USA G. Pippin, Boeing, Seattle, USA E. F. Nikishin, M. V. Khrunichev State Space Scientific Production Center, Russia E. Grossman, Soreq NRC, Israel

xv

xvi

ORGANIZATION

Local Organizing Committee Z. Iskanderova, ITL Inc./ UTIAS, Toronto, Canada R. Worsfold, CRESTech, Toronto, Canada H. Pellegrini, MMO, Toronto, Canada P. Patnaik, NRC-IAR, Ottawa, Canada

Session Chairs Opening Session Moderator: Jacob Kleiman, ITL Inc./UTIAS, Canada

Session A: Space Environmental Effects: Radiation, Charging and UV Effects Moderator: Kim K. DeGroh, NASA, Cleveland, USA Session B: Space Environmental Effects: AO/VUV/TC/Micrometeoroids Effects (Ground Simulation and Flight Experiments) Moderator: Dave Edwards, NASA, Houston, USA Session C: Materials and Processes I Moderator: Gary Pippin-Boeing, Seattle, USA Session D: Materials and Processes II Moderator: Bruce Banks, NASA, Cleveland, USA Session E: Modeling and Computer Simulations Moderator: Tim Minton, Montana State University, USA Session F: Space Environmental Effects: Instrumentation and Calibration Moderator: Alan Chambers, University of Southampton, UK

RADIATION EFFECTS OF PROTONS AND ELECTRONS ON BACKFIELD SILICON SOLAR CELLS

ZHENGYU HU, SHIYU HE, AND DEZHUANG YANG Space Materials and Environment Engineering Laboratory, Harbin Institute of Technology, Harbin 150001, P. R. China

Abstract. Radiation effects of protons and electrons on the backfield silicon solar cells were investigated. The samples without cover glass were irradiated by the protons and electrons with 30–180 keV and a given flux of 1.2 × 1012 cm−2 · s−1 for various fluences at 77 K. Experimental results show that the short circuit current decreases gradually with increasing the proton fluence, while the open circuit voltage degrades severely under lower fluences. No obvious changes appear in the electric properties before and after the irradiation by electrons, and there exists a recovery effect in the in situ measurement for the irradiated samples. The effect of the combined radiation of protons and electrons does not show simple additivity. The damage extent of proton radiation is larger than that of combined radiation under lower electron fluences, while the combined radiation results in more severe damage under higher fluences. The DLTS analysis verified that the primary defects induced by protons were the H1 or [V+B] type at the energy level of +0.45 eV, which would result in formation of a resistance layer in the base region and degradation of the backfield Si solar cells. Key words: Radiation effects, Protons, Electrons, Sollar cells

1. Introduction Since solar cells are key elements to provide spacecraft with electric energy in orbit, the degradation in their properties could directly influence the working condition and lifetime of spacecraft. The electric properties of solar cells would be degenerated under the space radiations [1]. It is important to improve the radiation resistance of the solar cells [2–5]. In order to evaluate the performance of the solar cells in the geostationary Earth orbit, it is necessary to characterize the effects of charged particles with the energy less than 200 keV, which exist in large amount in the Earth’s radiation belts. Also, when spacecraft passes into the Earth shadow, the solar cells are subjected to the effect of low temperatures. The 1 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 1–8. C 2006 Springer. Printed in the Netherlands.

2

ZHENGYU HU ET AL.

aim of this study was to examine the change in electric properties of the silicon solar cells under the radiations of protons and electrons with <200 keV at 77 K, as well as the radiation damage mechanism.

2. Experimental The sample of the backfield silicon solar cell is schematically shown in figure 1, which is 250 μm thick and has the base layer resistivity of 10–12 ·cm. The samples with the size of 20 × 20 mm2 were not covered with glass. The phosphorus doped Si emitter junction depth from the surface is 0.2 μm. The parameters of samples before irradiation are shown in table 1. The ground-based simulation equipment used in this study can simulate the radiations of solar electromagnetic rays, electrons and protons in orbit, independently and simultaneously. The irradiation energy of protons and electrons was chosen as 60–180 keV, the flux 1.2 × 1012 cm −2 ·s−1 and the fluence from 2 × 1013 to 2 × 1016 cm−2 . The homogeneity of the irradiation flux is more than 95% in the area of 100 × 100 mm2 by scanning the proton and electron beam. During the irradiations, the chamber was kept at vacuum 10−4 Pa and temperature of 77 K. The electric properties of the samples were characterized by I –V curves, which were measured before and after the irradiations. In order to analyze the radiationinduced defects, DLTS analysis (deep level transient spectroscopy) was carried out in the temperature range of 77–450 K.

TABLE 1. The parameters of the samples before irradiation Isc (mA)

Voc (mV)

Pm (mV)

FF

η (%)

138.7

552

59.97

0.78

15

VI

VII

V

IV

III

II

I

I. Ohmic contact region on back surface; II. p+ doped region; III. p base region; IV. pn junction region; V. n+ top region; VI. Attenuating reflection coating; VII. Gate electrode

Figure 1. Schematic diagram of the samples

3

RADIATION EFFECTS OF PROTONS AND ELECTRONS

Figure 2. The I –V curves for the samples irradiated with (a) 60 and (b) 170 keV protons

3. Results and Discussion 3.1. CHANGES IN ELECTRIC PROPERTIES UNDER PROTON RADIATION

Figure 2(a) and 2(b) show the I –V curves for the samples irradiated with 60 and 170 keV protons, respectively. The changes in the normalized short-circuit current Isc /I0 and the normalized open-circuit voltage Voc /Vo are given as a function of fluence of protons with various energies, as shown in figures 3(a) and 3(b). With increasing both fluence and energy, the short-circuit current Isc decreases gradually. The irradiation of protons also leads to an obvious decrease of the open-circuit voltage Voc , but the degradation extent depends on the fluence and energy of the protons. After the irradiation by protons with 150–170 keV, the Voc is noticeably degraded at lower fluences, and then remains almost unchangeable 1.0

1.0

Normalized I sc

0.9 0.8 0.7 0.6 0.5 0.4

1013

77 k 60 keV 150 keV 170 keV

0.9 Normalized V oc

77 k 60 keV 150 keV 170 keV

0.8 0.7 0.6 0.5 0.4

1014

1015

1016 −2

Proton fluence (cm )

(a)

1017

0.3 1013

1014

1015

1016

1017

Proton fluence (cm−2)

(b)

Figure 3. Changes in the normalized Isc (a) and normalized Voc (b) with fluence for the samples irradiated with 60, 150, and 170 keV protons

4

ZHENGYU HU ET AL.

TABLE 2. Variation of the characteristics for the samples before and after 60 keV proton irradiation (AM0) Characteristics

Voc (V)

Isc (mA)

Jsc (mA.cm−2 )

FF

EFF (%)

Rs ()

Before irradiation 60 keV, 1×1016 cm −2

0.5299 0.2823

159.80 128.56

39.94 32.14

0.725 0.511

11.2 4.19

0.42 3.39

with the increase of fluence or even shows a recovery after a fluence higher than 5 × 1015 cm−2 . Table 2 shows that after the irradiation by protons with 60 keV to a fluence of 1 × 1016 cm−2 , all the characteristics including the Voc , Isc , Jsc , FF, and EFF decrease, while the internal resistance in series increases remarkably (almost 10 times higher). This implies that the concentration of carriers inside the samples is obviously reduced due to the proton irradiation. 3.2. CHANGES IN ELECTRIC PROPERTIES UNDER ELECTRON RADIATION

Figure 4 shows the I –V curves for the samples before and after irradiation by 60 and 180 keV electrons. It is obvious that there is very little change in the I –V curves. However, the situation is quite different, if the I –V curves are examined in situ during the exposure. With increasing the electron fluence, the normalized Voc gradually decreases (see figure 5(a)), while the normalized Isc increases until the fluence less than 1 × 1016 cm−2 (see figure 5(b)). Figures 6(a) and 6(b) show the changes in the Voc and Isc with time in vacuum after the electron irradiation, respectively. It is demonstrated that with increasing the time, the Voc rises and the Isc drops again. Both the Voc and Isc tend to return gradually to the original level for the unirradiated samples. This phenomenon implies that a recovery must be considered in evaluating the electron radiation effect for the backfield Si solar cells in orbit. 140 120

I (mA)

100 80

before irradiation 60 keV 1×1016 e/cm−2 180 keV 1×1016 e/cm−2

60 40 20 0

0

100

200

300

400

500

600

V (mV)

Figure 4. Variation of I –V curves for the samples before and after electron irradiation

5

RADIATION EFFECTS OF PROTONS AND ELECTRONS 1.5

170 keV

0.9

Normalized Isc

Normalized Voc

1.0

0.8 0.7 0.6 0.5

1.4 1.3 77 K 170 keV

1.2 1.1 1.0

0.0

4.0×10

15

15

8.0×10

16

1.2×10

16

1.6×10

15

15

1.2×10 4.0×10 8.0×10 Electron fluence (e/cm2)

0.0

2

Electron fluence (e/cm )

(a)

16

1.6×10

16

(b)

Figure 5. The normalized Voc (a) and normalized Isc (b) measured in situ as a function of fluence for the samples irradiated with 170 keV electrons

3.3. CHANGES IN ELECTRIC PROPERTIES UNDER THE COMBINED RADIATION OF PROTONS AND ELECTRONS

Figure 7 shows the normalized I –V curves for the samples under the proton, the electron and the combined irradiations with 170 keV. Notice that the damage effect of protons is much larger than that of the electrons (see curves 1, 2, and 4). The effect of the combined radiation of protons and electrons does not show simple additivity. The degradation extent due to the proton radiation is larger than that of the combined radiation under the same proton fluence and lower electron fluence (see curves 3 and 4). In contrast, the combined radiation results in more severe damage under the same proton fluence and higher electron fluence (see curves 4 and 5). 1.4

0.9 0.8

77 K 170 keV 1.5×1016 e/cm2 after irradiation

0.7 0.6 0.5

170 keV 1.5×1016 e/cm2 after irradiation

1.3

0

20

40

60 80 100 120 140 160 Time (min)

(a)

Normalized Isc

Normalized Voc

1.0

1.2 1.1 1.0 0.9

0

20

40 60 80 Time (min)

100

120

(b)

Figure 6. The normalized Voc (a) and normalized Isc (b) measured in situ as function of time in vacuum after electron irradiation with 170 keV

6

ZHENGYU HU ET AL. 1

1.0

170 keV 1- 5×1015 e/cm−2 2- 3×1015 p/cm−2 3- 3×1015 e/cm−2 and 3×1015 p/cm−2 4- 5×1015 p/cm−2 5- 5×1015 e/cm−2 and 5×1015 p/cm−2

I'

0.8 2

0.6 3 4 5

0.4 0.2 0.0

0.0

0.2

0.4

0.6

0.8

1.0

V' Figure 7. The normalized I –V curves for the samples under the proton, the electron, and the combined irradiations with 170 keV

3.4. ANALYSIS ON RADIATION-INDUCED DEFECTS

Figure 8 shows the DLTS spectra for the samples irradiated with 150 keV protons for various fluences. With increasing fluence, the DLTS signals are reduced and the peak width increases. The effect of proton energy on the DLTS spectrum is shown in figure 9. The DLTS spectrum shows a more symmetrical distribution for the 60 keV proton irradiation, while exhibits an obvious asymmetry after the irradiation with 150 keV protons. In addition, the DLTS signals in the temperature range of 125–230 K are increased noticeably. The change in the DLTS peak height can be related to the variation of the concentration of the defects at deep energy levels due to proton irradiation. According to the DLTS analysis results, the

Intensity

1

50

1- 150 keV 2×1016p+/cm2 2- 150 keV 2×1013p+/cm2

2

100

250

150

200

300

350

T (K) Figure 8. DLTS spectra for the samples irradiated with 150 keV protons for various fluences

RADIATION EFFECTS OF PROTONS AND ELECTRONS

7

1

Intensity

2

1- 60 keV 2×1016 p+/cm2 2- 150 keV 2×1016 p+/cm2

50

100

150

200

250

300

350

T (K) Figure 9. Variation of DLTS spectrum for the samples after proton irradiation with different energies and the same fluence

primary defects induced by the protons are believed to belong to the H1 type with the +0.45 eV energy level. The H1 type defects are generally referred to as the combination of radiation-induced vacancies with boron atoms, namely the [V+B] defects [6]. Under the <200 keV proton irradiation, a large amount of vacancies could be formed in the base region close to the pn junction of the backfield Si cell samples, and the vacancies would interact with the nearby boron atoms in the silicon, forming the [V+B] type defects. The interaction of the radiation-induced vacancies with boron atoms can provide the thermodynamic driving force for the segregation of boron atoms in the base region close to pn junction. As a result, a resistance layer with high concentration of the [V+B] defects near the pn junction could be formed due to the proton radiation. The formation of such a resistance layer can contribute to the decrease in electric properties for the backfield Si solar cells irradiated with <200 keV protons.

4. Conclusions Solar cells are key elements of the system providing the electric energy in spacecraft. It is of theoretical and practical significance to thoroughly study the damage effects of charged particles on solar cells under space environment. The samples for the backfield Si solar cells were irradiated by protons and electrons with energy of <200 keV. It is found that the proton irradiation leads to remarkable decreases in the short-circuit current and open-circuit voltage. No obvious changes appear in the electric properties after irradiation by electrons. The radiation effects of the electrons are recoverable after the irradiation. The effect of the combined radiation of protons and electrons does not show simple additivity. The damage extent of

8

ZHENGYU HU ET AL.

proton radiation is larger than the combined radiation under the same proton and the lower electron fluence. In contrast, the combined radiation could result in more severe damage if the electron fluence is higher. The damage of the backfield Si solar cells is primarily caused by the protons, and could be related to the formation of the [V+B] type defects at +0.45 eV level and the resistance layer in the base region close to the pn junction.

Acknowledgment The financial support of the National Basis Research Foundation of China under Grant # G19990650 is greatly appreciated.

References 1. Hisamatsu, T., Kawasaki, O., and Tsukamoto, K. (1998) Radiation Physics and Chemistry 53, 25. 2. Yeh, Y. C. M., Chu, C. L., Krogen, J., and Ho, F. F. (1996) In Proceedings of the 25th IEEE Photovoltaic Specialists Conference, IEEE, New York, p. 187. 3. La Roche, G. J., Schultze, W., Rizo, I., and Bogus, K. (1991) Assessment of Third Generation Solar Cells, NASA N92-13223, pp. 487–494. 4. Barabanenkov, M. Y., Gyulai, J., Leonov, A. V., and Mordkovich, V. N. (2001) Nuclear Instrument and Methods in Physics Research B 174(3), 304–310. 5. Hu, Z., He, S., and Yang, D. (2004) Nuclear Instrument and Methods in Physics Research B 1217, 321–326. 6. Hazdra, P., Brand, K., and Vobecky, J. (2002) Nuclear Instrument and Methods in Physics Research B 1192, 291–300.

SOLAR ARRAY ARCING IN LEO How Much Charge is Discharged? D. C. FERGUSON,1 B. V. VAYNER,2 AND J. T. GALOFARO1 1 NASA Glenn Research Center, 21000 Brookpark Road Cleveland, OH 44135, USA 2 Ohio Aerospace Institute, 22800 Cedar Point Road Cleveland, OH 44142, USA

Abstract. It is often said that only the solar array or spacecraft surfaces that can be reached by an arc plume are discharged in a solar array arc in LEO (low Earth orbit). We present definitive results from ground test experiments done in the National Plasma Interactions (N-PI) facility at the NASA Glenn Research Center that this idea is mistaken. All structure surfaces in contact with the surrounding plasma and connected to spacecraft ground are discharged, whether the arc plasma can reach them or not. Implications for the strength and damaging effects of arcs on LEO spacecraft are discussed, and mitigation techniques are proposed. Key words: solar arrays, low Earth orbit, arcing, plasma interactions, arc plumes, mitigation

1. Background Modern solar arrays have areas of tens of square meters, and they operate with bus voltages exceeding 100 V. Electrostatic discharges (arcs) are undesirable and detrimental events for spacecraft function, and preventing these events and/or mitigating their consequences are of primary importance for spacecraft designers. There are two types of arcs that may occur in space. The first, the primary or trigger arc, is a transient event that discharges some spacecraft capacitance. The second type, the sustained arc, can occur between two solar array strings, and may be powered by the array’s current generating capacity. In this paper, we consider only the transient trigger or primary arcs. Ground tests of small samples of large solar arrays have been used to provide the necessary information regarding arc inception voltages and expected arc damage for an entire array during its lifetime in space. However, the volume of the space plasma and the size of the test arrays that may be simulated in ground tests 9 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 9–19. C 2006 Springer. Printed in the Netherlands.

10

D. C. FERGUSON ET AL.

is limited by the size of the test vacuum chamber, and this fact necessitates the installation of additional capacitance between the sample and ground to simulate the actual capacitance of a spacecraft and its solar array discharging through the arc plasma. The magnitude of the capacitance to be added has been the subject of discussions for many years (see for example, [1]). If the discharge of a space solar array capacitance is caused by an arc plasma front propagating along the array surface (see [2, p. 31]), this magnitude is limited to about 1 μF, because the array capacitance is approximately 0.25 μF·m−2 , and the propagation distance of the dense arc plasma is less than a few meters under the conditions of a typical LEO plasma. On the other hand, if the entire array discharges through a current channel created by an arc, this capacitance can even reach 103 μF. The amplitude and width of an arc current pulse are both increasing functions of the capacitance discharged, and that is why the damage inflicted on the solar array by an arc depends on the capacitance discharged in the arc. Is it the capacitance reached by the dense arc–plasma front, or is it the much larger capacitance of the coverglasses of the entire array that is discharged? The experiments described below confirm that the entire array capacitance discharges through the arc current channel even under conditions when the arc– plasma front is prevented from propagating along the sample surface. Thus, the proper value for an additional capacitor must be high (∼103 μF) for ground tests of arrays in order to properly simulate the damaging effects that may occur for arcs on spacecraft with large arrays.

2. Experimental Setup All of our tests were done in the National Plasma Interactions Facility (N-PI Facility) at the NASA Glenn Research Center, Cleveland, Ohio. In our tests, the LEO space plasma was simulated in a large vacuum chamber (2 m in diameter and 3 m high) equipped with four oil diffusion vacuum pumps that provided a background pressure about 0.5 μTorr (66.6 μPa). One Kaufmann-type plasma source generated a xenon plasma with an electron temperature of 1–1.3 eV, an electron number density of (4–5) × 105 cm−3 , and a neutral gas pressure of about 50 μTorr (6.67 mPa). Two solar array samples (on fiberglass) were mounted on an aluminum sheet and installed vertically in the middle of the chamber (figure 1). One sample (strings 1, 2, and 3) represented a silicon solar array with UVR coverglasses of 300 μm thickness that corresponds to a capacitance of 4344 pF·string−1 . Another sample (strings 4, 5, and 6) had a capacitance of 7020 pF·string−1 because of its thinner coverglasses (150 μm). The additional capacitor was chosen to have a capacitance of C = 0.03 μF (±10%) for the convenience of measurements–such a choice provided comparable currents in all branches of the bias circuit. However,

SOLAR ARRAY ARCING IN LEO

11

Figure 1. Two solar array samples are mounted on an aluminum sheet and installed vertically in a large vacuum chamber. All dimensions are shown in mm

some measurements were done with a higher capacitance (0.25 μF) to reveal the dependence of the arc current pulse characteristics on the value of this capacitance. Four current probes provided measurements of discharge currents flowing in essential branches of the circuit (figure 2). For the second series of measurements, a grounded aluminum plate was installed between the samples to prevent the propagation of the arc–plasma front from one sample to the other. 2.1. EXPERIMENTAL RESULTS

All four current pulse waveforms were registered by a four-channel digital oscilloscope and stored in a computer for further processing. Twenty events (arcs) were observed for each configuration (positions of keys 1–4, and capacitance C). That amounted to 260 files, one of which is shown in figure 3. Each file was used to obtain the following data: (1)I p —peak arc current; (2) τ0.5 —pulse width at 0.5 of the peak current value; (3)qi —net electrical charge flowed through the corresponding branch; and (4) tij —time interval between current pulse peaks in the different circuit branches. The magnitude of the net electrical charge flowing through a branch was calculated as qi = (1) Ii (t) dt

12

D. C. FERGUSON ET AL.

Figure 2. Four current probes were used to measure the discharge currents in four different branches of the circuit. The additional capacitance was chosen to be 0.03 μF to obtain comparable current magnitudes for all probes. A few measurements were done with C = 0.25 μF

Then, the average and standard deviation over several measurements were calculated, and the resultant value was compared with the theoretically predicted value. For example, the ratio of charges for string #1 (q1 ) and capacitor (q2 ) was calculated by Cstring q1 = q2 C

(2)

Figure 3. One example of an arc pulse current sequence for C = 0.03 μF, and for switches K1 and K2 in the closed position. CP1 is the most positive current. The arc was initiated at time t = 0

13

SOLAR ARRAY ARCING IN LEO TABLE 1. Measurement results and theoretical estimates

No.

Key # closed

Arc on string #

1 2 3 4 5 6 7 8 9 10

None None K K1 K1&K2 K1&K2 K1&K2 K1&K2 All All

1 6 1/2 6 2/3 6 2/3 6 4/5 2/3,6

qi q2

qarc q2

Measured

Estimate

Measured

Estimate

0.25(0.03) 0.17(0.03) 0.25(0.024) 0.16(0.02) 0.25(0.026) 0.217(0.06) 0.26(0.034) 0.21(0.05) 0.19(0.05)

0.234(0.02) 0.145(0.015) 0.234(0.02) 0.145(0.015) 0.234(0.02) 0.434(0.04) 0.234(0.02) 0.434(0.04) 0.234(0.02)

1.186(0.11) 1.077(0.04) 1.41(0.08) 1.22(0.08) 1.4(0.08) 1.49(0.07) 1.43(0.1) 1.46(0.1)

1.379(0.05) 1.379(0.05) 1.52(0.06) 1.52(0.06) 1.67(0.07) 1.67(0.07) 1.67(0.07) 1.67(0.07)

1.85(0.06)

2.14(0.1)

Bias (V) 300 300 300 300 300 300 280 280 280 280

It should be noted that the possible errors in the calculations of string capacitances could not be estimated properly because of unknown errors in the corresponding geometrical and electrical parameters. However, the consistency of all or our final results is a very convincing argument that the calculations of array capacitances were done with an error of less than 10%. Also, the following ratio: Cstring + C qarc = (3) q2 C was verified for all events when the experimental setup made it possible to do so. This equality means that the array capacitance that discharged through the arc current channel is independent of the distance between the arc site and other strings, and the array discharged fully with or without an aluminum plate installed between the two samples. The results of our measurements and theoretical estimates are compiled in table 1. Standard deviations (1σ ) of the measurements are shown in parenthesis. The numbers shown in table 1 demonstrate a very good agreement between the measured parameters and their theoretical estimates. We believe that some insignificant differences can be explained by our poor knowledge of the string capacitances, possibly by a somewhat incomplete discharge of the panels, and possibly by a somewhat inhomogeneous plasma potential distribution. However, the considerably smaller-than-expected discharge of string #6 observed during two different runs (run numbers 6 and 8 above) cannot be explained to date. These results look particularly strange if one takes into account the very good agreement between the measurements for strings #2/3 and their estimated values (run numbers 5 and 7 above), because these two runs were supposed to be symmetrical to each other.

14

D. C. FERGUSON ET AL.

TABLE 2. Experimental results with aluminum panel installed qi q2

qarc q2

No.

Key # closed

Arc on string #

Measured

Estimate

Measured

Estimate

Bias (V)

1 2 3 4 5 6

K1&K2 K1&K2 K1&K2 K1&K2 K1&K2 K1&K2

2/3 1 2/3 1 2/3 1

0.21(0.03) 0.2(0.02) 0.03(0.004) 0.028(0.002) 0.191(0.02) 0.203(0.024)

0.234(0.02) 0.234(0.02) 0.028(0.003) 0.028(0.003) 0.234(0.02) 0.234(0.02)

1.32(0.07) 1.59(0.07) 0.956(0.02) 1.084(0.04) 1.39(0.06) 1.63(0.1)

1.67(0.07) 1.67(0.07) 1.08(0.05) 1.08(0.05) 1.67(0.07) 1.67(0.07)

450 450 400 400 400 400

In the second stage of the experiment, an aluminum panel was installed between the two samples to prevent the propagation of the arc plasma from the arc site to the other sample (figure 1). Measurements were done with the same additional capacitor (0.03 μF, ±10%) and with another capacitor (0.22 μF, ±5%) connected in parallel with the first one. The results are shown in table 2. It can be seen from the data in tables 1 and 2 that those strings that were not arcing discharged fully in both cases—with or without the aluminum panel between the samples. Thus, the mechanism of discharge of an entire array can only be explained by an electron current flowing from the negatively charged conductor (or semiconductor) to the surrounding plasma through the arc–plasma conductive channel. The positive charge of the coverglass is neutralizing by ambient plasma electrons that are attracted by the positive potential of the coverglass. The dependences of the arc current pulse width and amplitude on the net capacitance were found from the experiments shown in table 1. However, the narrow range of capacitances used (0.042–0.064 μF) and large deviations in the measured values did not allow us to confirm (or to reject) any expected square root dependence (figure 4). For a discussion of the square root dependence, see [3]. For some measurements (shown in table 2), a bigger additional capacitor (0.25 μF) was used, and this provided the opportunity to verify the expected dependence of the pulse width on the capacitance (figure 5). It turned out that this dependence is slower than an expected (about 0.3 in power-law index, rather than the expected 0.5 as Snyder depicted in [4]. One more interesting feature of the discharge process is a time delay between the instant of the peaks in the arc–current pulse and in the discharge current of those strings not arcing (figure 3). We believe that this delay is caused by a changing plasma potential during the discharge process (which corresponds to the spacecraft potential for LEO orbit). In actuality, in the simple situation when all keys (K1–K4) are open, and the arc occurs on string #1, the relaxation current

SOLAR ARRAY ARCING IN LEO

15

Figure 4. Arc current amplitudes and pulse widths vs. net capacitance are shown for the experiments without a conducting plate between the samples. Pulse width measurements are at top and peak current measurements are at bottom. Units are as in the legend at top

on string #6 satisfies the following equation (eq. (4)): dI4 (x) Cstr + I4 (x) = − I1 (x) dx C

(4)

where x = t/τstr , and τstr is the string relaxation time. The solution of the eq. (4) with the initial condition I4 (0) = 0 can be written as x Cstr Cstr exp(−x) t exp(t)dt (5) I1 I4 (x) = − C C 0

Figure 5. In spite of large deviations, a square root dependence of pulse width on capacitance can be excluded. The actual dependence is closer to a C 1/3 dependence

16

D. C. FERGUSON ET AL.

Figure 6. The theoretical time delay (in x units) between peaks of arc current and string discharge current is similar to the observed one

If the arc current pulse is simulated by two exponents or by a Gaussian curve, the solution of eq. (5) is shown in figure 6. The measured time delay between current peaks in strings 1 and 6 was 1. with 280 V, no plate, 0.03 capacitor—2.6 (0.8) μs; 2. with 450 V, plate, 0.03 capacitor—4.88 (1.68) μs.

3. Conclusion The results of our current experiments and their analysis confirm the necessity of using a large additional capacitance (0.3–0.5 of the expected entire spacecraft solar array capacitance) in ground tests in order to adequately simulate the consequences of arcing on solar arrays in orbit. Spacecraft solar arrays that have very large capacitances connected to spacecraft ground may be damaged by even transient arcing events unless mitigation techniques are used. For example, a spacecraft solar array with a 300 V power system and 10 μF surface capacitance may be subject to arcs of more than 1000 A peak strength and total energies greater than 0.4 J, even if the arcs do not become sustained by the array power. Miller [5] showed that even arcs of 40 A peak strength may severely damage solar cell interconnects.

4. Mitigation (reprinted from [2, p. 38]) The design of a solar array must consider the plasma environment and interactions with that environment. Arc prevention is extremely important. The following techniques have been shown in ground and flight tests to prevent arcs or minimize their damage:

SOLAR ARRAY ARCING IN LEO

17

1. If possible, use array string voltages of less than 55 V. No trigger arcs have been seen on LEO arrays of less than about 55 V string voltage even under simulated micrometeoroid bombardment. Solar arrays coming out of eclipse will generate more voltage than when they operate at their max power point. 2. If solar array cell edges or interconnects are exposed to the LEO plasma and string voltages are greater than 55 V, the strings should be laid out on the substrate such that no two adjacent cells have a voltage difference of greater than 40 V. Sometimes a leapfrog arrangement will be sufficient. In other highvoltage arrays, the strings should be arranged parallel to each other. Serpentine strings can be used to prevent the array width from becoming prohibitive. If the string layout cannot be modified to prevent cells with more than 40 V difference being adjacent to each other (anything less than about 1 cm may be considered adjacent) then the total string voltage must be kept low enough that the initial (trigger) arcs do not take place. The lowest known array trigger arcing has occurred on thin-coverglass cells at about 75 V ([6], PASP Plus results). 3. For array string voltages greater than about 75 V, trigger arcs in LEO can only be completely prevented by encapsulating the cell or array edges so they do not see the ambient plasma. If encapsulation is not possible, a thorough array bakeout on-orbit (1 week at 100 ◦ C or more) may get rid of contaminants and prevent trigger arcing up to about 300 V, or possibly more (see [7]). Recontamination may occur on “dirty” spacecraft (spacecraft with excessive venting, cold gas nozzles, etc.). Good encapsulation may prevent arcing up to 1000 V string voltage. 4. Sustained (or continuous) arcs may occur whenever trigger arcs occur and adjacent cells have more than 40 V potential differences. However, sustained arcs, in addition to this voltage threshold, have a current threshold, below which they will not occur. It is believed that the current threshold is greater than about 0.5 A. If the current produced by each cell is above this threshold, a single string may sustain arcs. If each cell is below this current threshold, then isolating separate strings of solar cells from each other will prevent other strings from “feeding” the arc site, and will prevent sustained arcs. This isolation can be achieved by using blocking diodes in each string (EOS-AM1, now called Terra, e.g., [8], for instance). Care must be taken that the power bus and/or other components do not have the conditions necessary for sustained arcing. On the Terra arrays, for instance, it was found that diodes used to block interstring currents did not prevent the bus power traces from having sustained arcing events. Covering all exposed bus conductors with Kapton® insulation finally solved the problem. Low-outgassing RTV may be used to cover bare conductors as well. 5. RTV grout between adjacent solar cells and strings that have a high voltage with respect to each other has been shown to effectively block sustained arcs between

18

D. C. FERGUSON ET AL.

cells and strings. The degree of coverage, etc., is important in determining the final voltage threshold for sustained arcing. 6. Arrays of 300 V and greater string voltage must be fully encapsulated in order to prevent arcing. 7. Finally, although design and construction are important in preventing trigger arcs and sustained arcs, each new solar array implementation must be tested in a simulated LEO plasma before it can be sure not to arc. This step must not be omitted. The test bias voltage relative to the plasma should include the maximum when the arrays come out of eclipse (or the highest potential expected on the “floating” spacecraft). The interstring voltage should be at least as great as that expected anywhere on the solar array on-orbit. Tests should ideally be conducted at sample temperatures as low as the eclipse-egress temperature.

5. Mitigation Techniques on Other Spacecraft On the International Space Station (ISS), ground testing of the solar arrays showed that the arrays did not arc at negative voltages less than about 250 V. This is much less than the system operating voltage (160 V), so in the absence of auroral charging, ISS is never expected to charge more than 160 V negative. And, if trigger arcs do not occur, sustained arcs will not. However, ISS did take the precaution of incorporating a system of PCUs (plasma contactor units) to prevent structure charging by the solar arrays and to prevent dielectric breakdown of anodized aluminum surfaces. The LEO design guidelines [2] referenced above were based partly on ground tests done for the SS/Loral satellite series and the EOS-AM1 spacecraft (now called Terra). Mitigation techniques called out above were partly the result of those tests, and therefore should be considered to be based on real spacecraft experience in failure modes and failure investigations. Kapton® is a registered trademark of Dupont Co.

References 1. Levy, L., Reulet, R., Sarrail, D., and Migliorero, G. (1998) In 6th Spacecraft Charging Technology Conference, Hanscom Air Force Base, MA, AFRL-VS-TR-20001578, pp. 43–48. 2. Ferguson, D. C. and Hillard, G. B. (2003) Low Earth Orbit Spacecraft Charging Design Guidelines, NASA/TP—2003-212287. 3. Galofaro, J., Vayner, B., Degroot, W., and Ferguson, D. (2002) The Role of Water Vapor and Dissociative Recombination Processes in Solar Array Arc Initiation, NASA TM -2002-211328. 4. Purvis, C. K., Ferguson, D. C., Snyder, D. B., Grier, N. T., Staskus, J. V., and Roche, J. C. (1986) Environmental Interactions for Space Station and Solar Array Design, Preliminary, NASA Lewis Research Center, OH.

SOLAR ARRAY ARCING IN LEO

19

5. Miller, W. L. (1985) In Proceedings of the Spacecraft Environmental Interactions Technology Conference, NASA Lewis Research Center, OH, pp. 367–377. 6. Soldi, J. D., Jr. and Hastings, D. E. (1996) Arc Rate Simulation and Flight Data Analysis for the PASP Plus Experiment, NASA Technical Report PL-TR-95-2126. 7. Vayner, B. V., Galofaro, J. T., and Ferguson, D. C. (2002) The Neutral Gas Desorption and Breakdown on a Metal-Dielectric Junction Immersed in a Plasma, AIAA Paper 2002–2244. 8. Snyder, D. B., Ferguson, D. C., Vayner, B. V., and Galofaro, J. T. (2000) In 6th Spacecraft Charging Technology Conference, Hanscom Air Force Base, MA, AFRL-VS-TR-20001578, pp. 297–301.

SELF-RESTORATION AS SEU PROTECTION MECHANISM FOR RECONFIGURABLE ON-BOARD COMPUTING PLATFORM LEV KIRISCHIAN,1 VADIM GEURKOV,1 IRINA TERTERIAN,1 AND JACOB KLEIMAN2 1 Embedded and Reconfigurable Systems Laboratory (ERSL) at Ryerson University, Canada 2 Integrity Testing Laboratory Inc., Markham, Canada

Abstract. The following paper presents multilevel protection mechanism for the digital processing circuits, based on partially reconfigurable FPGA devices and run-time reconfiguration of processing architectures in the FPGA device. This approach assumes that if a hardware fault occurs in one of task processors, it can be restored without interruption of any computing processes in the rest of FPGA circuitry. The novel procedures of function restoration with and without performance degradation are proposed. These procedures allow automatic restoration of Application Specific Virtual Processors inside the FPGA device with minimization of performance degradation when a fault occurs. Implementation of this method can dramatically increase lifetime of the on-board computing platform and at the same time reduce restoration time period. Key words: reconfigurable computing, FPGA, radiation protection, SEU mitigation, onboard computing platforms

1. Introduction Most of the recent space missions require real-time processing of complicated images and large amounts of data-flow information coming from numerous independent sources. These tasks employ many processors and powerful computing recourses that are integrated in the application specific computing platform with a fixed architecture. This approach, which was known to be the only possible solution before has however, many disadvantages. The major disadvantage is the lack of flexibility and ability for any hardware modifications, and remote repair when there is a fault. Then again, a flexible computing architecture based on field programmable gate array (FPGA) devices, allows development of the uniform task-optimized computing platforms with the ability of downloading those processing modules that are specialized for the task. After completing the task, the 21 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 21–33. C 2006 Springer. Printed in the Netherlands.

22

LEV KIRISCHIAN ET AL.

architecture of the processing platform could be reconfigured to be optimal for the next task. This approach allows to have the same processing hardware for different applications and simultaneously provides a very high performance as a result of architecture-to-task optimization. However, this type of devices can be affected by radiation in LEO or GEO environments. The most common radiation effect is known to be SEU (single event upset). That means that a charged particle can change the state of flip-flops in the configuration static random access memory (SRAM), and/or configurable logic blocks (CLB) or block memory modules in the FPGA device. Thus, a protection mechanism is needed to keep FPGA-based onboard computing platforms reliable in the space environment. A novel approach based on the idea of self-restoration of damaged FPGA circuitry by run-time reconfiguration of processing architecture in the FPGA device, is proposed and discussed in the paper. This approach assumes that if fault occurs in one of task processors, it can be self-restored without interruption of computing processes in the rest of FPGA circuitry. The novel method of function restoration with and without performance degradation is proposed. This method allows automatic architecture-to-fault adaptation with minimization of the performance degradation when at fault. Implementation of this method can dramatically increase the lifetime of a computing platform, and reduce restoration time period. This paper also presents the description of self-restoration mechanism based on the concept of self-assembling virtual hardware components (VHC). It is shown that implementation of this mechanism minimizes reconfiguration time, size of nonvolatile system memory, and power consumption of computing platform. The rest of the paper is organized as follows: section 2 presents a brief overview of on-chip architecture (micro-architecture) organization of data-stream reconfigurable processors based on Xilinx Virtex FPGA. It also describes self-assembling mechanism, which will allow run-time assembling of these task-optimized processors by Virtual Hardware Components (VHCs). Section 3 presents the concept of multilevel mechanism for hardware fault restoration. In section 4, the procedure for hardware restoration with performance degradation is discussed. Section 5 provides the description of reconfigurable parallel stream processor (RPSP)— a prototype platform used for experiments for hardware restoration procedures. Lastly, section 6 presents experimental results and their analysis, followed by the summary of the project in section 7.

2. On-Chip Architecture Organization of Data-Stream Processor It is known fact [1, 2] that data-stream processing architecture can reach the highest cost-effective parameters, if it reflects the Data Flow Graph (DFG) structure of the task. In [3] was proposed the approach where each DFG node called

SELF-RESTORATION AS SEU PROTECTION MECHANISM

Global Routing Lines

23

V H C

Processing Element (PEi) Local routing Interface Element (IEj)

Tristate Buffer

Figure 1. Virtual hardware component structure

macro-operator—MOi (e.g., FFT, FIR-filter, Matrix Multiplication, etc.) corresponds to the appropriate virtual hardware component—VHCi. Most of the datastream processing architectures can be represented as a pipeline reflecting the structure of data-flow graph (DFG). Thus, we took into consideration these FPGA devices that can easily implement long pipelines and allow reconfiguration of pipeline components (stages). The best candidate for these requirements was Xilinx “virtex” family of partially reconfigurable FPGA devices [4]. The structure of “virtex” FPGA as described in [5] consists of: arrays of CLBs (Configurable Logic Blocks); arrays of IOBs (Input Output Blocks), SRAM memory blocks (Block RAM); clock logic resources (DLLs, etc.), and routing resources (global and local routing). These resources can be configured into one or more datapaths for one or more pipelined data-stream processors. The configuration can be done by loading configuration data file into the configuration SRAM, which programs logic functions of look-up-table (LUT) of each CLB and interconnections between logic, I/O, clock and memory resources. The configuration data file for entire FPGA device can be divided into smaller configuration data files for partial FPGA reconfiguration. Each small configuration data file represents a virtual hardware component (VHC) to be downloading into addressable FPGA slots (CLB-columns). Each VHC consists of two major parts (figure 1): (a) processing element (PE) and (b) interface element (IE). The Xilinx “virtex” FPGA structure allows loading of VHC partially, because partial reconfiguration for this family of FPGAs allows addressable configuration of each frame (part of a CLB-column). As was shown in “virtex” FPGA device data sheet [4] the special tristate buffers (T-buffers) can be implemented to connect or disconnect Virtual Hardware Components (VHCs) to the Global Routing Lines. Those T-buffers can be dedicated to specific global routing lines. Thus, each VHC, which contains the Interface Element with T-buffers is associated with specific

24

LEV KIRISCHIAN ET AL. A

B

C

D

A

B

C

D

Y +

+ X X % %

Y Global Routing Lines (GRL)

a) Task DFG

b) ASVP On-Chip Architecture

Figure 2. ASVP on-chip architecture (micro-architecture) assembled from VHC “EGO” blocks in accordance to task data-flow graph

global routing lines, will be connected to those lines but initially tri-stated at the initial architecture loading state. In [6] we proposed an approach where virtual hardware components are LEGO-type building blocks that allow self-assembling of task optimized application specific virtual processor (ASVP), inside the partially reconfigurable FPGA. To illustrate this concept in figure 2—the on-chip architecture (micro-architecture) of Application Specific Virtual Processor assembled from virtual hardware components is presented (figure 2(b)) according to task data-flow graph (figure 2(a)).

3. Mechanism for ASVP Self-Restoration There are a few main reasons for hardware faults that can occur in FPGA-based devices: (i) physical defects in wafer; (ii) hidden manufacturing defects; and (iii) radiation effects. In our project, we considered faults occurring in small amount of CLBs in the SRAM based on FPGA devices. Mostly, this type of faults occurs as a result of radiation effects (SEU—single event upset, SEL—single event lutch-up, etc.) usual for aerospace applications [7] or applications for radiation intensive environment (nuclear power stations, etc). As an alternative approach based on protection of materials from radiation intensive environment, we proposed selfrestoration mechanism based on reassembling of damaged processing structures. This mechanism utilize so-called scrubbing procedure [7], but allows functional restoration to be much faster. In this paper we will give only a brief description of proposed method. To restore ASVP we proposed a three-level procedure:

SELF-RESTORATION AS SEU PROTECTION MECHANISM

25

First level of self-restoration is based on VHC-scrubbing: reloading of same VHC configuration data-file to the same address area of configuration SRAM, of the FPGA where this VHC was located. This can eliminate wrong state of Flip-Flops in the configuration SRAM and correct damaged VHC structure. This prosedure can be performed using the same ASVP assembling mechanism discussed in previous section. This approach allows dramatic reduction of restoration time logic resources and power consumption in comparison with the common approach [8]. That is when scrubbing procedure should be performed cyclically for the entire FPGA device. Experimental results and comparison analysis will be discussed in section 6. Another benefit is that there is no additional hardware expenses for this level of restoration. Necessity of second level of self-restoration appears when radiation effects make one of logic gates damaged. This case can be considered when VHC scrubbing does not restore ASVP functions. In this case, the damaged gate should be avoided. However, because any VHC is based on CLB-column organization we developed a mechanism for relocation of VHC from the slot, to the spare slot that would contain the damaged gate. In this approach extra hardware cost is involved to increase reliability of the system. However, it is possible that after some period of time all reserved slots will be used. Third level of restoration with performance degradation appears when all spare CLB-columns (slots) are used. In this case it is possible to load the existing CLB-columns with another variant of VHC, with a reduced area and functional parameters. It allows the computing system to overcome the damaged CLB in the column being corrupted. In this case we “pay” extra processing time to increase the reliability. 3.1. VHC SCRUBBING

The first level of self-restoration of ASVP consists of the following steps: 1. Determination of a VHC which doesn’t process data properly using Built-inSelf Test subsystem [9]. 2. Pausing the data-stream computation in the damaged pipeline. 3. Reprogramming (scrubbing) the faulty VHC using the same VHC configuration bitstream. 4. Continuing the data-stream computation in the restored ASVP. 5. Increment the number of faults in the restored CLB-column and if the number of hardware faults exceed the limit initiate the second level of VHC restoration— CLB-column replacement. Thus, if VHC scrubbing procedure does not help it means that hardware fault is constant. In the proposed restoration procedure we do not analyse the reason for which causes this fault. In this case an entire CLB-column where VHC was located

26

LEV KIRISCHIAN ET AL.

is considered to be damaged and has to be excluded from further utilization. To restore the functionality of ASVP, VHC that is where the hardware fault was detected has to be relocated to another CLB-slot(s). This is guiding us into the second level of self-restoration mechanism—CLB-column replacement. 3.2. CLB-COLUMN (SLOT) REPLACEMENT

The following can be considered to be the process of replacement of damaged slot by reserved ones. This process consists of the following steps: 1. Pausing the ASVP pipeline. 2. Disabling damaged virtual hardware component by loading “dummy” VHC (marked as * in figure 3) to tri-state all outputs and to prevent data contention on the bus; 3. Selecting the available CLB-column(s) for loading configuration bitstream of VHC to be restored; 4. Composing the configuration bit-stream by inserting the selected frame addresses and associated information into the VHC-configuration data file. 5. Loading the VHC configuration file to the selected column(s) while continuing data-stream processing. In a case of a lack of spare CLB-columns, the third level of self-restoration procedure has to be initiated to perform. In this case different variant of VHC with reduced area and performance should be loaded to the non damaged part of the same CLB-column. For example, a 16-bit VHC architecture can be replaced on 8-bit VHC, which can restore a ASVP, but with a reduced performance. This is discused in more detail in section 4. A

B C

D Y

A +

B C

D Y +

X

*

%

% X

Figure 3. VHC replacement using spare CLB-slots. In this example a fault is detected in the CLBcolumn where Multiplier VHC was located. The “dumm” VHC (indicated as ∗ ) is loaded to the faulty CLB-slot and then Multiplier VHC is loaded to one of spare CLB-slots

SELF-RESTORATION AS SEU PROTECTION MECHANISM

27

4. ASVP Functional Restoration with Performance Degradation In this situation let us consider that all spare CLB-columns were already used or were not reserved at all. We also should put in account the fact of hardware fault in some CLB-slots, which means that one or few transistors in one or few logic gates are damaged. A CLB-column consists of 20,000 up to 50,000 transistors (depending on FPGA device specifics). Therefore, damaged logic gates can be avoided by reloading other variants of virtual hardware component—Ri, which requires less logic gates. In result we have considered two possible solutions to restore ASVP with performance degradation: a) First approach—“Minimization of restoration time”—assume reloading another variant of the same VHC (e.g., Multiplier), which requires less area, to the same CLB-slot. This approach gives the shortest restoration period and a simplest restoration control scheme. However, this solution may not provide minimum of performance degradation. For example, if a fault appears in a CLB-slot where Multiplier is located it may be more efficient to relocate Multiplier as is to another slot and load the damaged slot by a smaller variant of another VHC (e.g., Adder). b) Second approach—“Minimization of performance degradation”—assume that before functional restoration an optimization of ASVP architecture to the task algorithm and data structure should be done. This procedure should select the variant of ASVP architecture which satisfies new (after fault) resource limitations and allows minimization of the performance degradation (e.g., latency, data processing rate, etc.).

4.1. ASVP RESTORATION WITH MINIMUM TIME

Firstly the approach which allows minimization of the restoration time was considered. The restoration procedure consists of the following steps: 1. Suspension of damaged ASVP pipeline. 2. Selection of one of VHC variants with reduced logic area and performance (usually in this VHC some part is “dummy” and thus different VHC variants has different “location” of the “dummy” component). 3. Loading configuration bit-stream of selected variant of VHC to the damaged CLB-slot. 4. Continuation the data-stream processing on restored ASVP. 5. If hardware failure still affects data-stream processing the next VHC variant with reduced area and performance should be selected from the VHC table. 6. Return to the point (iii) of presented procedure.

28

LEV KIRISCHIAN ET AL.

This is a very general description of the restoration procedure. The detailed description of this procedure is specific for Xilinx Virtex FPGA devices and requires detailed explanation of organization of this FPGA. 4.2. ASVP RESTORATION WITH MINIMUM PERFORMANCE DEGRADATION

In [3] the method of data-flow architecture to task adaptation was described. This method is based on the idea of partial arrangement of ASVP design space. It was shown that processor’s design space can be represented by Architecture Configuration Graph (ACG) presented in figure 4. It was assumed that each Virtual Hardware Component (VHC) can be represented in different variants of logic resource organization: Rij, where i—indicates the type of VHC and j—indicates variant number of the VHC. It is obvious that all variants of any VHC could be arranged in order of amount of logic resources required for the VHC-variant such as: Ri,( j − 1 ) would need less logic gates than variant Ri, j. On the other hand,

R1

Ri,

1

R1.1

R1,k

R1,m1

Ri

Ri

Ri

Ri,j-1

Ri,j

Ri,

2

3

4

Performance

1

Ri,j-1

5

Ri,j

Ri,

6

7

Ri,j-1

8

Ri,j

9

Area Limitation: a) After fault; b) Before fault

2

3

4

5

6

7

8

9

Number of variant of ASVP architecture

Figure 4. Architecture configurations graph (arranged design space) for a ASVP (top tree) and performance diagram associated with ASVP variants

SELF-RESTORATION AS SEU PROTECTION MECHANISM

29

performance of VHC-variant Ri,( j − 1 ) generally is lower than of Ri, j, variant of VHC (needs more clock cycles to process same data). Being an element of pipelined data processing structure it results in a performance of degradation of complete stream processor. However, using the hierarchical arrangement of the Architecture Configuration Graph (ACG) it is possible to minimize performance degradation by a replacement of resources which are located on the lowest level of the ACG (figure 4). For example, let us assume that ASVP architecture variant # 6 was determined as optimal for a task. Furthermore, let’s assume that a fault occurs in CLB-slot where R1 is allocated. If the R1 , k VHC-variant will be replaced for another (smaller) variant of R1 (for example, on R1 , 1 ) the architecture will change from variant # 6 to variant # 3. This will result performance degradation higher than if we change variant # 6 to variant # 5 changing Ri, j to Ri,( j − 1 ) instead of R1 . This example shows that replacement of VHC allocated on the lowest level of ACG can restore data processing with minimum performance degradation. Thus, the procedure of ASVP restoration with minimum performance degradation requires the following major steps: 1. Creation and partial arrangement of Architecture Configurations Graph in accordance to the method described in [3]. 2. Selection of the optimal variant of ASVP architecture in accordance to new (after fault) area limitations (see figure 4). 3. Suspension of data-stream processing in ASVP. 4. Reloading new variants of VHCs to the FPGA slots in accordance to new variant of ASVP architecture. 5. Continuation of data processing. This is a general procedure, which does not include many technical details such as automated mapping procedure, data-stream suspension procedure, etc. These details are chip specific and thus, are not reflected in the general procedure. This procedure up to now is not fully automated. However, all of the above mentioned steps were implemented on the prototype platform and tested manually.

5. Implementation All levels of self-restoration procedures were implemented on the base of prototype of multitask and multimode reconfigurable parallel stream processor (RPSP). Reconfigurable parallel stream processor (RPSP) architecture, adaptable for the multitask and multimode workload, implements the architecture-to-task adaptation mechanism described in [6]. Architecture of RPSP includes the following components shown in figure 5:

30

LEV KIRISCHIAN ET AL. Input

data-streams

Output data-

FPGA1: RFM (Re-configurable Functional Module) SRAM: VHC Cache Configuration bus

FPGA 2: HOS (Hardware Operating System)

PC-interface Card

Memory (ROM) for: i) Library of VHC variants (bit-streams) ii) ASVPs variants for each data-stream task

VHC-address bus

Host PC

Figure 5. Architecture of reconfigurable parallel stream processor (RPSP) for multitask and multimode workload

1. Reconfigurable Functional Module (RFM) is based on partially reconfigurable FPGA and contains all of the necessary circuits: power regulators, line buffers, etc. This module is a field of uniform logic resources that can be configured to a number of task-optimized data-stream processors (ASVPs). 2. VHC memory is based on ROM (read only memory) and contains the library of configuration data files (configuration bit-streams), of virtual hardware components to be utilized in all modes of all tasks in workload. 3. Task memory is based on ROM (read only memory) and stores the configuration data files of the initial architectures of ASVPs optimized for all tasks in workload. 4. VHC Cache is based on SRAM (static random access memory) and stores all VHC cores that can be used for active tasks (loaded in RFM). 5. Hardware Operating System (HOS) is based on FPGA and performs the following functions: (i) task initiation and termination; (ii) Mode switching by loading set of VHCs to certain CLB-slots in RFM; (iii) Switching data-streams; (iv) interface control; (v) RFM diagnostic and restoration functions. 6. Host-PC is based on Pentium III with reconfigurable high bandwidth (∼1Gb/s) interface card. It plays role of human-machine interface, high-level (software) operating system for RPSP and secondary storage for all ASVPs and associated VHC. It also contains all instrumentation CAD tools. First prototype of RPSP developed in ERSL at Ryerson University is based on Xilinx XCV-400 E (RFM) and XCV-50 E (HOS) FPGA devices. The bandwidth

31

SELF-RESTORATION AS SEU PROTECTION MECHANISM

of I/O interface was equal to 7.2 Gb·s−1 (LVDS) and 8.5 Gb·s−1 (LVTTL). Cache volume = 1 M × 16 bits. Configuration bus bandwidth = 528 Mb·s−1 (33 MHz × 16 bit·cc−1 ).

6. Experimental Results To test and adjust ASVP restoration procedures a set of “faulty” virtual hardware components (FVHCs) was developed for parallel video-stream processing tasks such as: edge detection, brightness-to-colour conversion and image combination. The experiment included 1. 2. 3. 4. 5. 6. 7.

Initiation of ASVP and processing test-data stream. Loading FVHC in selected CLB-slot at random moments of time. Measurement of hardware fault detection time. Measurement of scrubbing time till ASVP restoration. Repetition of steps (ii)–(iv) to initiate second level restoration. Initiation of CLB-slot replacement to the spare CLB-slot. Measurement of CLB-slot replacement time.

Results of the above experiments are summarized in table 1. We have measured restoration periods for level 1 (VHC-scrubbing) and level 2 (CLB-slot replacement). Based on this information we calculated the acceleration of the functional restoration comparing with common approach when an entire FPGA device has to be reprogrammed (see table 1). These results show that restoration time directly depends on size of VHC. Restoration time is minimal if VHC can be allocated in one CLB-column and increases nearly linear when VHC requires 2, 3, or more CLB-columns. Another result is that CLB-slot (column) scrubbing (1st level of restoration) works twice faster than CLB-slot replacement. This effect appears because of the extra time TABLE 1. Experimental results of functional restoration time and acceleration of functional restoration comparing with entire FPGA reprogramming (all results collected for Xilinx Virtex XCV-400E FPGA) Number of CLB-Slots in the ASVP Restoration time if VHC-scrubbing Restoration time if CLB-slot replacement Acceleration (in times) of restoration for scrubbing Acceleration (in times) of restoration for CLB-slot replacement

1

2

3

4

0.09 ms 0.18 ms

0.17 ms 0.34 ms

0.25 ms 0.50 ms

0.33 ms 0.66 ms

73.3

38.8

26.4

20

36.6

19.4

13.2

10

32

LEV KIRISCHIAN ET AL.

needed for a “dummy” VHC loading. Thus, for maximum acceleration (comparing with the scrubbing of entire FPGA [8]) designers should develop a virtual hardware components close to 1 CLB-column. This can dramatically increase the restoration time (up to 70–100 times depending on FPGA volume and organization), for the reason that it was mentioned in [8] most of hardware faults in LEO and GEO are caused by the SEU which can be restored by scrubbing procedure.

7. Summary The major goal of the project was to find a mechanism that would be able to restore FPGA-based computing platforms functionality quickly, when hardware faults occur to minimize data corruption. Ideally, this restoration should not cause performance degradation but it is assumed that if there is no other choice, the system is better to decrease its performance characteristics rather then stop functioning. Thus, a multilevel self-restoration mechanism was developed based on the idea of self-assembled task optimized application specific virtual processors (ASVP). These processors being just cores can be assembled on-chip from the LEGO-type sub-cores—virtual hardware components (VHC). Furthermore, because ASVP is VHC-based devices they can be reassembled inside the FPGA when hardware fault occurs in one of VHCs. Established on this idea special procedures were developed for temporal faults caused by radiation (e.g., SEU) or static faults (e.g., physical defects in the wafer). All procedures were implemented on the reconfigurable platform based on the Xilinx Virtex FPGA (XCV-400 E). It was shown and seen that these procedures can provide a much faster (in orders) restoration of FPGA-based computing systems in comparison to a common approach.

References 1. Micheli, G.D. (1994) Synthesis and Optimization of Digital Circuits, McGraw-Hill, Inc., New York, 580 p. 2. Rixner, S. (2002) Stream Processor Architecture, Kluwer Academic Publishers, Nozwel, MA, USA 120 p. 3. Kirischian, L. (2000) In Proceedings PARELEC 2000International Conference on Parallel Computing in Electrical Engineering (IEEE #PR00759), Trois-Rivieres, Quebec, Canada, August 2000, pp. 100–105. 4. XILINX Virtex II Platform FPGA Handbook VG002 v.1.0, 6 December, 2000. 5. XAPP151 v1.6. Virtex Series Configuration Architecture User Guide, Xilinx Inc., 2003. 6. Kirischian, L., Szajek, L., and Chayab, F. (2002) In Proceedings PARELEC 2002 International Conference on Parallel Computing in Electrical Engineering, Warsaw, September 2002.

SELF-RESTORATION AS SEU PROTECTION MECHANISM

33

7. XAPP 216 v 1.0. Correcting Single Event Upsets through Virtex Partial Configuration, Xilinx Inc., 1 June, 2000. 8. Carmichael, C. (2001) Triple Module Redundancy Design Techniques for Virtex FPGAs, Xilinx Application Note XAPP197 (v 1.0), November 2001. 9. Abramovici, M., Breuer, M., and Friedman, A. (1994) Digital Systems Testing and Testable Design, Wiley-IEEE Computer Society Press.

SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS ON RADIATION DAMAGE OF METHYL SILICONE RUBBER

LIXIN ZHANG, SHIYU HE, DEZHUANG YANG, AND QIANG WEI Space Materials and Environment Engineering Laboratory, Harbin Institute of Technology, Harbin 150001, P. R. China

Abstract. The synergistic effect of protons and electrons on radiation damage of methyl silicone rubber was investigated using a ground-based simulator for space radiation environment. The energy of protons and electrons was chosen as 180 keV, and the fluence 1016 cm−2 . Experiment results showed that discharged patterns and aged cracks appeared on the surface of the irradiated samples. The tensile strength and electrical insulation properties were degraded, and the mass loss rate increased after the combined irradiation of protons and electrons. The influence on properties of the silicone rubber induced by the combined irradiation was greater than the separate irradiation of protons or electrons. The synergistic effect of protons and electrons did not show additivity. Under the combined irradiation, degradation effect of the silicone rubber occurred, and the free radicals formed by the breakage of molecular chains due to the combined irradiation were the primary cause for the change in electrical properties.

1. Introduction Once a spacecraft is deployed in orbit, it will experience the complicated space environment. Depending on the orbit, the spacecraft experiences various space environment factors separately, sequentially and simultaneously. Whether there is an additivity for the impact on material properties induced by the combined space environment is an uncertain problem [1–3]. In recent years, the silicone rubber is widely applied in spacecraft because of its excellent properties such as electrical insulation and the resistance to radiation. It was used as a binder between the cover glass and solar cells as well as between the cells and baseplate. Also, the silicone rubber can be used as the sealant in spacecraft [4, 5]. Under the charged particle radiation in space environment, the silicone rubber would be aged, resulting in degradation of properties, which directly influences the reliability and lifetime of spacecraft [6, 7]. In this paper, the effect of combined radiation of protons and electrons on the methyl-silicone rubber was investigated. The following formula 35 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 35–41. C 2006 Springer. Printed in the Netherlands.

36

LIXIN ZHANG ET AL.

is introduced to discuss the additivity for the combined radiation effect of protons and electrons: K add =

P(p+e) Pp + Pe

(1)

where, K add is the additivity coefficient; P(p+e) is the property change due to combined radiation of protons and electrons; Pp is the property change due to proton radiation; Pe is the property change due to electron radiation. It can be deduced that there is an additivity for the influence of the combined radiation on material property when K add = 1, while the additiivity does not exist when K add = 1.

2. Experimental The silicone rubber with molecular mass of 68,000 was used as the experimental material, which was treated in vacuum, and then incorporated with 2–3 wt% Si(OH)4 and 3–5 wt% Be2 Sn(OCOC11 H23 )2 . After mixing homogeneously, the material was degassed for several minutes in vacuum, poured into a tetrafluoroethylene mould, and then vulcanized at room temperature. The irradiation experiment was done in a system that can simultaneously and separately simulate the radiation of electrons and protons with 30–200 keV. In the test, the energy of protons and electrons was chosen as 180 keV, the fluence of protons and electrons was 1016 cm−2 , the vacuum in the sample chamber was 10−5 Pa, and the surface temperatures of irradiated samples were controlled at 10 ± 5◦ C by a cold screen with liquid nitrogen. Sample thickness was 2 mm. The sample surface morphology was evaluated using an OLYMPUS BH2UMA type optical microscope. The samples were coated with aluminium in vacuum for measuring the volume resistivity ρv , the dielectric coefficient ε and the loss angle tangent tg δ. The ρv was measured using a ZC-36 type meter. The tests for ε and tg δ were conducted using a TR-10C transformer bridge. The samples were analyzed by a JES-FE3AX type electron spin resonance spectrometer made by JEOL, Japan.

3. Results and Discussion 3.1. SURFACE MORPHOLOGY

Figure 1(a) shows that the sample surface is smooth and flawless. After proton irradiation, the surface colour of the silicone rubber darkened and aging cracks

SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS

(a)

0.2mm

0.2mm

(b)

0.2mm

(c)

(d)

37

0.2mm

Figure 1. Change in surface morphology of silicone rubber before and after irradiation: (a) before irradiation, (b) p = 1016 cm−2 , (c) e = 1016 cm−2 , (d) e = p = 1016 cm−2

appeared, as shown in figure 1(b). Our previous work demonstrated that the aging degradation of the silicone rubber surface occurred under the protons irradiation with 180 keV to a fluence of 1016 cm−2 . After the electron irradiation, the surface colour of the silicone rubber was also darkened, and the arborescent discharge stripes appeared on the surface, as seen in figure 1(c). Measurement of the surface potential showed that the silicone rubber surface was negatively charged to 1–2 kV under the electron irradiation. It is believed that the discharge is a main mode of surface damage under the electron irradiation. After the combined irradiation by protons and electrons, the surface colour of the silicone rubber was darkened further and aging cracks and more arborescent discharge stripes were observed at the same time, as shown in figure 1(d). There is a combined effect of the surface damage modes for the two types of charged particle irradiations after the combined radiation. 3.2. CHANGE IN PROPERTIES

3.2.1. Mass loss Table 1 indicates that the mass loss ratios of silicone rubber are 1.48, 1.06, and 2.99% after the electron, proton, and combined irradiation respectively. The mass TABLE 1. The change in mass loss ratios and tensile fracture strength under different irradiation conditions. Irradiation condition Before irradiation After protons irradiation After electrons irradiation After combined irradiation K add

mg

m/m %

σf MPa

σ/σ %

5.745 5.660 5.684 5.573 —

— −1.48 −1.06 −2.99 −1.18

4.83 4.41 4.48 3.86 —

— −8.69 −7.25 −20.1 −1.26

38

LIXIN ZHANG ET AL.

70

73

60 50

a.u.

40

57

30 20

2 1

10 0

16 15 14

0

10

20

30

40

50

60

70

80

90

Mass to charge ratio, m/z Figure 2. Mass spectrum of atmosphere in the sample chamber during combined irradiation of protons and electrons for silicone rubber

loss ratio for the combined irradiation is the greatest; the second for the proton and the lowest for the electron irradiation. However, the additivity does not exist, because K add = 1.18. The in situ mass-spectrometric analysis showed that the micromolecule gaseous products such as H2 (m/z:2.1, 1.1) and CH4 (m/z:16.2, 15.2, 14.2) were formed during the proton, the electron and the combined irradiations [6, 7]. Also, the gaseous product CH3 SiOCH3 (m/z:73.3, 57.3) was found during both the proton and the combined irradiations, as shown in figure 2. This could be attributed to the protons participating in the destruction of macromolecular chains of silicone rubber. 3.2.2. Tensile strength Table 1 indicates that the change in tensile fracture strength of the silicone rubber is 8.69, 7.25, and 20.1% after the electron, the proton, and the combined irradiation, respectively. The change in tensile fracture strength of the silicone rubber induced by the combined irradiation is larger than those for the independent irradiations of protons and electrons. However, the additivity does not exist, because K add = 1.26. The aging cracks and arborescent discharge stripes might be primarily responsible for the decrease of tensile fracture strength.

SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS

39

TABLE 2. Changes in electric properties of silicone rubber under different irradiation conditions Irradiation condition

ρv × 10−13 cm

ρv /ρv %

ε

ε/ε%

tg δ ×103

tg δ/ tg δ %

1.54 1.44 1.42 1.17 —

— −6.49 −7.79 −24.0 −1.68

3.13 3.21 3.46 3.77 —

— 2.56 10.54 20.44 1.56

3.53 3.74 3.71 4.06 —

— 5.94 5.10 15.01 1.36

Before irradiation After protons irradiation After electrons irradiation After combined irradiation K add

3.3. ELECTRIC PROPERTIES

Table 2 shows the changes in volume resistivity ρv , dielectric coefficient ε and dielectric loss tangent tg δ for the silicone rubber after the proton, the electron, and the combined irradiation. After the irradiation, the ρv decreases, while the ε and tg δ increase. Usually, with increasing the molecular dipole moment, both ε and tg δ increase. The appearance of free radicals after the irradiation would enhance the dipole polarization. Also, the radiations could induce the degradation of macromolecular chains of the silicone rubber, decreasing the intermolecular forces. As a result, the orientation of polar groups becomes easier. Both factors above might contribute to the increase of the ε and tg δ. The change in the electric properties induced by the combined irradiation is much greater than the independent irradiations. There is not an additivity for the change in the ρv , ε, and tg δ. The K add values are 1.68, 1.56, and 1.36, respectively. 3.4. ELECTRON SPIN RESONANCE SPECTRA

Figure 3 presents the electron spin resonance (ESR) spectra for the silicone rubber after the proton, the electron and the combined irradiation. In the figure, the g value g=2.0036 g=2.0036

g=2.0036

50G (a)

(b)

(c)

Figure 3. ESR spectra for silicone rubber after the proton, electron and combined irradiation: (a) Φp = 1016 cm−2 , (b) Φe = 1016 cm−2 , (c) Φe = Φp = 1016 cm−2

40

LIXIN ZHANG ET AL.

was equal to 2.0036, which represents the overlapped peak for the free radicals of a a !Si · and !SiO ·. The fourfold split peak for the CH3 · and the peak from ternary split for the a !SiCH2 · do not appear, because of faster attenuation rates of these free radicals. Some macrochains of the silicone rubber could break down due to radiation. After the irradiations, these free radicals attenuate automatically. Figure 3 shows that the peak for the combined irradiation is the highest, implying that the concentration of surviving free radicals is much higher than those for the independent irradiation. Therefore, the decrease of the ρv is more obvious after the combined irradiation. 3.5. INFRARED SPECTRUM ANALYSIS

The changes in the infrared spectrum of methyl silicone rubber before and after the proton, the electron, and the combined irradiation are shown in figure 4. Although no new peaks appear, the relative intensity of the peaks changes after the irradiations. The peaks at the wave numbers of 800 cm−1 and 1259 cm−1 originate from the vibration of CH3 groups, and the overlapped peaks around the 800 cm−1 might be related to the nonplanar vibration absorption of C–H bonds. The absorption peaks at 2962 cm−1 and 2900 cm−1 are due to the stretch vibration of C–H bonds, and the 1070 cm−1 absorption peak was caused by the Si–O stretching vibration in the Si-O-Si structure. As seen in the figure, the height of all the peaks decreases after the proton, the electrons and the combined irradiation, showing that the degradation occurs for both the macromolecule main chain Si–O bonds and the lateral chains CH3 . Moreover, the decreasing extent for all the peaks is larger for the combined irradiation, indicating that the combined one would lead to more severe degradation effect on the methyl silicone rubber.

Figure 4. Changes in infrared spectrum of silicone rubber before and after the proton, electron, and combined irradiation: (a) before irradiation, (b) Φp = 1016 cm−2 , (c) Φe = 1016 cm−2 , (d) Φe = Φp = 1016 cm−2 , (e) Φe = Φp = 2 × 1016 cm−2

SYNERGISTIC EFFECT OF PROTONS AND ELECTRONS

41

4. Conclusions The above results show that after the combined irradiation of protons and electrons, the changes in the surface morphology for the silicone rubber show a combined feature of the two independent radiation effects. The changes of tensile strength, mass loss ratio, and electrical properties are not following simple additivity. The combined radiation results in a more severe degradation effect on the methyl silicone rubber, which is primarily responsible for the decrease in mechanical properties and the increase in mass loss ratios. Under the combined radiation, the formation of free radicals, due to the scission of molecular chains, leads to the decrease in electrical properties of the methyl-silicone rubber.

References 1. Denkins, P., Badhwar, G., and Obot, V. (2001) Acta Astronautica 49(3), 313–319. 2. Stevenson, I., David, L., Gauthier, C., Arambourg L., Davenas J., Vigier G. (2001) Polymer 42, 9287–9292. 3. Badhwar, G. D. and Oneill, P. M. (2001) Nuclear Instruments and Methods in Physics A 466, 464–474. 4. Yibing, Z., Chunxiao, X., Wenyan, L., and Weiquan, F. (2003) European Space Agency, (Special Publication) ESA SP 540, pp. 723–725 5. Chang-Su, H., Bok-Hee, Y., and Sang-Youb, L. (2000) In Proceedings of the IEEE International Conference on Properties and Applications of Dielectric Materials, Xian, China Vol. 1, 2000, pp. 367–370. 6. Zhang, L. X., Yang, S. Q., and He, S. Y. (2003) Chinese Journal of Polymer Science 21(5), 563–567. 7. Zhang, L. X. and He, S. Y. (2004) Materials Chemistry and Physics 83, 255–259.

INFLUENCE OF ELECTRON RADIATION ON OUTGASSING OF SPACECRAFT MATERIALS R. H. KHASSANCHINE,1 A. N. TIMOFEEV,1 A. N. GALYGIN,1 V. I. KOSTIUK,1 AND V. M. TSVELEV2 1 Joint-stock Company “Kompozit,” 4, Pionerskaya street, Korolev, Moscow region, Russia 2 Lavochkin Association, Moscow, Russia

Abstract. In this work mathematical models describing the influence of electron radiation on outgassing of spacecraft materials and condensation of generated volatile products are given. Data of experimental investigations of outgassing kinetics of materials irradiated with different electron fluxes and results of numerical analysis of the processes are presented. Key words: outgassing, electron radiation, desorption, diffusion, remission 1. Introduction Outgassing of the materials intended for application on outer spacecraft surfaces is the major source of volatile products (VP) that are able to condense on contamination-sensitive surfaces. Outgassing of polymeric materials being subjected to electron radiation is the result of their radiochemical decomposition and thermal desorption of generated and existing VP. Contamination by deposition of outgassing products is caused by condensation of high-molecular VP and polymerization of low-molecular VP on spacecraft surfaces. The factors mentioned below influence the radiation-enhanced adsorption, desorption, and diffusion, which are the basic mechanisms for VP outgassing and condensation processes that occur in the near-surface layer of a VP source and on condensation surface, of individual outgassing components: 1. Radiolysis of organic components in material and changes of VP diffusion, desorption, and adsorption coefficients; 2. Various radiation-induced defects generated on VP condensation surfaces under the action of ionizing radiation; 3. Structural changes in the near-surface layers, both for the VP sources (for example, loosening) and for condensation surfaces; and 4. Generation of electric fields due to accumulation of volume charge in materials. 43 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 43–50. C 2006 Springer. Printed in the Netherlands.

44

R. H. KHASSANCHINE ET AL.

Usually, absorbed dose in the near-surface layer of spacecraft materials is determined by the low-energy part of spectrum of space ionizing radiations. The profile of radiation load, to which the spacecraft external surfaces are subjected, can be simulated in laboratory by exposing them to electron fluxes with different energies.

2. Volatile Products Outgassing and Condensation Model From mathematical models discussed in [1–3] it follows that change of concentration of i-type outgassing component in VP source under the exposure to electron radiation can be described with the help of the differential equation: M ∂Ci (x, t) ∂ 2 Ci (x, t) rad rad rad = Di − σi→ j C i (x, t) − χi C i (x, t) + Si (x, t) ∂t ∂x2 j=1

(1)

for x ∈ (0, h − υt), t > 0, υt < h, that satisfy the initial condition Ci (x, t)|t=0 = Ri

x ∈ [0, h]

(2)

and boundary conditions ∂Ci (x, t) ∂Ci (x, t) rad Di +(ki + ai ) Ci (x, t)|x=h−υt = Di =0 t > 0 ∂ x x=h−υt ∂ x x=0 (3) rad where σi→ —weighting coefficient of radiation-enhanced destruction of i-type j VP through j-channel, s−1 ; χirad —chemical reaction rates with involvement of i-type VP, s−1 ; Di —effective diffusion coefficient of i-type VP, μm2 ·s−1 ; Ri —concentration of i-type VP in material at initial moment of time, g·μm−3 × 10−12 ; ki —effective desorption coefficient of i-type VP, μm·s−1 ; h—thickness of sample, μm; υ—evaporation rate of specified material, μm·s−1 ; airad —parameter defining influence of electron radiation on i-type VP desorption kinetics, μm·s−1 ; Sirad (x, t) – i-type VP source function that is determined by influence of electron radiation on material. To determine the form of the source function, we have made calculations, using the Monte Carlo method, of basic characteristics concerning interaction of electron radiation with materials under study. Figure 1 shows curves of distributions of absorbed energies in ECOM-1 polymer-based white conductive paint

45

INFLUENCE OF ELECTRON RADIATION 40

a

Ea (keV/electron)

30

2 3

20

E0=20; 30; 40; 50 keV 4

10

0

b

0.1

P (Gy*s −1)

1

0.01

1E-3

0

1

2

3

x (mg/sm2)

4

5

0

20

40

60

80

100

x (μm)

Figure 1. Distribution of the absorbed energy in ECOM-1: (a) for various values of electron energies in laboratory experiments: 1–20 keV, 2–30 keV, 3–40 keV, 4–50 keV; (b) in material that is in service in GSO

[4]. Curve 1(a) gives a distribution of the absorbed energy for various values of electron energies in laboratory experiments, curve 1(b) gives a rough distribution of the absorbed dose of electron radiation when material is in service in GEO. It was found from the computational results of absorbed energy of electron radiation in spacecraft coatings when in service that the VP source function can be represented in the form Sirad (x, t)i = Ai0 exp[α0 (x + υt − h)] + Ai1 exp[α1 (x + υt − h)]

(4)

where Ai0 , Ai1 —parameters depending on composition of material and electron radiation spectrum. They are equal in values to the ratio of absorbed dose rate in the near-surface layer of material to average energy required to generate an i-type VP molecule in material being under exposure to low-energy and high-energy spectrum parts of space electron radiation respectively; α 0 , α 1 —effective coefficients of linear reduction of absorbed energy in spacecraft coatings from low-energy and high-energy spectrum parts of space electron radiation respectively. Having found the distribution Ci (x, t) of i-type VP in a sample from (1– 3), dependences of VP mass in material and the flux through the unit surface of material-vacuum boundary on time t are respectively determined from expressions h−vt

Msi (t) = S0

Ci (x, t)dx

dFi (t)/dt = υ + ki − airad Ci (h + υt, t)

(5)

0

where Ci (h – υt, t)—concentration of i-type VP in the near-surface layer of material at time t; S0 —surface area of a sample of material (source).

46

R. H. KHASSANCHINE ET AL.

When studying outgassing kinetics of materials in on-board or laboratory experiments, the total VP flux Ftotal (t) = Fi (t) is usually observed. Part of this flux is deposited on a sensitive surface of the quartz microbalance that is intended to study outgassing kinetics of polymeric materials. Deposition rate depends on composition and energies of molecular flux components as well as on surface condition, temperature, and external factors. The rate of change of i-type VP mass Msi (t) in a sample, and the mass Mci (t) deposited onto a unit area of condensation surface by time t are connected by the following system of equations: dMsi (t) = −S0 υ + ki − airad Ci (h − υ t, t) dt dMci (t) dMsi (t) = −αcs − kci Mci (t) − χci Mci (t) dt dt

(6) (7)

where kci —effective remission coefficient of i-type VP from the condensation surface; χ ci —chemical reaction rates with involvement of i-type component on condensation surface; α cs —geometrical factor that depends on arrangement of VP source with respect to the condensation surface. The outgassing rate of the VP source and deposition and remission rates on/from the condensation surface respectively are defined mainly by the constants in (6) and (7).

3. Results of Numerical Analysis When calculating, the parameters were chosen in such a way that results would clearly show the outgassing processes described by the model. Figure 2 shows influence of electron radiation on distribution of VP concentration in material at different model parameters. Plots 2(a) and 2(b) show changes of VP concentration in material at equal values of D, k, β, and temperatures in material under thermal-vacuum action and when the thermal action is overlapped with the electron radiation influence respectively (β i —effective first-order reaction rate with i-component. βi =

M

rad rad σi→ j + χi

j

Here it is seen that in the second case VP concentration in material becomes greater with time. This is due to radiation-enhanced generation of VP in material. Plot 2(d) is a distribution of VP concentration generated in a material by electron

47

INFLUENCE OF ELECTRON RADIATION

C(x, t), rel. units

C(x, t), rel. units

α0 =0.1; A0 =0.01; α 1=0.01; A1=0.002

t (h

)

x

) ( μm

)

) μm

α0 =0.1; A0 =0.01; α 1=0.01; A1=0.002

α0 =0.1; A0 =0.01; α 1=-0.01; A1=-0.002 0.20

C(x, t), rel. units

1.0

C(x, t), rel. units

x(

t (h

0.8

0.15

0.6 0.4

0.10 0.05

0.2 0.0

) μm x(

t (h

)

t(

h)

x (μ

m)

Figure 2. Distribution of VP in polymeric composite (υ = 0, β = 1.E-7, α0 = 0.1, α1 = 0.01, k = 0.01, D = 0.001): (a) Ai0 = Ai1 = 0, R = 1.0; (b) Ai0 = 0.01, Ai1 = 0.002, R = 1.0; (c) Ai0 = 0.01, Ai1 = −0.002, R = 1; (d) Ai0 = 0.01, Ai1 = 0.002, R = 0

radiation when there was no such VP in initial material (R = 0), plot 2(c) gives the eventual case when radiation is responsible for the concentrations of some types of VP to be reduced. Figure 3 shows functions of mass loss δMs (t) and outgassing rate dF(t)/dt from the unit surface of VP source under thermal vacuum action and influence of electron radiation at various parameter values R = 1 (figure 3(a)), R = 0 (figure 3(b)) with other model parameters being equal. Outgassing rate is maximum at initial times in the first case and it grows up starting from zero value in the second case. The mass loss function can be determined from δ Ms (t) = S0

N n=1

h−vt

(h Rn − exp(βn t)

Cn (x, t)dx) 0

where hS0 Rn —outgassing potential for n-type VP.

(8)

48

R. H. KHASSANCHINE ET AL.

10

dMs

1

dF/dt

b

dMsrad

R =0

δMs, dF /dt

δMs, dF /dt

a

dMsrad

R =1

dF/dt

rad

rad

dF/dt

0.1 0

20

40

t (h)

60

80

100

0

20

40

t (h)

60

80

100

Figure 3. Outgassing kinetics at different initial VP concentrations at electron radiation: (a) R = 1; (b) R = 0 (D = 0.001, k = 0.01, β = 10−6 , α0 = 0.1, α1 = 0.01, Ai0 = 0.01, Ai1 = 0.002)

The effect of electron radiation that influences the source material at different remission coefficients kci on VP deposition kinetics is given in figure 4 showing curves of VP mass deposited on the unit condensation surface Mci (t) by time t. For large values of kci , the deposited mass of the VP initially grows due to the significant outgassing rate, reaches maximum, and begins to decrease (figure 4(a)). The VP mass chemically reacted on the condensation surface can be calculated with the help of eqs. (6)–(8). Curves in figure 4(b) represent the mass loss Msi (t), the mass deposited onto a unit of condensation surface Mci (t), and the outgassing rate dF(t)/dt functions responding to a stepwise change of the model parameters resulted from electron radiation for t0 = 40 h. Electron irradiation with time leads to the growth of VP concentration in the near-surface layer due to radiolysis of material that sets off, to some extent, the fall of the outgassing rate dF(t)/dt and leads to increase of the mass loss δ Msi in VP source.

Figure 4. (a) VP deposition kinetics under thermal-vacuum action for different remission coefficients kc ; (b) responses of δ Msi (t), dF(t)/dt, and Mci (t) functions on electron radiation started at t0 = 40 h

INFLUENCE OF ELECTRON RADIATION

49

Figure 5. Outgassing kinetics of ECOM-1 irradiated by 10-keV electrons of different fluxes: (1) = 0; (2) = 5 × 1015 cm−2 , electron flux density 5 × 1011 cm−2 ·s−1 ; (3) = 5 × 1016 cm−2 , electron flux density 5 × 1011 cm−2 ·s−1 ; (4) = 5 × 1016 cm−2 , electron flux density 5 × 1012 cm−2 ·s−1

4. Experimental Results Figure 5 shows the experimental curves of outgassing kinetics in ECOM-1 samples irradiated by electrons of different fluxes. To show the effect of electron radiation on the process under study, figure 5 also gives the experimental curve of outgassing kinetics for the thermal-vacuum influence ( = 0). Comparing the curves, one can see that there is no linear dependence between amount of emerged VP and fluence of electron radiation to which the samples of VP sources were subjected. When analyzing experimental data obtained for materials irradiated with the same fluxes of electrons it was found that for electron flux densities exceeding 1012 cm2 ·s−1 the mass loss of irradiated VP source due to outgassing process decreases (see curves 3 and 4, figure 5). 5. Conclusion Results of numerical analysis of electron influence on outgassing kinetics show essential dependence of the process on absorbed dose distribution function and parameters of the mathematical model of outgassing. Experimental results for samples irradiated with different electron flux densities confirmed the hypothesis that outgassing rate of materials depends on both the spectrum and the radiation flux densities. Analysis of experimental data of outgassing kinetics of electron-irradiated polymeric materials shows that recommended electron flux density in such experiments should not exceed 5 × 1011 cm2 ·s−1 .

50

R. H. KHASSANCHINE ET AL.

References 1. Khassanchine, R. H., Grigorevskiy, A. V., and Gordeev, Y. P. (2002) In Protection of Materials and Structures from Space Environment, ICPMSE-6, Toronto, Canada, May 2002, 327–334. 2. Kostiuk, V. I. and Khassanchine, R. H. (2002) Cosmonautics and Rocket Engineering 28, 155– 163. 3. Khassanchine, R. H., Grigorevskiy, A. V., and Galygin, A. N. (2004) Journal of Spacecraft and Rockets 41(3), 384–388. 4. Grigorevskiy, A. V., Gordeev, Y. P., Gurov, A. A., Kiseleva, L. V., and Shuiski, M. B. et al. (2000) In Proceedings of 8th International Symposium, Materials in a Space Environment and 5th International Conference on Protection of Materials and Structures from Space Environment, Arcachon, France, 5–9 June 2000, 81–86.

EFFECT OF SURFACE CHARGING ON THE EROSION RATE OF POLYIMIDE UNDER 5 EV ATOMIC OXYGEN BEAM EXPOSURE MASAHITO TAGAWA∗ , SHINSUKE SEIKYU, KEN-ICHI MAEDA, KUMIKO YOKOTA∗ , AND NOBUO OHMAE Department of Mechanical Engineering, Faculty of Engineering, Kobe University, Rokko-dai 1-1, Nada, Kobe 657-8501, Japan

Abstract. The effect of charging on the atomic oxygen-induced erosion of polyimide was investigated. A polyimide sample was spin-coated on the quartz crystal microbalance (QCM), and mass loss of the film was directly measured from the resonant frequency shift of QCM during atomic oxygen exposure. The experiment was carried out using the specially designed QCM, which allowed a sample bias voltage up to 1500 V during resonant frequency measurements. From the experimental results, we confirmed that the erosion rate of polyimide at ±1500 V was almost identical to that of grounded. It was, thus, concluded that the polyimide erosion by atomic oxygen is hardly affected by surface charging. A similar conclusion was also obtained with polysulfone. This conclusion was not consistent with that reported by King et al. [5]. The discrepancy of the experimental results could be due to the surface ionization yield of reaction products. Key words: atomic oxygen, charging, low Earth orbit, synergy, polyimide

1. Introduction Pyromelliticdianhydride-oxydianiline (PMDA-ODA) polyimide (Kapton-H, DuPont) has been used as a reference material for atomic oxygen fluence measurement both in flight and in ground-based experiments. In order to maintain the accuracy of atomic oxygen fluence measurement in exposure tests, erosion properties of PMDA-ODA polyimide in various synergistic exposure conditions have to be clarified. We have studied the effect of ambient air exposure [1], temperature [2], incident angle [3], and ultraviolet exposure [4] on the atomic oxygen-induced erosion rate of PMDA-ODA polyimide. Following results were obtained from a series of experiments mentioned above; (1) The amount of oxygen adsorbed during atomic oxygen exposure would be higher than that analyzed after ambient air exposure, (2) Due to high impact energy of atomic oxygen, the activation energy of gasification reaction of polyimide is on the order of 10−4 eV and no temperature 51 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 51–59. C 2006 Springer. Printed in the Netherlands.

52

MASAHITO TAGAWA ET AL.

dependence of erosion is noticed, (3) Erosion rate of polyimide follows a cosine function with respect to incident angle of atomic oxygen which suggests reaction yield of atomic oxygen is independent of the incident angle, and (4) Simultaneous 172 nm ultraviolet exposure promotes erosion of polyimide up to 400% depending on the relative intensity of ultraviolet. The results of synergistic testing listed above showed that polyimide erosion is influenced by other environmental factors beside atomic oxygen. On the other hand, spacecraft charging has been recognized as a serious problem for electronic systems aboard spacecraft. It sometimes seriously damages spacecraft system due to discharge. However, effect of charging on material erosion has not been studied intensively. The only literature reported was the effect of electron beam irradiation in the atomic oxygen erosion of polysulfone which was reported by King and Wilson [5]. They found that electron beam irradiation or bias voltages applied to the back plate of polysulfone increased the signal of reactive products; CO and CO2 . They examined only the polysulfone. If surface charging also influenced the erosion rate of polyimide, surface charging phenomenon during flights or ground-based tests needs to be considered to provide an accurate fluence measurement of atomic oxygen. In this study, we examined the effect of surface charging on the atomic oxygeninduced erosion of polyimide. A quartz crystal microbalance (QCM) technique, which was established to study synergistic effect of atomic oxygen and ultraviolet on polymer erosion [1–4], was used to measure the erosion rate of polyimide under biased or electron beam exposed conditions. Effect of bias potential and electron beam exposure on the atomic oxygen-induced mass loss phenomenon of polyimide witness sample was analyzed and discussed.

2. Experimental Details 2.1. ATOMIC OXYGEN BEAM SOURCE

The atomic oxygen source used in this study was a laser detonation atomic oxygen source. This type of source was developed by Physical Sciences Incorporation [6]. Details of the source used in this study were described elsewhere [7]. The PSItype atomic oxygen source was attached to the space environment simulation facility at Kobe University (figure 1) [1–4]. The translational energy of atomic oxygen beam was monitored by the time-of-flight (TOF) measurement system consisting of a quadrupole mass spectrometer (QMS) and a multichannel scalar. Mean translational energy of the atomic oxygen was calculated to be 5.0 eV, which corresponds to the orbital impact velocity of atomic oxygen. The atomic oxygen flux in a beam was measured by an Ag-coated QCM with an accommodation coefficient of 0.62 [8]. The principle of measurement is explained in following section. Since the reaction of atomic oxygen with Ag is a nonlinear

EFFECT OF SURFACE CHARGING ON THE EROSION RATE

53

X-ray Source

DP-CMA Polymer-coated QCM Nozzle PSV

Laser QMS Electron gun

Au Orifice Mirror

Figure 1. Atomic oxygen beam facility used in this study. Laser detonation atomic oxygen source and high-energy electron gun are equipped

phenomenon, only the initial reaction, which gave a linear mass gain, was used to calculate atomic oxygen flux [9]. 2.2. EROSION MEASUREMENT

The erosion rate of the polyimide film was measured by the resonant frequency of QCM, which was coated by a thin film of polyimide. The change in mass of the polyimide film was recorded during atomic oxygen exposure. The resonance frequency of QCM is given by the following expression: f = − f 02 W/NAρ

(1)

where, N is the frequency constant, A is the electrode area, ρ is the density of quartz, and f 0 is the resonant frequency. Since N , A, ρ, f 0 are known factors, one can calculate the mass change (W ) of the sensor crystal from the frequency shift ( f ) of the QCM. Resonant frequency was measured every 10 s with accuracy of 0.1 Hz. For a 5 MHz AT-cut QCM sensor crystal used in this study, frequency resolution of 0.1 Hz corresponds to mass resolution of 2 ng. The QCM system used in the study was modified to apply bias voltages to the front surface of a sensor crystal. Since a conventional QCM surface was grounded in order to avoid ion sputtering effect during PECVD processes, an electrically isolated manifold was used to float the QCM from the grounded vacuum chamber. Bias voltage up to 1500 VDC was superimposed on to the driving voltage of QCM sensor crystal (DC 8V) and the QCM holder. However, the computer interface of

54

MASAHITO TAGAWA ET AL.

the QCM driver was maintained grounded. This electric isolation was achieved by a specially designed circuit. The PMDA-ODA polyimide film was formed on a QCM sensor crystal by a process described below. Precursor of PMDA-ODA polyimide (Semicofine SP510, Toray) was spin-coated on a QCM sensor crystal (12,000 rpm, 10 s). Twostage cure treatment (150◦ C, 1 h followed by 300◦ C, 1 h) in N2 atmosphere was carried out to create PMDA-ODA structure. The PMDA-ODA film, thus formed, has a thickness of approximately 0.1 μm, and XPS spectrum of the film was similar to that of commercially available polyimide (Kapton-H). Polysulfone film was also spin-coated with a solution involving polysulfone pellet (Mw = 63,000) and N, Ndimethylformamide as solvent. The spin-coated solution containing polysulfone was dried in air at 70◦ C 1 h. 3. Results 3.1. SYNERGISTIC EFFECT WITH ELECTRON BEAM

Figure 2 shows the experimental setup to study the effect of simultaneous electron beam irradiation on the atomic oxygen-induced erosion of polyimide. Electron beam (acceleration voltage: 7 kV, filament current: 2.2 A) irradiated the sample during atomic oxygen exposure. The experimental results for polyimide are shown

Figure 2. Photograph of the experimental set up for the e-beam experiment. A high-energy electron gun is installed at the top flange of the reaction chamber. QCM was temporally attached to the source chamber in order to measure atomic oxygen flux

EFFECT OF SURFACE CHARGING ON THE EROSION RATE

55

20

Frequency Shift (Hz/s)

18

Elapsed time vs AO exposure alone Elapsed time vs AO and e-beam (grounded) Elapsed time vs AO & e-beam (floated)

16 14 12 10 8 6 4 2 0

0

100

200

300

400

500

Elapsed Time (s)

Figure 3. Frequency shift of the polyimide-coated QCM under simultaneous exposure of atomic oxygen and electron beam

in figure 3. Open circles indicate the resonant frequency shift of polyimide-coated QCM under atomic oxygen exposure alone, open triangles and solid circles present data obtained under simultaneous atomic oxygen and electron beam exposures when a sample was grounded or floated, respectively. There is some scattering in the experimental data in figure 3; however, it is clear that no significant change in the slope of the lines occurred. The slopes of the lines calculated by a least square fit are listed in table 1. As listed in table 1, the slopes of the resonant frequency at three exposure conditions are distributed within an error of 3%. Since electron beam exposure alone did not affect the mass of the polyimide (figure 4), it was concluded from a series of experiments that simultaneous electron beam exposure does not affect the mass loss phenomenon of atomic oxygen-induced polyimide. Similar result was obtained also for polysulfone. 3.2. SYNERGISTIC EFFECT OF BIAS VOLTAGE

Bias voltages from 0 to 1500 V (positive and negative) were applied to polyimidecoated QCM, and the influence on atomic oxygen-induced mass loss phenomenon TABLE 1. Slope of the frequency shift of QCM under various exposure conditions of atomic oxygen and electron beam Experimental conditions AO exposure alone AO plus e-beam (grounded) AO plus e-beam (floated)

Slope (Hz·s−1 ) 3.7 × 10−2 3.8 × 10−2 3.8 × 10−2

56

MASAHITO TAGAWA ET AL. 8

Frequency shift (Hz)

6 4 2 0 -2 -4

Grounded Floated

-6 -8 0

50

100

150

200

250

Elapsed time (s)

Figure 4. Frequency shift of the polyimide-coated QCM under electron beam exposure. Note that atomic oxygen beam is turned off. No erosion is detected both for grounded and floated samples

was examined. Measurements were carried out during the increasing and decreasing phases of bias voltages and the results were averaged. Figure 5 shows the rate of resonant frequency change of polyimide-coated QCM during atomic oxygen exposures with bias voltages from 0 to 1500 V. Figures 5(a) and 5(b) indicate the results for negative and positive bias, respectively. It was clear that no significant effect of bias voltages on atomic oxygen-induced erosion was observed. We have also examined the polysulfone in the same experiment, however significant effect was not observed. From the QCM experiments reported herein, we confirmed that the bias voltage does not affect the reaction yield of atomic oxygen with polyimide. This experimental finding provided a conclusion that effect of charging on polyimide is not necessary to consider when polyimide is used as a witness sample for atomic oxygen fluence monitor.

4. Discussion The effect of electron beam irradiation and bias voltage on atomic oxygen-induced erosion of polysulfone was reported by King and Wilson [5]. They used a laser detonation atomic oxygen beam, which is the same type of atomic oxygen source used in this study, and detected the reactive products of CO and CO2 by quadrupole mass spectrometer during the experiment. They reported no significant effect of electron beam irradiation when spin-coated polysulfone sample was grounded, in contrast, significant increase in CO and CO2 signals were detected when the sample was electrically floated. Since similar effect was observed when sample was simply biased, they concluded that the increase in CO

EFFECT OF SURFACE CHARGING ON THE EROSION RATE

57

80

Frequency Shift (Hz)

(a) 60

40

0V −500 V −1000 V −1500 V

20

0

0

50

100

150

200

250

Elapsed Time (s)

Frequency Shift (Hz)

80

(b)

60

40 0V +500 V +1000 V +1500 V

20

0

0

50

100

150

200

250

Elapsed Time (s)

Figure 5. Frequency shift of the polyimide-coated QCM under oxygen beam exposure with bias voltages of 0–1500 VDC: (a) neagative bias and (b) positive bias. Bias voltage of 0 V means that the sample was grounded

and CO2 production yields are due to surface charging. Actually, they indicated that CO2 signal increased when bias voltage was applied to the back plate of the film. However, the experimental results obtained here were inconsistent with King’s report; i.e., mass loss phenomenon was not affected by electron beam irradiation nor by application of bias voltages up to 1500 V (positive and negative). The origin of this discrepancy would be explained by a surface ionization phenomenon. In the King’s experiment, they detected CO and CO2 ions. These ions are ionized in the electron bombardment ionizer at the QMS. Ionization cross section of molecules in the electron bombardment ionizer has not been evaluated, but is usually low (10−4 or even lower). On the other hand, ions originally contained in the beam are detected with a cross section of almost 1. King’s analysis, where CO and CO2 production yields become greater with bias voltage, is based on the

58

MASAHITO TAGAWA ET AL.

assumption that the amount of ions in the beam is unchanged by the application of bias voltages. One of the mechanisms that explains two inconsistent experimental results simultaneously is the effect of bias voltages onto surface ionization cross section. Namely, if the fraction of ions in the reactive products (CO and CO2 ) becomes greater with bias voltages, the experimental results reported by King and in this study can be explained. This is because the detection probability of ion is much higher than that of molecule with QMS. Measuring of the surface ionization cross section under application of bias voltages will verify the proposed mechanism. 5. Conclusion Effect of charging on the atomic oxygen-induced erosion of polyimide was investigated. Polyimide sample was spin-coated on the quartz crystal microbalance (QCM), and mass loss of the film was directly measured from the resonant frequency shift of QCM during atomic oxygen exposure. From the experimental results, we confirmed that the erosion rate of polyimide at 1500 V was almost identical to that at grounded. It was, thus, concluded that the polyimide erosion by atomic oxygen is hardly affected by surface charging. It was confirmed that the effect of surface charging does not have to be considered in atomic oxygen fluence measurements using Kapton witness sample. Similar conclusion was also obtained with polysulfone. This conclusion for polysulfone was inconsistent with that reported by King et al. The discrepancy would be explained by the surface ionization cross section of the reaction products. Acknowledgments A part of this work was supported by the grant-in-aid of scientific research contract no. 14350511 and 15560686 from the Ministry of Education, Culture, Sports, Science, and Technology, Japan; Kawanishi memorial Shinmaywa Education Foundation. References 1. Tagawa, M., Yokota, K., Ohmae, N., and Kinoshita, H. (2002) Journal of Spacecraft and Rockets 39(3), 447–451. 2. Yokota, K., Tagawa, M., and Ohmae, N. (2003) Journal of Spacecraft and Rockets 40(1), 143–144. 3. Yokota, K., Tagawa, M., and Ohmae, N. (2002) Journal of Spacecraft and Rockets 39(1), 155–156. 4. Yokota, K., Ohmae, N., and Tagawa, M. (2004) High Performance Polymers, Vol. 16, No. 2 (2004) pp. 221–234.

EFFECT OF SURFACE CHARGING ON THE EROSION RATE

59

5. King, T. and Wilson, W. (1997) “Synergistic effects of atomic oxygen with electrons,” A Bound Collection of Papers, AIAA Defence and Space Programs Conference and Exhibit, AIAA paper No. 97-3901, pp. 11–15. 6. Caledonia, G. E., Krech, R. H., and Green, D. B. (1987) AIAA Journal 25(1), 59–63. 7. Tagawa, M., Muromoto, M., Hachiue, S., Yokota, K., Ohmae, N., Matsumoto, K., and Suzuki, M. (2003) In Proceedings of the 10th European Space Mechanism and Tribology Conference, ESA SP-524, pp. 311–314. 8. Tagawa, M., Yokota, K., Kinoshita, H., and Ohmae, N. (2003) In Proceedings of the 9th International Symposium on Materials in a Space, Noordwijk, The Netherlands, ESA SP-540, pp. 247–252. 9. Kinoshita, H., Tagawa, M., Yokota, K., and Ohmae, N. (2001) High Performance Polymers 13(4), 225–234.

INFLUENCE OF SPACE ENVIRONMENT ON SPECTRAL OPTICAL PROPERTIES OF THERMAL CONTROL COATINGS V. M. PROSVIRIKOV,1 A. V. GRIGOREVSKIY,1 L. V. KISELEVA,1 A. P. ZELENKEVICH,1 AND V. M. TSVELEV2 1 Joint-stock Company “Kompozit,” Korolev, Moscow region, Russia 2 Lavochkin Association, Khimki, Russia

Abstract. Measurements of the hemispherical reflectance for some thermal control coatings developed by the joint-stock company “Kompozit,” within the spectral range 0.2–2.5 μm were carried out. Samples under investigation have been exposed to electron, proton, and UV irradiation separately and combined. Peculiarities in changes of spectral reflectance of thermal control coatings under single and multiple influence of space environment were studied. Bidirectional reflectance at grazing incidence angle studies revealed features of material and its surface condition after exposure to environmental factors. Key words: Thermal Control Coatings, Optical Properties, Space Environment

1. Introduction The solar radiation absorption coefficient (αs ) and emissivity (ε) are the major parameters governing spacecraft surface temperature and its serviceability [1, 2]. It is known that αs of white paint coatings increases whereas ε doesn’t change under the influence of environmental factors. The process of applying the paint coatings to surfaces of various shapes is simple enough. The joint-stock company “Kompozit” develops new paint coatings that can be used as thermal control coatings (TCC) for prolonged exposures. To evaluate the effects of the space environment, samples of coatings are subjected to accelerated tests, which include both separate and combined influence of environmental factors, with the changes of αs being determined. Using the data obtained in such tests an analysis of applicability of these coatings is carried on. Studies of changes in reflectance spectrum of a coating due to the effect of environmental factors are necessary to prove the legitimacy of simulation of 61 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 61–69. C 2006 Springer. Printed in the Netherlands.

62

V. M. PROSVIRIKOV ET AL.

radiation conditions, to study the damage processes in TCC, and to support the choice of coating degradation model that will be used in calculation of αs to predict the long-term serviceability of a TCC.

2. Samples under Study Samples of new ECOM-1 enamel developed in the joint-stock company “Kompozit” as an antistatic TCC of the “solar reflector” class were chosen for this study. This enamel is applied to the outer skin of spacecraft and launch vehicles to prevent accumulation of static electricity and to maintain the optical properties. ECOM-1 enamel is a blend of ZnO pigment and acrylic-based binder. To increase electric conductivity of the enamel, ZnO was doped through the surface treatment of powder with Ga-containing solution using the method of thermal diffusion. ECOM-1 enamel coating was deposited with the layers up to 120 μm in thickness onto substrates, diameter 30 mm, made of AMg6 aluminum alloy.

3. Equipment Samples were exposed to different environmental factors in the UV-vacuum integrated test simulator. The stand consists of a solar simulator based on 5000 W xenon lamp, electron and proton accelerators with energies up to 50 keV each. The KG-500 cascade generator was used for irradiating coatings with protons with energies 150–500 keV. The simulator allows long-term continuous tests of materials and coatings in vacuum not less than 10−6 mbar with the sample temperature on thermostabilizing table maintained within the prescribed limits. Spectral hemispherical reflectance of coatings is measured using the stateof-the-art U-4001 double-beam spectrophotometer manufactured by “Hitachi” with RSA-HI-40 integrating sphere, 150 mm in diameter, manufactured by “Labsphere” using the relative method. The calibrated spectralon reflectance standard was used as comparative sample. Bidirectional reflectance was measured using the automated goniospectrophotometer that covers the spectral measurement range (0.2–2.5 μm) and incidence angles and radiation record ranges (−90◦ –+90◦ ) taken from normal to sample surface. Angular resolution is ±1◦ while recording the intensity of reflected radiation. 4. The Research Approach Reflectance spectra of ECOM-1 samples were measured in air before and after they have been exposed to environmental factors, separately and combined. The

INFLUENCE OF SPACE ENVIRONMENT

63

coefficient αs was calculated using spectral values of reflectance coefficients according to standardized procedure [3] with the use of data on spectral distribution of solar radiation [4]. Kinetic dependences of changes of αs during the tests were measured using the FM-59 integral photometer adjusted for calculating values of α s . Thus, the reflectance spectra of samples contain only irreversible changes resulted from intensive continuous action of environmental factors simulating exposure to GEO environment. We used the method of recovering optical properties of a sample upon its transfer to air to measure the integral coefficient αs , based on estimates of how it changes in air with time, immediately after withdrawal. The first measurement of this dynamic dependence is taken within 2 min after the sample has come in contact with air. Changes of bidirectional reflectance of samples were also conducted in air before and after they have been exposed to 45-keV electrons and 40-keV protons.

5. Results from UV Exposure The ECOM-1 coating samples were illuminated by a xenon lamp that provides an accelerating factor of 5. The samples were irradiated to a total of 2000 equivalent solar hours (ESH), with measurements of αs performed at intermediate points. Figure 1 shows reflectance spectra of ECOM-1 before and after exposure. Main changes in ECOM-1 reflectance spectrum were in the shortwave range near the fundamental absorption line of ZnO (380 nm). Zone of irreversible changes ends at about 800 nm wavelengths. The trends of coefficient changes in IR range, after UV tests, followed those in the air bleaching process.

R, ECOM-1 before, αs=0.280 ECOM-1 after UV, αs=0.365

λ (nm)

Figure 1. Reflectance spectrum of ECOM-1 TCC before and after UV exposure (2000 ESH)

64

V. M. PROSVIRIKOV ET AL.

R,

ECOM-1 before, αs=0.280 ECOM-1 after (Ee-45 keV), αs=0.440

λ (nm)

Figure 2. ECOM-1 TCC reflectance spectra before and after irradiation with 45-keV electrons, flux density ϕ = 5.0 × 1011 cm−2 ·s−1 up to fluence F = 1.5 × 1017 cm−2

6. Results from Electron Exposure Irradiation of coatings was carried out with 45-keV electrons at flux densities of 5.0 × 1011 cm−2 ·s−1 up to the fluence of 1.5 × 1017 cm−2 . According to calculations, the electron penetration into samples under study is about 15 μm. While irradiating, αs was measured in intermediate points. ECOM-1 reflectance spectra before and after electron irradiation are given in figure 2. Changes in reflectance spectra of electron-irradiated sample are similar to results from UV irradiation and occur mostly in the shortwave range. The zone of irreversible changes of the reflectance coefficient ends at about 900 nm. Figure 3

Figure 3. Bidirectional reflectances of radiation (λ = 570 nm) from initial and postexposed ECOM1 TCC samples: E e = 45 keV, F = 1.5 × 1017 cm−2 for incidence angle 70◦ with respect to surface normal

INFLUENCE OF SPACE ENVIRONMENT

65

shows angular dependences, in relative units, of sample reflecting radiation with wavelength of λ = 570 nm before and after electron action. Samples after irradiation show a reduction of reflection peak intensity in near-mirror direction that seems to be connected with partial evaporation of the surface layers of the polymeric film in vacuum, which is typical.

7. Results from Proton Exposure Proton radiation produces maximum degradation of samples. Figure 4 shows ECOM-1 reflectance spectra as measured before and after proton radiation with energies within the range 40–500 keV for different fluences from 1.2 × 1016 to 2.6 × 1016 cm−2 at the same flux density of 1.0 × 1011 cm−2 ·s−1 . According to estimate [5], penetration of protons into samples under study is from 0.7 μm for 40-keV protons to 4.2 μm for 500-keV protons. While irradiating by protons, the defects originate in a thin near-surface layer. From the spectra one can see that the obtained reflectance curves are essentially different. Reflectance lines move to longer wavelength with higher proton energy and fluence. As the particle energy and/or flux increase other defects emerge in Zn and O sublattices, and their complexes that are observed at longer wavelengths as compared with the F+ -centers. The F+ -centers make the basic contribution to light absorption when exposed to low-flux protons. At these fluences for 40-keV, 150-keV, and 300-keV protons,

R,

ECOM-1 before, αs=0.280 ECOM-1 after (Ep=40 keV), αs=0.628 ECOM-1 after (Ep=150 keV), αs=0.591 ECOM-1 after (Ep=300 keV), αs=0.729 ECOM-1 after (Ep=500 keV), αs=0.609

λ (nm)

Figure 4. ECOM-1 TCC reflectance spectra before and after irradiation with protons of different energies (ϕ = 1.0 × 1011 cm−2 ·s−1 ): E p = 40 keV, F = 1.5 × 1016 cm−2 ; E p = 150 keV, F = 1.2 × 1016 cm−2 ; E p = 300 keV, F = 2.6 × 1016 cm−2 ; E p = 500 keV, F = 2.0 × 1016 cm−2

66

V. M. PROSVIRIKOV ET AL.

Figure 5. Bidirectional reflectances of radiation (λ = 570 nm) from the initial and proton exposed ECOM-1 TCC samples: E p = 40 keV, F = 1.5 × 1016 cm−2 for incidence angle 70◦ with respect to surface normal

the concentrations of F+ -centers located near the ZnO fundamental absorption line grow and reach maximum in the near-surface layers. Figure 4 shows that 500-keV protons induce less near-surface defects in the wavelength range below 500 nm. At the same time one can see intensive broadband absorption for 300-keV protons with integral flux of 2.6 × 1016 cm−2 (the lowest curve in figure 4). The kinetic curves showing the dependence of αs on proton fluences also reveal peculiarities related to activation and generation of other defects. Bidirectional reflectance of the 40-keV proton irradiated sample became more stretched out (figure 5) in the near-mirror direction. This takes place probably because the effective scattering layer of the sample has decreased in thickness as a result of the damage of the near-surface layer by protons.

8. Integrated Tests In addition to exposure to single environmental factors, we studied influence of combined action of environmental factors on optical properties of ECOM-1 coating. For this purpose, three modes of combined action were chosen specified below K1, K2, K3, and that included K1—protons with energy E p = 30 keV and flux density of 1011 cm−2 ·s−1 , electrons with energy E e = 10 keV and flux density of 1012 cm−2 ·s−1 ; K2—protons with energy E p = 30 keV and flux density of 1011 cm−2 ·s−1 , electrons with energy E e = 10 keV and flux density of 1012 cm−2 ·s−1 ; solar radiation with 2.7 ESH exposure; K3—protons with energy E p = 30 keV and flux density of 1012 cm−2 ·s−1 , electrons with energy Ee = 10 keV and flux density of 2 × 1011 cm−2 ·s−1 .

INFLUENCE OF SPACE ENVIRONMENT

67

R,

ECOM-1 before, αs=0.280 ECOM-1 after K1, αs=0.566 ECOM-1 after K2, αs=0.617 ECOM-1 after K3, αs=0.652

λ (nm)

Figure 6. ECOM-1 TCC reflectance spectra before and after integrated tests: K1—E p = 30 keV, Fp = 9.4 × 1015 cm−2 ; E e = 10 keV, Fe = 9.4 × 1016 cm−2 ; K2—E p = 30 keV, Fp = 1.1 × 1016 cm−2 ; E e = 10 keV, Fe = 2.8 × 1016 cm−2 , UV – 67.2 ESH; K3—E p = 30 keV, Fp = 2.4 × 1016 cm−2 ; E e = 10 keV, Fe = 3.6 × 1015 cm−2

Figure 6 shows reflectance spectra before and after integrated tests conducted at the above conditions. The trends in behavior of reflectance coefficients after tests conducted at K1, K2, and K3 modes are different. The most significant changes were in K3 irradiation mode up to maximum proton fluence of about 2.4 × 1016 cm−2 . It is seen from figure 7 that beginning with fluences near 1.5 × 1016 cm−2

Figure 7. Dependence of change of αs on proton fluences for three irradiation modes: K1, K2, and K3

68

V. M. PROSVIRIKOV ET AL.

Δαs

F(cm−2)

Figure 8. Dependence of αs on proton fluence for K2 mode irradiation of ECOM series TCC samples (UV = 2.7 ESR, E p = 30 keV, and Ee = 10 keV)

the curve representing dependence of αs on fluences becomes nonlinear and the rate of change of αs decreases. It should be noted that the coating degradation rate at proton flux of 1012 −2 −1 cm ·s (K3 mode) is noticeably lower than at 1011 cm−2 ·s−1 (K2, K3 modes). Samples of ECOM-1P and ECOM-2 presenting the ECOM series of coatings with filler modifications were subjected to K2 mode integrated tests together with ECOM-1. Figure 8 shows these experimental results. The curves in figures 7 and 8 were created using the method for recovery of the optical properties; the accuracy of αs varies from 0.5% for the values measured in initial measurement point to 3% in the end point.

9. Conclusion Exposure of ECOM-1 coating samples to environmental factors, separately and in combination, demonstrated different dependence of spectral reflectance coefficient on the exposure that is associated with various processes of defect generation. At initial stage of irradiation by electrons, protons, and UV irreversible changes take place in the shortwave area of reflectance spectrum near the absorption edge of ZnO. If proton fluences exceed 1016 cm−2 , spread of absorption bands in the longwave range takes place that is associated with generation of other defects

INFLUENCE OF SPACE ENVIRONMENT

69

and their complexes. Absorption lines are shifted to longwave range area with an increase of proton energy. The most significant irreversible changes of optical properties of ECOM-1 TCC are produced by protons. In this case the shape of bidirectional reflectance at sliding incidence angles of radiation changes appreciably. When simulating the space environment, one should take into account, by reason of nonadditivity, the influences of separate environmental factors and their combined action. ECOM-1P TCC is a prospective material showing good stability to environmental factors.

References 1. Heat Transfer and Spacecraft Thermal Control. (1974) Edited by John W. Lucas; Translated from English under edition of N.A. Anfimov. M.: Mir, pp. 22–28. 2. Gurevich, M. M., Itsko, E. F., and Seredenko, M. M. (1984) L. Chemistry, 120, pp. 79–85. 3. ASTM E 903-96, Standart Test Method for Solar Absorptance, Reflectance, and Transmittance of Materials Using Integrating Spheres, American Society for Testing and Materials, Philadelphia, USA. 4. ASTM E 490, Solar Constant and Air Mass Zero Solar Spectral Irradiance Tables. American Society for Testing and Materials, Philadelphia, USA. 5. Soloviev, G. G. (1979) Science-Exploration Institute of Technical, Economical Problems, NIITEChIM 9(159) pp. 1–39.

MITIGATION OF THRUSTER PLUME-INDUCED EROSION OF ISS SENSITIVE HARDWARE

COURTNEY PANKOP, JOHN ALRED, AND PAUL BOEDER The Boeing Company, 13100 Space Center Blvd., Houston, TX, 77059

Abstract. Optically sensitive surfaces on the International Space Station (ISS) can be damaged (or eroded/pitted) when impacted by high velocity unburned liquid propellant drops present in bipropellant thruster plumes. Surfaces with thin optical coatings, such as solar arrays and radiators, are of primary concern. Thruster plume-induced erosion/pitting of sensitive surfaces has been observed on Space Shuttle flight experiments. The Boeing ISS Environments Team in Houston has developed an approach to modeling thruster plume-induced erosion/pitting of ISS surface materials. The Boeing team has conducted analyses simulating bipropellant thruster droplets impacting ISS sensitive surfaces for various assembly stages. Thruster firings for ISS Reboost/Attitude Control as well as Visiting Vehicle thruster firings during approach or separation to ISS docking ports were simulated. The results of these analyses show that particle impingement angle greatly affects surface damage, with normal impacts being the most severe. Particles with highly oblique impact angles (∼75◦ off normal), however, will essentially skid off surfaces without causing any erosion/pitting. A mitigation technique has been developed to prevent plume erosion/pitting of solar array coatings. Prior to a thruster firing event, solar arrays may be rotated to a preestablished position that will eliminate plume particle impact damage to the surface. The preestablished positions are defined based on the geometry of the ISS thrusters to the solar array panels to ensure that plume particles will impinge at highly oblique angles (greater than 75◦ off normal). Operational constraints for plume erosion mitigation are being coordinated with other solar array operational constraints such as power, thermal, and plume-induced structural loads. An integrated operational solution is being implemented to support the ISS assembly flight sequence. This paper will discuss the plume erosion analyses and the implementation of operational mitigation. Key words: ISS, Optical Surfaces, Erosion Modelling

71 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 71–85. C 2006 Springer. Printed in the Netherlands.

72

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

1. Introduction During International Space Station (ISS) operations, a variety of rocket engines are used by the Space Station and its visiting vehicles (Space Shuttle Orbiter, Soyuz, Progress, etc.) for attitude control, orbit reboost, and docking/separation. These various engines are bipropellant thrusters using hypergolic components, either monomethyl hydrazine (MMH) or unsymmetrical dimethyl hydrazine (UDMH) as the fuel and nitrogen tetroxide (N2 O4 , or NTO) as the oxidizer. The exhaust plumes (see figure 1) from these engines have been recognized as a potential source of loads, heating, and contamination. Laboratory studies have revealed the presence of unburned propellant in the exhaust plume in the form of liquid particles [1, 2]. The origin of the particles, which can range from 1–100 microns in diameter, is commonly attributed to the incomplete combustion. Although the induced external contamination from the exhaust plume can come from the gaseous constituents, ISS sensitive surfaces do not operate at temperatures low enough to promote condensation of the gas phase. Hence, ISS plume-induced contamination is primarily due to drops of unburned propellant in the plume. The gases in the exhaust plume accelerate these propellant drops to high velocities (1–3 km·s−1 ) due to gas drag forces [1, 2, 9]. The effect of these high-velocity drops impacting onto ISS sensitive surfaces, such as the solar arrays and active radiators, is akin to the impact of micrometeoroid and orbital debris (MM/OD) particles. Given that the flux of drops in thruster plumes is much larger than the flux of MM/OD particles of comparable diameter [3] (see figure 2), this erosion/pitting effect is of great concern to the ISS program.

Figure 1. Exhaust plume of an orbiter primary reaction control system thruster

73

COURTNEY PANKOP ET AL. Number Density Comparison: Orbiter Primary Reaction Control System (PRCS) Engine Centerline vs. MM/OD Flux 1.00E+07 1.00E+06 1.00E+05

PRCS Centerline Liquid Droplets (20m from nozzle)

1.00E+04

MM/OD

1.00E+03 Number Density (#/cm^2-sec)

1.00E+02 1.00E+01 1.00E+00 1.00E-01 1.00E-02 1.00E-03 1.00E-04 1.00E-05 1.00E-06 1.00E-07 1.00E-08 1.00E-09 1.00E-10 1.00E-11 0

10

20

30

40

50

60

70

80

90

100

Particle Diameter (um)

Figure 2. Orbital debris environment

Three space flight experiments, which studied exhaust plume-induced contamination, are the shuttle plume impingement experiment (SPIE) on STS-52, the shuttle plume impingement flight experiment (SPIFEX) on STS-64, and the plume impingement contamination experiment (PIC) on STS-74, a mission to the Mir space station, which studied plume contamination from both American and Russian thrusters. Both SPIFEX and PIC demonstrated plume contamination and pitting from plume particles [4, 5]. A SPIFEX aluminum witness coupon, which was plumed by the space shuttle reaction control system thrusters, is shown in figure 3. A postflight examination of a glass camera lens on the PIC experiment also revealed impact craters on the surface [6]. An example of these impact craters is shown in figure 4. It should be noted that the craters on the SPIFEX and PIC samples were not visible with the unaided eye. Surface pits were observed using a scanning electron microscope.

2. Implication for ISS For many optically sensitive surfaces, special coatings are applied to enhance performance or for environmental protection. For ISS sensitive surfaces, mechanical

74

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

Figure 3. SPIFEX aluminum witness coupon

damage from thruster plume particle impacts has significant implications. The Boeing ISS Environments Team began to examine the influence of thruster firings on the degradation of ISS sensitive surfaces. Two sensitive surfaces that may see a great number of thruster firings during ISS operations are the U.S. photovoltaic (PV) solar arrays and the active thermal control system (ATCS) radiators. Most ATCS radiators, which provide heat rejection for the core modules, are coated with Z-93, a ceramic thermal control paint. Z-93 consists of a matrix of zinc oxide (ZnO) dispersed in a potassium silicate binder (K2 Ox SiO2 with x = 3–5, commonly referred to as waterglass). The minimum specified Z-93 paint thickness for the ISS ATCS panels is 5 mils. A typical application of Z-93 is shown in figure 5.

Figure 4. PIC glass camera lens

75

COURTNEY PANKOP ET AL.

Figure 5. Z-93 paint at magnification of 100x

A minimum thickness of 4 mils is necessary for Z-93 to maintain its optical properties, as shown in figure 6 [7]. Therefore, 1 mil of the Z-93 coating could be lost to thruster plume erosion before the ATCS radiator’s optical properties begin to degrade. Each ISS solar array wing is composed of 32,800 solar cells covering a deployed area of 115 ft long and 38 ft wide. The solar cells are mounted on a flexible backing of scrim cloth and Kapton. Each solar cell is topped with a CMX cover glass. The CMX cover glass is coated with an Ultraviolet Energy (UVE) filter coating for reflecting UV energy in the wavelength region below 350 nm. The UVE coating thickness is specified at 4.33 μm. The concern for thruster plume erosion is this UVE coating. The Kapton backing of the solar array wing is susceptible to atomic oxygen ˚ AO-resistant SiOx coating is (AO) erosion. To prevent such erosion, a 1300 A

Absorptance

0.4

0.3

0.2

0.1 0

2

4

6

Z-93 Thickness (mil)

Figure 6. Z-93 absorptance versus thickness

8

76

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

applied to the Kapton. While the SiOx coating performs no optical function, its loss due to thruster plume erosion would allow AO erosion of the Kapton.

3. Thruster Plume Erosion Model The Boeing ISS Environments Team developed an approach to modeling the mechanical erosion on surfaces due to the impact of particles in thruster plumes. A Los Alamos National Laboratory Smooth Particle Hydrodynamics (SPH) code, called SPHINX, was used to simulate thruster plume particle impact damage. SPHINX has been tested and verified on a number of projectile impacts and astrophysical problems [8]. In addition, the Boeing ISS Environments Team conducted ground tests to verify SPHINX in the regime utilized for thruster particle impact simulation. The results of this ground testing are being prepared for future publication. Boeing conducted parametric studies varying plume particle drop size, impingement angle, and velocity via SPHINX simulations. Analysis of the SPHINX output yielded a damage matrix for a specified impacted surface material. The damage matrix information was coded inside Boeing’s NASAN-II contamination computer model [9]. NASAN-II is an integrated computer model, utilizing inputted NASTRAN geometric models, view factor calculations, and transport routines to assess a given thruster’s effect on an ISS configuration, with results available in tabular or graphics formats. One result found during the parametric studies was that particle impacts at angles greater than 75◦ with respect to the normal produce no damage. A synopsis of these analyses is shown in figure 7. The application of this result to plume erosion mitigation will be discussed in section 5.

4. Plume Erosion Analysis All ISS vehicles which utilize thrusters must be considered as plume erosion sources. The NASAN-II thruster plume erosion model was used to assess the impact of thruster firing operations on the changing ISS configurations as radiators and solar arrays are relocated and deployed. Results of these analyses were used to determine if plume erosion mitigation would be required. 4.1. ATCS RADIATOR ANALYSIS

NASAN-II was used to conduct a parametric study of the thruster plume pitting effect on the ATCS radiators during nominal ISS operations. The analyses showed that the ATCS radiators’ Z-93 coating would be eroded but the total depth of

COURTNEY PANKOP ET AL.

77

Figure 7. SPHINX results for (a) 0◦ , (b) 45◦ , and (c) 75◦ impingement angles

erosion during nominal operations would be well under 1.0 mil. Since the loss of 1.0 mil of Z-93 does not compromise the heat rejection optical properties (per figure 6), the conclusion was made that the ATCS radiators were not adversely impacted by thruster plume erosion. Analysis of Z-93 ATCS radiators raised concern for ATCS radiators coated in Silver Teflon (AgFEP). To prevent Atomic Oxygen (AO) erosion, AgFEP radiators on ISS have a protective SiOx coating. If the SiOx were to be pitted by thruster plume particle impacts, the Teflon layer would become exposed to AO. A parametric study was conducted to determine possible degradation due to thruster plume pitting and subsequent AO erosion on AgFEP radiators. Results showed that although plume and AO erosion could degrade the AgFEP coating, the time to effect would exceed the mean time between failures due to penetrant MM/OD impacts. The Boeing ISS Thermal Team concluded that no mitigation would be necessary to protect AgFEP radiators from thruster plume erosion. 4.2. U.S. SOLAR ARRAY ANALYSIS

The NASAN-II code was used to analyze the effect of plume particle impacts on the U.S. Solar Arrays from all Russian and U.S. Vehicle thruster firings during reboost, attitude control, and proximity operations (which includes visiting vehicle approach and separation to ISS docking ports). Preliminary studies showed that solar array coatings would be highly susceptible to plume erosion damage.

78

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

P6 Solar Array P4 Solar Array

Orbiter

Progress

(PMA2)

(SM Aft)

Soyuz

(FGB Nadir)

Progress (DC1 Nadir)

Figure 8. ISS visiting vehicles and docking ports

A higher fidelity assessment was initiated to identify specific areas of concern. Because the effect of plume particle impacts on solar arrays is unknown, this assessment aimed to predict the percent of solar array surface area that would be pitted during nominal ISS operations. Solar array damage predictions, described as percent surface area pitted (with an average pit depth of 1 micron), were used to determine if plume erosion mitigation would be necessary. Analysis of Russian Segment thruster firings included visiting vehicles Progress and Soyuz. The Progress and Soyuz vehicles will fire thrusters during approach and separation to ISS docking ports at the Service Module (SM) aft, the Docking Compartment 1 (DC1) nadir, and the Functional Cargo Block (FGB) Nadir. A diagram of the various ISS docking ports is shown in figure 8. Progress vehicles may also provide ISS attitude control while mated to the Russian Segment. When docked to SM aft, Progress will control ISS pitch and yaw thruster firings. The Progress docked to DC1 nadir performs roll control thruster firings. Depending on ISS configuration, SM thrusters may perform pitch, yaw, and roll control for ISS. Analysis results for the early assembly phase of ISS predicted that the Progress and Soyuz separation flights were the major contributors of pitting on the solar arrays. Areas on the P6 Solar Array of up to 25% surface area pitted were predicted (see figure 9) for the ISS configuration up through ISS Flight 12A.1. Note: For analysis, the solar arrays were assumed to rotate as dictated by flight operations.

79

COURTNEY PANKOP ET AL. P6 Solar Array

P4 Solar Array

Figure 9. Predicted erosion on P6 solar array due to russian segment thruster firings from flight 4A to 12A.1

Russian vehicles were also analyzed for later configurations when new solar arrays will be deployed. The geometry of the ISS and visiting vehicle thrusters to the new solar arrays changes considerably, resulting in new areas of concern. For the P4 solar array, and all solar arrays deployed subsequently, thruster firings from Progress/Soyuz during approach and separation to ISS are no longer a plume erosion concern. The thruster pointing direction on these vehicles, though severe for the P6 array, is advantageous for the P4 array. Progress vehicles docked to the SM aft port, for ISS pitch and yaw control, are also in an advantageous position to mitigate plume erosion. In this position, Progress thrusters are so far aft that they do not have a centerline view to the P4 array. Progress docked to the nadir port of the DC1 for ISS roll control will, however, pose a threat for plume particle impacts. For the ISS assembly complete configuration, analysis predicted that Progress on DC1 nadir thruster firings would be the major Russian Segment contributor to pitting on the U.S. solar arrays. For this configuration, the analysis results show that some areas of the solar array wings would be 100% pitted (see figure 10). On the U.S. Segment, the Space Shuttle Orbiter was analyzed for plume erosion impacts to U.S. solar arrays. The Orbiter thrusters will fire toward ISS during approach and separation to the pressurized mating adapter 2 (PMA2) docking port. At close ranges, a different set of side-pointing thrusters may also plume solar arrays. In addition, the Orbiter may perform several reboosts and maneuvers while mated to ISS. These maneuvers, especially reboost, have the potential to cause severe pitting damage to solar arrays given the Orbiter thruster configuration and the high thruster firing time.

80

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

Figure 10. Predicted solar array erosion for ISS thruster operations through assembly complete due to Russian Segment thruster firings

Analysis results for both Russian Segment and Orbiter thruster firings indicated that a mitigation technique would be required to protect solar arrays from thruster plume pitting damage. 5. Mitigation With predictions of solar array surface pitting, the next step was to determine the extent of degradation to solar array performance resulting from erosion damage. The data available to perform such an assessment was found to be inadequate. A working group within the NASA ISS Program Office recommended a series of ground-based tests to produce data detailing performance degradation to solar cells due to surface pitting. The data from such tests, however, would not be available in time to affect on-orbit solar array integration issues in the near term. Therefore, the working group concluded that operational mitigation techniques must be put in place now to protect ISS optical surfaces. 5.1. CRITERIA FOR MITIGATION

The Boeing ISS Environments Team was tasked to determine the angles to which the solar arrays could rotate such that thruster plume erosion would be minimal. Based on previous analyses which showed that particle impacts at angles greater than 75◦ to normal produced no damage, the following mitigation condition were initially derived: (1) the solar array must be rotated such that the plume impingement angle to a solar array surface is greater than 75◦ from the normal, and (2) no thruster plume is allowed to contact the active side of the solar array.

81

COURTNEY PANKOP ET AL. 1.00E+05 Percent Area Damaged

1.E-05 1.E-04

1.00E+04

1.E-03 0.01 1.00E+03

0.1 1 10

1.00E+02

50

1.00E+01

1.00E+00

0

10

20

30

40

50

60

70

80

90

Angle From Plume Centerline (deg)

Figure 11. Damage contours for 130 N thruster

These criteria alone were found, in some cases, to be difficult or impossible to execute operationally. Therefore, an alternative criterion was added to allow more operational flexibility: Solar array positions that induced no greater than 1% surface area damage per year are assumed acceptable. This criterion is largely a function of the solar array’s position in degrees from the centerline of the thruster plume, since the majority of plume particles are located near the plume centerline. Typically, a solar array positioned 30 to 40◦ from a plume centerline (or farther) would be nearly free of plume particle impact damage. An example of this correlation is shown in figure 11 for the 130 N Russian thruster. In several ISS configurations, the U.S. solar arrays have two degrees of rotational freedom: about the ISS truss (alpha) and about the solar array wing centerline (beta). If the solar array alpha joint could be rotated away from the plume centerline (to meet the alternative criterion), the beta may be allowed to sun-track freely. Otherwise, both the solar array alpha and beta rotations must be fixed, or ‘feathered,’ to mitigate plume (per the first criteria set). 5.2. APPLYING THE MITIGATION CRITERIA

For the early ISS configuration, analyses found that the P6 solar array would receive pitting damage during Progress, Soyuz, and Orbiter proximity operations. Existing ISS Program Flight Rules for feathering the P6 solar array during Progress/Soyuz approach and Orbiter approach and separation (put into place for other requirements) met requirements for plume erosion mitigation, so no further action was required. For Progress/Soyuz separation, however, new flight rules were drafted to

82

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

Figure 12. P4-2A allowable feathering angles for Progress on DC1 nadir roll control thruster firings

ensure that flight controllers would position the P6 solar array to mitigate plume erosion prior to the separation operation. This flight rule is in effect for the current ISS configuration and will remain in effect until the P6 solar array is relocated to the main truss. Analyses show that additional thrusters will become a plume erosion concern when additional solar arrays are deployed. The thrusters of concern include Progress on DC1 nadir thruster firings for ISS roll control, Orbiter thruster firings during approach and separation to PMA2, and Orbiter thruster firings during mated ISS operations. Feathering angles to mitigate plume erosion must be defined for each of these thruster firing events. For Progress on DC1 nadir roll control firings, allowable solar array alpha/beta angle pairs to mitigate erosion were defined per the mitigation criteria. These alpha/beta pairs have been tabulated for inclusion into a flight rule to provide flight controllers with the proper settings to assure solar array protection. A sample table is shown for the P4 Solar Array in figure 12. It should be noted that allowable solar array alpha/beta angle pairs are shown in gray. Proximity operations have another factor adding complexity to the issue of plume erosion. Solar arrays may need to be positioned to mitigate plume erosion from the incoming (or departing) vehicle’s thruster firings. In addition, ISS

COURTNEY PANKOP ET AL.

83

Allowable P4 2A Solar Array Alpha/Beta Angle Pair Combinations

P4 Feathering Angles for Orbiter Approach Prox Ops Alpha locked at 0° 2A Beta = 70° 4A Beta = 270° (P6 Also feathered for Orbiter Proximity Operations)

Figure 13. P4-2A allowable feathering angles for Orbiter approach to ISS (with progress on SM Aft and progress on DC1 Nadir attitude control)

thrusters fire to maintain ISS attitude during proximity operations. Consequently, solar arrays must be positioned to mitigate plume erosion from both the visiting vehicle as well as the ISS vehicles firing thrusters for attitude control. Orbiter proximity operations provide a good example of this scenario. Typically, the Progress docked to SM aft will perform ISS pitch and yaw control and the Progress docked to DC1 nadir will perform ISS roll control during Orbiter approach and separation. The solar array position must protect against all thruster firings from these vehicles. Analyses were conducted to determine the allowable solar array alpha/beta angle pairs for which plume erosion during Orbiter proximity operations (with Progress on SM aft and Progress on DC1 nadir performing ISS attitude control) is mitigated. These alpha/beta pairs have been tabulated for inclusion into a flight rule to provide flight controllers with the proper settings to assure solar array protection. An example of this table is shown for the P4 solar array in figure 13. By comparison to figure 12 (the alpha/beta pairs which mitigate plume erosion from Progress on DC1 nadir alone), it is evident that the added element of Orbiter thruster firings severely limits allowable solar array positions.

84

MITIGATION OF THRUSTER PLUME-INDUCED EROSION

5.3. FLIGHT RULE INTEGRATION

As Boeing Environments completes plume erosion analyses for ISS thrusters pluming solar arrays, results must be incorporated into ISS Program Flight Rules. The tabulated alpha/beta combinations for allowable solar array positions are prepared for various thruster firing events (see figures 12 and 13). Before the feathering angle tables can be implemented into flight rules, plume erosion results must be integrated with all other applicable ISS requirements. These requirements may include feathering solar arrays to minimize thruster plume-induced structural loads or heating. Power requirements may also be a driver in determining where solar arrays may be positioned. An integrated solution must be found including all subsystem requirements; this solution then moves forward into the flight rules. Flight controllers use the proper solar array settings defined in flight rules to ensure system-wide performance during thruster firing events. 6. Future Work In upcoming years, ISS operations will include two new visiting vehicles: the European Space Agency’s Automated Transfer Vehicle (ATV) as well as then Japanese H-II Transfer Vehicle (HTV). ISS operational mitigation techniques for the U.S. solar arrays during ATV and HTV proximity operations must be developed as well as mitigation techniques for ATV mated operations. The Boeing ISS Environments Team will perform analyses to define the allowable alpha/beta angles for the solar arrays during these thruster firing events. The resulting requirements are to be integrated with similar requirements for thermal and structural effects of thruster plume impingement. Integrated requirements will be implemented through ISS Program Flight Rules. The Boeing ISS Environments Team will also continue updating previous analyses as changes to the ISS configuration are made. In the near team, the Service Module will be analyzed for unanticipated utilization during later ISS assembly sequences. Russian Segment and U.S. vehicle analyses may also be updated as new thruster firing data becomes available. In addition to conducting plume erosion analyses, the Boeing ISS Environments Team will continue efforts to improve plume erosion modeling and predictions. Ground-based tests, which will detail solar cell performance degradation due to surface pitting, are planned to take place this year. Results will be incorporated into the NASAN-II plume erosion model. 7. Conclusions Thruster plume particles can cause mechanical damage in the form of erosion/pitting to ISS sensitive surfaces. In general, any spacecraft sensitive surfaces

COURTNEY PANKOP ET AL.

85

exposed to the exhaust plumes of thrusters on-orbit should either compensate for thruster plume erosion via design or prepare mitigation techniques to minimize thruster plume particle pitting. In this paper, the development of the erosion model was outlined along with predictions of damage due to thruster firings on selected ISS sensitive surfaces. Analysis results indicate that plume erosion of ATCS radiator coatings would not affect heat rejection properties. Characterization of thruster plume inducederosion/pitting of U.S. solar arrays due to nominal operations of ISS vehicle thrusters was shown for an interim and final ISS assembly configuration. The solar array analysis results show that plume erosion mitigation is required to protect solar array optical coatings. A mitigation technique for ISS solar arrays, which minimizes thruster plume erosion during current and future ISS assemblies, was presented. Implementation of these mitigation techniques into ISS Program Fight Rules was discussed, as well as plans for future analyses and ground-based testing to characterize plume erosion effects. The Boeing ISS Environments Team pursues these activities to ensure a known environment around the ISS so as to guarantee the success of the ISS as a platform for scientific experiments in low Earth orbit.

References 1. Trinks, H. (1987) Experimental Investigation of the Exhaust Plume Flow Fields of Various Small Bipropellant and Monopropellant Thrusters, AIAA Paper 87–1607, June 1987. 2. Rebrov, S., and Gerasimov, Y.; “Investigation of the Contamination Properties of Bipropellant Thrusters,” AIAA 2001-2818, 35th AIAA Thermophysics Conference, Anaheim, CA, 11–14 June 2001. 3. Theall, J. (2000) JSC Space Science Branch, Private communication, 27 November 2000. 4. Koontz, S., Melendez, O., Zolensky, M., and Soares, C. (1996) SPIFEX Contamination Studies, JSC-27399, SPIE May 1996, 4–11. 5. Soares, C. E. and Mikatarian, R. R. (2002) In International Symposium on Optical Science and Technology, SPIE 4774-20, Seattle, July 2002, 2–13. 6. Orr, W. (2000) PIC Camera Lens Test Analysis, Lockheed Martin Science, Engineering, Test, and Analysis Report to NASA JSC EM2, 16 August 2000. 7. Daneman, B., and Thomlinson, H. (1993) Advancements in Long-Life Thermal Control Coatings for Low Earth Orbit Applications, McDonnell Douglas Aerospace, #93222. 8. Wingate, C. and Stellingwerf, R. (1993) Smooth Particle Hydrodynamics: The SPHINX and SPHC Codes, Technical Report LA-UR-9301938, Los Alamos National Laboratory, January 1993. 9. Alred, J., Boeder, P., Mikatarian, R., Pankop, C., and Schmidl, W. (2003) In Proceedings of the 9th International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands, 16–20 June 2003.

DEGRADATION OF THERMAL CONTROL COATINGS UNDER INFLUENCE OF PROTON IRRADIATION L. S. NOVIKOV,1 G. G. SOLOVYEV,1 V. N. VASIL’EV,2 A. V. GRIGOREVSKIY,3 AND L. V. KISELEVA3 1 Skobeltsyn Institute of Nuclear Physics Moscow State University 119992 Moscow, Russia 2 Central Research Institute of Machine Building, Korolev, Russia 3 Corporation “Kompozit,” Korolev, Russia

Abstract. A number of thermal control coatings were exposed to proton irradiation with particle energies 100–500 keV and particle fluences 1014 –2 × 1016 cm−2 in a proton accelerator. Experimental data on changes of αs of paint coatings and mirror coatings versus the proton fluence and energy are presented. Prediction of the αs changes under the impact of proton radiation with distributed energy spectrum in space flight based on the ground testing results with monoenergetic proton beams is discussed. Estimations of coating degradation in geosynchronous orbit are made using a mathematical model of degradation. Key words: thermal control coating, degradation, proton irradiation, simulation tests

1. Experimental Studies Tests of thermal control coating (TCC) samples under impact of protons with energies up to 500 keV and fluences from 1014 to 2 × 1016 cm−2 have been carried out on the KG-500 accelerator in Skobeltsyn Institute of Nuclear Physics Moscow State University (SINP MSU). The accelerator provides energy stability of 0.1%. The beam current of protons was set not to exceed 1 μA. At the maximum energy of protons (0.5 MeV), the power at the TCC sample was less than 0.5 W, and the temperature of the sample during the irradiation did not exceed 100◦ C. The proton fluence was measured with accuracy of at least 1%. The measurement of the TCC optical absorption coefficient αs was done with the photometer built in the vacuum chamber. The high value of the proton fluence (2 × 1016 cm−2 ) was chosen to determine the parameters in function αs = f (F, E) used in physical and mathematical model of the TCC optical degradation [1– 3]. It was shown [4] that protons produce the strongest irreversible changes in 87 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 87–91. C 2006 Springer. Printed in the Netherlands.

88

DEGRADATION OF THERMAL CONTROL COATINGS TABLE 1. Coating parameters No

TCC type

Pigment

Binding

Color

1 2 3 4

ECOM-1 TR-SO-TSM SS-1 ECOM-2

ZnO ZrO2 ZnO + Al paste Metal oxides + carbon black

Acryl pitch Liquid glass Acryl pitch Acryl pitch

White White Silver grey Black

optical thermal properties that could be associated with processes of lattice defect generation.

2. Test Results Data on the tested TCC are presented in table 1. Experimental results of the TCC αs coefficient changes are presented in figure 1 as a function of the proton fluence with energy of 300 keV. The data presented in figure 1 confirmed the relative radiation stability of the tested coatings. For definition of the model parameters, it is necessary to obtain similar dependences for TCC irradiation with protons with various energies. Such tests were performed and the dependences for ECOM-1 and TR-SO-TSM are presented in figures 2 and 3. The maximum amount of radiation defects is formed under proton irradiation at the end of the proton range. Concentration of the defects and the proton range increase with proton increasing energy. The value of αs grows in the energy interval from 0 to 200 keV due to the increase of concentration of the defects [4]. This dependence is not holding at higher energies due to an increase of the ΔAs 0.6

1 2

0.5 0.4 0.3 0.2

3

0.1

4

0 10

14

5 . 1015

1016

1.5 . 1016

F, cm−2

Figure 1. Change of the TCC αs Coefficient after the proton irradiation with energy 300 keV: (1) ECOM-1, (2) TR-SO-TSM, (3) SS-1, (4) ECOM-2

89

L. S. NOVIKOV ET AL. ΔAs

0.6

1

0.5

2

0.4

3

0.3 0.2 0.1 0 0

15

5.0 . 10

16

1.0 . 10

1.5 . 1016

16

2.0 . 10 F, cm−2

Figure 2. Change of the ECOM-1 coating As coefficient as function of the proton fluence with energies (1) 150, (2) 300, and (3) 500 keV

radiation defects localization depth, where the influence of the defects on the αs value decreases.

3. Model of Optical Degradation of Coatings The parameters of the physical and mathematical model describing the process of optical degradation of the TCC due to radiation are determined from the experimental data obtained during irradiation of the coatings by monoenergetic beams. These parameters are used in calculation of the forecast of the αs coefficient change under impact of corpuscular radiation with continuous energy spectrum in spacecraft orbit. In the given work, the model was chosen that describes the ΔAs 0.5

3 2 1

0.4 0.3 0.2 0.1 0 0

15

5.0 . 10

1.0 . 1016

1.5 . 1016

2.0.1016 F, cm−2

Figure 3. Change of the TR-SO-TSM coating As coefficient as function of the proton fluence with energies (1) 150, (2) 300, and (3) 500 keV

90

DEGRADATION OF THERMAL CONTROL COATINGS ΔAs

0.7

1

0.6

2

0.5 0.4 0.3 0.2 0.1 0 0

100

200

300

400

500 E, keV

Figure 4. Dependence of the As coefficient of (1) ECOM-1 and (2) TR-SO-TSM coatings on the proton energy at a fluence 2 × 1016 cm−2

optical degradation of the TCC as follows: As = a[1 − exp(−bE γ F β )]

(1)

where a,b, γ , ß—parameters of the model, E andF—energy and fluence of protons accordingly. The model parameters as determined from the experimental data presented in figures 2–4 are summarized in table 2. For definition of coating damage under the influence of radiation with a continuous energy spectrum and with a distribution function dϕ/dE, an effective flux density Feff of monoenergetic particles with energy E 0 is used [3] ⎛∞ ⎞ dϕ γ /β Feff (E 0 ) = ⎝ E γ /β (2) dE ⎠ E0 dE 0

Using eqs. (1) and (2), the changes in αs were calculated for different GEO mission durations. In table 3, the calculated values of the αs coefficient change influenced by proton radiation in geosynchronous orbit (GEO) are presented [5]. As it follows from the presented results, the resistance of TCC to radiation impact can vary significantly. 4. Conclusions

r Tests of various types of TCC under impact of protons with energies 150– 500 keV with fluence up to 2 × 1016 cm−2 revealed that theAsvalue variation TABLE 2. Parameters used in the model calculation TCC type ECOM-1 TR-SO-TSM

a

B

ß

γ

0.48 0.78

1.8. 10−14 1.7. 10−11

0.89 0.56

0.74 0.46

91

L. S. NOVIKOV ET AL. TABLE 3. Predicted As coefficient changes of the TCC under impact of proton radiation for various durations of spacecraft in GEO TCC type ECOM-1 TR-SO-TSM

0.5 year

1 year

3 years

5 years

10 years

0.01 –

0.02 –

0.05 0.001

0.08 0.002

0.14 0.003

does not exceed 0.05 for black coating ECOM-2 and 0.1 for grey SS-1 coating. The variation is 0.3–0.6 for white coatings ECOM-1 and TR-SO-TSM. r Radiation resistance of the ECOM-1 enamel coating is higher than that of the TR-SO-TSN ceramic coating. r The As value functions of the proton fluence and energy obtained for the ECOM1 coating and TR-SO-TSN coating were used the determine the predicted variations of the As value during the spacecraft flight in GSO up to 10 years. r ECOM-2 coating and TR-SO-TSM coating are recommended as the most preferable for TCC on spacecrafts in GSO.

References 1. Mikhailov, M. M., Dvoretsky, M. I., and Krutikov, V. N. (1982) Space Technology and Material Research, Science, Moscow, pp. 95–100. 2. Titov, V. I. and Tarasov, Y. I. (1984) Journal of Physics and Chemistry 58(5), 1212–1214. 3. Vasiliev, V. N., Grigorevskiy, A. V., and Gordeev, Y. P. (2002) In Proceedings of the 6th International Space Conference on Protection of Materials and Structures from Space Environment, Toronto, Canada, 2002, pp. 543–550. 4. Prosvizikov, V. M., Grigorievskiy, A. V., Kiseleva, L. V., Zelenkevich, A. P., and Tsvelev, V. M. (2004) In Proceedings of the 7th International Space Conference on Space Materials, Toronto, Canada, Springer, 61–69. 5. Garret, H. B., Schwank, D. C., and De Forest, S. E. (1981) Ions Planetary and Space Science 29(10), 1045–1060.

MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION (ISS) FROM WATER DUMPS

WILLIAM SCHMIDL, JAMES VISENTINE, AND RON MIKATARIAN The Boeing Company, 13100 Space Center Blvd. Houston, TX, 77059-3599

Abstract. The International Space Station (ISS) and Orbiter dump water overboard. This water is from the ISS condensate system, and from the Orbiter’s fuel cell (supply side) and wastewater (urine and condensate) systems. When water is dumped into a vacuum, some of it flashes into a vapor. The expanding vapor bursts the stream into vapor, and small and large liquid/ice particles. The large liquid/ice particles are approximately 2 mm in diameter and have nominal velocities of approximately 31 ft·s−1 (U.S. Lab) and 50 ft·s−1 (Orbiter). As these liquid/ice particles impact, they may cause mechanical damage due to erosion/pitting of sensitive surfaces, including solar array or radiator surfaces. Solar arrays are of particular concern because of the thin optical coatings on the surface of the cells. ˚ Damage to these coatings can cause The thickness of these coatings is 43300 A. degradation of the cells’ optical characteristics that can potentially reduce performance and shorten the life of the cells. To mitigate damage from water dumps, the characteristics of the water dumps were studied and the results used to develop the constraints needed to mitigate damage to ISS hardware from U.S. Lab and Orbiter water dumps. Key words: water dump, vent, purge, impact, damage, erosion, pitting

1. Introduction The International Space Station (ISS) and Orbiter dump water overboard into space. The phenomena of releasing water into a vacuum has been studied for many years [1–3]. Figure 1 shows a schematic of water dumping into a vacuum. It is known that as the liquid exits the nozzle it freezes. Then the expanding gas bubbles burst the stream into vapor, and small and large liquid/ice particles that can travel in various directions. When water is dumped overboard, there is a concern that direct contact from the liquid/ice particles on ISS hardware can cause mechanical damage to sensitive 93 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 93–105. C 2006 Springer. Printed in the Netherlands.

94 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION Nozzle

Water Jet

Burst Ice Particles

Ice Particles A collimated jet of water is ejected. The outermost layers of the liquid jet rapidly cool by evaporation. The inner core of liquid(at close to injection temperature and pressure) is superheated and boils forming bubbles.

As the bubbles of vapor expand and approach the diameter of the liquid jet they overcome the surface tension and break through the surface, i.e.,the jet bursts.

Particles of water freeze by evaporation and radiation, followed by sublimation due to absorption of solar radiation, Earth albedo, Earth blackbody radiation etc.

Figure 1. Schematic of water dumping into a vacuum. The outer layer of water freezes as the inner core of liquid boils and forms bubbles. The expanding bubbles burst the stream into liquid/ice particles

surfaces due to erosion/pitting of those surfaces. Solar arrays are of particular concern because of the thin optical coatings on the surface of the solar cells. ˚ Damage to these coatings can cause The thickness of these coatings is 43,300 A. degradation of the solar cells’ optical characteristics that can potentially reduce performance and shorten the life of the solar cells. To mitigate potential damage from water dumps, a methodology was developed that could be used to develop the constraints needed to protect ISS sensitive surfaces. To develop the methodology, the characteristics of water dumps were studied and based on the results, constraints were developed to define the select angles that the ISS solar arrays can be parked at to preclude damage to solar array and radiator surfaces. 2. Characteristics of Water Dumps 2.1. U.S. LAB WATER DUMPS

Condensate is produced aboard the ISS by the thermal heating and cooling system and stored in a tank that must eventually be emptied. The tank is fill restricted to 100 lbm (60.35% of capacity) to prevent hard-filling the tank and to maximize the tank bellows’ life. The U.S. Lab condensate system dumps the excess water at cabin pressure (14.7 psi) and with a mass flow rate that is approximately 14.6 g·s−1 . The condensate system dumps the excess water overboard through nozzles located on the U.S. Lab forward end-cone. These nozzles are located on the

WILLIAM SCHMIDL ET AL.

95

U.S. Laboratory dump nozzles Figure 2. ISS assembly complete configuration. The blue arrows indicate the position and direction of the U.S. Lab condensate dump nozzles. The nozzles are located on the starboard nadir and port zenith sides of the U.S. Lab

port-zenith side, and starboard-nadir sides, and have a diameter of 1.4 mm. Each nozzle is preheated to 120◦ C and heated while dumping to keep ice from forming close to the nozzle exit that could result in clogging the nozzle, forming an icicle, or diverting the dump stream. Figure 2 shows the relative position of the condensate dump nozzles and their direction. The plume from the nozzles is directed 37.4◦ from the zenith and nadir directions. The flight experiment SDTO 16004-A, performed on 7 September 2001, was to video a condensate dump using the ISS robotic arm, the space station remote manipulator system (SSRMS) and to measure the characteristics of the U.S. Lab water dump [4–6]. The images were captured using the SSRMS cameras of both the port and starboard dumps. The SDTO measurements included the plume cone angle, velocity of the large liquid/ice particles, and the duration of each of the dump phases. Figure 3 shows the distinct dump phases that were observed in the SDTO video. The phases observed were a startup, steady state, and shutdown phase. The startup phase occurs when the line valve is opened and the liquid first moves into the lower pressure environment. The steady state phase occurs for the majority of the dumping duration. The liquid is confined to a narrow directed cone. It lasts for the remainder of the dump, excluding the startup and shutdown phases. The shutdown phase occurs when gas becomes entrained in the condensate water being dumped. The shutdown phase was also divided into an initial shutdown phase and a sputtering shutdown phase. The initial shutdown phase occurs when gas becomes entrained in the condensate

96 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION

(a) Startup

(b) Steady state

(c) Shutdown

Figure 3. U.S. Lab water dump video captured during flight experiment SDTO 16004-A. Three phases were observed during the U.S. Lab condensate water dump: a startup, steady state, and shutdown phase. The startup phase lasts approximately 36 s, the steady state phase lasted approximately 564 s, and the shutdown phase lasted approximately 54 s

tank and the valve is still open. The shutdown sputtering phase occurs after the valve is closed. The plume cone angles were measured from the centerline for each of the dump phases. It was observed that the plume had a core region where the majority of the liquid/ice particles were concentrated; however, some particles were also observed outside the core region. For SDTO 16004-A, approximately 145 particles were ejected outside the core region. In figure 3, it can be observed that the steady state phase has a tight cone. The startup and shutdown phases, at the beginning and end of the dump event, have large cone angles. This is due to the bursting of the stream. The plume cone angle for the steady state phase was measured to be 10◦ . However, for engineering margin, the plume cone angle for the steady state phase was defined to be 20◦ . For the startup and shutdown phases, the plume cone angle was defined to be 60◦ . Figure 4 shows the angular distribution function that was developed for each of the phases observed during SDTO 16004-A. For the startup, shutdown initial, and shutdown sputtering phases, the distribution is flat to account for the larger plume cone angle that was observed. Figure 4(b) shows the angular distribution function for the steady state phase. It can be seen that the majority of large particles is concentrated near the plume centerline out to 10◦ . This region is defined as the impact zone. The ISS External Contamination Team defined the cone out to 20◦ from the plume centerline as the impact zone with engineering margin. Figures 5(a) and 5(b) shows the field of views from the port and starboard U.S. Lab condensate water dump nozzles for ISS assembly complete. It can be seen that if the solar arrays are not rotated and feathered for the water dump, they could rotate directly into the impact zone with engineering margin of the water dump. For each vent phase of SDTO 16004-A, the liquid/ice particle velocities were measured both outside and inside the core region of the plume. For the startup phase, velocities measured outside the 60◦ core region ranged from 5.8 to 15 ft·s−1 (measured velocities including the core region ranged from 2.8 to 30.6 ft·s−1 ). For

97

WILLIAM SCHMIDL ET AL. 1

1

Total

0.9

Startup Phase Angular Distribution vs Angle (degrees)

0.8

Large Particle

0.7

Total

0.6

Large Particle

0.5

Gas

0.4

Gas

0.3

Small Particle

0.2

0.7 0.6

Total

0.5

Large Particle

0.4

Gas

0.3 0.2

0.1

Steady State Phase Angular Distribution vs Angle (degrees)

Large Particle

0.8

Angular Distribution

Angular Distribution

Total

0.9

Small Particle

Gas

Small Particle

0.1

Small Particle

0

0 0

10

20

30

40

50

60

70

80

0

90

10

20

30

Angle (degrees)

50

60

70

80

90

Angle (degrees)

(a)

(b)

1

1

Total

0.9

Shutdown Initial Phase Angular Distribution vs Angle (degrees)

Total

0.9

0.8

Shutdown Sputtering Phase Angular Distribution vs Angle (degrees)

0.8

Large Particle

0.7

Angular Distribution

Angular Distribution

40

Total

0.6

Large Particle

0.5

Gas

0.4

Gas

Small Particle

0.3 0.2

Large Particle

0.7

Total

0.6

Large Particle

0.5

Gas

0.4

Small Particle

Gas

0.3 0.2

0.1

0.1

Small Particle

0

Small Particle

0 0

10

20

30

40

50

60

70

80

90

0

10

20

30

40

50

Angle (degrees)

Angle (degrees)

(c)

(d)

60

70

80

90

Figure 4. U.S. Lab water dump cumulative mass fraction distribution: (a) startup phase, (b) steady state phase, (c) shutdown initial phase, and (d) shutdown sputtering phase

the steady state phase (core region within 20◦ ), from average flow rate and nozzle diameter, the average velocity was determined to be 31 ft·s−1 . For the initial shutdown phase, measured velocities outside the core region ranged from 2.0 to 3.3 ft·s−1 (measured velocities including the core region ranged from 2.0 to 8.0 ft·s−1 ). For the shutdown sputtering phase, measured velocities outside the 60˚ core region ranged from 2.0 to 4.6 ft·s−1 (measured velocities including the core region ranged from 1.5 to 9.1 ft·s−1 ). The phase durations for the SDTO 16004-A dump, as shown in figure 3, were determined from the video by the NASA Imaging Science and Analysis Group (IS&AG). The startup phase took approximately 36 s, the initial shutdown phase took approximately 22 s, and the shutdown sputtering phase took approximately 32 s. The steady state phase lasted for the remainder of the dump, which for this experiment (SDTO 16004-A), lasted approximately 564 s. 2.2. ORBITER WATER DUMPS

The Orbiter dumps both waste water (urine and condensate) and supply water. Supply water is generated by the Orbiter’s fuel cells and is extremely pure, so it is

98 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION (a) Impact Zone

Impact zone with engineering margin (b)

Impact Zone

Impact zone with engineering margin Figure 5. (a) Field of view from port zenith U.S. Lab condensate water dump nozzle for ISS assembly complete. It can be seen that if the solar array is not feathered, it will be impacted broadside; (b) field of view from starboard nadir U.S. Lab condensate water dump nozzle for ISS assembly complete. It can be seen that if the solar array is not feathered, it will be impacted

not a molecular deposition concern. The condensate water is collected from the Orbiter cabin and has approximately 0.004% nonvolatile residues in the water. So, it does not leave a significant deposit. Pure urine contains approximately 4% residue and is a deposition concern. To protect the solar cells from urine deposits, the solar arrays are feathered so that only the backside is hit. The feathering angles defined to protect against erosion/pitting will also mitigate deposition concerns. Flight rules are currently in place to minimize urine dumps. To minimize urine dumps, urine and condensate water are separated, and the condensate water

WILLIAM SCHMIDL ET AL.

99

Figure 6. Orbiter water dump for ISS assembly complete. The Orbiter dumps water out to the port side of ISS

is stored in contingency water containers (CWCs). Only urine in excess of the tank capacity is dumped overboard. All urine dumping is to be discontinued when the Japanese hardware goes up, because it is in the direct line of sight of the Orbiter vent nozzles. Water released from the Orbiter is dumped through two nozzles located close to each other on the Orbiter’s port side and are directed to the port side of ISS. One nozzle is for supply side water and the other is for wastewater. The nozzles are 1.4 mm in diameter. Water released from each nozzle is dumped at 31 psia at 23.7 g·s−1 . The concern for Orbiter water dumps is with sensitive hardware on the port side of ISS. Figure 6 shows a schematic of the direction of the Orbiter dump for ISS assembly complete. Figure 7 shows an image of the Orbiter dumping water. The inset figure was captured from a video of the Orbiter water dump. The Boeing external contamination team developed a model for the Orbiter water dump [7]. Figure 8 shows the plume distribution for that model. It can be seen that the majority of large particles is concentrated near the centerline of the plume out to 10◦ . This region is defined as the impact zone. The ISS External Contamination Team defined the cone out to 20◦ from the plume centerline as the impact zone with engineering margin.

100 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION

Figure 7. Image of Orbiter water dump. The inset image was captured from a video of the Orbiter water dump

1 0.9

Angular Distribution vs Angle (degrees)

Angular Distribution

0.8 0.7 0.6

Total

0.5

Large Particle

0.4

Gas

0.3

Small Particle

0.2 0.1 0 0

10

20

30

40

50

60

70

80

90

Angle (degrees)

Figure 8. Boeing two-phase water dump plume model for Orbiter. The region from the plume centerline out to 10◦ is defined as the impact zone. The region out to 20◦ is defined as the impact zone with engineering margin

101

WILLIAM SCHMIDL ET AL. Impact Zone

Impact zone with engineering margin Figure 9. Field of view from Orbiter water dump nozzles for ISS assembly complete. It can be seen that if the solar array is not feathered, it will be impacted broadside

Figure 9 shows the field of view from the Orbiter dump nozzles for ISS assembly complete. Sensitive surfaces of concern are the Japanese Aerospace Exploration Agency (JAXA) payload sites and the P4/P6 radiators and solar arrays. The JAXA payload sites can be seen just below the center of the plume. It can be seen that if the solar array is not feathered, it will be impacted broadside. In this configuration, the radiator is rotated out of the field of view. As will be shown later, the solar arrays should be feathered so that the impacts occur on the backside of the array and at a shallow (15◦ ) angle. The Arnold Engineering Development Center (AEDC) measured the characteristics of the Orbiter water dump. These measured characteristics include the Orbiter water dump plume cone angle, the composition of the plume (the fractions of vapor/small particles/large particles), and the velocity of the large particles. The results from the AEDC ground tests have been compared with flight images and an Orbiter dump model was developed [7]. The results show that the velocity of the particles is nominally 50 ft·s−1 and the plume cone angle is 10◦ .

3. Damage Mitigation SPHINX is an impact code developed at the Los Alamos National Laboratory (LANL) [8–10]. SPHINX uses smooth particle hydrodynamics (SPH) to simulate impact phenomena. SPHINX has been applied, in the past, to modeling thruster droplet impacts onto sensitive surfaces to develop the appropriate feathering angles for the solar arrays [8]. The results from these studies showed that direct impacts normal to the surface would damage the solar cells. Impacts at a shallow angle

102 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION to the surface of less than 15◦ (or 75◦ from the surface normal) did not show any damage to the solar cells. Using these results as a starting point, the constraints for liquid/ice particle impacts were developed. Although, the results did not show any damage to the ˚ If the coating is damaged, solar cell, the UVE protective coating is thin, 43,300 A. it degrades the performance and lifetime of the solar cell. To be conservative, it was determined that the operational constraint will be to not allow impacts onto the active side. ˚ SIOx coating to protect the kapton The backside of the solar array has a 1300 A backing from atomic oxygen erosion. Below the kapton, there are additional layers that if lost will not affect the performance of the solar cell. However, to minimize damage to the solar array, the operational constraint developed for the backside of the solar array is that impacts will be at a shallow angle, less than 15◦ to the surface. In addition, to minimize the number of impacts, an operational constraint was developed so that the solar arrays will be rotated to remain outside the impact zone with engineering margin (the 20◦ half cone angle cone around the plume centerline). To mitigate damage to the solar array photovoltaic radiators (PVR), an operational constraint was developed to keep the radiators away from the plume centerline. The radiators operate cold and cannot be feathered. In addition, the effect of liquid/ice particle impacts on the radiators has not been well defined yet. The constraint is to keep a solar array between the impact zone with engineering margin and the radiator. This constraint will allow the radiators no closer than 50 to 60◦ from the plume centerline. 3.1. U.S. LAB WATER DUMP DAMAGE MITIGATION

Figure 10 shows the field of view from the port side U.S. Lab condensate water nozzle for ISS assembly complete. The solar arrays are rotated outside of the impact zone with engineering margin and feathered so that impacts from the water dump occur on the backside of the arrays. An example of the feathering angles defined, based on the defined constraints, to feather the solar array wings (SAWs) so that impacts from the liquid/ice particles occur on the backside of the arrays at a shallow angle is shown in figure 11. 3.2. ORBITER WATER DUMP DAMAGE MITIGATION

Figure 12 shows the field of view from the Orbiter dump nozzle for a solar array rotated out of the high impact zone and with the solar array feathered so that impacts are on the backside of the array at a shallow angle. The solar array is feathered so that there are no impacts on the active side of the array and so that impacts on the backside of the array are at a shallow angle.

103

WILLIAM SCHMIDL ET AL.

Impact Zone

Impact zone with engineering margin Figure 10. Field of view from port side U.S. Lab condensate water dump nozzle for ISS assembly complete. The solar array is feathered and rotated out of the impact zone with engineering margin

Figure 11. Example U.S. Lab port side condensate water dump allowable solar array feathering angles for the P4-4A solar array wing (SAW) for beta gimbal assembly (BGA) beta rotations from 0 to 238◦

104 MITIGATION OF DAMAGE TO THE INTERNATIONAL SPACE STATION Impact Zone

Impact zone with engineering margin Figure 12. Field of view from Orbiter water dump nozzle for ISS assembly complete. The solar array is feathered and rotated out of the impact zone with engineering margin

4. Summary The International Space Station (ISS) and Orbiter both dump water overboard into space. The dumped stream bursts into liquid/ice particles. The large liquid/ice particles are approximately 2 mm in diameter and have nominal velocities of approximately 31 ft·s−1 (U.S. Lab) and 50 ft·s−1 (Orbiter). As these liquid/ice particles impact, they can cause mechanical damage due to erosion/pitting of sensitive surfaces, such as the coatings used on the solar array or radiator surfaces. Solar arrays are of particular concern because of the thin optical coatings on the surface of the cells. Damage to these coatings can cause degradation of the cells’ performance and operational lifetime. To mitigate damage from water dumps, the characteristics of the water dumps were studied. The results were used to develop the constraints needed to mitigate damage to ISS hardware from the U.S. Lab and Orbiter water dumps. The results of these studies show that the ISS solar arrays can be parked at select angles during water dump operations to preclude damage to the solar array and radiator surfaces.

References 1. Fuchs, H. and Legge, H. (1979) Acta Astronautica 6, 1213–1226. 2. Kofsky, I. L., Rall, D. L. A., Maris, M. A., Tran, N. H., Murad, E., Pike, C. P., Knecht, D. J., Viereck, R. A., Stair, A. T., and Setayesh, A. (1992) Acta Astronautica 26(5), 325–347. 3. Mikatarian, R. R. and Anderson, R. G. (1964) AIAA Unmanned Spacecraft Meeting 12, 255–259. 4. Schmidl, W. D., Alred, J. W., Mikatarian, R. R., Soares, C., Miles, E., Howorth, L., Mishina, L., and Murtazin, R. (2002) Characterization of On-Orbit U.S. Lab Condensate Vacuum Venting, AIAA IAF-02-T.P.06.

WILLIAM SCHMIDL ET AL.

105

5. Schmidl, W. D., Alred, J. W., Mikatarian, R. R., Soares, C., Miles, E., Howorth, W., Mishina, L., and Murtazin, R. (2003a) In 33rd AIAA Fluid Dynamics Conference and Exhibit, AIAA paper 2003-4268, Orlando, FL, 23–26 June 2003. 6. Schmidl, W. D., Alred, J. W., Mikatarian, R. R., Soares, C., Miles, E., Howorth, W., Mishina, L., and Murtazin, R. (2003b) In 9th International Symposium on Materials in Space, ESTEC, Noordwijk, Netherlands, June 2003, 423–430. 7. Alred, J. W., Smith, L. N., Wang, K. C., Lumpkin, F. E., and Fitzgerald, S. M. (1998) In Proceedings of AIAA/ASME Joint Thermophysics and Heat Transfer Conference, Albuquerque, NM, June 15–18, 1998, 98–2588. 8. Alred, J., Boeder, P., Mikatarian, R., Pankop, C., and Schmidl, W. (2003) In 9th International Symposium on Materials in Space, ESTEC, Noordwijk, Netherlands, June 2003, 1–9. 9. Wingate, C. and Stellingwerf, R. (1993) Smooth Particle Hydrodynamics: The SPHINX and SPHC Codes, Technical Report LA-UR-9301938, Los Alamos National Laboratory, January 1993. 10. Stellingwerf, R. and Wingate, C. (1993) International Journal of Impact Engineering 14, 707– 718.

INVESTIGATION OF SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION ON THE DYEING OF OPTICAL QUARTZ GLASS HAI LIU,1 SHIYU HE,1 HONGBIN GENG,1 DEZHUANG YANG,1 AND V. V. ABRAIMOV2 1 Harbin Institute of Technology, Harbin, 150001, P. R. China 2 Kharkov National University, Kharkov, 60164, Ukraine

Abstract. Optical materials have been widely used in spacecraft. Many optical elements exposed in open space, such as mirrors of telescopes, fused quartz windows, lenses and cover plates on solar cells, are affected by various space environmental factors. These include high vacuum, thermal cycling, irradiation by charged particles, and solar electromagnetic irradiation. It is important to investigate the interaction of optical materials with space environmental factors. The synergistic effects of proton and electron radiation on optical properties of quartz glass for space applications were studied using a complex simulator for the space environment named “KIFK.” The energy of protons and electrons ranged from 60 to 180 keV, which is within the energy range of the Earth’s radiation belt. The changes in transmission of quartz glass under radiation were analyzed. Synergistic effects were found during combined irradiation. The synergistic effect of complex radiation indicates at the necessity of developing a complex space environment test. Key words: optical quartz glass, proton and electron radiation, synergistic effect, transmission

1. Introduction With increasing human activity in space, the reliability and lifetime of spacecraft is becoming an important restriction. Space environment affects the reliability and lifetime of a spacecraft. Great effort has gone into studying the space environment effects on spacecraft among which one can mention the “space environment and effects” (SEE) plan of USA. Research of space effects has two approaches, i.e., space exposure testing and ground based simulation testing. Although the two approaches do not conflict with each other, the ground based simulation testing is a more economic and feasible method available to most researchers, rather than more 107 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 107–113. C 2006 Springer. Printed in the Netherlands.

108 SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION expensive space activities. Since spacecraft are exposed to a complex environment in space, questions arise, for instance, which factors of the environment are mostly affecting the spacecraft? Will different environments act synergistically? Groundbased space environment simulation techniques have gained more attention since the first space flights. In past decades, many different systems were manufactured for environmental tests, covering a wide spectrum from materials to entire satellites. These activities have provided an important guarantee to the reliability of spacecraft. With increase in lifetime of spacecraft and rapid development of manned space flight, more reliability of spacecraft is required due to increased solar activity in recent years. The space-irradiating environment has become severely aggressive, leading to abnormal conditions and damage to many satellites and interstellar space exploration vehicles. This indicates at the limited ability of spacecrafts to withstand the extreme space environment and our insufficient understanding of the space environmental effects on spacecraft. Thus, an increase in research of space environmental effects is an essential path to improve the reliability and lifetime of spacecrafts, as well as to meet developing requirements for future space flight. The present work is aimed at a systematic study of the synergistic effects of vacuum, low temperatures and proton and electron irradiation with energy less than 180 keV on optical quartz glass. The space environment simulating system described in ref. [1] was used for simulating the complex space environment on the ground.

2. Basic Variation in Dyeing of Quartz Glass by Charged Particles Radiation The primary effect of charged particles on optical quartz glass is dyeing. In reference [2] the radiation induced dyeing of optical quartz glass by single particle was investigated. Figure 1 shows the optical spectral transmission of JGS3 quartz glass before and after radiation by protons and electrons. Radiation by protons and electrons results in dyeing of JGS3 optical quartz glass within the ultraviolet to visual wavelength range, while it has little influence within the infrared range. The critical dose of proton radiation that induces dyeing is c = 5 × 1014 part·cm−2 . The transmission is first decreased within the 200–250 nm of ultraviolet zone and then the variation in transmission is shifted to longer wavelengths with increasing radiation dose. Transmission decreases linearly with increasing proton radiation for doses less than 2 ×1016 part·cm−2 . The critical dose for electron radiation induced dyeing is c = 5 × 1014 part·cm−2 with similar variation in dyeing of that under proton radiation. However, the electron induced decrease in transmission is higher than that under proton radiation with the same energy and dose. A critical dose at 60 keV of simultaneous

109

HAI LIU ET AL. 100

90

1

Transmission(%)

Transmission(%)

100

2

80

3

70

4

Ep=80keV 1- before radiation 15 2 2- § = ¶ 2¡ 1 Á0part/cm 16 2 3- § = ¶ 10 part/cm 16 2 4- § = ¶ 2¡ 1 Á0part/cm

60 50 200

300

400

500

600

700

800

90 80 70

2 3

Ep=180keV 1- before radiation 15 2 2- § = ¶ 10 part/cm 15 2 3- § = ¶ 2¡ Á 10part/cm 16 2 4- § = ¶ 10 part/cm 16 2 5- § = ¶ 2¡ Á 10part/cm

4 5

60 50 200

900 1000

1

300

400

500

600

700

800

900 1000

Wavelength(nm)

Wavelength(nm) 100

100

90 80

4

Transmission, %

Transmission, %

e n t re te tx h e e r

1 2 3 Ee=100keV 1- before radiation 15 2 2- § = ¶ 10 /cm 15 2 3- § = ¶ 2¡ 1 Á0/cm 16 2 4- § = ¶ 10 /cm

70 60 50 200

300

400

500

600

Wavelength, nm

700

800

90

1 2 3

80 70

4 Ee=140keV 1- before radiation 14 2 2- § = ¶ 5¡ Á 10/cm 15 2 3- § = ¶ 10 /cm 15 2 4- § = ¶ 5¡ Á 10/cm

60 50 200

300

400

500

600

700

800

Wavelength, nm

Figure 1. Optical spectral transmission of JGS3 quartz glass before and after radiation by protons and electrons

radiation is pc = ec = 1014 part·cm−2 that decreases with increasing particle energy. Under proton and electron irradiation, the variations in optical spectral properties of JGS3 quartz glass share the common characteristic of separated radiation by the two types of particles. The simultaneous radiation affects the optical spectral properties of quartz glass in the 200–800 nm range with little influence in the infrared zone. The critical dose of simultaneous radiation to change the spectrum for JGS3 quartz glass is lower than that found for separate radiation by each single particle, that is, effect on properties has close relationship with particle energy. With increasing energy, the critical radiation dose shows a decreasing trend. Under the simultaneous radiation, there are two distinct absorption bands within the 430–450 nm range and at 540 nm, as well as including the strong absorption band within 200–250 nm that exist simultaneously when the critical dose is reached. The absorption band at 540 nm has no ability to recover. Under a certain energy and radiation dose, the decrease in transmission caused by simultaneous radiation by both type of particles is larger than found under radiation by each type of particles.

Change in transmittance, %

110 SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION 16 15

2

1- combined irradiation, Φ p=Φ e=10 part/cm 2- sum of curres 3 and 4 15 2 3- electron irradiation Φ e=10 part/cm 15 2 4- proton irradiation Φ p=10 part/cm

12

1

8

2

4

3 4 0 200

300

400

500

600

700

800

25 15

20 15

1- combined irradiation, Φ p=Φ e=2*10 part/cm 2- sum of curres 3 and 4 15 2 3- electron irradiation Φ e=2*10 part/cm 15 2 4- proton irradiation Φ p=2*10 part/cm

2

1

10

2

5 0 200

3 4

300

400

500

600

700

800

Change in transmittance,%

Change in transmittance, %

Wavelength, nm 35 16

1- combined irradiation, Φ p=Φ e=10 part/cm 2- sum of curres 3 and 4 16 2 3- electron irradiation Φ e=10 /cm 16 2 4- proton irradiation Φ p=10 /cm

30 25

2

20 2

15 10

4

5

3

0 200

300

Wavelength, nm

1

400

500

600

700

800

Wavelength, nm

Figure 2. Synergistic effect of simultaneous irradiation of quartz glass by protons and electrons with energy of 100 keV (a) = 1015 part·cm−2 ; (b) = 2 × 1015 part·cm−2 ; (c) = 1016 part·cm−2

3. The Synergistic Effect of Complex Radiation Whether there is a synergistic effect from a simultaneous irradiation is always a debatable question. A detailed analysis showed that the variation in transmission of quartz glass is not only dependent on radiation dose but also on the type of radiation. The result from a simultaneous radiation is not equal to the simple sum of effects caused by each type of particles. Figure 2 clearly shows the synergistic effects of proton and electron radiation on quartz glass. Under radiation dose less than = 2 × 1015 part·cm−2 , the variation in transmission caused by simultaneous radiation is obviously larger than the sum of each change induced by the single factor acting with the same radiation dose. Under the radiation dose of about = 2 × 1015 part·cm−2 , they are of comparable level. When this dose is exceeded, the synergistic effects are less than the sum of two single factors. This implies that the synergistic effects of proton and electron radiation on quartz glass are dependent not only on the irradiation type, but also on dose. In other words, the effects of the proton and the electron radiation are enhanced by each other when the dose is less than the critical value, whereas they are weakened by each other, for doses great than the critical value.

111

HAI LIU ET AL.

The existence of synergistic effects due to complex radiation indicates that complex environmental simulation using only one type of particle dose would not produce accurate results. This proves the necessity for developing complex space environment simulation systems.

4. Synergistic Effect Mechanism Under the proton and electron radiation, the main radiation induced absorption bands in quartz glass are at 203–215 and 220–240 nm, corresponding to the E center which is the main radiation defect [3–5]. Here we are going to analyze the mechanism of synergistic effects of simultaneous radiation by protons and electrons and explain the experimental data using figure 3. Assuming the particle flux irradiates the specimen from left to right, Rp and Re are the shot penetration distances of protons and electrons, respectively. Within the shot range, simultaneously applied protons and electrons will act with the outer electron shells of the atom resulting in the atom ionization and formation of secondary electrons and a cavity. Since the energy of the incident protons and electrons is far higher than ionization energy of atoms in target material, the ionization is a complicated cascade process. One incident particle will interact with many atoms in sequence to produce multi-ionization until it looses its energy. On the other hand, the ionized electron Complex interaction region Hydrogen-diffuse region

sample

Beams p++e-

Rp

Re

Figure 3. Schematic presentation of the material irradiated simultaneously by protons and electrons. Rp —penetration depth of protons; Re —penetration depth of electrons

112 SYNERGISTIC EFFECTS OF PROTON AND ELECTRON RADIATION has enough energy to produce secondary ionization. Ionization is an important mechanism to create defect sites, leading to the formation of electron–cavity pairs, which are the key step to form color centers. In case of a simultaneous irradiation by two types of particles, the degree of ionization of an atom in the target material is highly increased, resulting in an exponential increase in the amount of defect sites and electron–cavity pairs. Thus the rate of formation of the color centers under simultaneous irradiation is higher than found under separate irradiations by each type of particles. Before the saturation, a certain concentration of defect sites is reached and the concentration of color centers is mainly dependent on the electron–cavity pairs. For this reason, the number of color centers formed under simultaneous irradiation is larger than that under individual irradiations, which shows an enhancement effect during formation of color centers, as shown in figure 2(a). When radiation dose reaches = 2 × 1015 part·cm−2 , the concentration of color centers is essentially saturated, resulting in the concentration of color centers under simultaneous radiation being at a similar level with the sum result under separate radiations of two types of particles. In other words, the effect of simultaneous radiation is similar to the sum result from each single factor, as shown in figure 2(b). With further increase in radiation dose, the concentration of various color centers is in dynamic equilibrium, with no change in the density of color centers. However, hydrogen atom concentration is continually increased and will participate in the radiation induced physical and chemical damage process according to the above described mechanism. Such a process will damage the structure of the color centers and/or will lead to formation of other types of defects. As a result, the simultaneous irradiation effect is weakened, which shows as a lower variation in transmission under simultaneous irradiation when compared to the sum of the separate irradiation by protons and electrons, as shown in figure 2(c). The hydrogen effect on color centers is not limited to the shot range of the protons. Having good diffusion properties in quartz glass, hydrogen can diffuse deeper influencing the electron affected zone and resulting in more severe damage effect for simultaneous irradiation when the critical radiation dose is exceeded.

5. Conclusions Simultaneous irradiation by protons and electrons induces synergistic effects in the mechanism of quartz glass damage, i.e., the results from simultaneous irradiation are not equal to the simple summation of results from single factor irradiation. The synergistic effect has a strong dependence on dose. When radiation dose is less than the critical value of c = 2 × 1015 part·cm−2 , the synergistic effects result in the variation in transmission being larger than from simple summation of each separate radiation factor with the same dose. When the critical radiation

HAI LIU ET AL.

113

dose is exceeded, the synergistic effect has less affect on the transmission, with the total transmittance being less than the sum result of the two single factors. The existence of synergistic effects justifies the approach of simultaneous irradiation space environmental testing. Hydrogen has particular effects during formation of color centers. Although no hydrogen is contained in JGS3 quartz glass, the implanted proton reacts with some bonds to produce disturbances or transformations for color center of E 1 , reducing the amount of E 1 center and nonbridging oxygen cavities and improving the stability of radiation resistance for quartz glass within the ultraviolet range. The particular effect of hydrogen is the main reason for the difference in results obtained from single proton irradiation and simultaneous proton–electron irradiation. The existence of synergistic effects in simultaneous irradiation indicates that it is difficult obtaining convincing results in any simultaneous realistic conditions using results from irradiation by single factors. This proves the necessity to develop complex space environment simulation systems.

References 1. Abraimov, V. V., Shiqin, Y., Shiyu, H., Dezhuang, Y., Kolybaev, L. K., and Verhovtseva, E. T. (2000) In the Fifth Sino-Russian-Ukrainian Symposium on Space Science and Technology Held Jointly with the First International forum on Astronautics and Aeronautics Symposium Proceedings, Harbin Institute of Technology, Harbin, P.R.China, 2000, pp. 706–713. 2. Liu, H., Geng, H., He, S., Yang, S., Yang, D., Abraimov, V. V., and Wang, H. (2002) In Sixth International Space Conference on Protection of Materials and Structures from Space Environment, Toronto, Canada, 1–3 May 2002, pp. 81–90. 3. Hosono, H., Kawazoe, H., Oyoshi, K., and Tanaka, S. (1994) Journal of Non-Crystalline Solids 179, 39–50. 4. Griscom, D. L. (1980) Physical Review B 22(9), 4192–4201. 5. Marshall, C. D., Speth, J. A., and Payne, S. A. (1997) Journal of Non-Crystalline Solids 212, 59–73. 6. Belostotskyi, V. I. (1998) Journal of Inorganic Materials 34(6), 738–741. (Russian Translation)

THE ROLE OF “ABNORMAL” ELECTRON FLUXES WITH ENERGY <1 MeV IN THE SURFACE CHARGING DOSE OF SPACECRAFT O. R. GRIGORYAN,1 L. S. NOVIKOV,1 V. N. SHEVELVA,1 K. KUDELA,2 V. L. PETROV,1 AND I. V. TCHURILO3 1 Skobeltsyn Institute of Nuclear Physics, Moscow State University Vorobyevy Gory, Moscow, 119899, Russia 2 Institute of Experimental Physik Slovak Academy of Science, 04353, Kosice, Slovakia 3 Rocket-Space Corporation “Energia,” 4a Lenina street, Korolyev 141070, Moscow Region, Russia

Abstract. An analysis of the influence of low energy charged particles onto the surface charging doses of spacecraft was conducted using data from ACTIVE (InterCosmos 24) satellite, SPRUT-VI experiment on board MIR station and the SAMPEX satellite. The surface doses were calculated in the materials up to 10–20 μm in depth. The obtained results were compared against the AE 8 model and a conclusion was reached that the contribution of “abnormal” (i.e., unaccounted in the models electron peaks at L < 2) electron fluxes in the surface dose is insignificant at altitudes less than 600 km. Key words: Surface Charging, Electron Fluxes.

1. Introduction In the present study we continue to analyze the influence of low energy charged particles (mostly electrons with energy up to 1 MeV) on surface layers of materials and film coverings. These particles are regularly registered in the inner radiation belt of the Earth by different spacecraft. We analyze electron fluxes data obtained from ACTIVE (INTERCOSMOS-24) satellite experiment (E e < 500 keV, orbit altitude H = 500–2500 km, orbit inclination I = 81◦ ), SPRUT-VI experiment onboard MIR station (E e = 0.3–1.0 MeV, orbit altitude H = 350 km, orbit inclination I = 51.6◦ ) and SAMPEX satellite experiment (E e > 150 keV, orbit altitude H = 520–670 km, orbit inclination I = 82◦ ). These electrons were measured at geomagnetic index L < 2. The electron flux data were used for calculations of surface dose (thickness of material layer up to 10–20 μm). The distributions of surface dose and electron spectra for different altitudes for 115 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 115–122. C 2006 Springer. Printed in the Netherlands.

116

O. R. GRIGORYAN ET AL.

altitudes 500–1500 km were constructed based on obtained results. The results were compared with results of calculations based on AE 8 model. It is concluded that the contribution of “abnormal” (i.e., unaccounted in the models electron peaks at L < 2) electron fluxes in the surface dose is insignificant at low altitudes (<600 km).

2. Instrumentation In this paper a detailed analysis of electron flux data obtained onboard ACTIVE (INTERCOSMOS-24) satellite is presented. Electrons were registered by single Si-surface barrier detectors. The diameter and the thickness of each detector were 8 mm and 300 μm, correspondingly. The electron detector’s geometric factor was 0.03 cm2 .sr. There were three pairs of detectors that measured electron fluxes at different angles (99◦ , 69◦ , 39◦ ) with respect to zenith axis of the satellite. All detectors were protected by a Mylar foil to stop protons with energy E p < 500 keV. The period between subsequent electron flux measurements was 10 s. The peculiarities of the satellite orbit allowed analysis of electron distribution in a wide range of altitudes. The ACTIVE satellite was launched into a 500–2500 km with an inclination of 81◦ . The detectors installed on the ACTIVE satellite measured electrons in seven energy channels from 30 to 500 keV.

3. Results Electron fluxes under the inner radiation belt were observed in different experiments since 1980s [1–5]. Previous experiments revealed the existence of electrons at L = 1.2–1.8. Figure 1 presents the experimental data obtained in four different experiments a follows: A. SPRUT-VI experiment onboard MIR station, altitude H ∼ 350 km, electron energy Ee > 75 keV [1], B. CORONAS-I satellite experiment, altitude H ∼ 500 km, electron energy E e > 0,5 MeV [2], C. OHZORA satellite experiment, altitude H = 350–850 km, electron energy E e = 0.19–3.2 MeV, D. ACTIVE satellite experiment, altitude H = 500–2500 km, electron energy E e = 30–500 keV. Shaded areas correspond to the flux level of approximately 100 (cm2 .s.sr)−1 . A comparison of the data from mentioned experiments shows that locations of observed electron precipitation zones are similar. These zones have temporal and spatial stability.

THE ROLE OF “ABNORMAL” ELECTRON FLUXES

117

Figure 1. Areas of registration of the increased electron fluxes at L < 2

Figures 2 and 3 present the examples of electron flux registration at SAMPEX satellite at different L-values. Passes of SAMPEX for different days are shown. The existence of electron flux enhancements at L = 1.2–1.8 is obvious from these figures. On figure 3 we see a broad weak peak around L = 1.3, a relatively narrower peak near L = 1.7 and another broader peak near L = 2.5 that are indicative that multiple areas of localized precipitation persisted for at least tens of minutes. The variation of the electron precipitation with L-values on figure 2 shows the same general form from 1.7 < L < 2.5. The narrow peak could be produced

Figure 2. An example of electron flux registration at SAMPEX satellite

118

O. R. GRIGORYAN ET AL.

Figure 3. A south-to-north pass of SAMPEX satellite

by electron interaction with the lightning generated waves as they first cross the magnetic equatorial plane (at relatively low L) and before they are reflected. The differences would be due to the latitude of the lightning, to the density gradients in the ionosphere and the dependence of precipitation flux on L value.

4. Experimental Data from the Satellite “Active” Figure 4 presents the time history of the electron intensity for each energy channel on the ACTIVE satellite. Two narrow peaks at L region from 1.2 to 1.8 with maximums at L = 1.5 and 1.78 are observed.

Figure 4. The time history of the electron flux intensity at ACTIVE satellite

THE ROLE OF “ABNORMAL” ELECTRON FLUXES

119

Figure 5. Electron flux distribution at different altitudes

From the results of experimental data processing it can be suggested that the longitudinal boundaries and shape of electron flux registration are stable in time and space for the different year experiments. This also proves the existence of a constantly operating mechanism. Electrons are registered at low and middle latitudes at L = 1.2–1.8 with maximum intensities at L-values around 1.3–1.4 and 1.7–1.8. The next step was the investigation of the dependence of the electron flux distribution on altitude. The wide altitude range allowed comparing zones of electron registration and electron spectra at different altitudes. Figure 5 presents the electron flux distribution in geographical coordinates at L = 1.2–1.8 for different altitudes. These maps were constructed using ACTIVE satellite data. The position of electron flux precipitation at selected L-values and the dependence of these zones on altitude (500, 700, 900, 1100, and 1300 km) are shown, not taking into account all experimental data at L > 2 and in the region of South Atlantic Anomaly. The value of electron energy is 30–500 keV (every section of the figure is plotted for whole energy range). The following observations could be made analyzing figure 5: (a) zones of electron flux registration under the inner radiation belt exist constantly in an altitude interval from 350 km (see previous figure based on other satellite experiments) up to 1300 km; (b) electrons are observed at L = 1.2–1.8 both in northern and in southern hemispheres; (c) the longitudinal size of the zones practically does not depend on altitude up to 1300 km,

120

O. R. GRIGORYAN ET AL.

Figure 6. Electron spectra at (a) L = 1.2–1.5 and (b) L = 1.6–1.9 plotted for different altitudes from 500 km with an increment of 100 km

at higher altitudes the electron flux becomes larger and it is difficult to locate one. Figure 6 presents electron spectra at different altitudes from 500 to 1000 km in two L-zones; L = 1.2–1.5 (figure 6(a)) and L = 1.6–1.9 (figure 6(b)). As can be seen, the electron flux value increases with the altitude (10 times more at 1000 km altitude than at 500 km). Figure 7 compares results obtained in ACTIVE satellite experiment with other experimental data and with AE8 MIN and AE8 MAX models. Electron spectra obtained onboard MIR station (SPRUT-VI experiment), SAMPEX satellite and ACTIVE satellite experiments are presented. These satellites and the MIR station were launched to different orbits, so there is experimental data for a wide altitude range. SPRUT-VI data give the minimal value of electron flux, because the data were obtained of 350 km altitude. At 1000 km altitude the electron flux has the value of AE8 model flux.

5. Absorbed Radiation Dose Based on the data above, the observed radiation doze values were computed for the ACTIVE satellite orbit for two altitude values (500 and 1000 km). The results of computations of the annual absorbed radiation dose in Si as a function of thickness of an infinite flat Al shield for electron fluxes for 500 km (bottom) and 1000 km altitude (top) are presented in figure 8. As can be seen from figure 8, the contribution of the “abnormal” electron fluxes under the inner radiation belt at L = 1.2–1.9 in the absorbed dose increases with altitude and it is important to take it into account.

THE ROLE OF “ABNORMAL” ELECTRON FLUXES

Figure 7. Electron energy spectra comparing with AE8MAX/MIN models

121

122

O. R. GRIGORYAN ET AL.

Figure 8. Absorbed electron dose for various spectra of electron flux at 500 and 1000 km

6. Conclusions Spacecraft were exposed to impact of charged particles at low and middle latitudes (L-shell value from 1.2 to 1.8). The contribution of “abnormal” electron formations increase with the altitude and are ten times more at 1000 km comparing with the value at 500 km. For more correct estimation of absorbed radiation dose value, the modified AP-8 and AE-8 models for low energy values are necessary.

References 1. Grachev, Y. A., Grigoryan, O. R., Novikov, L. S., and Tchurilo, I. V. (2003) In J. Kleiman and Z. Iskanderova (eds.), Proceedings of ICPMSE-6, 2002, Toronto, Canada. Kluwer Academic Publishers, 2003, pp. 123–130. 2. Biryukov, A., Grigoryan, O., Kuznetsov, S., Ryaboshapka, A., and Ryabukha, S. (1996) Advance in Space Research 17, 189. 3. Bashkirov, V. F., Denisov, Y. I., Gotseluk, Y. V., Kuznetsov, S. N., Myagkova, I. N., and Sinyakov, A. V. (1999) Radiation Measurements 30, 537. 4. Seltzer, S. M. (1979) IEEE Transactions on Nuclear Science NS26, 21–60. 5. Blake, J. B., Inan, U. S., Walt, M., Bell, T. F., Bortnik, J., Chenette, D. L., and Christian, H. J. (2001) Journal of Geophysical Research 106, 733–743.

VACUUM ULTRAVIOLET RADIATION EFFECTS ON DC93-500 SILICONE FILM JOYCE A. DEVER,1 BRUCE A. BANKS,1 AND LI YAN2 1 NASA Glenn Research Center 21000 Brookpark Road, Cleveland, OH 44135 2 University of Nebraska-Lincoln, P.O. Box 880511, Lincoln, NE 68588-0511

Abstract. A space-qualified silicone polymer, Dow Corning (DC) 93-500, has been used as a spacecraft solar cell adhesive and has been proposed for use in a Fresnel lens solar concentrator for space power applications. Future applications of DC93-500 for exterior spacecraft surfaces require an understanding of its overall space environment durability. Vacuum ultraviolet (VUV) radiation is among the space environment elements that can be hazardous to the properties of DC93-500. This paper describes investigations into the effects of VUV radiation on DC93-500 silicone film. Vacuum ultraviolet ellipsometric optical measurements were made on DC93500 silicone to determine the depth of absorption of vacuum ultraviolet light as a function of wavelength. These data indicate the depth within which VUV radiation can cause material degradation. Laboratory VUV exposures were used to examine effects of various VUV exposure wavelength ranges and various VUV exposure intensities to determine whether there exist wavelength or intensity dependencies of degradation. In one set of experiments, transmittance degradation of DC93-500 was examined as a function of exposure to narrow wavelength bands (∼20 nm bandwidth) of VUV in the 140 to 200 nm wavelength range. In another set of experiments, broad spectrum VUV exposures (to wavelengths greater than 115 nm) were used to examine effects of VUV intensity on rates of optical and mechanical properties degradation. Correlations between observed degradation and the measured depth of VUV penetration will be discussed. Key words: Silicone, vacuum ultraviolet radiation, mechanical properties, transmittance 1. Introduction Dow Corning DC93-500 is a space grade silicone elastomer that is commonly used as an adhesive for spacecraft components such as solar cells and optical 123 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 123–140. C 2006 Springer. Printed in the Netherlands.

124

JOYCE A. DEVER ET AL.

solar reflectors and as a general encapsulant or potting material for spacecraft electronic components. Its wide service-temperature range (−115–+200◦ C), flexibility, and low outgassing properties make it desirable for spacecraft applications [1]. Recently, DC93-500, in free-standing film form, has been incorporated into the design of a refractive Fresnel lens solar concentrator for a patented spacecraft solar cell array referred to as the Stretched Lens Array (SLA) [2]. Simulated space environment testing of the DC93-500 material has shown that it is susceptible to ultraviolet radiation degradation [3]. Such degradation requires that DC93-500 surfaces have an ultraviolet protective surface for long-duration durability for space power applications. As part of ongoing studies examining vacuum ultraviolet (VUV) radiation effects on polymer films for spacecraft applications [4–6], this paper describes laboratory investigations on DC93-500 for VUV penetration depth, and wavelength dependence and intensity dependence of vacuum ultraviolet degradation. The air mass zero solar (AM0) spectrum contains ultraviolet (UV) radiation of wavelengths extending down to the Lyman-alpha emissions of hydrogen at 121 nm [7]. The high-energy portion of the ultraviolet spectrum containing wavelengths below approximately 200 nm is generally referred to as vacuum ultraviolet (VUV) radiation. Polymer surfaces on spacecraft are vulnerable to degradation due to incident solar radiation in this wavelength region, which contains short enough wavelengths and thus high enough energies to break bonds in organic molecules [8]. Photochemical reactions within organic molecules may result in effects such as discoloration of the material, which can result in increased solar absorptance, or loss of mechanical properties due to chemical changes in the material. The depth within which degrading effects occur depends upon the depth of penetration of the VUV light. In general, organic polymers absorb VUV within a shallow depth on the order of micrometers or less, although the depth is dependent upon the polymer [9]. Only surface degradation can occur for wavelengths that penetrate a shallow layer of the polymer, whereas for wavelengths that are more penetrating, bulk degradation is possible. Ground testing is important for predicting long-duration durability of spacecraft materials, especially for materials or components that have not been previously tested for long-duration exposure in space. However, differences between the space environment and ground laboratory environment lead to complexities in interpreting the ground test results. Two important differences between space and laboratory vacuum ultraviolet exposure conditions are light intensity, and irradiance spectra. It is desirable to be able to conduct accelerated testing, or testing at intensities greater than those in space, especially for long-duration mission durability predictions. However, there is no conclusive information on the maximum acceleration factor that will produce damage that is realistic compared to the space environment. It is likely that the acceleration factors are material-specific, depending on material chemistry. It is important to determine the maximum

VACUUM ULTRAVIOLET RADIATION EFFECTS

125

intensity that can be used to produce realistic damage in a reasonable amount of time. Differences between the air mass zero solar irradiance spectrum and that of a laboratory vacuum ultraviolet radiation source, when considered along with a materials spectral absorption curve, can also lead to complexities in interpreting results of ground tests. For example, a typical laboratory vacuum ultraviolet source is a deuterium lamp with a magnesium fluoride window. This source has a peak at approximately 160 nm, which does not exist in the air mass zero solar spectrum. If a material is highly absorbing in this wavelength region, the laboratory source might produce damage that is unrealistic compared to space, especially for high intensity levels. This paper describes results of testing to investigate wavelength dependence and intensity dependence of VUV degradation for DC93-500 silicone. The goals of this work were to determine the wavelength ranges that are most damaging to silicone DC93-500 and to determine whether accelerated testing of up to six time the sun’s intensity (i.e., six VUV suns) will produce realistic degradation by comparing results to those obtained using lower intensity levels. In order to best interpret the results relative to the properties of DC93-500 silicone, spectral measurements were made of VUV penetration depth.

2. Experimental Methods 2.1. DC93-500 FILM SAMPLES

Dow Corning® (DC) 93-500 Space Grade Encapsulant two-part silicone elastomer was cast and cured by ENTECH Inc. into a film of approximately 150 μm thickness following a proprietary process for mixing, casting, and curing via a proven temperature/time schedule in an environment free of cure inhibitors. Samples tested were cut from this “stock” film. 2.2. OPTICAL PROPERTIES MEASUREMENTS

Total transmittance of the silicone films was measured using a Perkin-Elmer Lambda-19 ultraviolet-visible-near infrared spectrophotometer equipped with a Labsphere reflectance accessory, which includes an integrating sphere of 150 mm diameter. Measurements were made over a wavelength range of 210–2500 nm. 2.3. MECHANICAL PROPERTIES MEASUREMENTS

Tensile testing was conducted using a DDL Inc. Model 200Q electromechanical test system. Tensile test specimens were punched using a die fabricated to the specifications defined in American Society for Testing and Materials (ASTM)

126

JOYCE A. DEVER ET AL.

Standard D-638 for type V tensile specimens [10]. Whereas the tensile die produces an overall sample length of 63.5 mm, samples for these experiments were trimmed to an overall length of 44.5 mm to better accommodate the VUV exposure chamber dimensions. The trimmed length was simply the excess material that would fall outside the grips of the tensile tester. Sample width is 9.52 mm with a narrow section of 3.18 mm width and 7.62 mm length. The initial grip separation distance was 25.4 mm. Samples were tested using a speed of 63.5 mm·min−1 and load vs. displacement data were obtained. From the load vs. displacement data, ultimate tensile strength (UTS) and elongation at failure were obtained. 2.4. VACUUM ULTRAVIOLET ELLIPSOMETRY

Spectroscopic ellipsometry is a surface-sensitive, nondestructive optical technique widely used to determine film thickness and optical constants. Reflection ellipsometry measures the change in the polarization state of light upon reflection from a sample surface. Measurement results are expressed as psi (Ψ) and delta (), which are related to the complex Fresnel reflection coefficients (R) as shown in eq. (1), ρ ≡ tan (ψ) ei = Rp /Rs

(1)

where p and s correspond to electric field component directions parallel and perpendicular to the plane of incidence, respectively. In this work, the optical constants of index of refraction (n) and extinction coefficient (k) of DC 93-500 silicone in the ultraviolet (UV), including vacuum ultraviolet (VUV), were determined using variable angle spectroscopic ellipsometry (VASE® ). Theory and methods of determining optical constants for materials using VASE have been described in detail elsewhere [11]. A commercial VUVVASE® system (J. A. Woollam Co. Inc.), covering a spectral range from 140–1100 nm, was used. Measurements were performed at two angles of incidence (60 and 70◦ ) and over a spectral range from 140 to 400 nm. Prior to measurement, the silicone film was backside-abraded to eliminate back surface reflections, and the sample was then placed on a glass slide for measurement. Raw VUV-VASE data were fit to optical models for an 8-mil (nominal) silicone film in order to determine the optical constants, n and k. Silicone was represented in the optical model by a sum of four Gaussian oscillators to account for interband absorptions. Surface roughness was modeled by a Bruggeman effective medium approximation (BEMA) layer, assuming 50% material and 50% void. The penetration depth (Dp ) for a material at a given wavelength is a measure of how far a beam of light of that wavelength will penetrate into the material. A propagating beam will be attenuated to 1/e (e is the value at which ln e = 1, approximately 2.7) of its original intensity after propagating a distance equal to

VACUUM ULTRAVIOLET RADIATION EFFECTS

127

the penetration depth. Dp is related to the optical constant k by λ (2) 4π k Using eq. (2), UV light penetration depth in silicone DC93-500 was determined as a function of wavelength from 140 to 400 nm. Dp =

2.5. VACUUM ULTRAVIOLET EXPOSURE FACILITY

The facility used for vacuum ultraviolet exposure has been described in detail elsewhere [6]. The facility uses a cryogenic vacuum pumping system. The test chamber contains four individual VUV exposure areas separated by water-cooled copper walls to minimize cross interactions between compartments. Each exposure area contained a 30-watt VUV deuterium lamp with a magnesium fluoride end-window (Hamamatsu model L7293) which provided broad spectrum VUV with a lower cut-off wavelength of 115 nm. Sample stages, one per area, were motor controlled so that the intensity of VUV light could be adjusted by changing the distance between the VUV light source and the sample. Each exposure area was equipped with a cesium iodide (CsI) phototube, calibrated to a NIST-measured deuterium source, to make measurements of lamp intensity (in the wavelength range of sensitivity of the detector, 115–200 nm) at the sample distance. Lamp intensity measurements were used to determine the “number of VUV suns,” defined as the ratio of lamp intensity to air mass zero solar intensity the 115–200 nm wavelength range. For reference, the air mass zero solar intensity in the 115–200 nm wavelength range is 1.073E-5 W·cm−2 [7]. Multiplying the number of suns by test exposure hours provides the number of “equivalent sun hours,” or ESH, which is the equivalent time for direct air mass zero solar exposure. Additionally, multiplying the lamp intensity (in units of W·cm−2 ) by test exposure duration (in seconds) provides incident energy fluence (in units of J·cm−2 ). These exposure conditions will be reported with material degradation results. Prior to exposure, samples were installed in the facility and the chamber was brought to high vacuum. An operating pressure of approximately 4 × 10−6 Torr was achieved within 24 h of establishing high vacuum prior to commencing ultraviolet exposure. The test chamber was brought to atmosphere using nitrogen in order to remove or replace samples. Separate from the sample exposure tests, thermocouple temperature measurements were made which indicated that the VUV lamps did not cause the temperature of the sample stage at the testing distances to rise above room temperature. A comparison of the air mass zero solar irradiance spectrum and the deuterium lamp irradiance spectrum is shown in figure 1 [12]. The lamp irradiance spectrum was obtained by the National Institute of Standards and Technology using a distance of 25.4 cm from the source to the detector. For the results described in this

128

JOYCE A. DEVER ET AL. 1 E-03

Irradiance (W/cm2 nm)

1 E-04 1 E-05 1 E-06 1 E-07 1 E-08

Sun: ASTM E490 L7293 VUV Lamp

1 E-09 1 E-10 100

200

300

400

Wavelength (nm)

Figure 1. Air mass zero solar irradiance compared to deuterium VUV lamp irradiance at a 25.4 cm source-to-detector measurement distance

paper, the VUV exposure facility used distances that were significantly greater than 25.4 cm, between 50 and 92 cm, so the absolute irradiance values shown in figure 1 are much higher than would be used for typical test conditions. However, from figure 1 it is evident that the spectral shapes are very different between the air mass zero sun and the deuterium lamp. For example, the deuterium lamp shows a peak around 160 nm that is absent in the solar spectrum. Additionally, beyond approximately 170 nm, the deuterium lamp irradiance is much less than that of the sun. Test conditions using this facility for wavelength dependence and intensity dependence studies for silicone DC93-500 degradation are described in the sections below. 2.5.1. Wavelength dependence of silicone DC93-500 degradation Samples of DC93-500 silicone films of approximately 152 μm thickness were exposed to VUV using narrow bandpass filters to isolate various wavelength bands provided by a broad spectrum deuterium lamp in the VUV exposure facility. For each of four DC-93-500 film samples, a magnesium fluoride window was placed over the silicone samples, and the narrow bandpass filter was placed over the magnesium fluoride window with the coating surface of the filter facing the magnesium fluoride window. This configuration was used so that organic contaminants would not have a view of the coating on the filter. Samples were located at a distance of 50.4 cm from the deuterium lamp source. Samples were measured for total transmittance in the 210–2500 nm wavelength range. Filter wavelength ranges and exposure conditions are provided in table 1. Intensity was measured

129

VACUUM ULTRAVIOLET RADIATION EFFECTS TABLE 1. VUV exposure conditions for wavelength dependence of DC93-500 degradation

Filter peak wavelength (nm)

VUV exposure wavelength range (nm), i.e. filter range at halfmaximum transmittance

151

143–159

170

157–184

182

171–193

188

179–197

Test duration (h) 74 408 74 408 74 408 74 408

Avg. intensity for test duration (μW·cm−2 )

Cumulative incident energy fluence (J·cm−2 )

2.7 ± 0.5

0.76 4.0 1.1 5.7 0.17 0.92 0.11 0.6

3.9 ± 0.8 6.2 ± 1 4.0 ± 0.8

Avg. number of suns in filter wavelength range for test duration 17 ± 3 2.0 ± 0.4 0.14 ± 0.03 0.07 ± 0.01

Total equivalent sun hours in filter wavelength range for test duration 1360 7190 150 800 11 57 5.7 30

approximately every 24 to 72 h of exposure. Whereas table 1 indicates the timeaverage intensity and number of suns, additional error exists due to the intensity variation across the sample area, which is approximately ±15% of the measured value. 2.5.2. Intensity dependence of silicone DC93-500 degradation Samples of DC 93-500 silicone film of approximately 152 μm thickness were exposed to VUV above 115 nm wavelength using three different intensity levels to determine rates of transmittance degradation and of mechanical properties degradation. Three exposure compartments in the VUV exposure facility were used in order to obtain the various intensity levels using different source-to-sample distances. These source-to-sample distances include the maximum and minimum distances achievable with the VUV exposure facility and one additional distance in between. Tensile samples were exposed for three exposure durations. Exposure conditions, measured in the VUV wavelength range of 115–200 nm, are shown in table 2. Whereas the deuterium lamp also provides radiation above 200 nm, it is a small fraction of the air mass zero solar intensity above 200 nm and is not accounted for here. Approximately every 24 h of testing, intensity values were measured, lamp windows were cleaned, intensity was remeasured, and exposure resumed. It is necessary to frequently clean lamp windows because even trace amounts of organic contaminants in the test chamber can be easily deposited and fixed on surfaces in the presence of high intensity UV light, and the highest intensity

130

JOYCE A. DEVER ET AL.

TABLE 2. VUV exposure conditions for intensity dependence of DC93-500 degradation

Test area

Source to sample distance (cm)

1

51.3

2

71.6

4

91.6

Exposure duration (h) 118 356 238 118 379 262 117 379 262

Avg. intensity (μW·cm−2 ) 64 ± 5.4 59 ± 6 56 ± 4 34 ± 2 32 ± 3 30 ± 2 18 ± 1 16 ± 2 15 ± 0.7

Cumulative incident energy fluence (J·cm−2 ) 27 75 47 14 43 29 7.6 22 14

Avg. number of suns

Cumulative equivalent sun hours

6.0 ± 0.5 5.5 ± 0.6 5.2 ± 0.4 3.2 ± 0.2 3.0 ± 0.2 2.8 ± 0.15 1.7 ± 0.1 1.5 ± 0.2 1.4 ± 0.1

700 1900 1200 370 1100 740 200 570 370

UV light is at the lamp output window. Because the VUV lamps are located on ports that can be isolated from the main sample chamber, lamp maintenance was conducted while the samples remained under high vacuum conditions. Whereas table 2 indicates the time-average intensity and number of suns, there is an intensity variation across the sample area, which is approximately ±15% of the measured value.

3. Results and Discussion 3.1. EFFECTS OF NARROW VUV WAVELENGTH BANDS ON OPTICAL PROPERTIES DEDRADATION

Figure 2 shows the total transmittance spectra of 152 μm thick DC93-500 films that were exposed to VUV from beneath various narrow bandpass filters, where test conditions are fully described in table 1. Data in figure 2 are shown only for samples exposed to VUV for 408 h, because spectra for samples exposed for 74 h were difficult to distinguish from the spectrum of the pristine sample. The legend for figure 2 indicates the filter peak wavelength and equivalent space exposure in units of ESH for each wavelength band from the exposure conditions listed in table 1. The most significant changes to the spectra are in the UV wavelengths below 300 nm as shown in figure 2(a). Significant degradation was observed for all four exposure wavelength ranges. This indicates that degradation is not exclusively caused by exposure to an individual wavelength or narrow wavelength band. However, data can be further analyzed to establish whether any of these wavelength bands is more effective than others in causing degradation. Table 3 shows data for transmittance at 250 nm for all exposed samples and a pristine

131

VACUUM ULTRAVIOLET RADIATION EFFECTS

Transmittance (%)

90

a b

80

c

d e

70

a: Pristine

60

b: 188 nm, 30 ESH c: 170 nm, 805 ESH

50

d: 151 nm, 7190 ESH e: 182 nm, 57 ESH

40 200

210

220

230

240

250

260

270

280

290

300

Wavelength (nm) (a) 100

Transmittance (%)

90

80 a: Pristine

70

b: 188 nm, 30 ESH c: 170 nm, 805 ESH

60

50 200

d: 151 nm, 7190 ESH e: 182 nm, 57 ESH

700

1200

1700

2200

Wavelength (nm) (b)

Figure 2. Effect of VUV exposure of various wavelength ranges on spectral transmittance of 152 μm DC93-500 silicone film for (a) the wavelength range of 200–300 nm where the most significant degradation occurred and (b) the overall spectral measurement range of 210–2500 nm

sample obtained from the spectral data shown in figure 2. Figure 3 shows a plot of transmittance at 250 nm, a wavelength where significant degradation is evident, vs. incident energy fluence (listed in table 1) for all exposures. It is evident from this figure that the VUV filters providing longer VUV wavelengths cause a higher rate

132

JOYCE A. DEVER ET AL.

TABLE 3. Effect of VUV exposure of various wavelength ranges on transmittance at 250 nm of 152 μm DC93-500 film VUV exposure conditions Filter peak wavelength (nm)

VUV exposure wavelength range (nm)

Cumulative incident energy fluence (J·cm−2 )

Equivalent sun hours in filter wavelength range

Transmittance measured at 250 nm

151

143–159

170

157–184

182

171–193

188

179–197

0.76 4 1.1 5.7 0.17 0.92 0.11 0.6 0

1360 7190 150 800 11 57 5.7 30 0

85.80 75.71 87.61 78.55 87.56 75.87 86.97 79.42 87.51

Pristine Sample

of transmittance degradation than the shorter VUV wavelengths. The most likely explanation for this is that the longer wavelengths are more deeply penetrating into DC93-500 and, therefore, affect more of the bulk of the material. Depth of penetration of VUV radiation relative to these observations will be further discussed in section 3.2. 100

Transmittance

80 60 151 nm filter 170 nm filter 182 nm filter 188 nm filter 151 nm filter 170 nm filter 182 nm filter 188 nm filter

40 20 0 0

1

2

3

4

5

6

Incident Energy Fluence (J/cm 2 ) Figure 3. Transmittance of 152 μm DC93-500 silicone film at 250 nm as a function of incident VUV energy fluence provided by exposure to various wavelength ranges through narrow bandpass filters

133

VACUUM ULTRAVIOLET RADIATION EFFECTS

Penetration depth, μ m

1000

100

10

1

0.1

0.01 125

150

175

200

225

250

275

Wavelength, nm

Figure 4. Depth of penetration of VUV in pristine DC93-500 Silicone Film

3.2. VUV ELLIPSOMETRY

Figure 4 shows VUV penetration depth as a function of wavelength for DC93-500 film. Because radiation must be absorbed in order to cause degradation, wavelength dependence of VUV degradation, observed in figure 3, can be related to the depth to which VUV can penetrate as a function of wavelength. Below 175 nm, penetration is limited to tenths of μm, but increases rapidly with increasing wavelength above 175 nm. Based on the very shallow penetration depth for wavelengths below 175 nm, it is evident that optical degradation caused by VUV below 175 nm would be limited to surface degradation. Table 1 indicates the wavelength bands for the narrow bandpass filters through which DC93-500 was exposed to VUV radiation. As shown in the table, the 151 nm filter and the 170 nm filter transmit VUV wavelengths of no greater than 185 nm. Based on the data in figure 3, indicating the slow rate of degradation for DC93-500 exposed beneath filters which transmit up to 185 nm, and the significantly more rapid rate of degradation for samples exposed beneath filters which transmit up to approximately 200 nm, it can be concluded that the wavelengths between approximately 185 and 200 nm are more effective in causing DC93-500 transmittance degradation. These wavelengths can penetrate from 1 to 3 μm in depth in DC93-500. 3.3. EFFECTS OF VUV OF VARIOUS INTENSITIES ON OPTICAL AND MECHANICAL PROPERTIES OF DC93-500

Figure 5 shows the total transmittance spectra of 152 μm thick DC93-500 films exposed to VUV of various intensity levels and for various equivalent sun hours.

134

JOYCE A. DEVER ET AL. 100 (a)

90

Transmittance (%)

80 (b)

70

(c) (d)

60 50

(e) (a) Pristine

(f)

40

(b) 200 ESH, 1.7 suns

(g)

(c) 370 ESH, 3.2 suns

30

(d) 570 ESH, 1.5 suns

20

(e) 700 ESH, 6.0 suns (f) 1120 ESH, 3.0 suns

10

(g) 1930 ESH, 5.5 suns

0 200

300

400

Wavelength (nm) (a)

100 90

Transmittance (%)

80 70 60 (a) Pristine

50

(b) 200 ESH, 1.7 suns

40

(c) 370 ESH, 3.2 suns (d) 570 ESH, 1.5 suns

30

(e) 700 ESH, 6.0 suns

20

(f) 1120 ESH, 3.0 suns (g) 1930 ESH, 5.5 suns

10 0 200

400 600

800 1000 1200 1400 1600 1800 2000 2200 2400 Wavelength (nm)

(b)

Figure 5. Effect of VUV exposure of various intensities on spectral transmittance of 152 μm DC93-500 silicone film for (a) the 200–400 nm spectral range and (b) the 210–2500 nm spectral range

Samples were exposed to the unfiltered deuterium lamp output, wavelengths above 115 nm, and output is reported in the 115–200 nm band, which contains the majority of lamp output as shown in figure 1. It is evident that the most significant degradation occurred in the ultraviolet wavelengths, below 400 nm, as is shown in

135

VACUUM ULTRAVIOLET RADIATION EFFECTS TABLE 4. Effect of VUV exposure (>115 nm) of various intensities on transmittance at 300 nm of 152 μm DC93-500 film Intensity (number of VUV suns)

Exposure level (Thousands of ESH)

Transmittance at 300 nm (%)

1.5

0.2 0.37 0.57 0.7 1.12 1.93 0

79.10 70.77 66.24 64.50 55.43 48.76 92.50

3 5.5 0

figure 5(a). It is also evident that degradation increases with increasing equivalent sun hours of exposure. Table 4 and figure 6 show transmittance at 300 nm (a wavelength at which significant transmittance degradation of DC93-500 is evident, based on figure 5) as a function of exposure, represented by thousands of ESH, for various intensities (number of suns). In the figure, symbols indicate the measured data points, and lines indicate exponential decay curve fits. In addition to examining data for individual intensities, data were considered all together, independent of intensity, as one of the curves shown in figure 6. 100

Transmittance, %

80 60 All Data 1.5 suns 3.0 suns 5.5 suns All Data 1.5 suns 3.0 suns 5.5 suns

40 20 0 0.0

0.5

1.0

1.5

2.0

Thousands of ESH

Figure 6. DC93-500 film transmittance at 300 nm wavelength as a function of VUV exposure duration expressed in thousands of equivalent sun hours

136

JOYCE A. DEVER ET AL.

When transmittance data for all intensities are considered together, the best curve fit equation is given in the format shown in eq. (3): τ = y0 + a exp(−kx)

(3)

where τ is transmittance (%), x is exposure duration (thousands of equivalent sun hours), and y0 , a, and k are constants. Eq. (3) represents an exponential decay reaching an asymptotic value, which is evident from examination of figure 5. We can further determine the expression for transmittance of unexposed materials, or τ0 , by setting x = 0: τ0 = y0 + a

(4)

Using the expression for “a” as a function of y0 and τ0 from eq. (4), and by expressing x (thousands of equivalent sun hours) as the product of equivalent suns, S, and thousands of exposure hours, h, eq. (3) can be rewritten as τ = y0 + (τ0 − y0 ) exp(−k Sh)

(5)

In order to determine whether the rate of transmittance decay shows a dependence upon S, the intensity (expressed as number of equivalent suns), eq. (5) is solved for k, the constant in the exponential expression, to give: 1 τ − y0 k=− (6) ln Sh τ0 − y0 The curve fit for all measured transmittance data at 300 nm (figure 6) produces a value of y0 = 46.7, which is the asymptote being approached by the decay in transmittance and which is assumed to be constant, independent of intensity. If we examine transmittance data at 300 nm for all intensities and plot k vs. S, it is possible to determine whether k is constant, which would mean it is independent of intensity, or whether it varies as a function of intensity, indicating intensity dependence. The values for S and h are obtained from the VUV exposure conditions (table 2), and τ and τ0 values are the measured transmittance data (shown in table 4). The plot of k vs. S is shown in figure 7. Based on the data shown in figure 7, it is evident that there is no statistically significant dependence upon exposure intensity between approximately 1.5 and 5.5 VUV suns. Table 5 and figures 8(a) and 8(b) show mechanical properties of 152 μm DC93500 silicone films as a function of exposure (equivalent sun hours). Both ultimate tensile strength (figure 8(a)) and elongation at failure (figure 8(b)) decrease with increasing exposure and indicate the approach of an asymptotic value near the 2000 ESH exposure level. An exponential decay curve fit is shown in each figure. Based on these data, especially because of similar degradation for similar ESH, regardless of intensity, there is no clear trend indicating an intensity dependence upon the rate of mechanical properties degradation.

137

VACUUM ULTRAVIOLET RADIATION EFFECTS 2.5 2

k

1.5 1 0.5 0 0

1

2

3

4

5

6

S (Number of Suns)

Figure 7. Coefficient, k, described by eq. (6), as a function of VUV exposure intensity for DC93-500 film samples

4. Conclusions DC93-500 silicone has been found to undergo degradation in optical and mechanical properties upon exposure to a laboratory deuterium lamp providing VUV radiation of wavelengths greater than 115 nm. In one experiment, samples of DC93-500 films were exposed to narrow bands of VUV radiation (∼20 nm) in wavelength ranges between approximately 140 and 200 nm by using a broad spectrum deuterium lamp as the VUV source and narrow bandpass filters over TABLE 5. Effect of VUV exposure of various intensities on mechanical properties degradation of DC93-500 Avg. number of VUV suns 5.5

3

1.5 0 (pristine)

Exposure duration (h)

Cumulative VUV equivalent sun hours

No. test samples

Avg. UTS (MPa)

118 356 238 118 379 262 117 379 262 0

700 1900 1200 370 1100 740 200 570 370 0

2 3 3 2 3 3 3 3 3 5

5.9 ± 0.1 3.8 ± 2.0 3.6 ± 0.4 6.2 ± 0.8 4.3 ± 0.6 5.3 ± 0.9 6.2 ± 0.4 5.0 ± 1.0 5.6 ± 1.6 9.2 ± 0.5

Avg. elongation (%) 79.0 ± 2.8 65.3 ± 2.0 69.3 ± 2.1 86.0 ± 2.8 69.0 ± 0 76.3 ± 2.1 91.3 ± 2.5 75.0 ± 4.4 82.3 ± 6.1 139 ± 6.9

138

JOYCE A. DEVER ET AL. 10.0 5.5 suns 3 suns 1.5 suns

9.0 8.0

UTS (MPa)

7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 0

200

400

600

800

1000

1200

1400

1600

1800

2000

VUV Exposure (ESH)

(a) 160.0 5.5 suns 3 suns 1.5 suns S i 1

Elongation at Failure (%)

140.0 120.0 100.0 80.0 60.0 40.0 20.0 0.0 0

200

400

600

800

1000

1200

1400

1600

1800

2000

VUV Exposure (ESH)

(b) Figure 8. Mechanical properties of (a) tensile strength and (b) elongation at failure for 152 μm DC93-500 silicone as a function of VUV exposure equivalent sun hours

the DC93-500 samples. Results indicated that each wavelength range used for the exposures produced degradation in DC 93-500 optical properties. However, degradation per incident energy fluence indicated the highest rate of degradation for samples exposed from beneath filters which included wavelengths above 185 nm. Vacuum ultraviolet ellipsometric optical measurements were made on DC93-500 silicone to determine the depth of penetration of vacuum ultraviolet light as a function of wavelength. Data showed that vacuum ultraviolet of wavelengths below

VACUUM ULTRAVIOLET RADIATION EFFECTS

139

185 nm penetrate DC 93-500 to depths no greater than 1 μm, indicating that VUV degradation of these wavelengths can only occur in a shallow layer compared to typical spacecraft polymer film applications which use films on the order of tens to over a hundred micrometers in thickness. Compared to VUV exposures to wavelengths below 185 nm, a significantly more rapid rate of transmittance degradation was observed in DC93-500 for VUV exposures which included wavelengths between 185 and 200 nm, which correspond to depths of VUV penetration between 1 and 3 μm. In another experiment, the rates of optical and mechanical properties degradation for DC93-500 films were examined for exposures to broad spectrum VUV (above 115 nm) of various intensities. It was found that for both transmittance degradation and mechanical properties (ultimate tensile strength and elongation at failure) degradation, loss of these properties followed exponential decay functions approaching asymptotic values. Examination of the data indicated no clear dependence of degradation on the intensity of exposure within a range of intensities between 1.5 and 5.5 VUV suns. The lack of intensity dependence in these data indicates that DC93-500 can be tested using VUV intensities as high as about 5.5 suns without causing significantly different degradation rates compared to near real-time exposure rates. It remains to be determined whether these rates of degradation are similar to those caused by actual space exposure, especially considering the significantly different spectra between the Sun and the laboratory VUV source. It is hoped that space exposure data will eventually be available to make such comparisons:

Acknowledgments The authors gratefully acknowledge the technical support of Frank Lam and James Mazor (Akima), Michael DePauw (NASA), Scott Panko and Edward Sechkar (QSS Group, Inc.), Michael Piszczor (NASA) and Mark O’Neill, Don Spears, and A. J. McDanal (ENTECH, Inc.).

References 1. Dow Corning Corp. (2001) Product Information for Dow Corning® Space-Grade Silicone Sealants, 2001. 2. O’Neill, M. J., Piszczor, M. F., Eskenazi, M. I., McDanal, A. J., George, P. J., Botke, M. M., Brandhorst, H. W., Edwards, D. L., and Hoppe, D. T. (2003), In International Symposium on Optical Science and Technology, SPIE’s 48th Annual Meeting, SPIE Paper no. 5179–17, August 2003. 3. Edwards, D. L., Finckenor, M. M., O’Neil, M., and McDanal, A. J. (2000) In 8th International Symposium on Materials in a Space Environment, Arcachon, France, 2000.

140

JOYCE A. DEVER ET AL.

4. Dever, J., Messer, R., Powers, C., Townsend, J., and Wooldridge, E. (2001) High Performance Polymers 13(3), S391–S399. 5. Dever, J., Semmel, C., Edwards, D., Messer, R., Peters, W., Carter, A., and Puckett, D. (2002)Radiation Durability of Candidate Polymer Films for the Next Generation Space Telescope Sunshield, AIAA 2002-1564, April 2002. 6. Dever, J. and McCracken, C. (2004) High Performance Polymers 16(2), 289–302. 7. American Society for Testing and Materials (Reapproved 1992). Solar Constant and Air Mass Zero Solar Spectral Irradiance Tables, ASTM-E 490-73a. 8. Dever, J. A. (1994) Flight-vehicle materials, structures, and dynamics—Assessment and Future Directions, Vol. 2, Advanced Metallics, Metal-Matrix and Polymer-Matrix Composites, American Society of Mechanical Engineers, New York, 1994, pp. 422–433. 9. Adams, M. R. (1993) The Degradation of Polymeric Spacecraft Materials by Far-UV Radiation and Atomic Oxygen, UMI Dissertation Services, Ann Arbor, MI, 1993, p. 138. 10. American Society for Testing and Materials. (1995) Standard Test Method for Tensile Properties of Plastics, ASTM D 638–95. 11. Herzinger, C. M., Snyder, P. G., Johs, B., and Woollam, J. A. (1995) Journal of Applied Physics 77(4), 1715–1724. 12. Dever, J. A., Pietromica, A. J., Stueber, T. J., Sechkar, E. A., and Messer, R. K. (2001) Simulated Space Vacuum Ultraviolet (VUV) Exposure Testing for Polymer Films, AIAA Paper no. 2001– 1054, American Institute of Aeronautics and Astronautics, January 2001.

ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION OF SPACECRAFT POLYMERIC MATERIALS BY SIMULTANEOUS ULTRAVIOLET EXPOSURE KUMIKO YOKOTA,∗ NOBUO OHMAE, AND MASAHITO TAGAWA∗ Department of Mechanical Engineering, Faculty of Engineering, Kobe University, Rokko-dai 1-1, Nada, Kobe 657-8501, Japan

Abstract. Synergistic effect on atomic oxygen-induced erosion of polyethylene and polyimide with 172 nm vacuum ultraviolet was investigated using a quartz crystal microbalance. In order to adjust the relative intensity of atomic oxygen and vacuum ultraviolet, the sample was rotated with an axis perpendicular both to the axes of atomic oxygen and ultraviolet. The erosion rate of polymers by ultraviolet exposure alone is independent of the incident angle of ultraviolet, whereas that by atomic oxygen alone follows cosine function. It was observed that the erosion rate of polyethylene increased 30–100% by a simultaneous exposure of 172 nm vacuum ultraviolet and 5 eV atomic oxygen depending on the relative intensity. The erosion of oxygen covered-polyethylene was three times greater than that of nonoxidized polyethylene. These erosion properties suggested that two independent erosion pathways exist in a simultaneous atomic oxygen and ultraviolet exposure condition; the atomic oxygen-induced erosion and the ultraviolet-induced erosion associated with oxidation. Key words: atomic oxygen, ultraviolet, low Earth orbit, synergy, space environment effect

1. Introduction There are many environmental factors in low Earth orbit (LEO) such as microgravity, thermal cycling, plasma environment (ions and electrons), ultraviolet, radiation, neutral gas, high-energy charged particles, and space debris. However, it has been recognized that atomic oxygen is one of the most important hazards to

∗ Corresponding

authors

141 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 141–152. C 2006 Springer. Printed in the Netherlands.

142

KUMIKO YOKOTA ET AL.

the spacecraft polymeric materials in LEO. Moreover, synergistic effect of atomic oxygen and other space environmental factors on polymer erosion is unsolved problem in space engineering. Since atomic oxygen fluence in a material exposure testing is measured by the erosion depth or mass loss of a polymer witness sample, the enhancement of atomic oxygen-induced erosion of witness sample by simultaneous ultraviolet or other environmental factors spoils the accuracy of atomic oxygen fluence measurement. Polyimide has been widely used as a witness sample, but polyethylene is also a candidate of a reference material because of its simplest possible chemical structure in all polymers [1]. The reaction efficiencies of polyethylene and polyimide with atomic oxygen in LEO have been established to be 3.7 × 10−24 and 3.0 × 10−24 cm3 ·atom−1 , respectively [2]. The fluence of atomic oxygen in LEO is calculated from the polyimide erosion with the established reaction efficiency. This analysis is based on the hypothesis that polyimide is not influenced by the synergistic effect of atomic oxygen and ultraviolet [3, 4]. On the other hand, in a ground based experiment a 30% increase in erosion of a polymer by simultaneous exposure to ultraviolet radiation and atomic oxygen is reported [5]. The erosion properties of polyethylene and polyimide under various space environmental factors need to be well understood because they are used as a reference material. In this paper, we focused on the effect of the relative intensity between atomic oxygen and 172 nm ultraviolet on the erosion of polyethylene and polyimide in ground-based experiments. The relative intensity of atomic oxygen and ultraviolet was adjusted by changing the incident angle of atomic oxygen and ultraviolet with respect to the sample surface. The erosion rate of atomic oxygen-exposed polymers was measured from the change in resonant frequency of quartz crystal microbalance (QCM) during atomic oxygen and/or ultraviolet exposures. The surface properties of atomic oxygen-exposed polymers were analyzed by X-ray photoelectron spectroscopy (XPS).

2. Experimental Details The samples used in this experiment were low-density polyethylene (LDPE) and polyimide films. Both of the films were spin-coated on a QCM sensor crystal. The polyethylene solution containing 0.3 g of LDPE (average molecular weight: 6500) in 40 ml xylene was prepared for polyethylene film. The pyromelliticdianhydride (PMDA)-oxydianiline (ODA) polyimide, which was supplied by Toray Industries Inc. (Semicofine SP-510), was used as a polyimide sample. Precursor of PMDAODA polyimide was spin-coated on a QCM sensor crystal and then annealed at 150◦ C and at 300◦ C. Details of the sample preparation are reported in [6]. The polyimide film, thus prepared, showed an XPS spectrum similar to that of Kapton-H film, which is commercially available polyimide film.

ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION

143

XPS

Temperature-controlled Rotatable QCM

Nozzle AO

PSV Laser

QMS TOF System

Au Mirror UV Source (excimer 172 nm)

Sample

Figure 1. A schematic drawing of the space environmental simulation facility using a laser detonation atomic oxygen beam source

A space environment simulation facility at Kobe University was used in this study. The schematic drawing of the facility is shown in figure 1. This facility equipped with a laser detonation atomic oxygen beam source, which was originally designed by Physical Sciences Inc., as a hyperthermal atomic oxygen source. The atomic oxygen beam was produced using a carbon dioxide laser (wavelength: 10.6 μm, output power: 5–7 J·pulse−1 ). The translational energy of the atomic oxygen was approximately 5 eV. The flux of the atomic oxygen beam was calculated to be 2 × 1014 atoms·cm−2 ·s−1 at the sample position. These values are almost equivalent to those in LEO at the altitude of 200–300 km. An excimer light source with a wavelength of 172 nm was used as an ultraviolet UV source in this study. This ultraviolet source was attached to the atomic oxygen source chamber. The sample was exposed to the UV through an evacuated light guide in order to avoid absorption of vacuum ultraviolet by air (wavelength of <200 nm). The axes of the atomic oxygen beam and the ultraviolet crossed at 90◦ and the sample was rotatable with the axis perpendicular both to atomic oxygen and ultraviolet in order to change the relative intensities of atomic oxygen and ultraviolet (see figure 1). The erosion rate of polymer film was calculated from the change in the resonant frequency of QCM during the atomic oxygen beam and/or ultraviolet exposures. The sample temperature was kept at 311 K during the experiment by the temperature-controlled circulating water system. Before the mass loss measurements, the polyethylene and polyimide films were exposed to atomic oxygen to saturate the surface oxygen content of the sample. This is in order to avoid the nonlinear effect in mass loss, which appears at the beginning of the atomic oxygen exposures at pristine polymer surfaces [7].

144

KUMIKO YOKOTA ET AL. 25 0 deg. 10 deg. 20 deg. 30 deg. 40 deg. 50 deg. 60 deg. 70 deg. 80 deg. 90 deg.

Frequency Shift (Hz)

20

15

10

5

0

0

50

100

150

200

250

Exposure Time (s)

Figure 2. Resonant frequency shift of the polyethylene-coated QCM during atomic oxygen exposures at impingement angles from 0 to 90◦ . Atomic oxygen flux: 5.9 × 1014 atoms·cm−2

3. Results and Discussion 3.1. DEPENDENCE ON INCIDENT ANGLE OF ATOMIC OXYGEN EROSION

In this study, the relative intensity of atomic oxygen and ultraviolet was adjusted by rotating QCM sample along the axis that is perpendicular to the axes of both the atomic oxygen and the ultraviolet. However, in this experimental configuration, the incident angle of atomic oxygen at polymer surface was changed when relative atomic oxygen/ultraviolet intensity was adjusted. Therefore, it was necessary to investigate the effect of incident angle of atomic oxygen and ultraviolet on the polymer erosion prior to the quantitative analysis of synergistic effect of atomic oxygen and ultraviolet. Figure 2 shows the frequency shift of polyethylene-coated QCM during atomic oxygen exposures at incident angles from 0 to 90◦ . The incident angle of 0◦ means that atomic oxygen hit the sample surface at normal incidence. As can be seen in figure 2, a good linear relationship between the frequency shift and exposure time was observed at all incident angles. The slope of the frequency shift, i.e., the mass loss rate of polyethylene film, was calculated by a least square fit and plotted against each incident angle. The result is shown in figure 3. It is observed that the erosion rate decreases with increasing incident angle and follows a cosine distribution, which is indicated by the solid line. The data point at 80◦ did not fit the cosine curve because the QCM holder partially blocked atomic oxygen beam. The experimental result on incident angle dependence of atomic oxygen-induced erosion of polyethylene showed a similar tendency with those reported on polyimide film [8]. The fact that the incident angle dependence of the erosion rate follows a cosine law indicates that the erosion rate of polyethylene is proportional to the effective flux of atomic oxygen; namely, the reaction yield of atomic oxygen with polyethylene is constant at all incident angles.

ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION

145

0.07

Erosion Rate (Hz/s)

0.06 0.05 0.04 0.03 0.02 0.01 0.00

0

10

20

30

40

50

60

70

80

90

100

Impingement Angle (degree)

Figure 3. Erosion rate of the polyethylene-coated QCM as a function of impingement angle

Figure 4 shows an atomic force microscopy (AFM) image of a polyethylene film exposed to atomic oxygen fluence of 3.0 × 1018 atoms·cm−2 . The surface of polyethylene was roughened by atomic oxygen exposure. The peak-to-valley height of the surface was in the order of 150 nm. Therefore, an actual incident angle of oxygen atom to a surface moiety is not directly related to macroscopic incident angle, because surface is not atomically flat due to the presence of microscale roughness even though the macroscopic incident angle is fixed. In addition, the “multiple bouncing effect” also fosters the independency of incident angle. Therefore, due to the microscopic roughness and the multiple bouncing effect at the polyethylene surface, the macroscopic erosion phenomena of polyethylene is simply reflected the effective fluence of atomic oxygen which follows cosine law with the macroscopic incident angle [7]. It should also be mentioned that the tilting of the sample reduces not only the effective flux of atomic oxygen, but also the effective incident energy (normal component). Lowering the incident angle from

Figure 4. The atomic force microscope image of atomic oxyge-exposed polyethylene surface. Atomic oxygen fluence: 3.0 × 1018 atoms·cm−2

146

KUMIKO YOKOTA ET AL.

0 to 60◦ leads to the reduction of effective incident energy from 5.0 to 1.25 eV; it is proportional to the cosine square of the incident angle. The activation energies of gasification reaction of polymers in this incident energy range are expected as low as 10−3 eV [9]. It was, thus, concluded that the effect of activation energies in the incident angle dependence of erosion rates could be negligibly small. From the experimental results shown here, it is concluded that the flux control of atomic oxygen can be achieved by tilting the sample with respect to the beam axis. 3.2. DEPENDENCE OF EROSION BY ULTRAVIOLET ON THE INCIDENT ANGLE

Effect of the incident angle of ultraviolet radiation on polyethylene erosion was also analyzed. The results are shown in figure 5. The flux of 172 nm ultraviolet radiation was 0.55 mW·cm−2 . The incident angle of ultraviolet was taken to the surface normal. Solid and empty circles indicate the results for polyethylene and polyimide, respectively. Note that the sample surface was exposed to atomic oxygen (fluence: 1.2 × 1018 atoms·cm−2 ) prior to ultraviolet irradiation in order to avoid the nonlinear phenomena. It is observed in figure 5 that the effects of incident angle of ultraviolet on the erosion rate of polymers are not significant compared to those of atomic oxygen, even though the ultraviolet flux geometrically follows cosine distribution when the sample is rotated. These results are probably due to the fact that the penetration depth of ultraviolet is greater than the film thickness (0.1 μm) and ultraviolet could reach the polymer/gold interface and being scattered [10]. The surface and interfacial roughness of polyethylene may play a critical role to cancel the angular dependence of ultraviolet and atomic oxygen. 0.10

Erosion Rate (Hz/s)

0.08

UV, AO q

0.06

0.04

0.02

0.00

0

10

20

30

40

50

60

70

80

90

Impingement Angle (degree)

Figure 5. Erosion rate of the polyethylene-coated and polyimide-coated QCM. : polyethylene erosion by AO, •: polyethylene erosion by UV, : polyimide erosion by AO, ◦: polyimide erosion by UV. AO Flux: 6.0 × 1014 atoms·cm−2 /s (polyethylene), 6.8 × 1014 atoms·cm−2 /s (polyimide) UV Flux: 0.55 mW/cm2 (polyethylene and polyimide)

147

ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION

AO q 90°

UV

Figure 6. The set up of AO beam and UV: The axes of AO and UV were crossed at 90 degree, so that relative intensity of AO and UV can be adjusted

3.3. ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION BY SIMULTANEOUS ULTRAVIOLET EXPOSURE

A quantitative analysis on the synergistic effect of atomic oxygen and ultraviolet was conducted using the crossed oxygen atom and ultraviolet beam configuration shown in figure 6. A 172 nm ultraviolet source was installed perpendicular to the atomic oxygen beam line, and the sample was rotatable with the rotating axis perpendicular both to the atomic oxygen and ultraviolet beam axes. In this configuration, the effective atomic oxygen and ultraviolet fluxes can be changed by rotating the sample as indicated in the previous section. The maximum atomic oxygen flux and ultraviolet flux at the sample position were 6.0–14.0 × 1014 atoms·cm−2 ·s−1 and 0.55 mW·cm−2 , respectively. The rate of frequency shift of polyimide-coated QCM and of polyethylene as a function of incident angle of atomic oxygen under simultaneous atomic oxygen and ultraviolet exposures are shown in figures 7(a) and 7(b), respectively. The abscissa is the incident angle of atomic oxygen. Namely, the incident angle of 0◦ means 100% AO and 0% ultraviolet exposure. 0.10

(a)

0.08

0.06

0.04

0.02

0.00

(b)

0.08

Erosion Rate (Hz/s)

Erosion Rate (Hz/s)

0.10

0.06

0.04

0.02

0

10

20

30

40

50

60

70

Impingment Angle of AO (degree)

80

90

0.00

0

10

20

30

40

50

60

70

80

90

Impingment Angle of AO (degree)

Figure 7. Erosion rate of the polyimide-coated (a) and polyethylene-coated (b) QCM during only atomic oxygen (·) and atomic oxygen & ultraviolet exposures (·). (a) AO: 6.8 × 1014 atoms·cm−2 /s, UV: 0.40 mW/cm2 , (b) AO: 6.0 × 1014 atoms·cm−2 /s, UV: 0.55 mW/cm2

148

KUMIKO YOKOTA ET AL.

Normalized Erosion Rate

4

3

2

1

0 0

1

2

3

Relative Intensity of UV//AO (x 10-15 mJ//atom)

Figure 8. The relationship between the relative intensity of UV/AO and the normalized erosion rate of polyethylene and polyimide

As reported elsewhere [11], the synergistic effect of atomic oxygen and 172 nm ultraviolet on polyimide was obvious when ultraviolet intensity is high compared with atomic oxygen; (UV/AO ratio >1 × 10−14 mJ·atom−1 ). However, this synergistic effect was not observed in the range of UV/AO ratio < =10−15 mJ·atom−1 , as demonstrated in figure 7(a); atomic oxygen-induced erosion data with ultraviolet (empty circle) and without ultraviolet (solid circle) are identical. In contrast, polyethylene shows clear effect of simultaneous ultraviolet irradiation even in the range of UV/AO ratio < =10−15 mJ·atom−1 as shown in figure 7(b); i.e., the erosion rates with ultraviolet exposure (empty circle) are 30–100% greater than those without ultraviolet exposure (solid circle). From the result of figure 7, the relationship between the relative intensity of UV/AO and flux normalized erosion rate of polyethylene was replotted in figure 8. The ordinate was normalized erosion rate, namely that means the erosion rate in the case of atomic oxygen and ultraviolet exposures was divided by those in the case of only atomic oxygen exposure. The erosion rate of atomic oxygeninduced polyethylene was enhanced 150–180% at the relative intensity of UV/AO 0.5–2.6 × 10−15 mJ·atom−1 by simultaneous ultraviolet exposure. It was observed that the erosion rate increased 300% when the ultraviolet intensity was high (not shown). In the case of polyimide, the similar effect by simultaneous ultraviolet exposure was observed at relative intensities one order greater than that of polyethylene [11]. This finding suggests that polyimide is a better material as a witness sample for measuring atomic oxygen fluence in LEO since synergistic effect was only in high UV/AO conditions.

ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION

149

0.025 Residual VUV only least square fits

Erosion Rate (Hz/s)

0.020

0.015

0.010

0.005

0.000 0

10

20

30

40

50

60

70

80

90

100

Incident Angle (deg.)

Figure 9. Erosion Rate of polyethylene-coated QCM under VUV exposure (◦), and the contribution of VUV in the synergistic effect (•). Note that the abscissa represents the incident angle of atomic oxygen

The erosion rates of polyethylene by atomic oxygen beam exposure (solid circle in figure 7(b)) were subtracted from those in simultaneous exposures of atomic oxygen and ultraviolet (empty circle). The results are shown in figure 9. As shown in figure 9, the result is independent of the incident angle at angles greater than 20◦ where the sample surface is fully exposed to UV. The empty circles in figure 9 indicate the erosion rates of polyethylene only by ultraviolet; no atomic oxygen exposure (same data in figure 5). Again, the erosion rates are constant with the incident angle of ultraviolet. Beside the independency of the incident angles, it is obvious that the absolute erosion rates in figure 9 are not consistent, i.e., the results are three times greater than the erosion by ultraviolet only. Even though the structure of low-density polyethylene does not absorb the ultraviolet at 172 nm, it is clear that the presence of atomic oxygen enhanced ultraviolet-induced erosion of polyethylene. Oxidized radicals such as carbonyl or carboxyl group formed by the atomic oxygen-induced oxidation of polyethylene may have a critical role on synergistic effect of atomic oxygen and ultraviolet at 172 nm. From the experimental results shown in figures 8 and 9, it is suggested that two independent erosion pathways exist in a simultaneous atomic oxygen and ultraviolet exposure condition; atomic oxygen-induced erosion which follows cos θ dependence with incident angle and ultraviolet-induced erosion which is independent of the incident angle of ultraviolet. The erosion rate of ultraviolet-induced erosion pathway is accelerated three times under the presence of atomic oxygen

150

KUMIKO YOKOTA ET AL.

in the present experimental condition. Since the absolute reaction rate is greater in atomic oxygen-induced erosion, it was considered that the overall erosion was rate-limited by atomic oxygen. 3.4. EFFECT OF ULTRAVIOLET EXPOSURE ON ADSORBED OXYGEN

In the previous section, the importance of the oxygen atoms adsorbed at the polyethylene surface on the ultraviolet-induced erosion was suggested. XPS was used for analyzing the oxygen adsorbed polyethylene surface. The experiment was carried out with atomic oxygen fluence of 3.0 × 1018 atoms·cm−2 . The sample was exposed to 172 nm ultraviolet radiation (0.45 mW·cm−2 ) for 40 min in vacuum. Figure 10 indicates the C1s core level XPS spectra of atomic oxygen-exposed polyethylene before and after ultraviolet exposure. As clearly observed in figure 10, atomic oxygen-exposed polyethylene showed a high-energy shoulder at 288.5 eV besides the main peak at 284.8 eV. This shoulder is contributed by carboxyl groups. The high-energy peak disappeared under the ultraviolet irradiation. Similar results were obtained from polyimide samples. Table 1 shows the composition of the polyethylene and polyimide surfaces analyzed by XPS. After atomic oxygen exposure, the oxygen composition increased to 27% for polyethylene and to 33% for polyimide. Increase in oxygen composition corresponds to surface oxidation forming carboxyl groups. However, the high-energy shoulder disappeared after ultraviolet exposure, and the surface composition recovered to normal values. This spectral change in XPS is explained by the photodissociation of carboxyl group created by exposure to atomic oxygen [12]. Similar conclusion was also obtained in polyimide [12]. It was, thus, confirmed by XPS that ultraviolet irradiation promotes the decomposition of carboxyl group formed by atomic oxygen exposure at polyethylene and polyimide surfaces. This photochemical reaction is the

2500

(a (a) 2000

(b)

c/s

1500

(c) 1000

500

0

294

292

290

288

286

284

282

280

Binding Energy (eV)

Figure 10. C1S XPS spectra of atomic oxygen-exposed polyethylene before and after ultraviolet exposure. (a): Pristine, (b): AO exposed, (c): AO·UV exposed

ENHANCEMENT OF ATOMIC OXYGEN-INDUCED EROSION

151

TABLE 1. The composition of polyethylene and polyimide before/after atomic oxygen and ultraviolet exposure

Samples Pristine AO exposed∗ AO->UV exposed∗∗ UV exposed∗∗

Polyethylene

Polyimide

Atomic percent (%)

Atomic percent (%)

C

O

C

O

N

92.8 72.8 89.7 92.5

7.2 27.2 10.3 7.5

75.5 61.2 66.9 75.6

17.8 33.3 26.6 16.6

6.7 5.5 6.5 7.8

Polyethylene: ∗ AO fluence: 3.0 × 1018 atoms/cm2 , ∗∗ UV intensity: 0.45 mW/cm2 for 40 min. Polyimide: ∗ AO fluence: 9.6 × 1017 atoms/cm2 , ∗∗ UV intensity: 0.6 mW/cm2 for 4 hr.

origin of the oxygen-enhancement of ultraviolet-induced mass loss of polyethylene observed in figure 9.

4. Conclusion Synergistic effect on atomic oxygen-induced erosion of polyimide and polyethylene with 172 nm vacuum ultraviolet was investigated using QCM crystals coated with polyethylene and polyimide. In order to change the relative intensity of atomic oxygen and vacuum ultraviolet, the sample on QCM was rotated around an axis perpendicular both to the axes of atomic oxygen and ultraviolet. It was found that the erosion rate of polymers by ultraviolet exposure is independent of the incident angle of ultraviolet, whereas that by atomic oxygen follows cosine function. It was observed that the erosion rate of polyethylene increased 30–100% by a simultaneous vacuum ultraviolet exposure depending on the relative intensity of ultraviolet and atomic oxygen. The increase of mass loss is due to the creation of new reaction pathway of ultraviolet-induced decomposition of carboxyl group, which was created by atomic oxygen exposure. Since the synergistic effect of atomic oxygen and ultraviolet is obvious in polyethylene at one order lower relative intensity of UV/AO, polyimide is a better material as a witness sample for measuring atomic oxygen fluence in LEO.

Acknowledgments This study was partially supported by the Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology, Japan

152

KUMIKO YOKOTA ET AL.

under contact no. 13750842, 14350511, and 15560686; and the Space Utilization Promotion from the Japan Space Forum. Financial support from Kawanishi Memorial Shinmaywa Education Foundation is also acknowledged. The authors express their acknowledgments to S. Seikyu and K. Maeda of Kobe University for their help with this experiment.

References 1. Standard Practices for Ground Laboratory Atomic Oxygen Interaction Evaluation of Materials for Space Applications, ASTM Designation, E2089-00. 2. Atomic Oxygen Effects Measurements for Shuttle Missions STS-8 and 41-G, NASA TM-100459, 1988. 3. Rutledge, S. K., Banks, B. A., and Kitral, M. (1998) A Comparison of Space and Ground Based Facility Environmental Effects for FEP Teflon, NASA TM 207918. 4. Brinza, D. E., Chung, S. Y., Minton, K. T., and Liang, R. H. (1994) Final Report on the NASA/JPL Evaluation of Oxygen Interactions with Materials-3 (EOIM-3), JPL Publication, pp. 94–31. 5. Koontz, S. L., Leger, L. J., Albyn, K., Cross, J. (1990) Journal of Spacecrafts and Rockets 27, 346–348. 6. Kinoshita, H., Tagawa, M., Umeno, M., and Ohmae, N. (1998) Transaction of the Japan Society for Aeronautical and Space Science 41(132), 94–99. 7. Tagawa, M., Yokota, K., Ohmae, N., and Kinoshita, H. (2002) Journal of Spacecraft and Rockets 39(3), 447–451. 8. Yokota, K., Tagawa, M., and Ohmae, N. (2002) Journal of Spacecraft and Rockets 39(1), 155– 156. 9. Tagawa, M., Yokota, K., Kida, T., and Ohmae, N. (2003) In J. Kleiman and Z. Iskanderova (eds.), Protection of Materials and Structures from Space Environment Space Technology Proceedings, Vol. 5, Kluwer Academic Publishers, Dordrecht, pp. 391–400. 10. http://www1.ushio.co.jp/tech/ 11. Yokota, K., Ohmae, N., and Tagawa, M. (2004) High Performance Polymers 16, 221–234. 12. Scnabel, W. (1992) Polymer Degradation-Principles Practical Applications, Carl Hanser Verlag, Munich.

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS ON POLYMERS R. VERKER,1,2 E. GROSSMAN,1 N. ELIAZ,2 I. GOUZMAN,1 S. ELIEZER,3 M. FRAENKEL,3 AND S. MAMAN3 1 Space Environment Division, Soreq NRC, Yavne 81800, Israel 2 Department of Solid Mechanics, Materials and Systems, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel 3 Plasma Physics Department, Soreq NRC, Yavne 81800, Israel

Abstract. Hypervelocity space debris impacts can lead to degradation of satellite performance and, in extreme cases, might cause a total loss of a spacecraft. The increase in space debris population provides the motivation for this study, which focuses mainly on the mechanical behavior of space-qualified polyimide Kapton films impacted by simulated hypervelocity debris. Kapton is used extensively on spacecrafts, especially in thermal control blankets. Kapton films 25, 50, and 125 μm-thick were studied at different impact velocities of up to 2900 m·s−1 generated by a laser driven flyer (LDF) system. The Kapton-impacted sites revealed ductile-type fractures for low-velocity debris, which changed gradually into mixed ductile–brittle fractures with crack formation when debris impact velocity was increased. Fractures created by impacts at velocities above 1700 m·s−1 showed central impact regions which experienced the highest strain rate and revealed a ductile-type fracture, while the outer regions which experienced a lower strain rate failed through brittle cracking. A model explaining this phenomenon, based on the temperature profile developed within the impacted region at the time of impact, is presented. Key words: Hypervelocity Impact Simulation, Polymer damage, Space Debris

1. Introduction Many of the satellites nowadays are being launched into low Earth orbit (LEO), ranging from 200 to 700 km. The LEO space environment possesses many obstacles to a successive spacecraft mission. These obstacles include ambient space conditions such as ultrahigh vacuum, as well as man-made obstacles such as spacecraft debris. In LEO, spacecrafts are subjected to various destructive environmental components, such as ionizing radiation (electrons, protons), 153 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 153–165. C 2006 Springer. Printed in the Netherlands.

154

R. VERKER ET AL.

vacuum ultraviolet (VUV) radiation, hyperthermal atomic oxygen (AO), and other factors such as extreme temperature variations, micrometeoroids, and orbital debris. Due to either singular or synergistic interactions with these space components, structural materials—in particular polymer-based materials—suffer a relatively rapid erosion (mass loss), structure modification, and surface roughening, leading to irreversible degradation of their physical characteristics (optical, thermal, electrical, and mechanical) [1, 2]. Therefore, a careful selection of surface materials, namely polymer films and paints, is required [3]. Micrometeoroids originate naturally from planetary or asteroidal collisions and cometary ejecta [4]. The large debris population at LEO altitudes comprises the waste products of spacecraft operations. Typical velocities of debris particles range from few kilometers per second up to 16 km·s−1 , making these particles a threat to spacecrafts. The debris issue must be quantified over the projected lifetime of a space system to determine the life expectancy of exposed systems and to quantify necessary shielding requirements [5]. Artificial space debris consists of large objects such as spent satellites and rockets, and mostly of small objects such as aluminum oxide fuel particles, paint chips, and fragmentation objects from collisions of these bodies in orbit [4]. The recovery of several spacecrafts in the last decade offers information concerning the directionality of the LEO meteoroids and space debris fluxes [6]. Such recovered spacecrafts and spacecraft’s parts include one of the Hubble space telescope solar arrays (retrieved in 1993), the European retrievable carrier—EURECA (retrieved in 1993), and the long duration exposure facility—LDEF (retrieved in 1990 after 69 months in LEO). Spacecraft debris impact damage can degrade the performance of exposed spacecraft materials and, in some cases, destroy a satellite’s ability to perform or complete its mission [3]. The Hubble space telescope solar array, for example, suffered impacts at ultrahigh velocities ranging from 2.9 to 11.5 km·s−1 from particles 7 to 98 μm in diameter [7]. Particles traveling at ultrahigh velocities generate temperatures in the range of 5000 K and pressures of several megabars when they collide with a surface [8]. Accumulation of impacts over the large surface area of solar panels leads, in some cases, to degradation in efficiency [9]. Impacts into metals form craters, which have diameters averaging about 5 times the impact diameter. These craters are of concern because they can prevent impacted components from operating. In the case of composites, if a complete penetration occurs, this can lead to further breakdown of the composite during subsequent exposure to AO or VUV. Debris impacts into polymer films occurs quite often, since they are used extensively onboard spacecrafts, mainly as thermal blankets. Mostly, these materials are thin laminated layers; thus, the impacts cause delamination of these layers into many times the diameter of the crater [3].

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 155

Thermal control materials on the LDEF have demonstrated a significant synergism of orbital debris with other space environmental components. Such synergism further expanded the damaged areas caused by impacts. For example, the top surface of a metalized Mylar sample aboard the LDEF was completely eroded, exposing the interior surfaces to VUV radiation, AO, and thermal cycling. As the number of missions sent into LEO is increasing, the frequency of debris impacts is expected to increase as well [3]. Such an increase may lead to further complications in operation of satellites in LEO environment. These complications may be in the form of (i) accelerated development of molecular and particle contamination, and (ii) an increased change in optical and mechanical properties due to debris impact. Thermal blankets that cover large parts of a spacecraft, will particularly be subjected to these changes. The expected increase in impacts frequency and the amount of polymeric thermal blankets onboard spacecrafts provides the main motivation of this study. The study deals mainly with the mechanical behavior of space-qualified polymer films subjected to ultrahigh velocity impacts.

2. Experimental 2.1. THE LASER DRIVEN FLYER METHOD

The laser driven flyer (LDF) method was used for simulating space hypervelocity debris with dimensions ranging from 10 to 100’s μm and velocities of up to 2900 m·s−1 . LDF is attractive as an acceleration technique for debris simulation due to its relative simplicity, relatively low cost, ease of incorporation into a vacuum facility, and high shot rate capability [8, 10, 11]. Figure 1 shows a schematic diagram of the LDF process. In this method, a high-intensity laser beam is shot into a metal foil (3–25 μm-thick) tenaciously bonded to a glass substrate (hereafter referred as target). The beam passes through the glass and hits the aluminum–glass interface. At the interface, high-pressure

Plasma (Shock wave) Laser beam

Glass/Aluminum laminate

Figure 1. Schematic description of the laser driven flyer (LDF) process

156

R. VERKER ET AL.

plasma is formed at the range of giga-Pascal (GPa). The instantaneous high pressure generates a compressive shock wave that propagates into the aluminum faster than the speed of sound. As the shock wave reaches the aluminum surface and the laser pulse ends, two (tensile) rarefaction waves are generated propagating one toward the other. When the rarefaction waves, and their tension pressure is higher than the spall pressure, a spall is produced. A pressure gradient between the plasma pressure on one side and the vacuum zero pressure on the other side causes the spalled layer acceleration which results in an aluminum layer to fly away at ultrahigh velocity. The aluminum flyers in this work were accelerated toward polymer samples located at a selectable distance of 2–12 mm from the laminate structure. The whole system was placed inside a vacuum chamber operating at a base pressure of 20 mTorr. Soreq’s LDF system is using a Titanium:Sapphire laser operating under a single-pulse mode at 810 nm wavelength. The length of each pulse is 300 ps, and the energy of the pulse can be controlled within the range of 250–750 mJ. The experiments were carried out with single laser shots. After each shot, a new sample was positioned and a new unexposed area of aluminum–glass target was placed into the laser beam path. 2.1.1. Flyer velocity The size of the formed aluminum flyer is identical to the beam spot size, which is controlled by a focusing lens. Changing the laser spot size leads to a change in the laser’s surface intensity, thus affecting the flyer velocity and size. The flyer velocity is affected also by the laser pulse energy. By changing the latter parameter, velocities of up to 2900 m·s−1 were attained. Figure 2 shows the theoretical and measured flyer velocity as a function of the laser’s pulse energy. The theoretical flyer velocity is the maximum possible velocity obtained when assuming a system’s

Velocity (m/s)

5000

Theoretical Measured

4000

3000

2000

1000 200

300

400

500

600

700

800

Laser pulse energy (mJ) Figure 2. Flyer measured and theoretical velocities as a function of the laser pulse energy

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 157

Intensity (arb. units)

0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000 0.0

−6

3.0x10

−6

6.0x10

−6

9.0x10

−5

1.2x10

−5

1.5x10

Time (sec)

(a)

(b)

Figure 3. (a) Schematic description of the LDF velocity measurements, and (b) a scope display during a velocity measurement experiment

hydrodynamic coefficient ηH = 1.0. This value means that the whole laser pulse energy is transferred into flyer kinetic energy. The hydrodynamic coefficient is defined as ηH =

E KF V2 = M2 E LP VT

(1)

where E KF is the flyer kinetic energy, E LP is the laser pulse energy, and VM and VT are the measured and theoretical velocities, respectively. According to the theoretical and measured velocities presented in figure 2 and eq. (1), the Soreq’s LDF system hydrodynamic coefficient was calculated to be ηH = 0.23. The flyer velocity was measured using the system shown schematically in figure 3(a). A continuous He:Ne laser beam was set orthogonal to the flyer’s trajectory, so that the beam crossed the flyer path twice in the presence of a prism. The two parallel beams were set at a known distance of 13 mm from each other. A photodiode attached to a scope received the continuous laser signal. As the flyers crossed the continuous laser’s path, two peaks were detected by the scope, allowing the velocity calculation. Figure 3(b) shows a typical scope display with peaks time difference of 6 μs obtained at laser pulse energy of 650 mJ, indicating a flyer velocity of 2000 m·s−1 . 2.1.2. Flyer size In order to estimate the flyer dimensions, series of experiments were conducted using a 1.6 mm-thick BK7 glass as the impacted sample. All experiments were done using similar parameters: pulse energy of 250 mJ, vacuum pressure of 100

158

R. VERKER ET AL.

100 μm Figure 4. A typical ESEM image of an impact-induced crater on a glass surface. Impact conditions were 250 mJ pulse energy, 300 ps pulse length, 100 mTorr system pressure, and flyer velocity of 1400 m·s−1

mTorr, and pulse length of 300 ps. The flyer velocity was 1400 m·s−1 . Sample morphology was characterized using an environmental scanning electron microscope (ESEM, model Quanta 200 from FEI), which allows characterization of nonconductive samples (e.g., Kapton and glass) without the need for a conductive coating. Figure 4 shows a typical SEM image of an impact-induced crater on a glass surface. Flyer velocities and crater diameters, as determined from the ESEM images, were applied into the “conchoidal cracking diameter equation [7]”: −0.5 Dco = 5 × 10−4 ρS ρP0.71 dP1.13 vp0.754

(2)

where Dco is the concentric cracking region diameter, ρs and ρp are the sample and flyer densities, respectively, dp is the plate diameter, and vp is the plate velocity. The flyer dimensions were calculated to be 23–29 μm in diameter. This calculation was conducted only for a velocity of 1400 m·s−1 . In practice, each laser shot at the target produced several major flyers being part of a cloud of such flyer fragments, all traveling at ultrahigh velocity. In order to evaluate the cloud’s dimensions, a 12 μm-thick aluminum foil was used as the impacted sample. The cloud of flyers punctured the foil, creating a hole of 1.5 mm in diameter. The diameter of the hole resembles the dimensions of the cloud. 2.2. POLYMER SAMPLE CHARACTERIZATION

The morphology of the fractures created by the debris impacts was analyzed using the same ESEM.

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 159

3. Materials The materials studied in this work were commercial KaptonR HN Polyimide (Du-Pont Inc.) films, 25, 50, and 125 μm-thick. Kapton possesses a unique combination of properties that makes it suitable for a variety of applications onboard of spacecrafts. Its main use is as the outer layer of multilayer thermal control insulation blankets, and also as flexible substrates for high-power solar arrays. Among its main properties are inherent strength, temperature stability, excellent insulation properties, and stability under ionizing and UV radiation. Kapton is also known for its superior optical properties including low solar absorbance and high thermal emittance.

4. Results and Discussion 4.1. FLYER VELOCITY EFFECTS

Figure 5 demonstrates the effect of flyer velocity on the extent and nature of damage developed in impacted 25 μm-thick Kapton films. The fractures were created using final flyer velocities VF of 1400 (figure 5(a)), 1650 (figure 5(b)),

Figure 5. ESEM images of 25 μm-thick Kapton films impacted by debris at velocities of (a) 1400 m·s−1 , (b) 1650 m·s−1 , (c) 1730 m·s−1 , and (d) 2900 m·s−1 , respectively

160

R. VERKER ET AL.

1730 (figure 5(c)), and 2900 (figure 5(d)) m·s−1 . All ESEM images were taken from the impact exit side. The volcano-like puncture sites obtained at the lowest velocity (figure 5(a)) may indicate ductile rupture of the polymer. At this relatively low velocity, only few flyers could penetrate the film and create the punctures. At a higher flyer velocity of 1650 m·s−1 (figure 5(b)), ductile rupture is still dominant, but some cracks begin to form around these volcano-like punctures. A further increase in the flyer velocity resulted in radial cracking around the central impact zone (figure 5c). The results indicate also that all flyers in the cloud had sufficient energy to penetrate the Kapton film. At the highest tested velocity of 2900 m·s−1 , these radial cracks completely developed into a brittle fracture of the polymer (figure 5(d)). The transition from ductile to brittle fracture may be expected because such transitions are strongly dependent on the strain rate. As the flyer velocity increased, the strain rate also increased. Brittle fractures are associated with less energy absorbance compared to ductile fractures [12]. At relatively low strain rates (figures 5(a) and 5(b)) the kinetic energy lost by the flyers was transferred into pronounced deformation energy. At relatively high strain rates (figures 5(c) and 5(d)), on the other hand, the successive kinetic energy was transformed into crack propagation energy and the associated formation of new surfaces. 4.2. FILM THICKNESS EFFECT

The effect of film thickness on the extent and nature of damage introduced into the Kapton film is demonstrated in figure 6. Laser driven flyers with velocity of 1730

Figure 6. Impacts at velocity of 1730 m·s−1 into (a) 25 μm, (b) 50 μm, and (c) 125 μm-thick Kapton films. Note the different scale bar in (c)

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 161

m·s−1 were shot against 25, 50, and 125 μm-thick Kapton films. All images in figure 6 were taken from the impacted sample exit side. It is evident that as the film thickness increases, a transition from brittle (figure 6(a)) to ductile (figure 6(c)) fracture occurs and the overall extent of damage caused by the impact is reduced too. The 25 μm-thick Kapton film (figure 6(a)) experienced significant damage with radial brittle-like cracks emanating from a central impact zone. The 50 μmthick Kapton film (figure 6(b)) experienced less damage, lacking any radial cracks; only few punctures were noticed. The least damage was introduced into the 125 μm-thick Kapton film (figure 6(c)); only a single penetration zone is observed, exhibiting a volcano-like puncture. The following argument may be given to explain the results aforementioned. As the Kapton film becomes thicker, its ability to absorb energy and slow the flyer increases. Consequently, the strain rate associated with the impact process is reduced. For the 25 μm-thick film, the strain rate is high enough to catalyze the formation of brittle radial cracks. For the 50 μm-thick film, intermediate strain rates probably existed, leading to a semiductile fracture of the polymer. In this case, no sufficient energy was left to allow radial cracking. Finally, in the case of the 125 μm-thick film, only a single puncture was formed—most likely by a single flyer. The strain rate under which this process took place was low enough to enable ductile fracture.

4.3. A THERMAL MODEL

Close examination of the sample impacted at the highest velocity of 2900 m·s−1 is shown in figure 7 at a higher magnification. This figure shows also the fracture surface morphology (fractography) around the circumference of the penetrating hole. It is clear that this morphology changes significantly, indicating the possible involvement of different mechanisms of fracture. The characteristics of fracture morphology seem surprising at first—while the central penetration zone that experienced the highest strain rate failed in a fairly ductile manner, the radial cracks that formed subsequently under lower strain rates (i.e., as a secondary process) exhibit a more brittle fracture morphology. We believe that this behavior may be explained in terms of a high temperature gradient that is established within the polymer sample as the flyer hits its surface and penetrates through the film. It is well known that ductile–brittle transitions depend strongly on the local temperature. Whereas a ductile fracture is expected above the glass transition temperature, Tg , below this temperature brittle fractures are most likely to occur. It should be noted that in Kapton, a second-order transition occurs within the temperature range of 360–410◦ C, which is assumed to be the glass transition temperature [13]. This temperature dependence of fracture mode also reminds the deformation map suggested by Spaepen for metallic glasses [14]. Ultrahigh velocity impacts generate temperatures in the range of 1727–6727◦ C

162

R. VERKER ET AL.

Figure 7. (a) A 25 μm-thick Kapton film impacted at hypervelocity of 2900 m·s−1 . Two distinct modes of fracture are evident: (b) a fairly brittle fracture in the radial cracking region, and (c) a more ductile fracture around the central penetration region

and shock pressures of 30–100 GPa when striking ceramic surfaces [9, 15]. Highdensity polyethylene projectiles shot at an average velocity of 5 km·s−1 were observed to generate temperatures of up to 7927◦ C in the case of head-on impacts on aluminum targets [16]. Hence, the following model can explain the phenomenon shown in figure 7. Due to the high temperature generated at the penetration zone and despite the ultrahigh strain rate involved in impact, the Kapton film exhibits a fairly ductile fracture in this zone. On the other hand, the significantly lower temperatures of T < Tg are not sufficient to compensate for the still high strain rates, thus the secondary cracks far from the impact point exhibit brittle fracture characteristics. 4.4. IMPACT TEMPERATURE EVALUATION

In order to support the above model, an attempt was made to estimate the mean temperature developed within a sample at the time of penetration. Figure 8 shows a 25 μm-thick Kapton film impacted at an entrance flyer velocity Vi of 2200 m·s−1 . The flyer velocity at the exit, Vo , was also measured and found to be 1540 m·s−1 . The sample was divided into three regions: (a) the penetration hole, at which a piece of Kapton was sheared-off, (b) a region where ductile fracture was observed, and (c) a region where brittle fracture was noticed.

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 163

Figure 8. A 25 μm-thick Kapton film impacted at a hypervelocity of 2200 m·s−1 . The sample is divided schematically into three regions: (a) the penetration hole, at which a piece of Kapton was sheared-off, (b) a region where ductile fracture was observed, and (c) a region where brittle fracture was noticed

The system energy balance was expressed as follows: E K,F = E K,S + E Sh + E C + E H

(3)

where E K,F is the flyer kinetic energy change due to the impact, E K,S is the kinetic energy provided to the piece of Kapton that was sheared-off, E Sh is the energy required for the shearing process, E C is the energy required for crack formation, and E H is the energy transformed into heat which results in temperature increase. The following assumptions were made in the calculation: (i) the impact was taken as an equilibrium process. Although this assumption is inaccurate, it greatly simplifies the calculation and provides the ability to calculate a mean temperature within a certain region of the sample; (ii) the velocity of the sheared Kapton piece is equal to the flyer velocity at the exit side; (iii) the energy required for deformation of the Kapton surface is relatively small, and may be neglected. E K,F and E K,S were calculated according to 1 E K = mv 2 2

(4)

Using a flyer mass m = 5.7 × 10−5 g, entrance velocity Vi = 2200 m·s−1 , and exit velocity Vo = 1540 m·s−1 , one obtains E K,F = 70.2 mJ, E K,S = 40.0 mJ. E Sh was calculated according to PS = 0.85 · h · 1 · (UTS) E Sh = 0.5 · PS · h

(5) (6)

164

R. VERKER ET AL.

where PS is the shearing force, h is the film thickness, l is the length of cut, UTS is the ultimate tensile strength of the material being sheared, and E Sh is the shearing energy. For h = 25 μm, l = 3.5 × 10−3 m, UTS = 2.31 × 108 N·m−2 , and PS = 17 N, we get E sh = 0.2 mJ. E C was evaluated according to E C = 2 · δa · h · (γS + γP )

(7)

where δa is the total length of cracks which were formed by the impact, γs is the specific surface energy, and γp is the plastic deformation energy. Using the measured value δa = 0.37 cm, and the typical values γs = 5.8 × 10−4 N·cm −1 and γp = 0.5 N·cm−1 , we find E C = 9.3 × 10−3 mJ. Now, by substituting these values into eq. (3), we find the amount of energy transformed into heat, E H = 30 mJ. Using the common expression for adiabatic heat transfer equilibrium process E H = m · Cp · T

(8)

where m is the mass of Kapton confined within region (b) in figure 8 (m = 5.7 × 10−5 g), CP is the Kapton’s specific heat (CP = 1.09 J·(g·K)−1 ), and T is the temperature difference, a mean temperature increase of T = 920◦ C was calculated. This calculation thus shows that a flyer penetrating at a velocity of 2200 m·s−1 causes a significant temperature increase in the penetration zone (figure 8, region (b)), raising the local temperature to as high as approximately 947◦ C. This estimated temperature is much higher than the glass transition temperature of Kapton, supporting the suggested model and explaining the ductile-like fracture within the penetration zone. This large temperature gradient also predicts a temperature lower than Tg in region (c), where the high strain rate results in a brittle-like fracture.

5. Summary and Conclusions A laser driven flyer (LDF) system was developed at Soreq NRC for simulation of space debris impacts. The system produces a cloud of flyers with diameters of up to 30 μm and measured velocities of 1400–2900 m·s−1 . The effect of simulated hypervelocity debris impacts on space-qualified Kapton films was studied for different film thickness (25–125 μm) and impact velocities. As the Kapton film thickness was increased from 25 to 125 μm at a fixed impact velocity, a transition from brittle to ductile fracture was observed. At a constant thickness of 25 μm, low impact velocities of 1400–1600 m·s−1 resulted in a ductile-like fracture. Increasing the flyer velocity resulted in cracks formation and a fracture that was mostly brittle-like. At impact velocities higher than 1700 m·s−1 , the central impact region (which is exposed to the highest strain rate) was characterized by a ductile-like fracture, while remote radial crack regions were

GROUND SIMULATION OF HYPERVELOCITY SPACE DEBRIS IMPACTS 165

characterized by a brittle-like fracture despite the lower strain rates. A model explaining this phenomenon was suggested based on the high impact temperature (T > Tg ) developed at the central impact region and the low temperatures (T < Tg ) at remote regions.

Acknowledgment This work was partially supported by Israeli Space Agency.

References 1. Grossman, E., Gouzman, I., Viel-Inguimbert, V., and Diguirard, M. (2003) Journal of Spacecraft and Rockets 40, 110. 2. Houdayer, A., Cerny, G., Klemberg-Sapieha, J.E., Czeremuszkin, G., and Wertheimer, M.R. (1997) Nuclear Instruments and Methods in Physics Research B 131, 335. 3. Silverman, E. M. (1995) Space Enviromental Effects on Spacecraft, LEO Material Selection Guide, NASA Contractor Report 4661, Langley Research Center, pp. 4–1. 4. Tennyson, R. C. and Shortliffe, G. (1997) In A. Paillous (ed.) Proceedings of the 7th International Symposium on Materials in Space Environment, Toulouse, France, ESA Publication, ESTEC Noordwijk, The Netherlands, 16–20 June 1997, p. 485. 5. Hastings, D. and Garrett, H. (1996) Spacecraft-Environment Interactions, Cambridge University Press, Cambridge, U.K., pp. 45–99. 6. Miao, J. P. W. (2001) Stark Planetary and Space Science 49, 927. 7. Paul, K. G., Igenbergs, E. B., and Berthoud, L. (1997) International Journal of Impact Engineering 20, 627. 8. Stein, C., Roybal, R., and Tlomak, P. (2000) In E. Werling (ed.) Proceedings of the 8th International Symposium on Materials in Space Environment, Arcachon, France, CNES Publication, Toulouse, France, 5–9 June 2000. 9. Medina, D.F., Wright, L., and Campbell, M. (2001) Advances in Space Research 28, 1347. 10. Tighe, A., Gabriel, S., and Van Eesbeek, M. (2000) In E. Werling (ed.) Proceedings of the 8th International Symposium on Materials in Space Environment, Arcachon, France, CNES Publication, Toulouse, France, 5–9 June 2000. 11. Roybal, R., Tlomak, P., Stein, C., and Stokes, H. (1999) International Journal of Impact Engineering 23, 811. 12. Callister, W. D. (2000) Materials Science and Engineering, John Wiley and Sons Inc., New-York, pp. 124–134. 13. Du-Pont Inc. Technical bulletin, http://www.dupont.com/kapton/general/H-38492-2.pdf. 14. Spaepen, F. (1977) Acta Metallurgica 25, 407. 15. Akins, J.A. (2003) Dynamic Compression of Minerals in the MgOFeO-SiO2 System, Ph.D. thesis, California Institute of Technology, CA. 16. Ramjaun, D., Shinohara, M., Kato, I., and Takayama, K. (2001) In Proceedings of the 23rd International Symposium on Shock Waves (ISSW23), Fort Worth, TX, 22–27 July 2001.

TESTING OF SPACECRAFT MATERIALS FOR LONG FLIGHTS IN LOW EARTH ORBIT L. S. NOVIKOV,1 V. N. CHERNIK,1 S. F. NAUMOV,2 S. P. SOKOLOVA,2 T. I. GERASIMOVA,2 A. O. KURILYONOK,2 AND T. N. SMIRNOVA3 1 Skobeltsyn Institute of Nuclear Physics Moscow State University, 119992 Moscow, Russia 2 RSC “Energia,” Korolev, Russia 3 Khrunichev State Space Scientific Production Center, Moscow, Russia

Abstract. Results of simulation tests of protective and functional coatings influence on resistance of polymeric constructional spacecraft materials to impact of atomic oxygen with fluences up to 3.5 × 1022 cm−2 are presented. It was demonstrated that oxygen plasma beams can be used in accelerated tests of carbon-based and polymeric materials (with the exception of fluorinated hydrocarbons) to evaluate their resistance to the AO impact in LEO. For the unprotected materials sharp fall of mechanical properties and optical characteristics deterioration were observed. Application of protective coatings had shown to reduce their degradation. Key words: atomic oxygen, protective coating, spacecraft materials

1. Introduction One of the principal damaging factors of space in low Earth orbits (LEO) is the atomic oxygen (AO). In prolonged exposures of materials on the outside surfaces of spacecraft, their resistance to AO attack is of major importance. Since polymerbased materials are widely used on spacecrafts and most of them are susceptible to AO attack, their protection by various means, including protective coatings, is necessary. The long-term LEO flight simulation requires irradiation of materials with high-fluence AO up to 1022 –1023 cm−2 . Simulator beam intensities do not exceed 1017 cm−2 ·s−1 (usually 1015 –1016 cm−2 ·s−1 ) [1] that results in practically unacceptable test duration. The reduction of test duration is achieved by an increase of the particle beam energy within the limits of conservation of the interaction mechanism with the test material [2]. 167 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 167–174. C 2006 Springer. Printed in the Netherlands.

168

L. S. NOVIKOV ET AL.

In our work, the accelerated simulation material tests were carried out in oxygen plasma beams, formed by a plasma accelerator, with the atomic oxygen particle energy of 20 eV. The changes of weight, mechanical and optical properties of prospective materials were investigated at equivalent AO fluences up to 3.5 × 1022 cm−2 .

2. Test Materials Various purpose materials were investigated: structural coating (black reinforced plastics, fabric, threads, films), functional coatings (protective, thermo-control (TCC), colored). The black reinforced plastic is composed of carbon fibers and epoxy matrix. Samples in form of plates 1 mm thick without coating and with TCC white conductive enamel EKOM-1 were used. The enamel EKOM-1 consists of ZnO pigment with acrylic binder, 0.1 mm thick. One more test object was a fragment of the protection shield of external ISS equipment used to protect from influence of space shuttle jets. The shield consists of the Terlon fabric (Kevlar-type polyamide) separate parts overcasted by the polyimide fibers threads. Another synthetic fiber-based material was a sennit “PARML.” It is an electrically conductive whisker material applied to cable shielding. It is twisted with polyimide fibers threads that are the sennit load-bearing basis. The threads are winded by a copper tinsel ribbon for high electroconductivity. Tensile, stress–strain measurements were used to evaluate the changes in the breaking strength of Terlon fabric and sennit “PARML.” For evaluation of the efficiency of different coatings, polyimide films 20 and 40 μm thick were chosen that were coated with silica layers 0.2–0.4 μm thick, deposited by plasma deposition. Silicone coatings with a thickness of 10–20 μm were made by spraying of varnishes of two types. Color enamels with epoxy-based binders and silicone pitches were studied. The efficiency of the epoxy enamels protection by a silicone varnish layer was studied also. The red, dark blue, white, black, and yellow coatings were applied to fiber glass fabric samples 20 × 40 mm2 in size. 3. The Test Technique The experiments were carried out in the plasma accelerator [3] in oxygen plasma beams consisting of ions, atoms, and molecules of oxygen with mean velocity of 16 km·s−1 (mean atom energy of 20 eV) and a flux density of (2.5–3.5) × 1016 cm−2 ·s−1 . Due to the dissociation of fast molecules and neutralization of the ions

TESTING OF SPACECRAFT MATERIALS

169

at collision with the surface, only the atoms with mean velocity of 16 km·s−1 impact the material. The AO equivalent fluence was evaluated using the witness film mass loss in an assumption of the erosion yield Y = 4.4 × 10−24 g·atom−1 O. The fluence is equal to a fluence of a fictitious 5 eV-AO beam that initiates the same polyimide mass loss. We followed the standard technique of AO fluence determination in ground-based facilities [4]. When material properties were studied the changes of weight and thickness were measured, and the reflection spectra in the range 0.2–2.5 μm were registered. The stress–strain properties were determined on a tensile test machine.

4. Experimental Results 4.1. PROTECTION EFFICIENCY OF TCC WHITE ENAMEL COATING OF BLACK REINFORCED PLASTIC

80 70 60 50 40 30

1.2 no coating with coating

20 10 0

1 2 3 4 0 AO equivalent fluence, 1022 1/cm2

Thickness, mm

Specific mass loss, mg/cm2

The specific mass loss and thickness of the black reinforced plastic samples uncoated and with the coating as a function of the AO equivalent fluence are given in figure 1. As can be seen, the mass loss and thickness of the uncoated samples are linear functions of the AO fluence. At the same time, the mass loss of the coated sample is lower by a factor of 6.5, and thickness change is negligibly small. Reactivity of the unprotected black reinforced plastic consisting of carbon fibers in epoxy matrix, measured in the tests, was 1.2 × 10−24 cm3 per O atom. It corresponds to the flight data known for the composite components: 1 × 10−24 cm3 per O atom for carbon and 1.7 × 10−24 cm3 per O atom for epoxy [3] that confirms that the plasma accelerator conditions can imitate adequately the LEO conditions. The polymer binder on the coated samples was etched away after plasma impact

1 0.8 0.6 0.4 no coating

0.2 0

with coating

0

1 2 3 4 AO equivalent fluence, 1022 1/cm2

Figure 1. Specific mass loss and thickness of samples of the black reinforced plastic uncoated and with the coating as functions of the AO equivalent fluence

170

L. S. NOVIKOV ET AL. 1400

Load, N

1200 1000 800 600 400 200 0

0

10

20

30

40

Elongation, % before exposure

after exposure

Figure 2. Tensile stress–strain diagrams of the sennit “PARML” samples before and after the AO exposure

exposing the fibers that were also partially destroyed with an observed increase in the surface roughness [5]. The picture is similar to observations in flight experiments onboard the space shuttle spacecraft, Salute-6 and Mir space stations [6]. 4.2. DEGRADATION OF MECHANICAL PROPERTIES OF SYNTHETIC FIBER MATERIALS

The results of measurements of specific mass loss of the Terlon fabric samples and the sennit “PARML” samples show, that the Terlon fabric is close to polyimide in AO resistivity, as the materials mass losses are close to 5.4 and 5.6 mg·cm−2 accordingly. A complete etching of the polyimide fibers in a significant part of bundles in the sennit “PARML” was observed. The tensile stress–strain diagrams of the sennit “PARML” samples before and after the AO exposure to an equivalent fluence of 3.7 × 1021 cm−2 are given in figure 2. The breaking load reduction by a factor of 5, and a drop of relative elongation at rupture by a factor of 1.7 was observed. After AO exposure the destruction of seam strings and Terlon fabric was observed on the fragment of the Terlon screen with the seam. The tensile stress– strain diagrams of reference and AO exposed samples of the shield fragments are presented in figure 3. When loading up to approximately 20-fold working load identical elongations were observed in both samples. By further increasing the loading, the exposed sample fails that manifests itself as a sharp break in the curve and an increase in elongation is observed without an increase or even on reduction of loading. The elongation on rupture of the exposed sample also decreases by a factor of 2.

171

TESTING OF SPACECRAFT MATERIALS

Load, N

800 600 400 200 0

1 2

0

2

4

6

8

10

12

14

Elongation, mm Figure 3. Tensile stress–strain diagrams of reference (1) and AO exposed (2) samples of the shield fragments

4.3. EFFICIENCY OF PROTECTIVE SILICON BASED COATINGS

In figure 4, the dependences of the mass loss of the polyimide films protected by deposited silica layers and by silicone coatings of two varnish types of AO fluence are presented. We quantify the AO protection efficiency γ as a ratio of bare film mass loss dm 0 to the protected film one dm i under equal fluence increment dF, i.e. γ = dm 0 /dm i |d F = const.

Mass loss, mg/cm2

The protection efficiency is greatest for silica layer (γ = 830) and decreases up to γ = 250–430 for varnish coverings [6]. The exposure to AO causes formation of microcracks in the silica layers. On the contrary, no cracking occurs in the silicone coatings. Thermo-optical characteristics of the samples changed very little. For the silica coating an insignificant and uniform in wavelength range reflectance reduction is observed that can be associated with scattering from the grid of microcracks. Thus, the solar absorptance αs increases from 0.366 to 0.380. The reflectance increase is typical for varnish coatings, especially appreciable in long wavelength part of the spectrum. 60 50 40 30 20 10 0

silicon 3 0

oxide

5 10 AO equivalent fluence, 1020 1/cm2

silicon 5 15

Figure 4. Dependences of the mass loss of the polyimide films with protection by silica layers deposited and silicone coatings of two varnish types on AO fluence

172

L. S. NOVIKOV ET AL.

4.4. TESTING THE RESISTANCE OF COLOR PAINTS

The paint test results have shown different AO resistance for various types of enamels [7]. Color change and significant mass losses were observed for the epoxy enamels. The erosion yields of the color epoxy enamels (0.3–0.5 × 10−24 g·atom−1 O) are much lower than that of the binder and of the witness polymers

Reflection coefficient

KO-811 blue 0.15 0.1

after before

0.05 0 300

400

500

600

700

Wavelength, nm

Reflection coefficient

EP-140 blue 0.35 0.3 0.25 after

0.2 0.15

before

0.1 0.05 0 300

400

500 Wavelength, nm

600

700

Reflection coefficient

EP-140 + KO-008 blue 0.2 after before

0.15 0.1 0.05 0 300

400

500

600

700

Wavelength, nm

Figure 5. Visible spectrum reflectance for three coatings after oxygen plasma exposure with AO equivalent fluence 1.4 ×1021 cm−2

TESTING OF SPACECRAFT MATERIALS

173

(3–4.4 × 10−24 g·atom−1 O) that is explained by protective action of pigments. The impact of oxygen plasma on silicone enamels almost does not change their color and mass. Protection of the epoxy enamel by silicone varnish layer increases the coating resistance up to the level of the silicon enamels. The erosion yields of these coatings are less than this of the witness film by two orders of magnitude. Figure 5 shows the reflection spectra of the dark blue paints in the visible wavelength range for three coatings: epoxy EP-140, silicon KO-811 K, and epoxy EP-140 with silicone varnish coating KO-008. Comparing the spectra, it is evident that the dark blue color peak seen in all initial spectra, almost completely disappears for epoxy enamel that is accompanied by increase of reflectance in the whole range. The color and the spectra of other dark blue paints practically do not vary.

5. Conclusions It was demonstrated that oxygen plasma beams can be used in accelerated tests of carbon-based and polymeric materials (with the exception of fluorinated hydrocarbons) and inorganic coatings to evaluate their resistance to the AO impact during the simulation of the long flights in LEO. The resistance of prospective spacecraft materials: polyimide films, synthetic Terlon fabric, sennit “PARML,” black reinforced plastic, polymeric paints, and thermo-control coatings under oxygen plasma beams simulating the AO fluxes in LEO was investigated. For the unprotected materials sharp fall of mechanical properties (manifesting in lower failure loads and relative elongation at rupture) was observed. Optical characteristics deteriorated as well. Application of protective coatings had shown to reduce the degradation of mechanical and optical properties. The protection efficiency is the greatest for coatings containing silicon.

References 1. Kleiman, J., Iskanderova, Z., Gudimenko, Y., and Horodetsky, S. (2003) In Proceedings of the 9th International Symposium on Materials in Space Environment, ISMSE-9, Nordwijk, 2003, pp. 313–324. 2. Rutledge, S.K., Banks, B.A., Dever, J., and Savage, W. (2000) In Proceedings of the 5th International Conference on Protection of Materials and Structures from the LEO Space Environment, ICPMSE-5, Arcachon, France ESA Publishing, Noordwijk, The Netherlands, 2000. 3. Akishin, A. I., Novikov, L. S., and Chernik, V. N. (2000) In L.S. Novikov and M.I. Panasyuk (eds.), New High Technologies and Technics, Vol. 17, Moscow, ENTSITEH, p. 100 (in Russian). 4. Protocol for Atomic Oxygen Testing of Materials in Ground-Based Facilities, JPL Publication, pp. 95–17.

174

L. S. NOVIKOV ET AL.

5. Novikov, L. S., Chernik, V. N., Babaevskij, P. G., Kozlov, N. A., Chalyh, A. E., Balashova, E. V., and Smirnova, T. N. (2001) Perspektivny materialy 5, 20–26 (in Russian) (Journal of Advanced Materials, Cambridge Interscience Publication). 6. Chernik, V., Naumov, S., Demidov, S., Sokolova, S., and Svechkin, V. (2000) In 5th International Conference on Protection of Materials and Structures from the LEO Space Environment, ICPMSE-5, Arcachon, France, 2000. 7. Chernik, V. N., Naumov, S. F., Sokolova, S. P., Gerasimova, T. I., Kurilyonok, A. O., Poruchikova, Y. V., and Novikova, V. A. (2003) In Proceedings of the 9th International Symposium on Materials in Space Environment, ESTEC, Noordwijk, 2003, pp. 281–285.

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODULE Post Flight Inspection Results JAMES L. HYDE,1 RONALD P. BERNHARD,1 AND ERIC L. CHRISTINSEN2 1 Lockheed Martin Space Operations, Johnson Space Center, Houston, TX 77258 2 NASA, Johnson Space Center, Houston, TX 77258

Abstract. The International Space Station (ISS) program has manifested a Multipurpose Logistics Module (MPLM) on five launch packages since 2001, with the next flight scheduled for the STS-114 mission. The MPLM has been deployed each time on the nadir docking port of node 1 by the space shuttle for a period of 5–6 days, and then returned to Earth for refurbishment prior to the next resupply mission. MPLM flight module 1 (Leonardo) and flight module 2 (Raffaello) have accumulated about 700 h of low Earth orbit (LEO) exposure time on the ISS. Through five missions, there have been two perforations of the aluminum outer wall of the MPLM shielding and 24 craters due to meteoroid/orbital debris impact. This paper will document the results of an ongoing postflight inspection campaign that identifies hypervelocity impact (HVI) damage to the meteoroid and debris protection system (MDPS) of the MPLM through observations, data collection and analysis. Residual projectile materials are collected at each impact site and subjected to scanning electron microscope/energy dispersive X-ray spectrometric analysis to identify the elemental composition of the impactor and presumed source (meteoroid or orbital debris) of the impact damage. The observed impact damage exhibits a marked directionality on the module, with the majority of damages occurring to the forward half of the MPLM. Postflight predictions of damage to MPLM from the BUMPER code, which is utilized by the aerospace community in meteoroid/orbital debris risk assessments, are compared to the observed damages. Results from BUMPER very clearly show the directionality of the expected impact damage as was observed in the postflight MPLM inspections. Observations and analytical data from MPLM demonstrate the meteoroid/orbital debris impact hazards that the ISS, shuttle, and other spacecraft face in low Earth orbit environment. Implications of this work on the design of adequate meteoroid/orbital debris protection for future vehicles will be provided. 175 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 175–191. C 2006 Springer. Printed in the Netherlands.

176

JAMES L. HYDE ET AL.

Key words: orbital debris, meteoroids, hypervelocity impact, international space station 1. Introduction 1.1. PURPOSE

One of the primary reasons for tracking on-orbit damage is the fact that NASA has used observation data from returned spacecraft surfaces to calibrate the ORDEM 2000 engineering model of the low Earth orbit (LEO) environment [1]. Postflight inspections of shuttle surfaces such as the Crew Module Windows, Payload Bay Door Radiators has produced a database with thousands of entries [2]. The long duration exposure facility (LDEF), retrieved by the shuttle in 1990 after nearly 6 years on orbit, contributed a rich set of data [3] that has been used in the formulation of the orbital debris environment. Postflight inspections also provide an opportunity to validate meteoroid/orbital debris (M/OD) threat assessment codes such as NASA’s BUMPER-II [4, 5] or ESABASE/Debris. The ability to compare predictions with observations is a valuable component in the development life cycle of M/OD analysis software. The performance and adequacy of meteoroid/debris protection systems can be demonstrated with postflight damage inspections. The flight history of the MPLM has established the protection capability of the shielding system. 1.2. OBJECTIVE

This paper will document the results of the five postflight inspections that have been performed on the MPLM, providing data on damage sizes and sources of the impact. The results will also include an estimate of the equivalent diameter of the impacting particle. Due to the nature of hypervelocity impacts, projectile shape and impact angle data are not readily discernable. 2. MPLM Overview 2.1. DESCRIPTION

Three flight units, built by Alenia Spazio for the Italian Space Agency (ASI) as part of an agreement with NASA, were named after famous Italian engineers: Leonardo da Vinci, Donato di Niccolo, di Betto Bardi, and Raffaello Sanzio. The MPLM functions as a cargo container for ISS outfitting and resupply missions and as a pressurized module while attached to the ISS. The modules can be configured as “active,” when a payload requires power, data or cooling resources.

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL

177

Figure 1. MPLM installation sequence

The MPLM can also be manifested in a “passive” configuration when the contents require no power, data or cooling resources (dry cargo). Each module has a design life of 10 years or 25 flights. Each module has an overall length of about 6.2 m and a diameter of about 4.5 m [6], giving a surface area of approximately 100 m2 . 2.2. INSTALATION

MPLM modules are carried into orbit and placed on the ISS by the shuttle arm as shown in Figure 1. At the end of a mission, the module is restowed in the payload bay of the orbiter and returned to Earth. 2.3. M/OD PROTECTION SYSTEM

Figure 2 provides an illustration of the meteoroid/debris protection system (MDPS) that is included on each MPLM. The MDPS consists of a sacrificial outer layer of 0.8 mm aluminum alloy (known as a “bumper”) mounted at a distance of ∼128 mm from the aluminum alloy pressure shell. The bumper-pressure wall cavity also contains a multilayer insulation (MLI) blanket loosely fastened to the 3 mm pressure wall of the MPLM. The MDPS consists of 8 forward cone panels, 48 side panels and 12 aft cone panels. Some of the module’s external features are highlighted in figure 3. MDPS panels on the side and forward cone are also shown. Some of these components will be referenced in a later section.

178

JAMES L. HYDE ET AL.

Figure 2. Meteoroid debris protection system (MDPS)

2.4. MISSION HISTORY

There have been five MPLM missions to date, with a cumulative docked exposure time of 701 h. Three missions used flight module (FM) 1 and two missions used FM2. Details of the missions are shown in figure 4. The modules were always SUPPORT BRACKET FOR ELECTRICAL PDA CBM RING

STABILIZER TRUNNION

FRGF

MDPS PANEL

MAIN TRUNNION

SUPPORT BRACKET FOR FLUID PDA STABILIZER TRUNNION

FRGF

MAIN TRUNNION

MDPS PANEL

Figure 3. External features of MPLM, including MDPS panels

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL MPLM

Exposure

Name

Flight

FM1 Leonardo FM2 Raffaello FM1 Leonardo FM2 Raffaello FM1 Leonardo

1 1 2 2 3

STS Mission Vehicle 102 100 105 108 111

5A.1 6A 7A.1 UF1 UF2

OV-103 OV-105 OV-103 OV-105 OV-105

Docking Port

Hours

145.52 97.84 166.07 145.22 146.02 701 hrs 29.2 days

Node1 - nadir Node1 - nadir Node1 - nadir Node1 - nadir Node1 - nadir

179

ISS Attitude RPY sequence

0°, 23.5°, 0° 0°, 23.0°, 0° 0°, 23.0°, 0° 0°, 21.0°, 0° 0°, 23.0°, 0°

Figure 4. MPLM flight data

mated with the ISS at the nadir docking port on node 1. This placed the port edge of the MPLM (relative to its stowed position in the shuttle) in the ram direction when docked to the ISS. The predominant flight attitude of the ISS when the shuttle and MPLM are attached is a positive pitch bias of about 23◦ , as shown in figure 5, the front of the station is angled upwards ∼23◦ .

3. Impact Survey The following section gives details of the impact features observed during postflight processing. Figure 6 illustrates the measurement terminology that was used to characterize the impacts.

Figure 5. Typical MPLM flight attitude

180

JAMES L. HYDE ET AL. Spall Lip Crater

Depth

Figure 6. Impact crater measurement terminology

3.1. PROCEDURE

As part of the standard postflight processing workflow, all MDPS panels are inspected by technicians at a processing facility at the Kennedy Space Center (KSC). They are tasked with the documenting defects on external and internal surfaces. Instances of scratches, scuffs, ding, and rub marks have all been recorded, along with hypervelocity impact craters. Specialists from the Johnson Space Center with hypervelocity impact (HVI) inspection experience travel to KSC and independently examine as much as possible of the exterior for HVI damage. When an impact site is identified, the location is recorded and a graduated handheld magnifier is used to measure dimensions of the impact features (crater, lip, and spall diameters). For impacts into monolithic aluminum MDPS surfaces, a soft wooden probe is used to collect potential impactor residue from the crater or hole. Dental mold impressions are another technique that is used when surface damage needs to be sampled. Finally, when “soft goods” such as beta cloth are impacted, an adhesive agent is sometimes used to aid in data collection. Crater depth is obtained back at the lab by measuring the effected region of the sampling device.

Figure 7. FM2 at the KSC space station processing facility (SSPF)

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL

181

Figure 8. MPLM inspection at the KSC SSPF

The figure above illustrates the inspection access limitations at the space station processing facility (SSPF). The mounting fixture has to be rotated at least once to allow a complete inspection of all MDPS panels. 3.2. FM1 FLIGHT 1 (STS-102/ISS 5A.1)

There were three hypervelocity impact sites observed after the first MPLM mission. One of the impacts produced a perforation of an MDPS bumper. The impact site, designated 1.1.1, had a 1.44 mm diameter hole with a 2.45 mm diameter lip. There was no discernable damage to the MLI or to the pressure wall of the MPLM. In an attempt to determine the origin of each impactor, samples collected from the three impact sites and subjected to a scanning electron microscope (SEM) and energy dispersive X-ray (EDX) analysis. The results are shown in the following table. Analysis indicated that the impactor that produced the bumper perforation was a particle of spacecraft paint. There was no discernable impact residue in the sample from site 1.1.2. When the particle type is known, an equivalent diameter can be estimated using eq. (1) with an assumed impact velocity of 9 km·s−1 and an impact angle of 45◦ . D = [(P/5.24 × H 1/4 × (ρB /ρP )1/2 × (S/Vn )2/3 ]18/19 IMP A C T D E TA IL S Impact Number Map ID Type A6 MDPS bumper 1.1.1 B5 MDPS bumper 1.1.2 D5 MDPS bumper 1.1.3

Location Region port/lower port/lower port/lower

Estimated

Diam (mm) Material Spall 0.8mm Al 6061-T6 --0.8mm Al 6061-T6 --0.8mm Al 6061-T6 ---

Depth

Lip Crater Hole

2.45 1.25 0.88

(1)

--0.78 0.54

1.44 -----

Figure 9. Impacts on MPLM FM1 flight 1

SEM/EDXA

Particle

Results Dia. (mm) --- Orbital Debris: Paint 0.46 1 0.51 Unknown --0.3 Meteoroid 0.19

(mm)

182

JAMES L. HYDE ET AL.

Figure 10. Impact 1.1.3–1.44 mm diameter hole in MDPS

where D—estimated particle diameter (cm) P—crater depth (cm) H —Brinell hardness of bumper (MDPS = Al6061-T6, 95) ρB —density of bumper (MDPS = Al6061-T6, 2.713 g·cc−1 ) ρP —density of projectile, 2.5 g·cc−1 (for spacecraft paint) S—speed of sound in target (Al6061-T6 MDPS bumper, 5.1 km·s−1 ) Vn —normal component of impact velocity (assuming a velocity of 9 km·s−1 and an impact angle of 45◦ , 6.36 km·s−1 ) Note: in the cases where a bumper perforation occurs, the value of P is extrapolated from the crater diameter. 3.3. FM2 FLIGHT 1 (STS-100/ISS 6A)

No hypervelocity impact features were observed on this module during postflight inspection activities. This was the second MPLM mission.

Figure 11. Results of SEM/EDX analysis—Impact 1.1.3

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL IMP A C T D E TA IL S Impact

Location

Number Map ID 1.2.1 1.2.2 1.2.3

D2 B12 B11

Estimated Depth

Diam (mm)

Type

Region

Material

Spall

MDPS bumper MDPS bumper MDPS bumper

port/upper stbd/upper stbd/upper

0.8mm Al 6061-T6 0.8mm Al 6061-T6 0.8mm Al 6061-T6

-------

183

SEM/EDXA

Dia. (mm)

2

— — 0.14

(mm)

2.80 1.80 0.85

0.9 Unknown 2 0.75 Unknown 0.55 Orbital Debris: SS

2.0 1.5 0.8

-------

Particle

Results

Lip Crater Hole

Figure 12. Impacts on MPLM FM1 flight 2

3.4. FM1 FLIGHT 2 (STS-105/ISS 7A.1)

There were three hypervelocity impact craters observed on the aluminum MDPS bumpers after the third mission (Fig. 12). No perforations were detected. SEM/ EDX analysis of site 1.2.3 indicated that the crater was caused by a particle of stainless steel. The other two sites were not available for sampling. 3.5. FM2 FLIGHT 2 (STS-108/ISS UF1)

Eight impacts were observed after this mission, four on the MDPS bumpers and one each on a grapple fixture, trunnion brace, and scuff plate (Fig. 13). Impact 2.2.8 was the site of the second bumper perforation. The impact produced a 1.2 mm diameter hole with a lip diameter of 1.8 mm. As with the first perforation, there was no visible damage to the underlying structure or components. SEM/EDX analysis indicated that the impactor was a particle of stainless steel. Samples were taken at all eight impact sites. Two impacts were shown to be from orbital debris and two were meteoroids. The remaining four impact samples revealed no discernable impact source. 3.6. FM1 FLIGHT 3 (STS-111/ISS UF2)

Postflight inspections for the third mission of FM1 (Leonardo) recorded 12 impact sites (Fig. 16). Ten occurred on the MDPS panels, another in the material between IMP A C T D E TA IL S Impact

Location

Number Map ID LA4 2.2.1

Estimated Depth

Diam (mm)

Type scuff plate

Region port/lower

Material Spall yellow Sheldahl tape 1.4

Lip Crater Hole

SEM/EDXA

Particle

Results 0.15 Meteoriod: Fe,Ni,S

Dia. (mm)

(mm)

2.2.2 2.2.3 2.2.4 2.2.5

A6 FC9 LA4 D10

MDPS bumper MDPS bumper trunnion brace MDPS bumper

port/lower fwd cone port stbd/upper

0.8mm Al 6061-T6 0.8mm Al 6061-T6 Betacloth 0.8mm Al 6061-T6

---------

--0.70 0.60 --0.45

0.45 0.55 0.4 0.31 0.35

-----------

2.2.6 2.2.7 2.2.8

A2 A12 D12

grapple fixture MDPS bumper MDPS bumper

port stbd/upper stbd/upper

Al 6061-T6 0.8mm Al 6061-T6 0.8mm Al 6061-T6

2.1 -----

1.05 --1.80

0.85 0.45 ---

--0.95 Unknown 1 --0.50 Unknown 1.20 --- Orbital Debris: SS

0.25 0.50 0.10 0.50

Figure 13. Impacts on MPLM FM2 flight 2

1

Unknown 1 Unknown Orbital Debris: Al Meteoroid 1

0.14 ----0.10 0.24 ----0.19

Figure 14. Impact 2.2.8–1.2 mm diameter hole in MDPS

Figure 15. Results of SEM/EDX analysis—Impact 2.2.8 IMP A C T D E TA IL S Impact

Location

Number Map ID Type D6 MDPS bumper 1.3.1 D6 MDPS bumper 1.3.2

Estimated Depth

Diam (mm)

Region port/lower

Material Spall 0.8mm Al 6061-T6 ---

0.8mm Al 6061-T6 0.8mm Al 6061-T6

-----

Lip Crater Hole

Results 0.60 Meteoroid

(mm)

0.60 0.70 0.68

0.45 0.55 0.5

-------

0.60 Unknown 1 0.45 Unknown

1.3.3

D6

MDPS bumper

port/lower port/lower

1.3.4 1.3.5 1.3.6 1.3.7 1.3.8 1.3.9 1.3.10

D5 C6 C4 A2 B2 B2 3-2

MDPS bumper MDPS bumper MDPS bumper grapple fixture MDPS bumper MDPS bumper intercostal

port/lower port/lower port/lower port port/upper port/upper port/upper

0.8mm Al 6061-T6 --0.8mm Al 6061-T6 --0.8mm Al 6061-T6 --Al 6061-T6 4.25 0.8mm Al 6061-T6 --0.8mm Al 6061-T6 --0.8mm Al 6061-T6 ---

0.43 0.40 0.63 1.00 0.70 0.48 0.80

----0.5 0.82 0.55 0.4 0.65

---------------

----0.35 0.65 0.5 0.25 0.45

1.3.11 1.3.12

B2 C12

MDPS bumper MDPS bumper

port/upper stbd/upper

0.8mm Al 6061-T6 0.8mm Al 6061-T6

0.25 0.45

-----

-----

-----

-----

SEM/EDXA

Figure 16. Impacts on MPLM FM1 flight 3

Particle Dia. (mm) 0.28

1

— —

Unknown 1 Unknown Orbital Debris: Paint Orbital Debris: SS 1 Unknown Orbital Debris: Paint Meteoroid

1

— — 0.23 0.20 --0.17 0.21

1

-----

Unknown 1 Unknown

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL

185

Figure 17. Summary of MPLM impacts from known sources

panels and one impact was observed on the port grapple fixture. Samples were obtained at all 12 impact sites. SEM/EDX analysis did not reveal impact residue in seven of the samples. Of the five remaining, three were determined to be orbital debris.

4. Summary Through five MPLM missions, 26 hypervelocity impact sites have been observed during postflight inspections. Although more than 80% of the impacts occurred on the MDPS panels, other parts of the module also sustained impact damage. Samples were obtained at 24 impact sites. SEM/EDX analysis was performed on each sample and 12 yielded results that could be interpreted as either meteoroid or orbital debris in nature. Of the known impact sources, seven were determined to be orbital debris and five were meteoroid. The table in figure 17 provides a summary of the 12 impacts. The estimated diameter of the particle that caused the damage is shown in the last column of the table, sorted from least to greatest.

5. Bumper Code Predictions An analysis of the expected number of MDPS outer wall perforations was performed using the BUMPER code, the meteoroid/orbital debris risk assessment software tool used by NASA to determine risk for the International Space Station [7] and the space shuttle [8]. The intent was to determine if BUMPER code predictions for the number of MDPS perforations came close to matching observations. Another type of “prediction vs. observation” activity that can be performed with the BUMPER code is comparison of observed impact locations to the predicted distribution on the MDPS.

186

JAMES L. HYDE ET AL.

1

2

3

4

Figure 18. Overview of BUMPER code

5.1. MDPS OUTER WALL PERFORATION CALCULATIONS

Using attitude information from the as-flown timeline in addition to the orbital and environmental parameters shown in the table below (Fig. 19), an analysis was performed with the BUMPER code to determine if the number of observed MDPS outer wall perforations agrees with predictions. Mission Duration Altitude Orbit Inclination Flight Year LVLH Attitudes - RPY Orbiter PYR Euler Sequence

(Exposure Hrs | % Duration) Orbiter Finite Element Model Orbital Debris Density Orbital Debris Environment Meteoroid Density Meteoroid Enviroment Meteoroid Velocity Distribution Meteoroid Showers

10 days 18 hours 258 hours 400 km 51.6° 2001 bias -XLV +ZVV = 0°, 113.5°, 0° 130 hrs bias +ZLV +XVV = 0°, 327°, 0° 5 hrs bias -XLV +YVV = 90°, 90°, 23.5° 5 hrs -ZLV -XVV = 0°, 180°, 0° 118 hrs ISS 5A.1 mated with orbiter, MPLM on Node 1 constant, 2.8 g/cc ORDEM2000 variable, 0.5 - 2.0 g/cc SSP30425 Rev. B variable, SSP30425 none

Figure 19. BUMPER code inputs

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL Pre & Post Dock

187

ISS TEA

Space

RPY = 0°, 180°, 0° Velocity

RPY = 0°, 113.5°, 0°

Space Velocity

Reboost

Waste Dump

RPY = 0°, 327, 0°

RPY = 90°, 90°, 23.5°

Space Space

Velocity Velocity

Figure 20. BUMPER finite element models showing typical MPLM mission attitudes

Risk is calculated from the BUMPER results using eq. (2) Risk—(1 − (e−N )),

(2)

where N is the expected number of perforations. It is usually expressed in percent notation. Perforation odds are the reciprocal of the risk, expressed in the familiar “1 in x” notation. Figure 21 presents the calculated expected number of MDPS outer wall perforations for a 258 h mission. MPLM REGION

Expected Number of Perforations M&D Deb Met

fwd cone Bay 1 Bay 2 Bay 3 Bay 4 Bay 5 Bay 6 Bay 7 Bay 8 Bay 9 Bay 10 Bay 11 Bay 12 aft cone

0.0057 0.0433 0.0358 0.0172 0.0166 0.0351 0.0428 0.0272 0.0090 0.0021 0.0022 0.0086 0.0271 0.0185

0.0028 0.0048 0.0070 0.0056 0.0057 0.0073 0.0053 0.0027 0.0013 0.0006 0.0006 0.0011 0.0023 0.0025 TOTALS

0.0085 0.0481 0.0428 0.0228 0.0223 0.0425 0.0481 0.0299 0.0103 0.0028 0.0028 0.0097 0.0294 0.0211 0.34

Risk 0.9% 4.7% 4.2% 2.3% 2.2% 4.2% 4.7% 2.9% 1.0% 0.3% 0.3% 1.0% 2.9% 2.1% 28.9%

Odds 1 in 118 1 in 21 1 in 24 1 in 44 1 in 45 1 in 24 1 in 21 1 in 34 1 in 97 1 in 361 1 in 352 1 in 104 1 in 35 1 in 48 1 in 3

% Total 3% 14% 13% 7% 7% 12% 14% 9% 3% 1% 1% 3% 9% 6% 100%

Figure 21. Expected number of MDPS outer wall perforations for a 258 h mission

188

JAMES L. HYDE ET AL.

Figure 22. MDPS panel locations—forward and aft cones

5.2. MDPS PANEL LOCATIONS

Figures 22 and 23 show the numbering scheme for the MDPS panels. The “top,” “bottom,” “forward,” and “aft” designations refer to the position of the module when it is stowed in the shuttle payload bay. Figure 22 above illustrates the 12 rows or bays of MDPS panels that are referenced in the analysis results table in figure 21.

Figure 23. MDPS panel locations—port and starboard cylinder

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL

189

Figure 24. MDPS impact risk distribution—port side

5.3. IMPACT RISK DISTRIBUTION

Figures 24–26 show the relative distribution of MDPS outer wall perforation risk from meteoroid and orbital debris impacts. The highest risk of bumper perforation can be seen in the red bands 30◦ –70◦ off of the velocity vector on the port and starboard sides. It can be seen in figures 24 and 25 that the MDPS panels on the

Figure 25. MDPS impact risk distribution—tarboard side

190

JAMES L. HYDE ET AL.

Figure 26. MDPS impact risk distribution—forward view (ISS and Orbiter removed for clarity)

end cones have considerably lower risk than the cylinder region. In figure 26, the meteoroid and orbital debris flux “shadowing” effect from the docked Orbiter can be seen in the forward view of the risk distribution. The correlation between observed and predicted impacts is shown in figure 27, where the values from the impact risk plots in figures 24 through 26 are mapped to the MDPS cylinder panels and observed impact locations. Of the 25 observed impacts in the cylinder region, 16 (64%) occurred in the top 2 “high-risk” areas.

6. Discussion This paper details the status of an ongoing inspection and analysis campaign for the MPLM. More flights are planned in the future and postflight inspections will be needed. One item for investigation is the determination of MDPS hole size at onset of significant MLI degradation/rear wall damage. Information of this type could aid in postflight serviceability requirements. The aluminum composition of the MDPS outer wall makes it difficult for the SEM/EDX analysis to discern aluminum projectile residue in the midst of the common background material. Many of the impacts in the database that were assigned to the “unknown” category are probably aluminum in nature. Since aluminum is one of the more common sources of orbital debris impacts, this discrepancy needs to be addressed.

M/OD IMPACTS ON THE MULTIPURPOSE LOGISTICS MODUL

191

Figure 27. MDPS cylinder region risk distribution with locations of observed impacts (dotted line indicates leading edge of module when docked at ISS)

References 1. Liou, J. C., Matney, M. J., Anz-Meador, P. D., Kessler D., Jansen M., Theall, J. R. (2002) The New NASA Orbital Debris Engineering Model ORDEM2000, NASA/TP – 2002-210780, May 2002. 2. Hyde, J. L. Christiansen, E. L., Bernhard, R. P., Kerr J. H., Lear, D. M. (2001) A History of Meteoroid and Orbital Debris Impacts on the Space Shuttle, Proc. Third European Conference on Space Debris, ESA SP-473, October 2001, pp. 191–196 3. See, T. H., Allbrooks, M. A., Atkinson, D. R., Simon, C. G. and Zolensky, M. (1990) Meteoroid and Debris Impact Features Documented on the Long Duration Exposure Facility, NASA documents published by the Johnson Space Center JSC-24608, August 1990. 4. Hyde, J. L. (2000) As-Flown Shuttle Orbiter Meteoroid/Orbital Debris Shield Assessment: Phase 1—Shuttle/Mir Missions, NASA documents published by the Johnson Space Center JSC-28768, January 2000. 5. Hyde, J. L. (2000) As-Flown Shuttle Orbiter Meteoroid/Orbital Debris Assessment: Phase 2 Flights, NASA documents published by the Johnson Space Center JSC-29070, September 2000. 6. MPLM Interface Definition Document (ISS-MPLM-IDD-006), December 2000. 7. Prior, T. G. (2003) International Space Station Meteoroid & Orbital Debris Integrated Threat Assessment #10, NASA documents published by the Johnson Space Center Revision C (JSC29951), April 2003. 8. Hyde, J. L. and Christiansen, E. L. (2001) Space Shuttle Meteoroid & Orbital Debris Threat Assessment Handbook, NASA documents published by the Johnson Space Center JSC-29581, December 2001. 9. Christiansen, E. L. (1991) Shield Sizing Equations, NASA inter-department communication SN-90-131, October 1991.

FUEL OXIDIZER REACTION PRODUCTS (FORP) CONTAMINATION OF SERVICE MODULE AND RELEASE OF N-NITROSODIMETHYLAMINE IN A HUMID ENVIRONMENT FROM CREW EVA SUITS CONTAMINATED WITH FORP WILLIAM SCHMIDL,1 RON MIKATARIAN,1 CHIU-WING LAM,2 BILL WEST,3 VANESSA BUCHANAN,4 LOUIS DEE,4 DAVID BAKER,5 AND STEVE KOONTZ6 1 Boeing, Houston, TX 2 Wyle Laboratories, Houston, TX 3 Hamilton Sundstrand, Houston, TX 4 Honeywell, White Sands, NM 5 NASA White Sands Test Facility,White Sands, NM 6 NASA Johnson Space Center, Houston, TX

Abstract. The U.S. Control Moment Gyros (CMGs) maintain the International Space Station (ISS) vehicle attitude by compensating for disturbances. It is preferred to use CMGs, instead of attitude control thruster firings. However, prior to an extravehicular activity (EVA) on the Russian Segment (RS), the Docking Compartment (DC1) must be depressurized, as it is used as an airlock. When the DC1 is depressurized, the CMGs’ margin of momentum is insufficient to compensate for the disturbance and the service module (SM) attitude control thrusters need to fire to desaturate the CMGs. The SM roll control thruster firings induce fuel oxidizer reaction products (FORP) contamination of the adjacent SM surfaces. One of the components present in FORP is the potent carcinogen N-nitrosodimethylamine (NDMA). Since the EVA crewmembers often enter the area surrounding the thrusters for tasks on the aft end of the SM and when translating to other areas of the Russian Segment, the presence of FORP contamination is a concern. This paper will discuss FORP contamination of the SM surfaces, the potential release of NDMA in a humid environment from crew EVA suits, if they happen to be contaminated with FORP, and the toxicological risk associated with the NDMA release. Key words: fuel oxidizer reaction products, FORP, NDMA, EVA

193 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 193–208. C 2006 Springer. Printed in the Netherlands.

194

WILLIAM SCHMIDL ET AL.

1. Introduction The U.S. Control Moment Gyros (CMGs) maintain the International Space Station (ISS) vehicle attitude by compensating for disturbances. However, when the docking compartment (DC1), the location of the Russian airlock, of the International Space Station (ISS) is depressurized for extravehicular activities (EVAs), the service module (SM) attitude control thrusters have to fire because the CMGs have an insufficient margin of momentum to compensate for the disturbance and must be desaturated. Thruster firings produce fuel oxidizer reaction products (FORP) that can contaminate adjacent surfaces. For EVAs on the aft end of the service module (SM) of the Russian Segment (RS), there is a concern that when the EVA crewmember translates around the FORP contaminated area they could inadvertently brush against the FORP and transfer some of it to their suit. FORP is composed of both volatile and nonvolatile components. How fast the volatile components leave varies. One of the components present in FORP that represents a toxicological risk to the crew is the potent carcinogen N-nitrosodimethylamine (NDMA). Although NDMA is volatile, it does persist long enough to be a concern. In addition, when dried FORP is reintroduced into a humid environment, like the ISS cabin, NDMA can reform from the components remaining in the FORP. So the concern is that when FORP (on the suit) is brought back into the humid environment of the ISS cabin, it can release NDMA into the atmosphere.

2. Background Discoloration around the SM zenith roll thrusters was observed during the ISS flight 5A Orbiter fly around of the ISS, as shown in the image in figure 1. In the image, the pitch thrusters are closest to the aft end of the SM (right side of the image). The roll thrusters are to the left of the pitch thrusters and raised above the surface of the SM. The discoloration (brown color) can be clearly seen close to the roll thrusters in the inset zoom image. The EVA handrails can be seen in the inset image close to the roll thrusters and on either side of the pitch thrusters. Contamination has also been observed around the SM nadir roll thrusters. Figure 2 shows the relative position and direction of the SM attitude control thrusters. It can be seen that the roll thrusters’ plume centerline is directed 47◦ away from the surface normal (lower diagram), while the pitch and yaw thrusters’ plume centerline is directed 13◦ away from the surface normal (top diagram). In addition, the pitch and yaw thrusters are recessed below the SM surface, while the roll thrusters are elevated above the surface to provide the roll control component. So the plumes from the roll thrusters are more likely to contaminate the adjacent

FUEL OXIDIZER REACTION PRODUCTS

195

Figure 1. Fight 5A observation of darkening near service module (SM) zenith roll thrusters. The inset image shows an enlarged image of the zenith roll and pitch thrusters. The brownish discoloration is visible near the role thrusters

SM surfaces than the plumes from either the pitch or yaw thrusters. The ISS External Contamination team has concluded that the discoloration is plume contamination due to the thrusters firings. In such areas, it must be assumed that FORP is present. Roll Thrusters

13°

Pitch Thrusters Yaw Thrusters

Contamination expected from roll Thrusters

Pitch Thrusters 47°

Roll Thrusters

Figure 2. Service module (SM) attitude control thrusters’ position and direction. The pitch and yaw thrusters point away from the vehicle surface (13◦ from the normal). The roll thrusters point 47◦ from the normal. contamination from the roll thrusters is expected on the adjacent SM surfaces

196

WILLIAM SCHMIDL ET AL.

Figure 3. (a) Service module (SM) Gas Dynamic Protection Unit (GZU) for the SM roll thrusters prior to flight and installation; (b) SM roll thrusters prior to installation of the GZU; and (c) SM roll thrusters (right side of the image) and pitch thrusters (left side of the image) prior to installation of the GZU

The image in figure 1 was acquired before shields, gas dynamic protection devices (GZUs), were installed on the thrusters in January 2002. Rocket and Space Corporation Energia (RSC-E) has designed the GZUs to constrain the thruster plume and limit contamination of the surrounding surfaces. Figure 3(a) shows the GZU for one of the roll thrusters prior to flight. It fits over the top of the roll thrusters. The handle on the top of the GZU is used during installation and is not used for translating during EVAs, as it becomes contaminated. The brackets on the sides of the GZU are used to install the GZU on fittings that were preinstalled on the SM surface prior to flight. Figure 3(b) shows an image of roll thrusters prior to installation of the GZU. Figure 3(c) shows the other side of the roll thrusters (right side of the image) and the corresponding pitch thrusters (left side of the image). It should be noted that the GZUs used for the pitch and yaw thrusters are different than the ones used for the roll thrusters because they are recessed below the SM surface. Images have been taken at regular intervals from the DC1 window of the SM nadir attitude control thrusters and the Russian Kromka experiment. Figure 4 shows one of these images. Kromka is an experiment to measure how well the GZUs are performing and to test how well some material samples perform in space [3]. The Kromka experiment is the material tray visible in the middle of the image. It has some small material samples mounted on it. In front of the Kromka are the pitch thrusters with its GZU installed. In front of the pitch thrusters are

FUEL OXIDIZER REACTION PRODUCTS GZU for SM pitch thrusters

197

Kromka experiment

GZU for SM roll thrusters

Figure 4. Image taken from Docking Compartment 1 (DC1) window. Kromka is visible in the middle of the image

the roll thrusters with its GZU installed. It can be seen that the GZUs are heavily contaminated compared to the preflight GZU in figure 3(a). Figure 5 shows a diagram of the nadir side of the SM and the relative position of the Kromka experiment and the thrusters. The Kromka experiment is visible near the SM pitch thrusters. EVA handrails can be seen on either side of the pitch thrusters.

3. Eva Constraints Due to the presence of FORP on the SM surfaces adjacent to the roll thrusters, additional EVA constraints were required to be implemented in these areas. The constraints were initially established through a nonconformance report (NCR) that discussed the removal of the Kromka 1-0 experiment and installation of the Kromka 1-1 experiment and in subsequent ISS Program Safety Review Panel (SRP) discussions. The constraints were later confirmed in the protocol from a

Figure 5. Nadir side of service module (SM). The Kromka experiment is visible near the SM pitch thrusters

198

WILLIAM SCHMIDL ET AL.

joint U.S./Russian FORP technical interchange meeting held in Houston, TX, April 15–26, 2002. The EVA constraints were initially developed because the Kromka experiment is in close proximity to the SM thrusters, as can be seen in figure 5, and the EVA crewmembers would need to enter that area. The constraints included establishing a one meter keep-out zone (KOZ) around the thrusters for 2.5 h after the last SM thrusters fired before the EVA crew members could enter the area, procedures for inspecting the EVA suits prior to ingress back into the airlock, and procedures for wiping the gloves and suit with towels that are jettisoned to retrograde. Also, once inside the ISS, the EVA gloves are bagged to mitigate any potential risk from FORP. Since EVAs are generally very time constrained, the ISS Program approved a test program at the NASA White Sands Test Facility (WSTF) to obtain FORP test data that could be used to determine if those EVA constraints could be relaxed.

4. NASA White Sands Test Facility (WSTF) Laboratory Tests 4.1. INTRODUCTION

A test program was setup at the NASA White Sands Test Facility to obtain the data that would be needed to determine what EVA constraints would be required to mitigate the risk of an EVA crewmember inadvertently contacting a FORP contaminated surface and bringing the FORP back into the humid ISS cabin. The program included tests to determine the evaporation rate of FORP on the zenith (25◦ C, hot) and nadir (−40◦ C, cold) sides of the ISS, the evaporation rate of NDMA within the FORP on the zenith and nadir sides of the ISS, the quantity of NDMA that would be released in a humid pressurized environment from the dried FORP, and the rate at which NDMA reforms when dried FORP is introduced into a humid environment. Two groups of tests were performed. The first group of tests were performed during Calendar Year 2003. Results from this group of tests are designated “CY 2003.” Results from these tests included 100 h evaporation rate data for FORP for both the zenith and nadir cases, NDMA evaporation rate data for the corresponding zenith and nadir cases, and NDMA formation rate data. Based on these results, the time to remain outside the keep-out zone (KOZ) after the last SM thrusters fired was reduced from 2.5 to 2 h. Additional tests were requested to determine if the time to remain outside of the KOZ could be further reduced from 2 to 1 h. This was the series of tests performed during early Calendar Year 2004 and designated “CY 2004.” Results from these tests included 1 to 6 h evaporation rate data for FORP for the zenith

199

FUEL OXIDIZER REACTION PRODUCTS TABLE 1. FORP composition ion results Ionic species CY 2003 Residue (10%)b CY 2004 CY 2004 Residue (14%)b a b

Ammonium (%/wta )

Methylammonium (%/wta )

Dimethylammonium (%/wta )

Nitrate (%/wta )

Nitrite (%/wta )

0.05

1.1

7.9

36

0.08

0.6 0.3

2 1.7

20 13

9 48

20 0.6

%/wt—Weight percent Ionic Species were analyzed after UFORP was subjected to vacuum for 5 days

and nadir cases, NDMA evaporation rate data for the corresponding zenith and nadir cases, and NDMA formation rate data.

4.2. PREPARATION OF FORP

For each group of tests, a batch of FORP was generated using a permeation technique developed at the NASA White Sands Test Facility (WSTF) [2]. In this technique, separate Unsymmetrical Dimethylhydrazine (UDMH) and NO2 gas streams are concentrated in a small controlled area. The batch of FORP needed for the tests was prepared over a couple of weeks. For the formation test, a sample of FORP from each batch was evaporated for 5 days at 25◦ C in a vacuum to generate a sample of dried FORP. Since the FORP was prepared over a long duration, the composition of the two batches of FORP varied. The composition of the FORP batches used in the tests is shown in table 1. For the CY 2003 FORP, only the composition of the evaporated sample was measured. When comparing the two evaporated samples, it can be seen that the CY 2004 FORP has a higher concentration of dimethylammonium, nitrates, and nitrites, 7.9% vs. 13%, 36% vs. 48%, and 0.08% vs. 0.6% respectively, than the CY 2003 FORP. The higher concentration of dimethylammonium, nitrates, and nitrites in the CY 2004 FORP likely explains why it has more mass remaining after the 5 day evaporation, 14% vs. 10%, than the CY 2003 FORP. The nitrite levels of the CY 2004 FORP before evaporation and after evaporation, 20% of the initial mass and 0.6% of the 14% mass remaining, indicate that the nitrite concentration is decreasing. A lower nitrite level significantly decreases the NDMA formation rate. This was seen in the results that will be discussed later. The concentration of dimethylammonium is also higher in the CY 2004 dried FORP, 13% vs. 7.9%, than in the CY 2003 FORP. The higher concentration of dimethylammonium and nitrites indicates that a higher NDMA formation rate

200

WILLIAM SCHMIDL ET AL.

would be expected for the CY 2004 FORP than the CY 2003 FORP. This was seen in the results. 4.3. MEASUREMENT TECHNIQUE

The technique used to measure the evaporation rate for FORP and NDMA and the NDMA formation rate is discussed in the WSTF report “Evaporation rate study and NDMA formation from UDMH/NO2 Reaction Product,” WSTF-IR-0188001-03 (2003) [1]. 4.3.1. Evaporation test The FORP evaporation rate was determined by measuring the difference between the initial weight of FORP on a slide before it was put in a vacuum chamber and the final weight of FORP on the slide after it was removed from the vacuum chamber. To determine the initial weight of FORP placed on the slide, an empty syringe was weighed, then ∼0.200 ml of FORP was put in the capped syringe and the syringe was reweighed. This was necessary because of the presence of NDMA in the FORP. A blank slide was also weighed. The weight of the blank slide, and the difference in weight of the syringe with and without the FORP is the weight of the initial FORP and slide. The evaporation test was performed in a vacuum chamber at 10−3 Torr. Earlier tests performed by Dee [2] showed that more FORP and NDMA would be removed for tests performed at 10−6 Torr. So the results from these tests will be more conservative than what would be expected on-orbit. After the samples were removed from the vacuum chamber, they were allowed to warm up to room temperature in a dessicator before being reweighed. The NDMA evaporation rate was measured by using gas chromatographymass spectroscopy (GC-MS). The NDMA concentration in the initial sample was measured before the evaporation test. After the evaporation test, the FORP remaining was rinsed from the slide and the GC-MS test performed again. 4.3.2. NDMA formation test The NDMA formation test was performed using a solid phase microextraction (SPME) test. The NDMA formation rate was measured by sampling the headspace above the sample at selected time intervals. 4.4. TEST RESULTS

4.4.1. NDMA evaporation rate The test results in figure 6 show the concentration of NDMA relative to the initial mass of FORP decreases rapidly. The test results are for the nadir (−40◦ C, cold) case. The red squares are from the CY 2004 tests and the blue squares are from

201

FUEL OXIDIZER REACTION PRODUCTS μg NDMA/g FORP(t=0) vs Time (hrs) 100000 μg NDMA/g FORP(t=0)[2003(-40°C)] μg NDMA/g FORP(t=0)[2004(-40°C)]20

10000

μg NDMA/g FORP(t=0)

1000

100

10

1 0

1

2

3

4

5

6

7

Time (hrs)

Figure 6. Concentration of NDMA relative to the initial mass of FORP (μg NDMA/g FORP) for the −40◦ C case (nadir, cold case) decreases rapidly with time

the CY 2003 tests. It can be seen that the CY 2004 FORP starts (time = 0) with a higher concentration of NDMA than the CY 2003 FORP. However, by the time it reaches the 2 and 4 h data points, the concentration of NDMA relative to the initial amount of FORP is comparable and the CY 2004 and CY 2003 data points overlap. The results in figure 7 are for the zenith (25◦ C, hot) case. It can be seen that the NDMA concentration drops more rapidly for the 25◦ C case than for the −40◦ C case. After 1 h the concentration has dropped approximately 2 orders of magnitude, compared with the 1 order of magnitude for the −40◦ C case. It can also be seen that by 1 h the concentration drop has reached a plateau. 4.4.2. FORP evaporation rate Figure 8 shows that the FORP volatilizes rapidly to a stable mass that persists over a longer period of time. The results show that within 1 h the FORP has reached a stable mass and that the CY 2004 FORP settles at a higher mass than the CY 2003 FORP. For CY 2004, the FORP remaining after 1 h is ∼36%. The results for CY 2003 showed 12–22% of the initial mass remained. The higher mass remaining is likely due to the higher concentration of nitrites and nitrates present in the CY 2004 FORP. The results in figure 9 are for the zenith (25◦ C, hot) case. It can be seen that the FORP volatilizes rapidly to a stable mass. These results showed that the CY 2004 FORP remaining after 1 h was 22% and that for CY 2003 it was 10–11%. Again,

202

WILLIAM SCHMIDL ET AL. μg NDMA/g FORP(t=0) vs Time (hrs) 100000 μg NDMA/g FORP(t=0)[2003(25°C)] μg NDMA/g FORP(t=0)[2004(25°C)])

μg NDMA/g FORP(t=0)

10000

1000

100

10

1 0

2

1

3

4

5

6

7

Time (hrs)

Figure 7. Concentration of NDMA relative to the initial mass of FORP (μg NDMA/g FORP) for the 25◦ C case (zenith, hot case) decreases rapidly with time. Within 1 h the concentration level has reached a plateau

FORP Weight % vs Time (hrs) 100

FORP Weight %

(FORP Weight %)[2003(−40°C) (FORP Weight %)[2004(−40°C)

10

1 0

1

2

3

4

5

6

Time (hrs)

Figure 8. FORP weight (%) vs. time for the −40◦ C case (nadir, cold case)

7

203

FUEL OXIDIZER REACTION PRODUCTS FORP Weight % vs Time (hrs) 100 (FORP Weight %)[2003 (25°C)] (FORP Weight %)[2004 (25°C)]

FORP Weight %

3 pts

2 pts

10 2 pts

2 pts

1 0

1

2

3

4

5

6

7

Time (hrs)

Figure 9. FORP weight (%) vs. time for the 25◦ C case (zenith, hot case). FORP volatilizes rapidly to a stable mass that persists for a longer period of time

the higher mass remaining is likely due to the higher concentration of nitrites and nitrates. 4.4.3. FORP formation rate Table 1 shows the nitrite levels of the CY 2004 FORP before evaporation and after evaporation, 20% of the initial mass and 0.6% of the 14% mass remaining. The nitrite levels indicate that the Nitrite concentration in the FORP is decreasing. A lower nitrite level in the FORP will decrease the NDMA formation rate. This was observed in the results, as no NDMA formation was detected in the CY 2003 FORP when moisture was introduced into the sample of dried FORP. For the CY 2004 sample of dried FORP, the NDMA formation rate was negligible when moisture was introduced. It was determined that to form NDMA, nitrite has to be present in the sample. Both the CY 2003 and CY 2004 dried FORP samples were spiked with nitrite before the formation rate was measured. The nitrite introduced was 25% of the mass of nitrate present in the sample. Table 1 also shows that the concentration of dimethylammonium is higher for the CY 2004 FORP, 13% vs. 7.9%. The higher concentration of dimethylammonium indicates that a higher NDMA formation rate would be expected for the CY 2004 FORP. Figure 10 shows the NDMA formation rate that was measured. The plot shows the rate for both the CY 2003 and CY 2004 FORP. The solid symbols indicate FORP where nitrite is added, and the open symbols are FORP with no nitrite

204

WILLIAM SCHMIDL ET AL. μg NDMA / g FORP Formed vs Time

1.00E+05

μg NDMA / g FORP

1.00E+04

1.00E+03

1.00E+02

2004 μg NDMA/g FORP (Nitriteadded) 2004 μg NDMA/g FORP (Nonitriteadded) 2003 μg NDMA/g FORP (Nitriteadded)

1.00E+01

2003 μg NDMA/g FORP (Nonitriteadded)

1.00E+00

1.00E-01 0

1

2

3

4

5

6

7

Time (hrs)

Figure 10. NDMA concentration (μg NDMA/g FORP) vs. time measured during the NDMA formation test. It can be seen that the NDMA forms rapidly

added. It can be seen that when no nitrite is present the NDMA formation rate is very low. For the CY 2003 FORP, no NDMA formation was detected. For the CY 2004, a low rate of ∼100 μg NDMA/g FORP was measured. The results in figure 10 also show that for the CY 2003 FORP, the NDMA formation rate is 1800 μg NDMA/g FORP present after 2 h and that the formation rate for the CY 2004 FORP was higher, 18400 μg NDMA/g FORP present after 2 h. One cause for the higher formation rate is the higher Dimethylammonium concentration in the CY 2004 FORP. The mass of CY 2004 FORP remaining after the 5 days of evaporation was also higher. This might indicate that there were other components present in the FORP that might have resulted in a higher NDMA formation rate.

5. Methodology A methodology was developed to determine the FORP and NDMA remaining on the SM surface after the SM roll thrusters fire prior to an EVA and the subsequent release of NDMA in a humid environment due to inadvertent contact by an EVA crew member with the service module (SM) surface in the vicinity of the SM roll thrusters contaminated with FORP. The first step is to calculate how much FORP would be deposited on the adjacent SM surfaces by the thrusters firing prior to an EVA. To calculate the FORP remaining on the SM surface, the Russian plume model was used. The thruster

FUEL OXIDIZER REACTION PRODUCTS Roll Thrusters with GZUs

205

Roll Thrusters

Pitch Thrusters with GZUs

Figure 11. Zenith side of service module (SM). Gas Dynamic Protection Units (GZUs) constrain the thruster plume. The distance from the roll thruster to the closest SM surface outside the GZUs is ∼3.2 in (∼8 cm)

firing times were obtained from the ISS Program’s Guidance, Navigation and Control group (GN&C). The value of 45 s was used because it is the longest thrusters firing time that has been observed during the previous EVAs. The next step is to calculate how much FORP would remain 1 h after the SM thrusters have fired. This was determined using the WSTF laboratory evaporation test data for the Nadir case. The conservative value of 36% FORP remaining used. Figure 11 shows a map of the distribution of FORP from the thruster remaining on the SM thruster based on the plume model and laboratory data. It can be seen that the FORP concentration drops rapidly with distance from the thrusters. Next, to be conservative, it is assumed that all the FORP in a 100 cm2 area is transferred to the suit by the inadvertent swipe. Based on the amount of FORP transferred, the amount of NDMA that would be released from this FORP inside the cabin in the first 2 h was calculated using the rate of 400 μg NDMA/g FORP present. This value was obtained from the WSTF data based on the evaporation rate of NDMA for the zenith hot case. The amount of NDMA that would be formed from the dried FORP in the first 2 h was also calculated based on the more conservative CY 2004 rate, 18400 μg NDMA/g FORP present. Figure 12 shows that since the GZUs were installed, the closest point on the SM surface that can be touched is ∼3.2 in (∼8 cm) from the edge of the thruster. The diagram is of the zenith side of the SM. The figure on the left shows the roll thrusters with the GZUs installed. The figure on the right shows the distance from the edge of the roll thruster to the closest SM surface outside of the GZUs. The roll thruster diameter is 1.8 in. The distance from the roll thruster centerline to the closest SM surface is ∼5 in (12.7 cm). The predicted concentrations of NDMA with distance from the SM thrusters that would be released by an inadvertent swipe into the DC1 and ISS compartments

206

WILLIAM SCHMIDL ET AL.

Figure 12. Map of the FORP predicted to remain after 1 h on the surface around the SM roll thrusters after a 45 s roll thruster firing. It can be seen that the FORP concentration drops rapidly with distance from the thruster

are shown in tables 2 and 3. It can be seen that the concentration drops off rapidly. Also the closest point is at the GZU itself, which is defined as a “no touch” area by flight rule. This data was given to the Toxicology group for assessment.

6. Toxicological Assessment The quantity of FORP present and NDMA released into the ISS from an inadvertent swipe by an EVA crewmember of the contaminated area around the SM TABLE 2. NDMA concentration predicted to be released in the cabin for FORP transferred to the EVA suit one hour after the roll thruster firing at different distances from the roll thruster for the −40◦ C case (Nadir)

Distance from roll jet (m)

FORP present (g·cm−2 )

NDMA released in pressurized environment

0.08 0.15 0.23 0.30 0.37 0.44

1.96E-02 6.22E-03 2.96E-03 1.72E-03 1.11E-03 7.77E-04

3.69E-02 1.17E-02 5.57E-02 3.23E-02 2.09E-02 1.46E-02

NDMA concentration (ppb) DC1

ISS

965 306 146 84 55 38

34 11 5 3 2 1

at GZU

207

FUEL OXIDIZER REACTION PRODUCTS

TABLE 3. NDMA concentration predicted to be released in the cabin for FORP transferred to the EVA suit one hour after the roll thruster firing at different distances from the roll thruster for the 25◦ C case (Zenith)

Distance from roll jet (m)

FORP present (g·cm−2 )

NDMA released in pressurized environment (g)

0.08 0.15 0.23 0.30 0.37 0.44

1.20E-02 3.80E-03 1.81E-03 1.05E-03 6.81E-04 4.75E-04

2.25E-02 7.15E-03 3.40E-03 1.97E-03 1.28E-03 8.93E-04

NDMA concentration (ppb) DC1

ISS

589 187 89 52 33 23

21 7 3 2 1 1

at GZU

roll thrusters was calculated and provided to the NASA Toxicology group for an assessment. This data is shown in tables 2 and 3. Acute toxicity has not been reported at concentrations below 10 ppm. Therefore, it can be concluded that at concentrations below 1 ppm, it is very unlikely that NDMA will produce any acute toxic reactions. So, the NASA Toxicology Group concluded that for the concentration levels expected, it is unlikely that NDMA will produce acute toxicity. The NASA Toxicology Group also found that the highest calculated risk from the projected NDMA concentrations is less than 8.46 × 10−5 (−40◦ C, distance 0.08 m from the thrusters). The NASA Toxicology Group, with the concurrence of the National Research Council Spacecraft Maximum Allowable Concentrations (SMAC) Subcommittee, accepts a cancer risk of 1/10,000 (i.e., 10−4 ) in deriving SMACs on carcinogenic compounds, such as benzene.

7. Summary Prior to an extravehicular activity (EVA) on the Russian Segment (RS), the Docking Compartment (DC1) must be depressurized, as it is used as an airlock. When the DC1 is depressurized, the CMGs’ margin of momentum is insufficient to compensate for the disturbance and the Service Module (SM) attitude control thrusters need to fire to desaturate the CMGs. The thruster firings result in FORP contamination of the adjacent SM surfaces. The FORP contamination of the SM surfaces, the release of NDMA in a humid environment from crew EVA suits, if they happen to be contaminated with FORP, and the toxicological risk associated with the NDMA release were calculated. It was determined that the FORP and NDMA evaporate rapidly and that their concentration drops off rapidly with distance from the thrusters. The

208

WILLIAM SCHMIDL ET AL.

FORP remaining after 1 h for the Nadir case was found to be 36% of the initial mass. For the Zenith case the FORP remaining after 1 h was 22% of the initial mass. The NASA Toxicology Group found that the highest calculated risk from the projected NDMA concentrations is less than 8.46 × 10−5 (−40◦ C, distance 0.08 m from the thrusters). The NASA Toxicology Group, with the concurrence of the National Research Council Spacecraft Maximum Allowable Concentrations (SMAC) Subcommittee, accepts a cancer risk of 1/10,000 (i.e., 10−4 ) in deriving SMACs on carcinogenic compounds, such as benzene. Based on these results the time to remain outside the 1 m KOZ could be reduced from 2 to 1 h.

Acknowledgments The authors gratefully acknowledge the ISS Program office for supporting this study, and the NASA White Sands Test Facility (WSTF) and its personnel for conducting the study.

References 1. Buchanan, V. D. and Dee, L. A. (2003) White Sands Test Facility (WSTF) Investigative Report, WSTF-IR-0188-001-03. 2. Dee, L. A., Davidson, V. D., and Baker, D. L. (2000) Protocol External Contamination Technical Interchange Meeting Fuel/Oxidizer Reaction Products (FORP) Plume Induced Contamination, 15–26 April 2002. 3. Naumov, S. F., Gerasimov, Y. I., Sokolova, S. P., Rebrov, S. G., Gerasimova, T. I., Kalistratova, O. V., Prokofyev, M. A., Grigorevsky, A. V., Prosvirikov, V. M., Buryak, A. K., and Chernik, V. N. (2003) In 9th International Symposium on Materials in a Space Environment, ESTEC, Noordwijk, The Netherlands, 16–20 June 2003 .

EFFECT OF VACUUM THERMOCYCLING ON PROPERTIES OF UNIDIRECTIONAL M40J/AG-80 COMPOSITES YU GAO,1 DEZHUANG YANG,1 SHIYU HE,1 AND ZHIJUN LI2 1 Space Materials and Environment Engineering Laboratory, Harbin Institute of Technology, Harbin 150001, P. R. China 2 The 39th Institute China Electronic Science and Technology Groups Inc., Xi-an 710065, P. R. China

Abstract. The vacuum thermocycling in the temperature interval of 93–413 K under vacuum of 10−5 Pa was performed for unidirectional M40J/AG-80 composites, and changes in the bending strength and modulus, the mass loss ratios, the surface morphology and the internal structure transitions were examined. Experimental results show that with increasing the vacuum-thermal cycles, both the bending strength and modulus exhibit an increase trend followed by decreasing after 40 cycles and leveling off after 97 cycles. The mass loss ratio increases firstly and then tends to level off after 48 cycles. The vacuum thermocycling could cause the debonding in the interfacial layers between the AG-80 epoxy matrix and carbon fibers, as well as the post curing in the epoxy matrix. The changes in the bending strength and modulus, the mass loss ratio and the surface morphology of the M40J/AG-80 composites can be related to the debonding and the post curing due to the vacuum thermocycling. Key words: carbon/epoxy composites, vacuum thermocycling, bending properties, debonding effect, loss tangent (tan δ) 1. Introduction Carbon/epoxy composites are extensively used in structural components for spacecraft, such as in the truss structure, antennas, and solar cell panels, etc. [1]. When flying in orbit, the spacecraft goes repeatedly into and out the shadow region of the Earth, leading to a change in the surface temperature [2]. For example, the surface temperature of spacecraft can vary in the interval 113–393 K. Since the thermal expansion coefficients are quite different between carbon fibers and the epoxy matrix, the thermocycling would result in thermally cyclic stresses and strains, leading to a degradation in mechanical properties and chemical structure. 209 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 209–215. C 2006 Springer. Printed in the Netherlands.

210

YU GAO ET AL.

Also, under high vacuum, the mass loss of the epoxy matrix composites could take place. Therefore, it is important to study the behavior of the composites under vacuum thermocycling, which might influence the performance of spacecraft in orbit [3–5]. The AG-80 resin is a new type of thermosetting matrix for advanced polymetric composites, which exhibits a unique resistance to wet and heat [6], showing a good prospect for the usage in spacecraft. The aim of this study was to examine the effect of vacuum thermocycling on the unidirectional M40J/AG-80 composite, in order to provide the basic information for its further development and application in spacecraft. The AG-80 resin is a type of tetraglycidyl diaminodiphenyl methane (TGDDM) with diaminodiphenyl sulfone (DDS).

2. Experimental A commercial grade of AG-80 resin was used as the matrix material, which was supplied by Shanghai Institute of Synthetic Resins in China. The chemical structure is shown in figure 1. Its epoxy value is approximately 0.80, and the curing agent is diaminodiphenyl sulfone (DDS). The M40J carbon fibers were chosen as the reinforcement for the composites, which were supplied by TORAY Company in Japan. The prepreg tape for the M40J/AG-80 composite specimens was laid on a mould board and placed in an autoclave for curing. The curing cycles were: heating to 130◦ C from room temperature with the heating rate 1.5–2.0◦ C·min−1 , and holding for 40–60 min at 130◦ C under the pressure of 0.6–0.7 Mpa; and then heating to 180◦ C with the rate 1.5–2.0◦ C·min−1 and holding for 120 min; finally, cooling to room temperature in the autoclave. The volume fraction of M40J fibers in the manufactured composites was approximately 60%. The vacuum thermocycling regime is shown in figure 2. Before and after the vacuum thermocycling, the bending strength and modulus of the M40J/AG80 composites were measured using a MTS 810 type test machine. The mass loss ratios for the specimens after the thermocycling were characterized by a high precision microbalance with a sensitivity of 10–5 g. The change in surface morphology of specimens due to the vacuum thermocycling was observed using an atomic force microscope (AFM) of Nanoscope IIIa Dimension 3100 type. H C

H2 C

CH2

O H C

H2C O

CH2 N

CH2

CH2

N CH2

H C

CH2 O

H C

CH2 O

Figure 1. The chemical structure of AG-80 resin

211

EFFECT OF VACUUM THERMOCYCLING

Temperature/°C

200

(140°C,10min)

10−5Pa

100 0 −100 −200

(−180°C,10min) 0

10

20

30

40

50

60

70

Time/min

Figure 2. Temperature vs. time curve for the vacuum thermocycling

The laid direction of fibers was parallel to the observed surface of specimens. In order to reveal the effect of thermocycling on the chemical structure of the AG-80 matrix, the dynamic mechanics thermal analysis (DMA) was performed using a Rheometric Scientific DMTA Vtype spectrometer by means of a standard three-point mode. The loading frequency was 1 Hz, the temperature ranged from −130◦ to +300◦ C and the heating rate was chosen as 5◦ C·min−1 . The size of DMA samples was 40 × 7 × 1 mm3 , and the longitudinal direction of samples was parallel to the laid direction of the carbon fibers. 3. Result and Discussion 3.1. BENDING STRENGTH AND MODULUS

Bend strength/MPa

1800

(a)

1700

10−5Pa 93 413K

1600 1500 1400 0

50 100 150 200 250 300 350

Thermal cycles

Bending modulus/GPa

Figures 3(a) and 3(b) show the changes in bending strength and modulus with vacuum-thermal cycles for the M40J/AG-80 composites, respectively. The variation trends for the bending strength and modulus are similar. With increasing the 215 210

(b)

10−5Pa 93 413K

205 200 195 190 185 0

50

100 150 200 250 300

Thermal cycles

Figure 3. The bending strength and (a) modulus vs. (b) vacuum-thermal cycles for M40J/AG-80 composites

212

YU GAO ET AL.

Weight loss/%

0.4 0.3 −5

10 Pa

0.2

93

413K

0.1 0.0 0

40

80

120

160

Thermal cycles

Figure 4. The mass loss ratio vs. vacuum-thermal cycles for M40J/AG-80 composites

vacuum-thermal cycles, both the bending strength and modulus increase firstly and begin to decrease after approximately 40 cycles. After 97 cycles, the decrease tends to level off. 3.2. MASS LOSS EFFECT

The outgassing characteristics of the M40J/AG-80 composites were examined at 125 ± 0.5◦ C for 24 h under the vacuum 10−6 Torr. It was shown that the percentage of total mass loss (TML%), the water vapor reverse amount (WVR%) and the collected vapor condensate matter (CVC%) were 0.46, 0.19, and 0%, respectively. These values indicate that the M40J/AG-80 composite has a good resistance to outgassing under high vacuum environment, meeting the requirement for the materials to be used in spacecraft. Figure 4 shows the mass loss ratio vs. vacuum-thermal cycles. The mass loss ratio increases with the thermal cycles, and tends to level off after 48 cycles. The highest mass loss ratio is approximately 0.38% during the vacuum-thermal cycling. It is believed that the vacuum-thermal cycling could induce the formation of smaller molecules. Also, during the storage, the surface of M40J/AG-80 composites would absorb water molecules, which are vaporized under vacuum. The mass loss of M40J/AG-80 composites under vacuum-thermal cycling could be attributed to the smaller molecules and the water absorption. 3.3. DEBONDING EFFECT

Figure 5 shows the change in the surface morphology after vacuum-thermal cycling for the M40J/AG-80 composites, in which the carbon fibers, the epoxy matrix, and the interfacial layers are indicated by A, B, and C respectively. Before the vacuum thermocycling, the bonding of the matrix with carbon fibers is in good

EFFECT OF VACUUM THERMOCYCLING

213

Figure 5. AFM micrographs showing the change in surface morphology after vacuum-thermal cycling for M40J/AG-80 composites: (a) 0 cycles; (b) 40 cycles; (c) 274 cycles

condition, as shown in figure 5(a). After 40 cycles, debonding occurred in some interfacial layers, see the C areas in figure 5(b). But, with further increasing the thermal cycles, the debonding extent did not change a lot, figure 5(c). 3.4. DMA ANALYSIS

Figure 6 shows the change of the loss tangent (tan δ) vs. temperature spectrum after vacuum thermocycling for the M40J/AG-80 composites. The spectrum can be divided into two portions, as shown in figures 6(a) and 6(b), respectively. Figure 6(a) shows the change of the peaks for the primary or α-transition. The temperature at which the α-transition peak appears is corresponding to the glass transition temperature (Tg ). It can be seen that with increasing the vacuum-thermal cycles, the Tg value decreases slightly. After 40 cycles, the peak height for the αtransition drops obviously (almost by 18%), implying that the cross-linking extent increases for the M40J/AG-80 composites. The increase in the cross-linking extent demonstrates that the vacuum thermocycling would lead to a post curing effect on the epoxy matrix. However, when the thermal cycles increase from 40 to 274, the

214

YU GAO ET AL. 0.30

0.045

(a)

0.040 0.035 0.030

tanδ

tan Ä

0.25

0.20 original state vacuum thermo-cycling for 40 cycles vacuum thermo-cycling for 274 cycles

0.15

0.10

original state vacuum thermo-cycling for 40 cycles vacuum thermo-cycling for 274 cycles

(b)

165 170 175 180 185 190 195 200 205

Temperature/°C

Y

0.025 0.020

β

0.015 0.010 −100

−50

0

50

100

150

Temperature/°C

Figure 6. The tan δ vs. temperature spectra for the M40J/AG-80 composites after vacuum thermocycling for various cycles

peak height change a little, indicating that the post curing does not take place any more. Figure 6(b) shows the changes of the peaks for the secondary transitions, including the β and γ ones. The β peak height decreases with increasing the thermal cycles, and disappears after 274 cycles. The β transition occurring in the temperature range of –10–100◦ C can be related to the movement of benzene rings, which would weaken with increasing the cross-linking extent. The change in the β peak height could also demonstrate the post curing effect during vacuum thermocycling. Also, as can be seen in figure 6(b), the γ -transition takes place in the temperature range of –120–50◦ C, which can be related to the movement of the chain segment –CH2 CHOH(R)CH2 –. Such a chain segment could be formed by breaking the epoxy rings. The γ peak signals are stronger for the vacuum thermocycling of 40 cycles than those before the thermocycling. This phenomenon indicates that the amount of the –CH2 CHOHCH2 – segments in the epoxy matrix is larger in the former state than in the latter one, and thus the post curing would occur due to the thermocycling.

4. Conclusions M40J/AG-80 composite is a new type of carbon/epoxy materials, which could be used in spacecraft. Since the vacuum and thermocycling are important environment factors in space, it is necessary to study the damage effect of vacuum thermocycling on the M40J/AG-80 composites. It was found that with increasing the vacuum-thermal cycles, both the bending strength and modulus exhibited an increase trend followed by decreasing after 40 cycles and leveling off after 97 cycles. The mass loss ratio increased firstly and then tended to level off after

EFFECT OF VACUUM THERMOCYCLING

215

approximately 48 cycles. The vacuum thermocycling could cause the debonding in the interfacial layers between the AG-80 epoxy matrix and carbon fibers, as well as the post curing in the epoxy matrix. The changes in the bending strength and modulus, the mass loss ratio, and the surface morphology of the M40J/AG-80 composites can be related to the debonding and the post curing, which are caused by the vacuum thermocycling. The debonding and post curing mainly occur during the thermocycling less than 40 cycles.

References 1. Xiao, S. and Liu, Z. (1993) Aerospace Materials and Technology 4, 1. 2. George, P. E. and Dursch, H. W. (1994) Journal of Advanced Materials 25(3), 10–19. 3. Shin, K.-B., Kim, C.-G., Hong, C.-S., and Lee, H.-H. (2000) Composites Part B: Engineering 31(3), 223–235. 4. Seehra, S., Benton, D., Rosen, J., and Gounder, R. (1985) SAMPE Journal 21(2), 18–23. 5. Tennyson, R. C. and Matthews, R. (1995) Journal of Spacecraft and Rockets 32(4), 703–709. 6. Wang, R.-M. and Lan, L.-W. (2001) Thermosetting Resin 16(1), 36–38.

DAMAGE CHARACTERISTICS OF Zr41 Ti14 Cu12.5 Ni10 Be22.5 BULK METALLIC GLASS IMPACTED BY HYPERVELOCITY PROJECTILES C. YANG,1,2 C. Z. FAN,3 Y. Z. JIA,3 X. Y. WANG,1,2 X. Y. ZHANG,1 H. Y. WANG,1 Q. JING,1 G. LI,1 R. P. LIU,1 L. L. SUN,2 J. ZHANG,2 AND W. K. WANG1,2 1 Key Lab of Metastable Materials Science and Technology, Yanshan University, Qinhuangdao 066004, P. R. China 2 Institute of Physics, Chinese Academy of Sciences, Beijing 100080, P. R. China 3 School of Material Science and Technology, Harbin Institute of Technology, Harbin 150001, P. R. China

Abstract. The damage characteristics of Zr41 Ti14 Cu12.5 Ni10 Be22.5 bulk metallic glass under planar shock wave have been investigated by firing aluminum projectiles using a two-stage light gas gun. The SEM results show that radial and symmetric cracks formed on the shocked plane of the sample at the impacted location by aluminum projectile at the velocity of 2.7 km·s−1 . Parallel shear crack/bands in the sublayer under the shocked plane were formed. For a better understanding of the response features under shock wave, hypervelocity impact tests with conventional sphere aluminum projectiles were carried out. Besides the same adiabatic shear crack/bands and crack propagations, craters were formed and lamination cracks occurred. Key words: bulk metallic glass, impact, fracture

1. Introduction Multicomponent Zr41 Ti14 Cu12.5 Ni10 Be22.5 bulk metallic glass (BMG) with extremely high glass forming ability, one of the most widely studied BMGs [1], has recently gained considerable attention due to its low density, high strength and fracture toughness, good corrosion and wear resistance, excellent ductibility, and uniquely dynamical deformation characteristics such as adiabatic shear banding. In recent years, this specific alloy has already made commercial application potential such as headway in golf clubs. BMG is also a likely candidate as a 217 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 217–223. C 2006 Springer. Printed in the Netherlands.

218

C. YANG ET AL.

space vehicle material. The BMG used for the spacecrafts might face a danger of collision by space debris and micrometeoroids. In order to protect the space vehicles against the impact of space debris and micrometeoroids, related knowledge of high strain rate response must be obtained to evaluate the dynamic properties of the BMG impacted by hypervelocity projectiles. Extensive research on damage behavior of Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG under the uniaxial compression [2–4], tensile tests [5, 6], and shocking loading by a powder gun loading system [7] has been carried out in the past. However, these studies paid particular emphasis to low and medium strain rate response of the BMG. In this paper, we report the damage characteristics of a Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG under the planar shock wave treatment and the impact of hypervelocity projectiles by accelerating aluminum projectiles using two-stage light gas gun.

2. Experimental Procedure The ingots of the alloy with a nominal composition of Zr41 Ti14 Cu12.5 Ni10 Be22.5 were firstly prepared from a mixture of pure elements in an arc-melting furnace under Ti-gettered Argon atmosphere. The purity of Zr, Ti, Cu, Ni, and Be were 99.999, 99.9, 99.5, 99, and 99.5% respectively. The ingots were crushed and remelted in a quartz tube, and then were quenched into water to obtain a BMG rod. The structure of the BMG was identified by X-ray diffraction (XRD) to be fully amorphous and no crystalline phase was detected. Small cylinders 10 mm long, cut out from the alloy rod, were utilized as impacted target. Targets were recovered using a “soft” recovery device, as shown in figure 1. The amorphous BMG cylinder was embedded into a copper cylinder with a diameter of 40 mm.

Figure 1. Schematic diagram of the recovery device for high speed impact [(1) steel cylinder; (2) steel cylinder tube; (3) aluminum cylinder; (4) polyethylene cylinder; (5) copper cylinder; (6). Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG target]

DAMAGE CHARACTERISTICS OF Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG

219

A steel cylinder was mounted tightly beneath the copper cylinder and aluminum cylinder was mounted tightly beneath the steel cylinder. The polyethylene cylinder was mounted tightly beneath the aluminum cylinder. Finally, the assembly of these cylinders was inserted into a steel cylinder with 40 mm inside diameter and 120 mm outer diameter. The end facing the direction of the cylinder assembly was machined as shown in figure 1. The other end was fixed by bolts. The BMG cylinder samples of diameter 22 mm were impacted by the hypervelocity spherical aluminum projectiles of 5 mm in diameter at speeds of about 2.6, 3.2, and 3.7 km·s−1 , respectively. Though these speeds are not very high, they can cause a catastrophic failure of protected layers for space vehicles [8]. As compared with the damage behavior for hypervelocity impact, planar shock experiment was carried out using the same recovery device except that the BMG sample with a diameter of 18 mm was covered by a copper plate with a thickness of 2 mm. In the planar shock experiment, the cylinder aluminum projectile was fired at a speed of 2.7 km·s−1 . The projectile was designed to be 22 mm in diameter and 2 mm thick in order to avoid the influence of side rarefaction wave and reflected rarefaction wave on the sample. The axis of projectile is coincident with the steel cylinder assembly. The projectiles were fired by two-stage light gas gun. The velocities of the projectiles were measured by the electromagnetic induction method. The specimens were cut from the impacted recovered targets with an electric spark-cutting machine, and were grinded and polished mechanically after cutting. The morphology of the craters and microstructure changes in the region near the crater and deformation damage characteristics of cross section such as adiabatic shear bands and microcracks were observed by XL30 S-FEG scanning electron microscope (SEM).

3. Results and Discussion 3.1. TYPICAL DAMAGE MORPHOLOGY OF CRATERS FORMED IN Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG MEDIUM THICKNESS TARGETS

Figure 2 shows the craters that were formed in the Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG target after the hypervelocity impact with velocities of 2.6, 3.2, 3.7 km·s−1 , respectively. The samples’ projectile 22 mm in diameter and a thickness of 10 mm belong to the medium thick target, in which the impact wave can interact with the side and back faces, resulting in significant effect on the impact process. It can be seen that shape of the craters is almost spherical coronary without an obvious outer projecting fringe in the “mouth” of the crater, which is extruded from the crater by high pressure and commonly observed in the target impacted by hypervelocity projectiles. This may be related to the high strength and low ductility of the BMG

220

C. YANG ET AL.

a

b

c Figure 2. Cross-section morphologies of damage craters formed in the Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMGs medium thick target: (a) sample A: v = 2.6 km·s−1 ; (b) sample B: v = 3.2 km·s−1 ; (c) sample C: v = 3.7 km·s−1

materials. The shape of the crater becomes spherical from spherical coronary with the increase of projectile velocity.

3.2. FRACTURE DAMAGE CRACKS/BANDS UNDER PLANAR SHOCK LOADING

The resulting dimensions of the sample after the planar shock loading with an aluminum flyer at a velocity of 2.7 km·s−1 were 20.5 mm in diameter and 7.3 mm thickness compared to 17.5 mm in diameter and 10 mm thickness for the original sample. The upper surface facing the flyer exhibits an annular crack and many densely radial cracks initiating from a central point. The reason for this phenomenon is that the central part of aluminum flyer collided firstly with the central part of the BMG sample because the flyer was hindered by air in the target housing. The central part of the flyer deforms during the flight. The compressive shock wave therefore acts initially on the center point of the BMG target, thus initiating the failure in this part first. Then, radial cracks propagate from the point under the action of strong shock wave. Annular cracks formed naturally by interaction between the wave (tensile wave) reflected from the cylinder copper wall and compressive shock wave. The recovered aluminum flyer with a prominent

DAMAGE CHARACTERISTICS OF Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG

221

Figure 3. SEM micrographs showing cross-section microstructures of Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG under planar shock loading of aluminum flyer with a speed of 2.7 km·s−1

feature formed along the direction of flight demonstrates experimentally the above discussed phenomenon. Figure 3 shows a SEM microstructure micrograph of Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG target hit by aluminum flyer with a speed of 2.7 km/s. Several fracture planes (shear cracks/bands) 2 mm apart and approximately parallel to each other were seen in the upper part of the sample (that are not shown in figure 3 due to limitation of the view field of SEM). Another kind of shear cracks/bands with an angle of 40◦ between the fracture plane and horizontal direction result from the interaction between the shock wave reflected from the cylinder copper wall and compressive shock wave. This kind of cracks/bands cannot appear in usually uniaxial compressive and tensile tests because the exterior walls of the cylinder sample has not been restrained under these conditions. In our case, as discussed above, the sample was reduced in thickness and expanded in diameter after the impact. Once the shearing occurs along the direction shown by the left arrow, it results in the formation of primarily adiabatic shear band and induces a new secondary shear band along the direction shown by the left arrow. Thus a triangular region with an apex angle of 115◦ is extruded from the sample because of convergence between a reflected shock compressive wave along the radial direction of the sample from the interior wall of the recovery copper cylinder and a compressive shock wave along the projectile trajectory. The upper two shear cracks/bands has already intersected each other, while the subjacent two haven’t. The zigzag-shaped cracks (indicated by right arrow) on the micrograph of the BMG suggest that the spalling of the BMG is characterized by brittle damage resulting from the shear localization even though the sample is under tension during decompression process. Shear cracks/bands propagate and end at a point where several secondary radial microcracks (shear

222

C. YANG ET AL.

Figure 4. Typical SEM micrographs of shear bands/cracks observed on the cross section of the recovered specimen after shock loading with a velocity of v = 3.7 km·s−1

cracks/bands) initiate. It can be predicted that these secondary microcracks would continue to propagate with the increase of shock pressure. 3.3. SHEAR CRACKS/BANDS UNDER THE IMPACT OF HYPERVELOCITY PROJECTILES

Figure 4 shows a typical SEM micrograph of shear bands/cracks observed on the cross section of the recovered specimen after shock loading by a hypervelocity aluminum projectile 5 mm in diameter with a velocity of v = 3.7 km·s−1 . Because the target belongs to medium thick type and the projectile is spherical, the shock wave reflected from the back and side faces interacts with the subsequent compressive wave, leading to sophisticated microdamages. Zigzag-shaped edge side shear band/crack (as indicated by the thick arrow) along vertical direction can be observed clearly, which is a macro-damage feature caused by edge side shear stress induced by hypervelocity projectile. The observed spallation of the BMG is due to propagation of higher energy shear cracks/bands. It can be suggested, therefore, that a crater/hole would be left at the BMG target face by a projectile having a high enough velocity to posses a sufficiently high kinetic energy. Lamination crack phenomenon (as indicated by the thin arrow) with a distance of 2 mm away from back surface can also be observed at our experimental velocity. The reason for this phenomenon is that the reflected rarefaction wave, from recovery casket copper of sample, interacts with compressive waves and form tensile waves in the BMG target due to lower impedance than the BMG. Lamination crack appears spontaneously if tensile stress exceeds the yield strength of the BMG. In our cases shear bands/cracks are visible clearly in the SEM micrographs. Furthermore, the

DAMAGE CHARACTERISTICS OF Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG

223

sizes of these shear bands/cracks increase gradually with the increase of the impact velocity.

4. Conclusions Zr41 Ti14 Cu12.5 Ni10 Be22.5 BMG glasses under shock loading by two-stage light gas gun revealed shear fracture cracks/bands formation. Radial symmetric cracks were formed at the shocked plane. Shear crack/bands parallel to each other on the part close to the shocked plane were formed readily on the cross section of the recovered sample. The depth of crater in the target increases linearly with increase of the projectile velocity in the range of experimental velocity. Edge sides shear band/crack, together with lamination shear band/crack has been observed on the recovered samples impacted by hypervelocity projectiles. Based on these results, it can be concluded that BMG can be used as a candidate for a space vehicle material that resist the impact of small space debris.

Acknowledgment We wish to thank the National Natural Science Foundation for its support of this research through Grant 50171077, 50171059, and 50325103.

References 1. Tang, X. P., Geyer, U., Busch, R., Johnson, W. L., and Wu, Y. (1999) Nature 402, 160–162. 2. Subhash, G., Dowding, R. J., and Kecskes, L. J. (2002) Materials Science and Engineering A334, 33–40. 3. Gilbert, C. J., Ager III, J. W., Schroeder, V., and Ritchie, R. O. (1999) Applied Physics Letters 74, 3809. 4. Wright, W. J., Schwarz, R. B., and Nix, W. D. (2001) Materials Science and Engineering A 319–321, 229–232. 5. Li, J. X., Shan, G. B., Gao, K. W., Qiao, L. J., and Chu, W. Y. (2003) Materials Science and Engineering A 354, 337–343. 6. Flores, K. M. and Dauskardt, R. H. (2001) Materials Science and Engineering A 319–321, 511–515. 7. Zhuang, S. M., Lu, J., and Ravichandran, G. (2002) Applied Physics Letters 80, 4522. 8. Zukas, J. A. (1989) Impact Dynamics (translated into Chinese by Zhiyun Zhang et al.), Publishing Company of Military Industry, P. R. China, p. 182.

EFFECT OF VUV RADIATION ON PROPERTIES AND CHEMICAL STRUCTURE OF POLYETHYLENE TEREPHTHALATE FILM

GUIRONG PENG, DEZHUANG YANG, AND SHIYU HE Space Materials and Environment Engineering Lab, Harbin Institute of Technology, Harbin 150001, P. R. China

Abstract. The effect of VUV radiation on polyethylene terephthalate (PET) film was investigated. A gas-jet type of VUV source with the wavelengths of 5–200 nm was used. The experimental results show that under the VUV irradiation, both the tensile fracture strength and elongation decrease slightly. The spectral absorbance of the PET film increases noticeably with increasing VUV dose. The absorption band mainly forms in the near-ultraviolet region. The XPS, FTIR, and ESR analyses indicate that in the skin layer of PET film irradiated with VUV, the C–O bonds could be broken and decarbonylation occurs, leading to the formation of free radicals of benzene rings as well as a trend of carbonification. The scission of the macromolecule chains, the increase in radical concentration, and the carbonification would cause the degradation in optical properties for the PET film under VUV exposure. Key words: PTFE, Radiation Stability, VUV 1. Introduction Polymeric materials have been widely employed in spacecraft for parabolic antennas, solar wings, and thermal control systems [1–3]. Their performance in space will directly influence the reliability and lifetime of spacecraft. Vacuum ultraviolet (VUV) radiation is one of the major environment factors causing the degradation of polymers. Although the VUV energy percentage in the total spectrum of the Sun is very limited, its photon energy is high enough to break most chemical bonds in the polymers [4]. Therefore, it is important to study the effect of VUV radiation on the chemical structure and properties of polymeric materials, such as the polyethylene terephthalate (PET) film. Most published work about the effects of VUV on polymeric films was performed using various VUV lamps, with which it could be difficult to obtain a VUV spectrum similar to the Sun. A gas-jet type VUV source was used in this study, which could give a spectrum similar to the Sun in the wavelength range of 5–200 nm [4]. The aim of this work was to examine the changes in chemical structure and properties of the PET film under VUV exposure. 225 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 225–232. C 2006 Springer. Printed in the Netherlands.

226

GUIRONG PENG ET AL. O

O

C

C

O

CH2 CH2

O

Figure 1. Chemical structure of the PET film

2. Experimental The PET film was 60 μm thick. Figure 1 shows the chemical structure of the PET film. The film was annealed at 70◦ C for 3.5 h, rinsed with analytically pure acetone and ethanol, and dried in a desiccator for more than 48 h at ambient temperature before the VUV exposure. The VUV source used in this study operates with supersonic argon gas, which is excited by a high-energy electron beam to give the VUV wavelengths ranged from 5 to 200 nm. The VUV intensity of 0.24 w·m−2 was acquired at a distance of 70 cm from the source, corresponding to 10 times the VUV solar constant (VUV suns). In this study, the gas-jet pressure was 5 atm, the electron-beam energy was 1000 eV and the electric current was 15 mA. The vacuum before and after the injection of argon gas into the chamber was 10−5 and 10−3 Pa, respectively. The samples were located at 72, 34, and 22 cm away from the VUV source, corresponding to the 10, 40, and 100 VUV suns respectively. The tensile tests were carried out at room temperature after the VUV exposure. The engaged surface area of the film samples was 20 × 5 mm2 , and the crosshead speed was 2.4 mm·min−1 . The transmittance of the VUV irradiated film samples in the wavelength range of 200–3200 nm was measured with a UV-3101PC scanning spectrophotometer. The X-ray photoelectron spectroscopy (XPS) was performed using a VG-ESCALAB Mark-2 type spectrometer with Mg Kα source. The vacuum in the chamber was 10−6 Pa, and the pass energy 20 eV. IR spectra were acquired using a Perkin Elmer System 2000 Fourier Transform Infrared (FTIR) spectrophotometer. The resolution of the spectrophotometer was 1 cm−1 . The test wave numbers were in the range of 400–4000 cm−1 . The electron-spin resonance analysis (ESR) was performed at room temperature in a JES-FE3AX type spectrometer.

3. Results and Discussion 3.1. TENSILE AND OPTICAL PROPERTIES

Figure 2 shows the changes in tensile strength and elongation for the PET film irradiated with various doses of VUV. It can be seen that both the strength and elongation decrease slightly with increasing the VUV dose.

227

EFFECT OF VUV RADIATION 240

120

(a)

(b)

90 δ, %

σf , MPa

200

160

60

100 VUV suns

30

100 VUV suns

120 0

2000 4000 Irradiation dose, esh

0

6000

0

2000 4000 6000 Irradiation dose, esh

Figure 2. (a) The tensile strength σ f and (b) elongation δ vs. VUV dose for PET film

Figure 3(a) shows the change in spectral absorbance Aλ for the PET film after VUV irradiation. The absorbance limit is at the wavelength of 322 nm and changes little with VUV dose. The Aλ increases with the irradiation dose. Figure 3(b) illustrates the effect of VUV intensity on Aλ . The Aλ in the near-ultraviolet region changes remarkably with the VUV intensity; but less in the visible to near-infrared regions. Moreover, under the same VUV dose, the Aλ in the near-ultraviolet region increases less for the PET film irradiated with 100 VUV suns than that under 40 VUV suns (see the curves 3 and 4), but more than that under 10 VUV suns (see the curves 1 and 2). Figure 4 shows the change in Aλ at 322 and 600 nm with VUV dose. With increasing the dose, the A322 increases rapidly and then tends to level off, while A600 changes almost linearly. 3.2. XPS ANALYSIS

Figure 5 shows the C1s spectra of PET film irradiated with 100 VUV suns and various doses. Before the irradiation, the C1s spectrum would be composed of 0.5 0.6

ΔAλ

3

0.3 0.2

2 1

3, 2940 esh 2, 1600 esh 1, 680 esh 100 VUV suns

0.1

0.4

4

0.3 ΔAλ

0.5 0.4

(b)

(a)

3 0.2

2

4, 1600 esh @ 40 VUV suns 3, 1600 esh @ 100 VUV suns 2, 680 esh @ 100 VUV suns 1, 680 esh @ 10 VUV suns

0.1 0.0

1

0.0 400

600 800 1000 1200 1400 Wavelength, nm

400

600 800 1000 Wavelength, nm

1200

Figure 3. The change in absorbance Aλ vs. (a) VUV dose and (b) intensity for PET film

228

GUIRONG PENG ET AL. 0.6 100 VUV suns

0.5

ΔA 322

ΔAλ

0.4 0.3 0.2

Δ A 600

0.1 0.0 0

600 1200 1800 2400 3000 Irradiation dose, esh

Figure 4. The change in absorbance at 322 nm (A322 ) and 600 nm (A600 ) vs. VUV dose for PET film irradiated with 100 VUV suns

three peaks with binding energy at 284.6, 286.3, and 288.4 eV in turn, as shown in figure 5(a). The three peaks might be related to the carbon in the benzene rings, the carbon singly bound with oxygen in the ethylene, and that in the carbonyl groups, respectively [5–6]. After the VUV irradiation, the C1s spectra could be still characterized with the three peaks, as shown in figures 5(b) and 5(c). Table 1 shows the change in the corresponding area ratio for the three characteristic peaks

(c)

2940 esh

CPS

(b) 680 esh

(a) unirradiated

291

288 285 Binding energy, eV

282

Figure 5. C1s spectra of PET film irradiated with 100 VUV suns for (a) 0, (b) 680, and (c) 2940 esh

229

EFFECT OF VUV RADIATION TABLE 1. The change in area ratios of the fitting peaks for C1s spectra of PET film after 100 VUV suns irradiation Dose (esh) 0 680 1600 2940

Binding energy (eV) and Area ratio (%) 284.6 (69.6) 284.6 (84.7) 284.6 (86.3) 284.6 (89.4)

286.3 (17.1) 286.7 (11.3) 286.5 (10.3) 286.6 (8.4)

288.4 (13.3) 288.8 (4.0) 288.6 (3.3) 288.7 (2.2)

with VUV dose. With increasing the irradiation dose, the area ratio for the peak at 284.6 eV increases, but those at 286.3 and 288.4 eV decrease obviously. Figure 6 shows the change in O1s spectrum after VUV irradiation. Before the exposure, the O1s spectrum is fitted into two Gauss peaks with the binding energy at 531.8 and 533.3 eV, respectively. The former could be related to the oxygen in the carbonyl groups, and the latter to the oxygen singly bound with carbon [5–6]. After the irradiation, the intensities of the two peaks decrease. From the above XPS analysis, it is believed that the single bonds of carbon with oxygen (C–O) in the PET molecules can be broken under the VUV radiation, and further resulting in decarbonylation. 3.3. FTIR ANALYSIS

The FTIR spectra for the PET film before and after VUV irradiation are shown in figure 7. The absorption peaks at 3627 and 3545 cm−1 are related to the vibration of hydroxy1 groups [7–8]. After the VUV irradiation, these two peaks are

(c)

2940 esh

(b)

CPS

680 esh

(a) 0 esh

540

535

530

525

Binding energy, eV

Figure 6. O1s spectra of PET film irradiated with 100 VUV suns for (a) 0, (b) 680, and (c) 2940 esh

230

GUIRONG PENG ET AL. 50

80 40

(b) unirradiated 100 VUV suns, 2940 esh

60

20

T,%

T,%

30

40 (a) unirrradiated 100 VUV suns, 2940 esh

0 3600

3200

2800 −1

Wave number, cm

2400

20 10 0 1800

1700

1600

1500 −1

Wave number, cm

Figure 7. FTIR spectra for PET film before and after irradiation with 100 VUV suns for 2940 esh in the wave number ranges of (a) 3700–2300 and (b) 1800–1480 cm−1

enhanced, indicating that the content of the end hydroxy1 groups might be increased to some extent. In addition, several new absorption peaks appeared in the region from 1680 to 1530 cm−1 . These peaks originate from the change in the substitute groups in the benzene rings and the formation of ethylene groups such as –CH=CH– [7]. 3.4. ESR ANALYSIS

Figure 8 shows the ESR spectra of PET film after VUV irradiation. The g value is 2.004 and the line width 9G. On the basis of the ESR analysis, the free radical concentration in the PET film samples could be estimated using the formula in reference [9], as shown in figure 9. It is shown that the free radical concentration increases rapidly and then tends to level off with the increase of the VUV dose. This trend is similar to that of A322 vs. dose in figure 4, demonstrating that the change of optical properties has a relation with the free radicals.

Figure 8. ESR spectra for PET film irradiated with 100 VUV suns for (a) 680 and (b) 1600 esh

EFFECT OF VUV RADIATION

Spin number, a.u

25

231

100 VUV suns

20 15 10 5 0 0

600 1200 1800 2400 3000 Irradiation dose, esh

Figure 9. The relative radical amount vs. VUV dose for the PET film irradiated with 100 VUV suns

4. Conclusion VUV radiation is a major factor of space environment. Under VUV radiation, the tensile properties of PET film did not change significantly, while the optical ones varied noticeably. With increasing the irradiation dose, the tensile fracture strength and elongation decreased slightly, and the spectral absorbance increased remarkably in the ultraviolet to visible regions. The XPS, FTIR, and ESR analyses indicated that in the skin layer of PET film irradiated with VUV, the C–O bonds could be broken and decarbonylation occurred, forming free radicals of benzene rings. As a result, the condensation extent of the benzene rings was increased, and then a trend of carbonification or enrichment in carbon appeared. The increase in the free radical concentration and the carbonification would cause the degradation in optical properties for the PET film under VUV radiation.

References 1. Nakayama, Y., Imagawa, K., Tagshira, M., Nakai, M., Kudoh, H., Sugimoto, M., Kasai, N., and Seguchi, T. (2001) Journal of High Performance Polymers 13(3), 433–451. 2. Dever, J. A., Pietromica, A. J., Stueber, T. J., Sechkar, E. A., Messer, R. K. (2002) “Simulated Space Vacuum Ultraviolet (VUV) Exposure Testing for Polymer Films”, NASA Report TM2002-211337. 3. Iwata, M., Tohyama, F., Ohnishi, A., and Hirosawa, H. (2001) Journal of Spacecraft and Rockets 38(4), 504–509. 4. Verkhovtseva, E. T., Yaremenko, V. I., and Telepnev, V. D. (1997) In The Proceedings of 7th International Symposium on Materials in Space Environment, Toulouse, France, 16–20 June 1997, p. 119. 5. Beamson, G. and Briggs, D. (1992) High Resolution XPS of Organic Polymers: The Scienta ESCA300 Database, John Wiley and Sons, Chichester, UK. 6. Ektessabi, U. and Yamaguchi, K. (2000) Thin Slid Films 377–378, 793–797.

232

GUIRONG PENG ET AL.

7. Grossetete, T., Rivaton, A., Gardette, J. L., Hoyle, C. E., Ziemer, M., Fagerburg, D. R., and Clauberg, H. (2000) Polymer 41(10), 3541–3554. 8. Miyake, A. (1959) Journal of Polymer Science 38, 497–512. 9. Yin, J. and Mo, Z. (2001) Modern Polymer Physics, Science Publisher, Beijing, 743–829, 679– 680.

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION AT NASA’S MARSHALL SPACE FLIGHT CENTER DAVID L. EDWARDS,1 CHARLES SEMMEL,2 MARY HOVATER,1 MARY NEHLS,1 PERRY GRAY,3 WHITNEY HUBBS,1 AND GEORGE WERTZ1 1 NASA/MSFC, ED31, MSFC, AL 35812 2 Qualis Corporation, ED31, MSFC, AL 35812 3 ICRC, ED31, MSFC, AL 35812

Abstract. Near term solar sail propelled science missions are targeting the Lagrange point 1 (L1) as well as locations sunward of L1 as destinations. These near term missions include the Solar Polar Imager [1] and the L1 Diamond [2]. The Environmental Effects Group at NASA’s Marshall Space Flight Center (MSFC) continues to actively characterize solar sail materials in preparation for these near term solar sail missions. Previous investigations indicated that space environmental effects on sail material thermo-optical properties were minimal and would not significantly affect the propulsion efficiency of the sail [3–5]. These investigations also indicated that the sail material mechanical stability degrades with increasing radiation exposure. This paper will further quantify the effects of space environmental exposure on the mechanical and thermo-optical properties of candidate sail materials. Candidate sail materials for these missions include Aluminum coated MylarTM , TeonexTM , and CP1 (Colorless Polyimide). Experimental data will be presented on the response of sail material to charged particle radiation that indicates change in the mechanical and thermo-optical properties. Thermo-optical property data will also be presented indicating the effects of long-term Near Ultraviolet (NUV) exposure. Charged particle radiation with subsequent Hypervelocity Impact (HVI) results will be presented, indicating that sail material damage is primarily limited to the diameter of the incident projectile. Key words: solar sail, space environment, mylar, polyimide

1. Introduction The non-Keplerian mission opportunities offered by solar sail propulsion are driving many science missions to require the capabilities of solar sail propulsion. This 233 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 233–246. C 2006 Springer. Printed in the Netherlands.

234

DAVID L. EDWARDS ET AL.

renewed interest in solar sails has resulted in escalating development of solar sail materials, structures, and deployment mechanisms. Many challenges confront sail material technologists and among these is the development of low areal density sails. Present state-of-the-art sails have an areal density ranging between 5 and 15 g·m−2 . Sail materials investigated in this effort are on the order of 3 g·m−2 . The farterm goal of sail material manufacturers is a sail material on the orderof 0.01 g·m−2 [6]. As reported by McInnes [7], a solar sail is a thin membrane that uses the momentum carried by photons to propel spacecraft. These photons originate from the sun or can be beamed onto the sail with a laser. If the sail is a good reflector, the momentum transferred to the sail can be almost doubled. Since the momentum carried by a single photon is extremely small, the surface area of a sail must be large to produce a reasonable acceleration, as indicated by a = (2Ap t )/m.

(1)

where, a—nonrelativistic sail acceleration (m·s−2 ), A—surface area (m2 ), pt —incident radiation pressure (N·m−2 ), m—mass (kg). The radiation pressure ( pt ) varies as the inverse square of the distance from the Sun as shown by: 2 pt = 4.56 × 10−6 [(1 + R)/rAU ]

(2)

where, R—surface reflectivity (0 < R < 1), rAU —distance to the Sun in astronomical units (AU). A perfectly reflective sail (R = 1) at a distance of 1 AU from the Sun experiences a light pressure of 9.1 μN·m−2 . Substituting eq. (2) into (1) and solving for the acceleration gives: a = 9.12 × 10−6 [(1 + R)A/mr 2AU ]

(3)

For example, a 1000 kg spacecraft with a perfectly reflective (R = 1) sail area of 106 m2 at a distance of 1 AU from the Sun experiences an acceleration of approximately 1.8 cm·s−2 . As shown in eq. (3), the reflectivity, R, of a sail material is a primary factor in sail performance. A factor not specifically referenced in eq. (3) is the mechanical stability of the film. This property is implied in the variable defining the area, A. Area loss can occur through numerous mechanisms including rips, delaminating

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION 6

Acceleration (cm/s^2)

1 10 2

5

8 10

Sail Area (m^2) 5 5 6 10 4 10

235

5

2 10

Sail Reflectivity

1.5

1

0.5

0

Sail Area

1

0.9

0.8

0.7

0.6

0.5

Ref lectivity

Figure 1. Sail characteristic acceleration dependence on sail material reflectivity and sail area Calculations were made assuming an initial sail area of 1 × 106 m2

of reflective coating, and meteoroid impact. Figure 1 shows the dependence of sail acceleration on the sail material reflectivity and sail area. The Environmental Effects Group at NASA’s Marshall Space Flight Center (MSFC) is tasked with characterizing the material properties of newly developed sail materials and further characterizing these materials in emulated space environments. The ultimate goal of this work is to determine the effect of space environment exposure on sail performance. This paper serves as a status of a work in progress and will report data obtained to-date, and describes the future work planned through September, 2004.

2. Description of the Facilities The process of characterizing materials in a space environment requires the utilization of specialized test facilities. The Environmental Effects Group operates a number of highly specialized facilities dedicated to understanding the effects of the space environment on materials. The facilities described below include: the Pelletron test system for combined exposure effects, Long-Term Ultraviolet radiation exposure facilities, and the MSFC Impact Facility (MIF) for characterizing the effects of micrometeoroid and orbital debris impacts. Material characterization equipment and analysis techniques will be introduced in chapter 3.

236

DAVID L. EDWARDS ET AL.

2.1. PELLETRON TEST SYSTEM

The Pelletron test system provides the unique capability to irradiate material to a simultaneous or sequential exposure to an emulated space environment, and perform in-vacuum reflectance measurements. This facility possesses the capability to irradiate materials to: high energy electrons (100–2.5 MeV), low energy electrons (1–100 keV), protons (40–700 keV) Vacuum Ultraviolet (VUV) (photon radiation over the wavelength range from 121 to 200 nm) and Near Ultraviolet (NUV) (photon radiation over the wavelength range from 200 to 400 nm). The exposure coverage is nominally a 7.6 cm × 7.6 cm (3 inch × 3 inch) area. In-vacuum reflectance measurements can be obtained on two 1-inch diameter coupons or on one 1-inch × 2-inch rectangular sample. The reflectance measurements are taken over the wavelength range from 250 to 2500 nm. The solar absorptance (α s ) is calculated from this spectral reflectance data. The Pelletron system is shown in figure 2. 2.2. LONG-TERM ULTRAVIOLET RADIATION

Ultraviolet radiation is a critical component of the space environment to which sails will be continuously exposed. To characterize the sail material response to long-term Ultraviolet (UV) radiation, sail materials are irradiated, under vacuum, to a solar intensity of 2 UV suns. Solar intensity is measured using a spectral

Protons

Test Chamber VUV

High Energy Electrons Low Energy Electrons

Reflectance Measurement

Figure 2. Photograph of the Pelletron combined space environmental effects exposure system

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION

237

Figure 3. Long-term UV exposure test facility

radiometer with integrating sphere. Calibration on the Near UV (NUV) source is accomplished by measuring the source with a spectroradiometer over the 200–400 nm region. This bandwidth is then integrated to obtain the total energy in W·cm−2 . This value is divided by the sun’s output in that spectral region, as given in ASTM E-490 [8] to determine the number of UV suns. A thermopile that has a flat spectral response is placed at the same plane as the spectroradiometer’s integrating sphere and a millivolt reading is recorded that is compared to the total energy measured by the spectroradiometer. This provides a UV suns per millivolt reading that can be used to determine the source intensity inside the chamber at the sample plane using the thermopile voltage reading. This particular test facility has the capability to irradiate up to 20 1-inch diameter coupons. The long-term UV test system is shown in figure 3. 2.3. HYPERVELOCITY IMPACT

Hypervelocity Impact (HVI) testing was performed on sail material to determine the effect of micrometeoroid impact on sail material. The concern is that the sail material will become brittle by exposure to radiation over the mission lifetime and increase the probability of rip propagation. The HVI testing was performed using the MSFC Impact Facility’s (MSFC) Micro Light Gas Gun (MLGG). The MSFC MLGG has the capability to impact targets with a single projectile. Projectile sizes range from a diameter of 0.4–1.0 mm. Projectile velocity is selectable and ranges from 2 to 7 km·s−1 . Each projectile is encased in a sabot to keep the projectile centered in the barrel to assure shot accuracy. Prior to each HVI test series, calibration shots were performed using Nylon slugs. These calibration shots ensure system operation and validate experimental parameters for desired projectile velocities. Projectile velocity is measured with each test using

238

DAVID L. EDWARDS ET AL.

Figure 4. MSFC MLGG test facility

photodiodes located at each end of the flight tube. Sail materials, investigated for this work, were approximately 15.2 cm × 15.2 cm (6 inch × 6 inch) and mounted in a frame. Test coupons can be as large as 20.3 cm × 25.4 cm (8 inch × 10 inch). The MSFC MLGG is shown in figure 4.

3. Experimental Procedure Sail materials investigated in this on-going activity include: aluminized Mylar, aluminized Teonex, and aluminized CP1. Certain characteristics of the material designs are held as proprietary, so further descriptions will not be discussed within this paper. Test matrices were developed to guide the material characterization for each specific material exposure condition. These matrices will be discussed in the following sections. 3.1. CHARGED PARTICLE EXPOSURE

The charged particle exposure activity focuses on the determination of the mechanical and thermo-optical property response of sail material to various types of exposure conditions (e.g., electron, proton, combined electron and proton). The objective of this work is to determine the radiation tolerance of the candidate sail materials for each type of charged particle radiation. Each candidate

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION

239

10

10

Rads

9E15 e-/sq.cm @ 50 keV 2E14 p+/sq.cm @ 700 keV

9

10

8

10

0

0.5

1

1.5

2 2.5 Microns

3

3.5

4

Figure 5. Flat dose profile of 50 keV electrons and 700 keV protons in aluminized Mylar

material type will be exposed to specific doses of charged particle radiation, ranging from 100 Mrad to over 5 Grad. The dose for each type of radiation is determined by modeling the radiation transport and energy loss during propagation through the sail material. Electron propagation and energy loss is determined using the Integrated Tiger Series 3 (ITS 3) [9]. Proton propagation and energy loss is determined using the Transport of Ions in Matter (TRIM) code [10]. To accurately characterize the sail material response to a dose from charged particles, the dose must be uniform throughout the entire thickness of the sail material. This dose is term to be a “uniform” or “flat” dose profile. Figure 5 shows a typical “flat” dose depth profile for 50 keV electrons and 700 keV protons in aluminized Mylar. 3.2. LONG-TERM UV EXPOSURE

Candidate sail materials were characterized, prior to UV exposure, to determine the baseline thermo-optical properties (e.g., solar absorptance, emittance, and reflectivity). These samples are exposed to 2 suns of UV radiation and periodically removed from the vacuum test chamber to characterize thermo-optical properties. The solar absorptance (α s ) and the thermal emittance (ε) are determined by measurement instrumentation. The Laboratory Portable Spectroreflectometer (LPSR) measures the spectral reflectance of material surfaces over the wavelength range from 200 to 2800 nm. The LPSR calculates α s from this reflectance spectral data.

240

DAVID L. EDWARDS ET AL.

The Reflectivity is determined by using the simple relationship: T +R+A=1

(4)

where T is the transmittance, R is the reflectance, and A is the absorptance of the sail material. For our case, we assume that no light is transmitted through the sail material, thus reducing eq. (4) to: R =1− A

(5)

Thermal emittance is measured using the TEMP 2000 and the Laboratory Portable Infrared Reflectometer (LPIR) instruments. The TEMP 2000 measures thermal emittance over an integrated wavelength range from 2.5 to 20 μm, while the LPIR provide spectral data over the wavelength range from 2 to 20 μm. 3.3. HYPERVELOCITY IMPACT (HVI)

Sail materials characterized by HVI, to-date, are aluminized Mylar (Al/Mylar) and aluminized Teonex (Al/Teonex). These materials were exposed to a uniform dose of electron radiation. The dose levels were 100 Mrad and 1 Grad. Hypervelocity impact (HVI) tests were performed on control sail material. These tests established the baseline response of the material. Irradiated samples were exposed to 100 keV electrons. Four HVI samples were placed in the test chamber such that the electron flux would pass sequentially through each sample. When the desired dose levels in the materials were achieved, the HVI samples were removed from the vacuum chamber and delivered for HVI testing. Each sail sample was positioned in the HVI test chamber and individually impact characterized with a single projectile. Photographs of each HVI sample were taken subsequent to impact testing. The photographs indicated that several samples exhibited micro-cracks extending radially from the penetration site. To account for the damage presented by these micro-cracks a damage area was defined. Figure 6 shows a typical Penetration diameter for HVI on sail material. Measurements were obtained of the impact site including penetration diameter and total damage area surrounding the penetration site. The HVI analysis was initiated after the sail material sample was removed from the chamber. After the HVI test, a photograph was taken of the impact site. A microscope equipped with two micrometers was used for impact site measurements. The penetration diameter was measured as well as the surrounding damage diameter. These measurements were then compared to the projectile size to determine if the damage from HVI changes with radiation exposure. The ratios of Penetration area to projectile size and damage area to projectile size are referred

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION

241

Penetration Area

Damage Area Figure 6. Penetration site of HVI of a 1.0 mm diameter projectile with velocity of 7 km·s−1 . The photograph shows the penetration area and the surrounding damage area induced by the impact. Sail material was Al/Mylar

to as “Beta factors.” These Beta factors are defined in eqs. (6) and (7). Beta Penetration =

(Area of Penetration) (Cross-Sectional Area of Projectile)

Beta Damage Area =

(Damage Area) (Cross-Sectional Area of Projectile)

(6) (7)

4. Results Preliminary results on the effects of charged particle exposure have yielded some interesting, but not unexpected results. The mechanical properties, including, ultimate stress and ultimate strain, decrease with increasing radiation dose. Stressstrain data was obtained for each exposure condition. Three to five tensile coupons of each material were exposed to each exposure condition. The results, to date, are shown in figures 7 and 8. In addition to mechanical properties, thermo-optical properties are measured, when practical. Figure 9 shows the effect of electron radiation on the reflectance property of aluminized Mylar as a function of increasing dose. The solar absorptance values, shown in table 1, indicate little change with radiation exposure. The spectral data, shown in figure 9, shows a slight decrease in reflection with increased radiation exposure. This reflection loss occurs in the low wavelength region of the reflectance spectrum. Table 1 details the thermo-optical properties obtained to date for materials exposed to long term Near Ultraviolet Radiation and other materials exposed to high doses of electron radiation.

242

DAVID L. EDWARDS ET AL. 5

Ultimate Tensile Strength (psi)

10

4

10

Al/Teonex (e- exposure) Al/Mylar (e- exposure) Al/CP1 (e- exposure) Al/CP1 (p+ exposure) Al/CP1 (e- & p+ & UV exposure)

3

10 0.001

0.01

0.1

1

10

2

10

10

3

4

10

10

5

Dose (Mrad)

Figure 7. Relationship between ultimate strength of candidate sail materials and radiation dose

The data, shown in table 1, indicates the solar absorptance of the Aluminum surface of the sail material has not significantly changed with either UV exposure or electron exposure. The back surface (e.g., polymer surface) emittance has also remained relatively stable, showing, at most, a 4% change after 1 Grad of electron exposure. 1

Strain to Failure

0.1

Al/Teonex (e- exposure) Al/Mylar (e- exposure) Al/CP1 (e- exposure) Al/CP1 (p+ exposure) Al/CP1 (e- & p+ & UV exposure)

0.01

0.001 0.001

0.01

0.1

1

10

10

2

10

3

10

4

5

10

Dose (Mrad)

Figure 8. Relationship between strain to failure of candidate sail materials and radiation dose

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION

243

TABLE 1. Sail reflectance as a function of exposure A1/Mylar UV ESH 0 250 500 750 1000

AL/Teonex

A1/CP1

Alpha Emmitance Reflectance Alpha Emmitance Reflectance Alpha Emmitance Reflectance 0.08 0.08 0.08

0.23 0.24 0.25

0.92 0.92 0.92

0.08 0.08 0.08

0.25 0.25 0.25

0.92 0.92 0.92

Electron Dose 100 Mrad 0.09 1Grad 0.09 5Grad

0.24 0.24

0.91 0.91

0.09 0.09

0.25 0.29

0.91 0.91

0.09 0.09 0.09

0.27 0.27 0.27

0.91 0.91 0.91

The Hypervelocity Impact (HVI) testing was designed to determine the effects of HVI over the functional lifetime of the sail material. Impact testing was performed at dose levels of 100 Mrad and 1 Grad. Sail material specimens were first exposed to electron radiation, then subjected to HVI. Beta factors for penetration and damage areas for the candidate materials Al/ Teonex and Al/Mylar after radiation exposure and HVI characterization are shown in figures 10 and 11. These materials were exposed to radiation doses of 100 Mrad and 1 Grad and then subjected to HVI with projectile of 3 km·s−1 and diameters of 0.4 and 1.0 mm. The complete data set for tests completed to date is shown in table 2.

0.98 Pre-Exposure

0.96

100 Mrad 1 Grad

Ref lectance

0.94 0.92 0.9 0.88 0.86 0.84 0.82 200

300

400

500

600

700

800

900

1000

Wavelength (nm)

Figure 9. Reflectance of aluminized Mylar solar sail material as a function of electron radiation dosage

244

DAVID L. EDWARDS ET AL. 1.6 1.5

Beta Penetration

1.4 Projectile Velocity 3 km/s

1.3

Al/Teonex 0.4 mm projectile Al/Teonex 1.0 mm projectile Al/Mylar 0.4 mm projectile Al/Mylar 1.0 mm projectile

1.2 1.1 1 0.9 0.001

0.01

0.1

1

10

2

3

10

10

Dose (Mrad)

Figure 10. Beta factor (penetration area) relationship with radiation dose in candidate sail material

This data does not constitute a statistically significant data set. This data does consistently indicate the damage area is no larger than a factor of 2 of the projectile diameter. The tests also indicated that the impacts did not result in rip propagation. The data indicates little change in hypervelocity impact response with increased radiation exposure of these materials.

1.8 1.7

Beta Damage

1.6 1.5 1.4

Projectile Velocity 3 km/s

1.3

Al/Teonex 0.4 mm projectile Al/Teonex 1.0 mm projectile Al/Mylar 0.4 mm projectile Al/Mylar 1.0 mm projectile

1.2 1.1 0.001

0.01

0.1

1

10

2

10

3

10

Dose (Mrada)

Figure 11. Beta factor (damage area) relationship with radiation dose in candidate sail material

STATUS OF SOLAR SAIL MATERIAL CHARACTERIZATION

245

TABLE 2. Beta factors of candidate solar sail materials Exposure (Mard)

Projectile size (mm)

Projectile velocity (km/s)

Beta factor penetration

Beta factor damage

Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex Al/Teonex

0 0 0 0 100 100 100 100 1000 1000 1000 1000

0.4 1 0.4 1 0.4 1 0.4 1 0.4 1 0.4 1

7 7 3 3 7 7 3 3 7 7 3 3

1.35 1.11 1.49 1.11 1.08 1.07 1.22 1.03 1.73 1.05 1.36 1.06

1.53 1.11 1.49 1.14 1.59 1.07 1.68 1.2 1.97 1.11 1.74 1.13

Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar Al/Mylar

0 0 0 0 100 100 100 100 1000 1000 1000 1000

0.4 1 0.4 1 0.4 1 0.4 1 0.4 1 0.4 1

7 7 3 3 7 7 3 3 7 7 3 3

1.35 1.27 1.58 1

1.9

1.33 1.02

1.78 1.18

1.02

1.15

Material

1.58 1.18

5. Future Plans Upcoming testing on the candidate solar sail material will focus on missionspecific applications. Test plans are being developed to expose the material to emulated Solar Polar Imager (SPI) and L1 Diamond radiation environments. The mechanical and thermo-optical properties will be evaluated over mission lifetime dose exposures. Hypervelocity impact tests will be conducted on sail materials exposed to mission middle and end-of-life radiation doses. The goal of this research is to select the optimum material for these specialized missions. References 1. Solar Polar Imager, http://umbra.nascom.nasa.gov/SEC/secr/missions/polarimg.html, January 2004. 2. Personal communication with Dr. Greg Garbe/NASA/MSFC.

246

DAVID L. EDWARDS ET AL.

3. Edwards, D., Hubbs, W., Gray, P. Wertz, G., Hoppe, D., Nehls, M., Semmel, C., Albarado, T., and Hollerman, W. (2003) In Proceedings of the 9th International Symposium on Material in a Space Environment, Noordwijk, The Netherlands, ESA Publications, Noordwijk, The Netherlands June 2003, pp. 16–20. 4. Albarado, T., Hollerman, W., Edwards, D., Hubbs, W., and Semmel, C. (2003) In Proceedings of ISEC 2003: 2003 International Solar Energy Conference, Hawaii, 15–18 March 2003. 5. Edwards, D., Hubbs, W., Stanaland, T., Hollerman, A., and Altstatt, R. (2002) In Proceedings of SPIE Photonics for Space Environments VIII, Vol. 4823, 2002. 6. Tech ICP-Solar Sails: Potential Mission Applications. www. inspacepropulsion.com/tech/ sails missionapps.html 7. McInnes, C. R. (1999) Solar Sailing: Technology, Dynamics, and Mission Applications, Praxis Publishing, Chichester, UK, pp. 1–50. 8. ASTM E-490-00a. (2000) Standard Solar Constant and Air Mass Zero solar Spectral Irradiance Tables, 200. 9. Halbleib, J. A., Kensek, R. P., Mehlhorn, T. A., Valdez, G. D., Seltzer, S. M., Berger, M. J. “ITS Version 3.0: The Integrated TIGER Series of Coupled Electron/Photon Monte Carlo Transport Codes”, SAND91-1634, (March 1992). 10. Ziegler, J. F. (2000) The Stopping and Range of Ions in Matter, SRIM-2000.40, www.SRIM.org.

ATOMIC OXYGEN DURABILITY EVALUATION OF A UV CURABLE CERAMER PROTECTIVE COATING BRUCE A. BANKS,1 CHRISTINA A. KARNIOTIS,2 DAVID DWORAK,3 AND MARK SOUCEK3 1 NASA Lewis Research Center, 21000 Brookpark Rd., M. S. 309-2 Cleveland, Ohio 44135 2 QSS Group, NASA Glenn Research Center, Mail Stop 309-2, 21000 Brookpark Road, Cleveland, OH 44135 3 University of Akron, Dept. of Polymer Engineering, 250 South Forge St., Akron OH 44325-0301

Abstract. The exposure of most silicones to atomic oxygen in low Earth orbit (LEO) results in the oxidative loss of methyl groups with a gradual conversion to oxides of silicon. Typically, there is surface shrinkage of the oxidized silicone protective coatings which leads to cracking of the partially oxidized brittle surface. Such cracks widen, branch and can propagate with continued atomic oxygen exposure ultimately allowing atomic oxygen to reach any hydrocarbon polymers under the silicone coating. A need exists for a paintable silicone coating that is free from such surface cracking and can be effectively used for protection of polymers and composites in LEO. A new type of silicone-based protective coating holding such potential was evaluated for atomic oxygen durability in an RF atomic oxygen plasma exposure facility. The coating consisted of a UV curable inorganic/organic hybrid coating, known as a ceramer, which was fabricated using a methyl substituted polysiloxane binder and nanophase silicon-oxo-clusters derived from sol-gel precursors. The polysiloxane was functionalized with a cycloaliphatic epoxide in order to be cured at ambient temperature via a cationic UV induced curing mechanism. Alkoxy silane groups were also grafted onto the polysiloxane chain, through hydrosilation, in order to form a network with the incorporated silicon-oxo-clusters. The prepared polymer was characterized by 1 H and 29 Si NMR, FT-IR, and electrospray ionization mass spectroscopy. The paper will present the results of atomic oxygen protection ability of thin ceramer coatings on Kapton H as evaluated over a range of atomic oxygen fluence levels. Key words: atomic oxygen, silicones

247 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 247–263. C 2006 Springer. Printed in the Netherlands.

248

BRUCE A. BANKS ET AL.

1. Introduction Silicones are one of the few polymers that can be applied by painting or spraying over composite or other organic spacecraft materials which have afforded reasonable protection from low Earth orbital (LEO) atomic oxygen attack. The gradual oxidation of the silicones in LEO results in an oxidized silicone surface which becomes a silicate or silica [1–3]. This surface conversion from silicone to silica also tends to produce tensile stresses in the surface of the oxidized silicone. An increase in the surface microhardness also results due to the atomic oxygen conversion reaction with the silicone [4]. A variety of approaches have been or are now being explored to identify silicones, silicone copolymers or silicone-hydrocarbon blends that could provide flexibility as well as atomic oxygen protection [5–8]. Results to date indicate that hybrid polymers composed of inorganic and organic polymers hold potential to survive LEO atomic oxygen attack. The silicones which are dominated by a oxygen-to-silicon ratio of 1.5 have shown greater resistance to atomic oxygen attack than the silicones with a ratio of 1.0. Polyhedral oligomeric silsesquioxane (POSS) contains covalently bonded reactive functionalities appropriate for polymerization or grafting. It can be blended or copolymerized with many aerospace polymers and is being considered for atomic oxygen durability [8]. The resistance to atomic oxygen attack of silicone blended or copolymerized polymers has been dependent not only on the oxygen-to-silicone ratio but the fractional fill of the silicone. The challenge to make functional use of such blends has been to find an adequately silicone-filled polymer that contains the appropriate protective silicone such that it has mission dependent properties that are acceptable. Examples of some of these are atomic oxygen durability, volatility, optical, thermal, mechanical and ease of application. Because of their ability to provide atomic oxygen protection, thermal stability, flexibility, and stability; polysiloxanes are an attractive candidate solution to achieving ideal protection from the elements of space. However, this is just part of the solution. The vacuum ultraviolet (VUV) radiation and high energy particles can still damage and degrade the composite material. Therefore, to incorporate protection from those components as well, ceramer coatings; which are inorganic/organic hybrid materials, can be utilized. Ceramers are part ceramic (inorganic) and part polymer (organic) and can offer protection from atomic oxygen as well as UV radiation and high energy particles via the in situ fabrication of nanophase silicon-oxo-clusters [9, 10]. The silicon-oxo-clusters are formed through a series of hydrolysis and condensation reactions between sol-gel precursors. The intention of a ceramer approach is to acquire a synergistic effect between the inorganic and organic phases on a nanoscale through the use of phase coupling agents, which for this system are alkoxy silanes pendant from the polysiloxane chain. There is confirmation of a synergy between the phases and this approach affords a uniformly distributed nanophase within a continuous organic phase [11].

249

ATOMIC OXYGEN DURABILITY O O

OH O

O O

O

OH

Atomic Oxygen O O O O O

O

HO

O Si

O

DUV

O Si

O

O Si

O

OH

Dissipated Heat

Si O

High Energy Particle

Metal Oxide

O

– SiO2 Layer Coating Composite Substrate

Figure 1. Depiction of the formation and function of protective silicon oxide layer and silicon-oxoclusters

Once the coating is exposed to atomic oxygen, a protective layer of silicon oxide is formed and, with the incorporation of silicon-oxo-clusters, the coating should protect the composite material against further atomic oxygen erosion, high energy particles, and VUV radiation. Figure 1 is an overall diagram of the coating’s function [12]. This paper summarizes an investigation of the high fluence atomic oxygen durability of ultraviolet radiation curable ceramer protective coating consisting of methyl substituted polysiloxane and compares the results with the commonly used silicone coating DC93-500. 2. Materials: Methyl Substituted Polysiloxane 2.1. SYNTHESIS

2.1.1. Experimental Materials. Octamethylcyclotetrasiloxane,1,3,5,7-tetramethylcyclotetrasil-oxane, 1,1,3,3-tetramethyldisiloxane, and vinyl triethoxysilane were purchased from Gelest, Inc. and used as supplied. Wilkinson’s catalyst (chlorotris (triphenylphosphine) rhodium(I), 99.99%), Amberlyst 15 ion-exchange resin, and 4-vinyl-1-cyclohexene 1,2-epoxide were purchased from Aldrich and used as supplied. Toluene, supplied by Aldrich Chemical Co., was distilled in order to eliminate any impurities. Irgacure 250 was supplied by Ciba Specialty Chemicals and used as received. Air sensitive materials were transferred and weighed in an inert atmosphere dry box under argon. Synthesis of poly(dimethylsiloxane-co-methylhydrosiloxane), hydride terminated. To a three neck, round bottom flask, equipped with a reflux condenser and

250

BRUCE A. BANKS ET AL. H3C O

H3C

O

CH3

O H3C

Si

+

H Si

CH3

O

Ion Exchange Resin

CH3

CH3

Si Si

CH3

H3C +

O H

CH3

HSi O SiH CH3

H3C

CH3

H

H

O

Si H3C

CH3

O

O

H

Si

Si H3C

H3C

CH3 Si

N2 Blanket

CH3 CH3

SiO (SiO)n (SiO)m Si

H

CH3 H CH3 CH3 m >> n

O H3C == CHCH3Si(OEt)3

CH3 CH3 O

Rh Catalyst N2 Blanket

CH3

CH3

SiO (SiO)n (SiO)m Si CH3 (CH2)3 CH3

O

CH3

Si(OEt)3

Figure 2. Synthesis of polydimethylsiloxane-co-methylhydrosiloxane, hydride terminated

nitrogen inlet/outlet, was added octamethylcyclo-tetrasiloxane (90.00 g), 1,3,5,7tetramethylcyclotetrasiloxane (5.33 g), 1,1,3,3-tetramethyldisiloxane (0.67 g), and Amberlyst 15 (20 wt%). The mixture was stirred at 90◦ C, under nitrogen, for 15 h. The viscous solution was then filtered to obtain poly(dimethylsiloxane-comethylhydrosiloxane), hydride terminated of various molecular weight ranges. Vacuum filtration was performed (<1 mm Hg) in order to remove low molecular weight oligomers and unreacted starting materials. Weight average molecular weight was obtained from gel permeation chromatography (GPC) analysis, Mw = 42, 000, PDI = 1.66. Polymer characterization and Si-H functionality was confirmed/analyzed through 1 H NMR, FT-IR, and titration. [1 H NMR (δ (ppm), CDCl3 ): 4.675 (s, CH3 -Si-H). FT-IR (cm−1 , KBr Plate): 2150 (s, Si-H).] Cycloaliphatic epoxide and alkoxy silane functionalization of prepared poly(dimethylsiloxane-co-methylhydrosiloxane), hydride terminated. To a three neck, round bottom flask, equipped with a reflux condenser and nitrogen inlet/outlet, was added was added poly(dimethylsiloxane-co-methylhydrosiloxane), hydride terminated (30 g), 4-vinyl-1-cyclohexene diepoxide (20 g), vinyl triethoxysilane (2 g), and Wilkinson’s catalyst (0.004 g). Via a canula, distilled

ATOMIC OXYGEN DURABILITY

251

Figure 3. Structure of cross linked polysiloxane phase with interconnected silicon-oxo-clusters

toluene (30 g) was added. The reaction was held at 75◦ C by means of an oil bath and mechanically stirred under a nitrogen blanket. The disappearance of the Si-H functionality was monitored by FT-IR and using the disappearance of the peak at ∼2160 cm−1 as an indication that the reaction was complete. Any solvent and unreacted starting materials were removed under vacuum (3–5 mm Hg). Cycloaliphatic epoxide and alkoxy silane functionalization was confirmed/analyzed through 1 HNMR, FT-IR analysis, and titration. 2.2. STRUCTURE

Once cured the coating should form a strong interlocking network consisting of a cross linked polysiloxane phase with interconnected silicon-oxo-clusters (figure 3). The silicon-oxo-clusters will be connected to the polysiloxane backbone though hydrolysis and condensation reactions with the tethered alkoxy silane. 3. Apparatus and Procedure 3.1. COATING APPLICATION

The polysiloxane was diluted with dry toluene (25% w/w) in order to reduce the viscosity. Sol-gel precursor (5% w/w) and photo initiator (3% w/w) was also

252

BRUCE A. BANKS ET AL.

added to the solution and thoroughly mixed. A piece of Kapton H (∼4 in. in diameter) was mounted onto a spinning stage and spun at a very high speed. The polysiloxane solution was dropped onto the center of the spinning Kapton sample. The sample was removed from the stage and passed through a UV curing chamber with a belt speed of 25 ft·min and an average intensity of 150 mW·cm−2 . The coating thickness was measured with a coating thickness gauge and atomic force microscopy (AFM). Both methods confirmed an average of a two micrometer thickness. 3.1.1. Instruments Viscosity measurements were taken on an AR 500 Rheometer (Thermal Analysis) equipped with a cup and bob sample holder and operated at 21.1◦ C. Pencil hardness tests were conducted according to ASTM method D3363-00. Taber scratch test was performed using a Taber Shear/Scratch Tester model 502 and conforming to ASTM method G171-03. Taber Abrasion tests were conducted on a Taber Industries 5130 Abraser using a CS-10 test wheel. Taber abrasion studies corresponded to ASTM method D552-93a. Thermogravimetric analysis was performed on a TGA Q 500 (Thermal Analysis). X-ray photoelectron spectroscopy was completed on a Kratos Model ES3000 with a non-monochromatic 120 W Al K-Alpha radiation source under 10−8 Torr. Scanning electron microscopy was performed on a Hitachi S-2150 operating at 15 kV. Atomic force microscopy was performed on a multimode scanning probe microscope (Digital Instruments) using the tapping mode. 3.2. ATOMIC OXYGEN EXPOSURE

Samples of the ceramer silicone-coated Kapton H polyimide (with silicone coat on both sides) were compared with samples of DC 93-500 silicone-coated Kapton H (with silicone coat on one side) for atomic oxygen durability. The same coatings were also applied to fused silica substrates for the purposes of obtaining changes in optical properties as well as noting evidence of tensile cracking. Changes in optical properties (reflectance, transmittance and absorptance) and mass loss were documented at atomic oxygen effective fluence levels of 2.22 × 1021 and 1.38 × 1022 atoms·cm−2 . Kapton H witness samples were used to determine the effective atomic oxygen fluence as described in ASTM E 2089-00, “Standard Practices for Ground Laboratory Atomic Oxygen Interaction Evaluation of Materials for Space Applications” [13]. All Kapton H substrates used for coating evaluation and fluence witnesses were made of 2.54 cm diameter by 0.127 mm thick Kapton H polyimide. An additional set of ceramer and DC93-500 silicone coated samples were made that were scratched prior to exposing to atomic oxygen using finger wiping with laboratory dust. This was done to see if minor abrasion of the silicone surface would cause preferential cracking of the silicone coatings with atomic

ATOMIC OXYGEN DURABILITY

253

1 cm

Figure 4. Sample holder to prevent curling of 2.54 cm diameter samples when they were exposed to atomic oxygen

oxygen exposure. Samples of silicone-coated Kapton H were punched out and vacuum dehydrated for 48 h prior to weighing to minimize mass uncertainty due to weight loss as recommended by ASTM E 2089-00 [13]. Atomic oxygen testing was performed on samples that were placed in an SPI Plasma Prep II 13.56 MHz radio frequency plasma asher. The ashers are typically operated on air at a pressure of 12.7–16 Pa (95–120 mTorr). The samples were each held down by fine wires attached to a metal frame (as shown in figure 4) laying on a glass plate to minimize curling of the samples with atomic oxygen exposure from only one side. Curling typically occurs for silicone-coated samples that are coated on one side and could allow atomic oxygen to attack the uncoated back of the samples which would compromise the sample weight loss data. The plasma asher was operated at a Kapton effective flux of 4.69 × 1015 atoms·cm−2 ·s−1 [13]. Because many silicones used on LEO spacecraft have a history of causing contamination on spacecraft as a result of the evolution of volatile silicones and with the subsequent oxidation and conversion to silica on neighboring spacecraft surfaces, cross contamination witness samples were placed in the plasma ashers next to the silicone-coated samples to assess the degree of silicone transport and resulting contamination. Tests were performed prior to sample exposures to validate that any contamination deposited would be as a result of the samples contained within the plasma asher. Thicknesses of deposited contaminants were measured using a Dektak 6M stylus profilometer which scanned the contamination coated fused silica slide from the deposited area to an area that was protected from contamination deposition by means of a tightly fitted aluminum foil mask. Optical properties prior to and after atomic oxygen exposure were measured using a Perkin Elmer Lambda-19 spectrophotometer.

254

BRUCE A. BANKS ET AL.

4. Results and Discussion 4.1. METHYL SUBSTITUTED POLYSILOXANE CHARACTERIZATION

The abrasion and scratch resistance of the cured coating were studied to determine how susceptible it is to physical damage. The Taber abrasion and scratch tests yielded a value of 183 wear cycles per mil and a scratch value of 50 g. Both of these values are low and show that the coating has poor abrasion resistance. These values were cross-checked using the pencil hardness test, which gave a value of 2B/B. This value is also low and this trend could be a result of the very low glass transition temperature of the coating, which is approximately −130◦ C. Such a low glass transition temperature makes the coating soft and vulnerable to damage. Varying the pendant group to raise the glass transition of the coating could be a potential answer in improving the abrasion resistance. Thermal gravimetric analysis was performed in order to observe the thermal stability of the cross-linked polysiloxanes. Irreversible changes to the cross-linked structure of silicone polymers unavoidably occur at high temperatures due to chain scission or oxidative cross-linking [14]. In an inert atmosphere, depolymerization occurs with the loss of volatile products, mostly low molecular weight cyclic oligomers; but is often catalyzed by traces of acids, bases, water, or residual catalyst used in the polymers original production [15]. Typically, depolymerization occurs near 400◦ C for reasonably pure polydimethylsiloxane [16]. Thermal gravimetric analysis of the cured coatings illustrates the loss of small molecular weight oligomers in the early stages of the analysis (figure 5). As 100

Weight (%)

80

60

40

20

0 0

200

400

600

800

1000

Temperature (°C)

Figure 5. Thermal gravimetric analysis of cured coating with 5% sol-gel precursor

ATOMIC OXYGEN DURABILITY

255

Figure 6. AFM images of crosslinked methyl polysiloxane substituted with 5% sol-gel precursor

expected, the depolymerization occurred near 400◦ C for the sample tested. The range of molecular weights give way to the multiple slopes the curve exhibits. It is also important to note that the sample generated a small amount of residue (∼13%), which can be attributed to the silicon-oxo-clusters formed during the polymerization process and high molecular weight chains that may not have completely volatized/degraded. One of the most important aspects of the coating is the presence of the siliconoxo-clusters. By utilizing the AFM’s tapping mode, it will be possible to detect “hard” (silicon-oxo-clusters) and “soft” (polymer) regions within the cross-linked polymer network. These clusters provide additional protection against high-energy particles and deep UV light. Figure 6 is an AFM image of a sample with 5% (w/w) sol-gel precursor added prior to casting. The silicon-oxo-clusters are clearly visible in the subjected sample. The average size for the methyl substituted polysiloxane nano phase is 125 nm. Figure 6 shows a more disperse and uniformly sized nanophase, which could be attributed to the small size of the pendant methyl group allowing more freedom to the growing nano clusters. Conductivity tests were performed on the cross-linked coatings. The coatings showed no signs of electrical conductivity, which is expected due to their insulating nature [14]. 4.2. ATOMIC OXYGEN EXPOSURE RESULTS

Photographs of the silicone coated Kapton H samples and silicone coated fused silica samples after two different levels of atomic oxygen exposure are shown in figure 7.

256

BRUCE A. BANKS ET AL.

(a)

(b)

(c)

Figure 7. Photographs of silicone coated Kapton H and silicone coated fused silica after atomic oxygen exposure to moderate and high fluence levels (a) Ceramer coated samples at a Kapton effective fluence of 2.22 × 1021 atoms/cm2 ; (b) DC93-500 coated samples at a Kapton effective fluence of 2.22 × 1021 atoms/cm2 ; (c) Ceramer coated samples at a Kapton effective fluence of 1.38 × 1022 atoms/cm2 ; (d) DC93-500 coated samples at a Kapton effective fluence of 1.38 × 1022 atoms/cm2

ATOMIC OXYGEN DURABILITY

257 (d)

Figure 7. (Continued)

As can be seen in figure 7, atomic oxygen exposure of the ceramer and DC93500 provide excellent protection for moderate (2.22 × 1021 atoms·cm−2 ) fluence levels. The ceramer coating appears to be a significant improvement at moderate fluence levels in that there is no sign of microcracking as occurs for DC93-500. However, at high fluences (1.38 × 1022 atoms·cm−2 ), both the ceramer and DC93500 develop microcracks. Unlike the DC93-500, the ceramer tends to detach from its substrate causing greater coating shrinkage due to atomic oxygen attack on both surfaces of the coating. This coating shrinkage and opening of the cracks allows atomic oxygen to attack the underlying Kapton. Thus, the ceramer coating is better for moderate fluences and could potentially be used for coating optical polymers such as Fresnel lenses for concentrators over solar cells. In such applications, protective coatings that form microcracks would not be suitable due to loss in specular transmittance. The mass loss of coated Kapton as a function of atomic oxygen fluence for both the ceramer and DC93-500 coatings is shown in figure 8. As can be seen from figure 8, the ceramer coating as well as the DC93500 coating does provide significant atomic oxygen protection for most fluences. However, at fluences above 1 × 1022 atoms·cm−2 , the ceramer coating develops apertures in it due to microcracking and the rate of oxidation of the underlying Kapton greatly increases. There did not seem to be significant differences in atomic oxygen protection resulting from the laboratory dust abrasion of the coatings. This is thought to be due to the very shallow scratches resulting without any going all the way through the coatings. Atomic oxygen exposure of the ceramer coatings causes an increase in optical absorptance, and therefore a reduction in transmittance for wavelengths <800 nm with little change in reflectance, as shown in figure 9.

258

Mass lost (mg)

BRUCE A. BANKS ET AL. 100 90 80 70 60 50 40 30 20 10 0 0

5E+21

1E+22 2

Fluence (atoms/cm )

1.5E+22 Ceramer - scratched Ceramer - unscratched DC - scratched DC - unscratched untreated kapton

Figure 8. Mass loss of coated Kapton as a function of atomic oxygen fluence for both the ceramer and DC93-500 coatings

Similar results were observed for DC93-500 and the ceramer as shown in figure 10. However, if one considers applications such as protective coatings for Fresnel concentrator solar cell arrays there is an important distinction between the two coatings for moderate fluences. The specular transmittance degradation caused by atomic oxygen exposure is in part due to increased absorption for wavelengths <800 nm as well as microcracking, as shown in figure 11. As can be seen by examining figures 11a and 7 (a and b), the absence (or presence) of microcracking at a fluence level of 2.22 × 1021 atoms·cm−2 makes a significant impact on specularly transmitted light. The ceramer coating does allow much greater specular transmittance than the DC93-500 for fluences up to 2.22 × 1021 atoms·cm−2 . X-ray photoelectron spectroscopy was performed in order to confirm the presence of a protective oxide layer (figure 12). It is important to note that sputtering was not performed during the analysis. This ensures that only the surface of the samples was analyzed. The initial XPS spectrum shows high amounts of both silicon and oxygen, which is expected since these elements are present in the polymer backbone. After atomic oxygen exposure, the oxygen peak increased while the silicon peaks decreased. This is anticipated due to the protective oxide layer possessing a high amount of oxygen compared to silicon. The oxide layer should be composed of silicon atoms whose valences are filled by oxygen atoms, yielding a Si-O4 network. The presence of carbon after exposure to atomic oxygen is due to impurities on the surface of the film such as dust, dirt, etc. and is, therefore, always present [12]. Cross contamination tests performed separately on the ceramer and DC93500 coating using fused silica witness slides adjoining the silicone samples

259

ATOMIC OXYGEN DURABILITY

Absorptance (%)

110 90 70 50 30 10 −10

0

500

1000 Wavelength (nm)

1500

2000

2500

original 2.22E=21 atoms/cm2 1.38E+22 atoms/cm2

a. Total absorptance

Transmittance (%)

100 80 60 40 20 0

0

500

1000 Wavelength (nm)

1500

2000

2500

original 2.22E+21 atoms/cm2 1.38E+22 atoms/cm2

b. Total transmittance 50

Reflectance (%)

40 30 20 10 0 −10

0

500

1000

Wavelength (nm)

1500

2000

2500

original 2.22E+21 atoms/cm2 1.38E+22 atoms/cm2

c. Total reflectance Figure 9. Effects of atomic oxygen on total optical properties for ceramer coated fused silica

260

BRUCE A. BANKS ET AL.

Absorptance (%)

100 80 60 40 20 0 −20

0

500

1000

1500

2000

2500

original 2.22E+21 atoms/cm2 1.38E+22 atoms/cm2

Wavelength (nm)

Transmittance (%)

a. Total absorptance 100 90 80 70 60 50 40 30 20 10 0

0

500

1000

1500

Wavelength (nm)

2000

2500

original 2.22E+21 atoms/cm2 1.38E+22 atoms/cm2

Reflectance (%)

b. Total transmittance 50 45 40 35 30 25 20 15 10 5 0

0

500

1000 Wavelength (nm)

1500

2000

2500

original 2.22E+21 atoms/cm2 1.38E+22 atoms/cm2

c. Total reflectance Figure 10. Effects of atomic oxygen on total optical properties for DC93-500 coated fused silica

Transmittance (%)

100 80 60 40 20 0 −20

0

500

1000

1500

2000

Wavelength (nm)

2500

3000

original 2.22 E+21 atoms/cm2 1.38 E+22 atoms/cm2

a. Ceramer Transmittance (%)

120 100 80 60 40 20 0

0

500

1000

1500

2000

Wavelength (nm)

2500

original 2.22 E+21 atoms/cm2 1.38 E+22 atoms/cm2

b. DC93-500

Figure 11. Effects of atomic oxygen on specular transmittance of ceramer and DC93-500 coated fused silica 9000

Si

8000

C

O

7000

Intensity (arb. units)

6000

5000

4000

3000

2000

1000

0 0

100

200

300

400

500

600

700

Energy (eV) Before AO Exposure

After AO Exposure

Figure 12. XPS spectrum of the cross-linked methyl substituted polysiloxane before and after atomic oxygen exposure

262

BRUCE A. BANKS ET AL.

indicated that there was transport of short chain silicones to the fused silica that resulted in a silica deposit 857 Angstroms for a Kapton effective fluence of 1 × 1022 atoms·cm−2 for the ceramer, but no measurable contamination from the DC93-500. This is probably due to the lack of vacuum stripping of the ceramer whereas the DC93-500 is vacuum stripped. A “no sample” test of the facility did, in fact, result in no deposit of oxidized silicone.

5. Conclusion Atomic oxygen exposure of ceramer and DC93-500 silicone coated Kapton H and fused silica slides indicates that the ceramer coating has superiority over DC93500 coatings for moderate (up to 2.22 × 1021 atoms·cm−2 ) Kapton effective atomic oxygen fluences. The ceramer coatings at this fluence resulted in low mass loss of coated Kapton samples and did not show evidence of the extended microcracking that occurred for DC93-500. This results in superior specular light transmittance for the ceramer coatings. This may allow its use as an atomic oxygen protective coating over silicone Fresnel concentrators for solar arrays. At high fluence levels (1.38 × 1022 atoms·cm−2 ), the ceramer coating develops microcracks that result in detachment of the coating causing exposure of the underlying Kapton which does not occur for DC93-500. Thus at high fluences DC93-500 would be a better choice of protective coating. The ceramer coating produced cross contamination of silica on witness slides which suggests that the ceramer should be vacuum stripped to prevent the transport of short chain silicones that can result in contamination.

Acknowledgments The authors gratefully acknowledge Justin Tokash and Dr. Rex Ramsier of the University of Akron’s Physics Department for their XPS contributions.

References 1. Banks, B. A., Dever J. A., Gebauer L., and Hill, C. M. (1991) Paper presented at the 1st LDEF Post-Retrieval Symposium, Kissimmee, Florida, 2–8 June 1991. 2. Banks, B. A., Rutledge, S. K., de Groh, K. K., Mirtich, M. J., Gebauer, L., Olle, R., and Hill, C. M. (1991) Paper presented at the 5th International Symposium on Materials in a Space Environment, Cannes-Mandelieu, France, 16–20 September 1991. 3. Banks, B., Rutledge, S., Sechkar, E., Stueber, T., Snyder, A., Hatas, C., and Brinker, D. (2000) In Proceedings of the 8th International Symposium on Materials in a Space Environment and the 5th International Conference on Protection of Materials and Structures from the LEO Space

ATOMIC OXYGEN DURABILITY

4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16.

263

Environment cosponsored by the CNES, Integrity Testing Laboratory, ESA, ONERA and the Canadian Space Agency, Arcachon, France, 4–9 June 2000. de Groh, K., Banks, B., and Ma, D. (2004) In 7th International Conference on Protection of Materials and Structures form Space Environment, Toronto, Canada, 10–13 May 2004. Rutledge, S., Cooper, J., and Olle, R. (1990) In Proceedings of the Space Operations Applications and Research Symposium, NASA CP-3103, Albuquerque, NM, 26–28 June 1990. Hung, C. (1994) Reaction and Protection of Wire Insulators in Atomic Oxygen Environment, NASA TM106767, 1994. Zhang, C., Babonneau, F., Bonhomme, C., Laine, R. M., Soles, C. L., Hristov, H. A., and Yee, A. F. (1998) Journal of American Chemical Society 120, 8380–8391. Brunsvold, A. L., Minton, T. K., Gouzman, I., Grossman, E., and Gonzalez, R. I. (2004) Journal of High Performance Polymers, V. 16, No. 2, pp. 303–318. Soucek, M. D. and Tuman, S. J. (1996) Journal of Coatings Technology 68(854), 73. Tuman, S. J., Chamberlain, D., Scholsky, K. M., and Soucek, M. D. (1996) Progress in Organic Coatings 28, 251. Wold, C. R. and Soucek, M. D. (1997) Journal of Coatings Technology 70, 43. Dworak, D. P. and Soucek, M. D. (2003) Progress in Organic Coatings 47, 448. ASTM E 2089-00, Standard Practices for Ground Laboratory Atomic Oxygen Interaction Evaluation of Materials for Space Applications, June 2000, American Society for Testing and Materials, Philadelphia, USA. Smith, A. L. (1991) The Analytical Chemistry of Silicones, Wiley, New York. Kang D. W., Rajendran, G. P., and Zeldin, M. (1986) Journal of Polymer Science, Part A: Polymer Chemistry 24, 1085. Grassie, N. and Macfarlane, I. G. (1978) European Polymer Journal 14, 875.

CERMET THERMAL CONVERSION COATINGS FOR SPACE APPLICATIONS

BENJAMIN W. WOODS, DANIEL W. THOMPSON, AND JOHN A. WOOLLAM Department of Electrical Engineering, University of Nebraska, 209N WSEC, Lincoln, Nebraska 68588-0511

Abstract. Cermet films of gold and aluminum oxide were sputter deposited onto silicon substrates. These films consisted of nanometer-sized gold grains embedded in a matrix of aluminum oxide, and were designed to maximally absorb visible radiation and minimally emit infrared radiation. Ex situ and in-situ spectroscopic ellipsometry were used to characterize film optical constants from vacuum ultraviolet to middle infrared. The optical constants, for cermet films with gold volume fractions ranging from 0 to 1, were used in predictive optical models. Thermal performance was optimized using a new algorithm based on energy balances. This algorithm weights the parameter fit in predictive optical models to an ideal reflectance spectrum. This weighting factor emphasizes fitting in more significant spectral ranges and deemphasizes fitting in less significant wavelength ranges. Absorptivity, emissivity, and their ratio were calculated for optimized film structures. Key words: Cermet Coatings, Space Applications

1. Introduction Cermet films contain metallic and ceramic materials. Because they can be tailored to have a wide range of optical, electrical, mechanical, and chemical properties, cermet films have many potential applications [1, 2]. In the present work, the focus was using spectroscopic ellipsometry to engineer optical coating thickness and concentration for optimal photothermal energy conversion [3]. These cermet films absorb energy in the visible range and reflect infrared (IR) radiation. The desired films should have maximum absorptivity in the visible spectral range and minimum emissivity in the infrared. One application for these films is to provide energy for a Stirling engine by absorbing solar energy in the space environment [4]. A Stirling engine converts heat into mechanical energy. Another application is photothermal energy conversion for terrestrial heating applications [2]. 265 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 265–276. C 2006 Springer. Printed in the Netherlands.

266

BENJAMIN W. WOODS ET AL.

In the literature, articles on thermal control coatings in the space environment are usually focused on degradation minimization or optical optimization. Degradation has been reported in detail elsewhere [5–9]. Since the materials of interest contain only gold and alumina, we expect they will be unaffected by atomic oxygen. But these materials might be susceptible to micrometeroid impact damage. However, degradation is not a focus of this work. Optical performance optimization is the major emphasis [10–12]. Visible absorptance and IR emittance were derived indirectly from measured normal reflectance data, using eqs. (1) and (2) [13]. 1 = Reflectance (λ) + Absorptance (λ) + Transmittance (λ)

(1)

and Emittance (λ) = Absorptance (λ),

(2)

where λ is wavelength. Coating substrates with an optically thick gold layer eliminated transmittance. Assuming the film is a blackbody, Planck’s radiation formula is used to approximate blackbody spectral irradiance, as expressed in eq. (3),

B(λ, T ) = λ5

2c2 hπ , hc −1 exp kλT

(3)

where c is the speed of light in a vacuum, h is Planck’s constant, λ is wavelength, k is the Boltzmann constant, and T is temperature [14]. Figure 1 illustrates the optimization problem in the space environment [15, 16]. The transition, from low to high reflectance, can be shifted to different wavelengths by changing the thickness, geometry, and/or composition of the cermet layers. However, in practice it is difficult to increase the visible absorptance without also increasing the infrared emittance. For best performance, the reflectance transition should have a steep slope resembling a step function near 1700 nm. The overlap of the AM0 and 450◦ C blackbody irradiance spectra prevents the films from being 100% efficient. For some terrestrial applications, there is negligible overlap because the solar irradiance spectrum (AM1, AM1.5, etc.) is less intense than the AM0 spectrum and the blackbody spectrum is shifted to longer wavelengths (lower temperatures). For all cases, the optical properties can be engineered for optimum performance. This paper introduces a meaningful and convenient way to perform the optimization. Multi-layer cermet stacks have been studied for use as optical coatings with controllable performance [3, 17–22]. These stacks consist of several layers, each with a different composition and thickness. The substrate is coated with a highly reflective metal, and then layers with decreasing metal compositions. The top layer

267

CERMET THERMAL CONVERSION COATINGS 3.0

1.0

0.8

0.6 1.5 0.4 1.0

AM0 Solar Irradiance 450°C Blackbody Irradiance Ideal Reflectance

0.5 0.0

0

1000

2000

3000

Reflectance

2.0

2

Irradiance (W/m /nm)

2.5

0.2

0.0 5000

4000

Wavelength (nm)

Figure 1. AM0 Solar irradiance spectrum, 450◦ C blackbody spectrum, and ideal photothermal film reflectance spectrum

is a ceramic without embedded metal particles that serves as an anti-reflection coating. An example of a four-layer cermet stack with two cermet layers is shown in figure 2. To simplify the optical modeling and clearly illustrate the optimization process, this work examines structures with one cermet layer. 2. Experimental All films were sputter deposited on flat, polished silicon wafer substrates. Deposition occurred in a Kurt J. Lesker Co. vacuum chamber with base and deposition pressures being ∼7 × 10−5 and ∼4 ×10−1 Pa, respectively. The chamber had a rotating platten, positioned by a stepper motor under computer control, as shown in figure 3. Cermet films were deposited via a cyclic process of: positioning the substrate above one of two sputter guns, dwelling for a chosen time period, positioning the substrate above the other sputter gun, and dwelling for a different Alumina Alumina

Cermet B

Cermet

Alumina

Cermet

Cermet A

Gold

Gold

Gold

Gold

Substrate

Substrate

Substrate

Substrate

Figure 2. Film structure layering for some photothermal films

268

BENJAMIN W. WOODS ET AL.

A. B. C. D.

A B G

C

G E F

D

Computer Controlled Platten Silicon Substrate DC Sputter Gun with Gold Target Pulsed-DC Sputter Gun with Aluminum Target E. Argon Feed Tube F. Oxygen Feed Tube G. Conical Sputter Chimney

Figure 3. Profile of the sputter deposition system

time period. The sputter guns were fitted with conical metal chimneys to prevent overspray. In other words, the chimneys prevented gold from being deposited on a substrate positioned above the aluminum target and vice versa. These chimneys also served as inlets for argon and oxygen. Gold was DC sputtered with a controlled power of 15W. A low wattage was chosen to minimize the gold deposition rate. The idea was to have a low deposition rate and short deposition times to minimize grain size. Initially, alumina was reactively sputtered using RF power. This was later replaced by pulsed-DC power. For RF alumina deposition, an RFX 600 power supply was used with a controlled bias of 700V which resulted in 250–300W power output. For DC alumina deposition, an MDX 1K power supply was put in series with a Sparc-le V pulse generator. DC power was set at 150W. The DC pulse generator was purchased to add stability and reproducibility compared to the RF deposition process. During deposition, oxygen gas was introduced through the chimney above an aluminum target at rates between 0.8 and 1 sccm. Before cermet deposition, each silicon wafer substrate was coated with ∼20 nm of chromium (to promote adhesion) followed by an optically thick (∼120 nm) gold layer. A set of nine cermet samples were made with gold compositions of 10%, 20%, 30%, . . . 90% (by volume) along with samples of pure gold and pure alumina. Composition estimation was based on the deposition rates of alumina and gold. Composition was controlled by the amount of time a substrate spent above a sputter gun. Since each cermet layer was composed of many sublayers of gold and alumina, an additional constraint was put on the deposition process. Each sublayer of gold was deposited with a thickness of 1 nm. This constraint was chosen so that gold grains would be small and roughly spherical. TEM and AFM micrographs of two cermet films are shown in figure 4. We believe that the dark regions in the TEM micrograph are gold which would give a gold grain size of ∼5 nm and smaller.

CERMET THERMAL CONVERSION COATINGS

269

Figure 4. Left) AFM micrograph of 45% gold 55% alumina cermet film (500 × 500 nm) Right) TEM micrograph of 30% gold 70% alumina cermet film (20 nm scalebar)

In-situ spectroscopic ellipsometry was used to monitor depositions at a 75◦ angle of incidence, with a spectral range between 400 and 1000 nm. Using in-situ ellipsometry, optical constants for gold and alumina were determined independently and used to determine deposition rates. These depositions rates were then used to set time values for each component (gold and alumina) during cermet deposition. In-situ ellipsometry was also used to monitor deposition of cermet layers. These ellipsometric data were used to determine preliminary thicknesses and optical constants which would later be used in optical models for ex-situ ellipsometric data. Ex-situ spectroscopic ellipsometry was used to find optical constants for gold, alumina, and each of the cermet films in the series from 10 to 90% gold. The ex-situ measurement had the benefit of greater accuracy and larger spectral range, 0.2–30 microns, compared to in-situ ellipsometry. 3. Theory and Results One goal of the present work was to find the best means to compare performance of films used for photothermal energy conversion. Currently, absorptivity, emissivity, and their ratio are the most commonly used figures of merit. Absorptivity and emissivity are defined by ∞ A(λ) [1 − R(λ, θ)] dλ ∞ α= 0 (4) 0 A(λ) dλ

270

BENJAMIN W. WOODS ET AL.

and

∞ ε=

0

B(λ, T ) [1 − R(λ, θ)] dλ ∞ 0 B(λ, T ) dλ

(5)

respectively, where A(λ) is the solar irradiance spectrum, B(λ, T ) is the blackbody spectrum, and R(λ, θ) is spectral reflectance. Angles of incidence are near normal for all calculations. In practice, the range of wavelengths for integration is set by the intensity of the blackbody and solar irradiance spectra. Thus, a range is chosen which includes all wavelengths for which the blackbody and solar spectra have significant magnitude. In this work, the range of integration is 0.2–30 microns. Absorptivity and emissivity determine how efficiently a film absorbs and emits radiation. Since high solar absorption and low blackbody emission are desired it is logical to create a single figure of merit α/ε. Unfortunately, α/ε is more sensitive to changes in ε than α. It is also difficult to use the two parameters separately because one must choose if high solar absorption or low infrared emission is more important. When there is significant overlap between the solar irradiance and blackbody spectra, there is an additional complication. The ideal spectral reflectance is zero for wavelengths where the solar irradiance is more intense and 100% for wavelengths where the blackbody spectrum is more intense. This implies that an ideal film with this spectral reflectance could not have 100% absorptivity or 0% emissivity. To visualize this, notice the area beneath the AM0 curve and at wavelengths higher than 1700 nm in figure 1. None of this area could contribute to the numerator in the calculation of α in eq. (4). In fact, for the AM0 solar spectrum and 450◦ C blackbody the absorptivity calculated from the ideal reflectance spectrum is ∼90%. We define a new figure of merit that incorporates the solar and blackbody spectra and spectral reflectance. The parameter, “photothermal conversion efficiency,” PCE, is defined by ∞ [X A(λ) − B(λ, T )] [1 − Rreal (λ, θ )] dλ PCE = 0∞ (6) [X A(λ) − B(λ, T )] [1 − Rideal (λ, θ )] dλ 0 where X is the solar concentration factor, A(λ) is the solar irradiance spectrum, B(λ, T ) is the blackbody spectrum, Rreal (λ, θ ) is the measured spectral reflectance, and Rideal (λ, θ ) is the ideal spectral reflectance. Ideal spectral reflectance, for absorbers, is zero for wavelengths where solar irradiance is greater than blackbody and 100% for the opposite case. The solar concentration factor accounts for the use of lenses or mirrors to enhance solar irradiance on the film. In this work, the solar concentration factor is assumed to be unity unless otherwise specified. PCE is a ratio of the net power absorbed divided by the net power available. It incorporates both absorption and emission. It has meaningful values ranging from 0 to 1.

271

CERMET THERMAL CONVERSION COATINGS

2.0

14

1.8

12

1.6

Extinction Coefficient 'k'

Index of Refraction 'n'

2.2

1.4 1.2

100% alumina 20% gold 40% gold 60% gold 80% gold 100% gold

1.0 0.8 0.6 0.4 0.2 0.0 200

400

600

800

1000

1200

100% gold 80% gold 60% gold 40% gold 20% gold

10 8 6 4 2

1400

1600

1800

200

0 200

400

600

800

1000

1200

1400

1600

1800

200

Wavelength (nm)

Wavelength (nm)

Figure 5. Optical constants for cermet films. The extinction coefficient of pure alumina is not shown because it is effectively zero over this range

It is possible to calculate a negative value for PCE, which simply means that the film is a better emitter than absorber. In the event, that PCE is negative, inverting the ideal spectral reflectance so that 100% becomes 0 and vice versa, then recalculating PCE will yield an emission efficiency instead of an absorption efficiency. Ellipsometric data (in-situ and ex-situ) for all 11 samples were analyzed. Insitu ellipsometric data were optically modeled to yield preliminary thicknesses and optical constants for each sample. These thicknesses and optical constants were used to set up optical models for the ex-situ data. In the end, oscillator models were used to characterize optical constants for each sample for the spectral range between (0.2 and 30 microns). These optical constants, shown in figure 5, were subsequently used to predict the optimum thicknesses for cermet films, given a solar irradiance spectrum and a blackbody temperature. To find the optimal structures (with respect to PCE) in the modeling software, a weighting factor was created to emphasize fitting in parts of the spectrum where solar irradiance or blackbody emission is more intense. This weighting factor is calculated at each wavelength with the equation σi =

1 (X Ai − Bi )λi

(7)

where X is the solar concentration factor (unity unless specified), Ai and Bi are the solar irradiance and blackbody spectra, respectively. λi accounts for the variable spacing of wavelengths of the measured solar irradiance spectrum. This makes σ a function of the areas beneath A and B, and not the absolute values of A and B for a given wavelength. The optical analysis software seeks to minimize a mean squared error, MSE, defined by 2 N ideal Ri − Rimod 1 MSE = (8) 2N − M i=1 σi

272

BENJAMIN W. WOODS ET AL.

TABLE 1. Calculated figures of merit as a function of solar and blackbody spectra. All simulated films were optimized with respect to PCE. The simulated reflectance was derived from the optical constants and layer thicknesses using Fresnel equations Solar spectrum

Blackbody spectrum (◦ C)

PCE

Absorptivity α

Emissivity ε

α/ε

450 450 50 650

.58 .52 .94 −.01

.66 .69 .95 .24

.014 .014 .046 .013

47 49 21 18

AM0 AM1.5 AM0 AM0

where R mod is the reflectance predicted by the optical model, R ideal is the reflectance value from the ideal reflectance file, N is the total number of wavelengths, M is the number of variables in the model (a small number compared to N ), and σ is a weighting factor. Combining eqs. (7 and 8), the MSE calculation becomes MSE =

N ideal

2 1 Ri − Rimod [X Ai − Bi ] λi 2N − M i=1

(9)

Data points which correspond to wavelengths where there is a large difference between the solar irradiance and blackbody spectra are weighted most heavily because they make the largest contribution to the number being minimized, MSE. Ideally, analysis software would optimize by maximizing PCE directly. However, the optimization described above was easy to use with existing software. The reflectance spectra, generated from the fit optical model, yielded the largest PCE for a given set of conditions (solar and blackbody irradiance). This means that attempts were made to modify variables from the values found by the software, but none of these modifications made the film structure perform any better with respect to PCE. Surprisingly, the optimal structure (with respect to PCE) for the AM0 solar spectrum and 450◦ C blackbody was a single cermet layer with 20% gold (77 nm thick) atop an optically thick gold layer, as shown in table 1. One might expect that the optimal structure would have an antireflection layer, but this is not the case. The optimal structure for the AM1.5 solar spectrum at the same blackbody temperature has the same structure, but the thickness of the cermet layer was 82 nm, see table 1. Note that models with multiple layers including antireflective coatings were tested. If additional protection from the space environment is desired, a 20 nm layer of alumina can be deposited on top of the cermet without significantly changing PCE. Optimizations were performed giving PCE for gold films with a single cermet layer on gold, and secondly for a cermet layer with an alumina layer on gold for two blackbody temperatures with the AM0 solar spectrum. These simulations were based on the optical constants obtained from deposited samples. Results of these simulations are summarized in figure 6. Notice that the optimal structures at both temperatures have a 30% gold cermet with an alumina over-layer. Since

273

Photothermal Conversion Efficiency, PCE

CERMET THERMAL CONVERSION COATINGS 1.0 0.9 0.8 0.7 0.6 0.5 0.4

50C Cermet Only 50C Cermet and Alumina 250C Cermet Only 250C Cermet and Alumina

0.3 0.2 0.1 0.0 10

20

30

40

50

60

70

80

Cermet Gold Composition (% volume) Figure 6. Optimum photothermal conversion efficiency as a function of cermet composition for two temperatures. At each temperature, a model with a cermet on gold and a model with alumina on a cermet on gold are shown. These results are based on simulations using optical constants from deposited samples

the optimized structures described in the previous paragraph were optimized for a higher temperature the alumina over-layer did not enhance performance. This is because the alumina layer enhances the absorption of light for wavelengths before the reflectance transition (∼200—1700 nm), but it also may increase blackbody emission for higher wavelengths. For high temperatures, the software puts more emphasis on maximizing infrared reflectance than minimizing visible reflectance because the area beneath the blackbody curve is larger than that of the solar irradiance curve. A set of simulations was done to show how a photothermal film structure would perform if the actual operating temperature was not the same as the blackbody temperature assumed for the design of the film structure. First, three multilayer optical models were constructed. Each model was optimized (with respect to PCE) for the AM0 solar spectrum and a different temperature. From each optimized optical model, a reflectance spectrum was generated. PCE was calculated for the optical models over a range of temperatures using the generated reflectance spectra. The results of these simulations are shown in figure 7. Notice how the conversion efficiency of a given structure asymptotically approaches a maximum value as operating temperature decreases. This means that there is a temperature range for which the efficiency doesn’t change significantly. Figure 7 also indicates that the film structures designed for lower operating temperatures have higher maximum efficiency values than structures designed for higher operating temperatures.

274

BENJAMIN W. WOODS ET AL.

Photothermal Conversion Efficiency, PCE

1.0

Optimized for 50°C Optimized for 250°C Optimized for 450°C

0.8

0.6

0.4

0.2

0.0 0

100

200

300

400

500

600

700

Operating Temperature (°C) Figure 7. Photothermal conversion efficiency for as a function of operating temperature for 3 samples optimized for different temperatures

Two samples were deposited which consisted of single cermet layers on gold. The first sample had a cermet layer with 20% gold and about 80 nm thick. This sample was deposited because the software indicated that it was the best structure for AM0 and 450◦ C blackbody spectra. The second sample had a cermet layer with 30% gold about 100 nm thick. It was not optimized for any solar or blackbody spectrum. Figures of merit were calculated for the reflectance generated by optical models and measured reflectance. These values are shown in table 2. All reflectance measurements were taken at near normal incidence and at room temperature. Temperature dependence was introduced using Planck’s equation for a blackbody. AM0 is the relevant solar irradiance for all calculations. Notice that model values and measured values are close, if not exact, for PCE. The differences in α, ε, and their ratio for modeled and measured values are understandable when one considers that none of the models were optimized for α or ε. 4. Conclusion In this work, cermet films made from gold and alumina were deposited and optically characterized with spectroscopic ellipsometry over the spectral range from 0.2 to 30 microns. Once the optical constants were determined, optical models

275

CERMET THERMAL CONVERSION COATINGS TABLE 2. Figures of merit calculated for simulated and measured reflectance PCE

α

ε

α/ε

Simulated Measured

.58 .54

.67 .74

.01 .02

47 32

Simulated

.80 .80 .80 .80 .56

.80 .79 .79 .79 .79

.02 .005 .006 .01 .03

44 163 130 75 23

Sample #1 Optimized for 450 C 80 nm 20% gold cermet on gold

#2 100 nm 30% gold cermet on gold

50◦ C 50◦ C 100◦ C 200◦ C 400◦ C

Measured

were set up to simulate and predict performance of cermet film structures. A new figure of merit, PCE, was used to quantify the net power absorbed with respect to the net amount of power available for absorption. A weighting factor was introduced so that optical analysis software would optimize film structures based on PCE. Simulations were performed to show how various cermet film structures would perform given blackbody temperatures and solar irradiance spectra. PCE decreases as blackbody temperature increases. Thus a new optimization problem is defined: A high blackbody temperature is desired to enhance heat transfer from a photothermal film to the Stirling engine is efficient as well as increase the Carnot efficiency of the Stirling engine. A low blackbody temperature is desired because photothermal conversion efficiency and net available power (the denominator in eq. (6) are higher. Thus, total energy conversion efficiency must be determined for each specific application. The predictive capability demonstrated in this paper should be helpful in this process.

Acknowledgments This research was generously supported by NASA Glenn Research Center, grant number NAG3-2219.

References 1. Kreibig, U. and Vollmer, M. (1995) Optical Properties of Metal Clusters, Springer, Berlin. 2. W¨ackelga˚rd, E., Niklasson, G. A., and Granqvist, C. G. (2001) Jeffrey Gordon (ed.) Published by James & James (Science Publishers) Ltd, London. Solar Energy: The State of the Art, London. 3. Adsten, M., Joerger, R., J¨arrendahl, K., and W¨ackelg˚rrd, E. (2000) Solar Energy 68, 325. 4. Jaworske, D. A. and Shumway, D. A. (2003) Space Technology & Applications International Forum, STAIF-2003, Albuquerque, NM, February 2003, pp. 65–70.

276

BENJAMIN W. WOODS ET AL.

5. Marco, J., Bhojaraj, H., and Hulyal, R. (2003) In Proceedings from the 9th International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands, 16–20 June 2003, p. 359. 6. Duvignacq, C., Hespel, L., Roze, C., and Girasole, T. (2003) In Proceedings from the 9th International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands, 16–20 June 2003, p. 399. 7. Itawa, M., Ohnishi, A., Saito, R., and Ohkubo, K. (1997) In Proceedings of the 7th International Symposium on Materials in a Space Environment, Toulouse, France, 16–20 June 1997, p. 59. 8. Marco, J., Paillous, A., and Gourmelon, G. (1994) In Proceedings of the 6th International Symposium of Materials in a Space Environment, ESTEC, Noordwijk, The Netherlands, 19–23 September 1994, p. 77. 9. Tenditnyi, V. A., Smirnov-Vasilev, K. G., Yevkin, I. V., and Mironovich, V. V. (1994) In Proceedings of the 6th International Symposium of Materials in a Space Environment, ESTEC, Noordwijk, The Netherlands, 19–23 September 1994, p. 113. 10. Hidden, G., Vega, J. M., Remaury, S., and Nabarra, P. (2003) In Proceedings from the 9th International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands, 16–20 June 2003, p. 49. 11. Triolo, J., Ackerman, N., Neuberger, D., and Harris, G. (1997) In Proceedings of the 7th International Symposium on Materials in a Space Environment, Toulouse, France, 16–20 June 1997, p. 411. 12. Guillamon, J. C. (1997) In Proceedings of the 7th International Symposium on Materials in a Space Environment, Toulouse, France, 16–20 June 1997, p. 427. 13. Granqvist, C. G. (1989) Spectrally Selective Surfaces for Heating and Cooling Applications SPIE, Bellingham, 1989. 14. Halliday, D. and Resnick, R. (1981) Fundamentals of Physics, 2nd ed., Wiley, New York, p. 777. 15. Wehrli, C. “Extraterrestrial Solar Spectrum”, Publication no. 615, PhysikalischMeteorologisches Observatorium + World Radiation Center (PMO/WRC) Davos Dorf, Switzerland, July 1985. Also available at time of publishing from http://rredc.nrel.gov/solar/ spectra/am0/wehrli1985.new.html. 16. Neckel, H. and Labs, D. (1984) Solar Physics 90, 205. 17. Ritchie, I. T. and Window, B. (1977) Applied Optics 16, 1439. 18. Trotter, D. M. and Sievers, A. J. (1980) Applied Optics 19, 711. 19. Zhang, Q. C. and Mills, D. R. (1992) Journal of Applied Physics 72, 3013. 20. Zhang, Q. C. (2000) Solar Energy Matererials in Solar Cells 62, 63. 21. Harding, G. L. and Craig, S. (1981) Solar Energy Materials 5, 149. 22. Farooq, M., Green, A. A., and Hutchins, M. G. (1998) Solar Energy Materials in Solar Cells 54, 67.

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS ROMAN V. KRUZELECKY,1 EMILE HADDAD,1 BRIAN WONG,1 WES JAMROZ,1 MOHAMED SOLTANI,2 MOHAMED CHAKER,2 DARIUS NIKANPOUR3 AND XIN XIAN JIANG3 1 MPB Communications Inc (MPB), 151 Hymus Blvd., Pointe Claire, Quebec, Canada 2 ´ INRS Energie et Mat´eriaux (INRS), Varennes, Quebec, Canada 3 Canadian Space Agency (CSA), St. Hubert, Quebec, Canada

Abstract. This paper describes a new multifunction smart coating that can provide atomic oxygen (AO) and electrostatic discharge (ESD) protection, while also improving the thermal control of space structures. The methodology is based on a passive thin-film structure employing VOn transition metal oxides that exhibit a metal to insulator transition. The coating, depending on its formulation, can provide a variable heat-transfer/emitter structure that operates passively in response to changes in the temperature of the space structure, by dynamically varying the ratio of solar absorptance (α) to thermal emittance (ε). This enhances self-heating of the structure at lower temperatures and cooling through thermal radiation at elevated temperatures. Work is currently underway to apply this coating to various polymers and membranes to improve their performance in space. In the space environment, such as low Earth orbit (LEO), the coating will be subject to various stresses including VUV radiation and AO. Atomic oxygen testing in a simulated environment at CSA indicated no resolvable change in the morphology or thickness of the coatings. The thermo-optic characteristics after AO exposure were similar to the “as deposited” films. Additional long-term radiation exposure at the Centre National d’Etudes Spatiales—France (CNES), equivalent to three years in a geostationary orbit (GEO) environment, resulted in a change in the coating ε and α of less than 0.002. Key words: smart material, thin-film coating, tuneable emittance, AO protection

1. Introduction Space represents a functional and/or durability challenge to many materials due to the high levels gamma radiation, charged particles, vacuum ultraviolet (VUV) 277 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 277–293. C 2006 Springer. Printed in the Netherlands.

278

ROMAN V. KRUZELECKY ET AL.

radiation and reactive atomic oxygen (AO) that are present to a greater, or lesser extent, in various orbits. These environmental factors can significantly modify the performance of the outer layers of a spacecraft, and for certain materials, especially various polymers, cause significant material erosion and degradation of their thermo-optic characteristics. The effects of the space environment on various materials has been examined through various studies including NASA’s Long Duration Exposure Facility (LDEF) set of experiments [1]. In particular, Kapton polyimide and Teflon FEP are widely employed in space as part of thermal blankets and for membrane antennas but suffer significant erosion due to the VUV AO that is present, especially in the low Earth orbit (LEO) environment [2]. The outer spacecraft coverings control the thermal balance between the spacecraft, incident solar and terrestrial radiation, and thermal emittance to dark space. The efficient thermal control of spacecraft structures is an important issue that impacts directly on the performance and longevity of attached and internal subsystems. While spacecraft can be subjected to external temperature swings from about −150 to +150◦ C; the corresponding internal temperature must be regulated over a nominal range, typically −10–30◦ C. The tighter the temperature control of the spacecraft; the better the performance ratings and the lifetime of the spacecraft subsystems and payloads. Thermal control systems typically require heaters with significant power requirements to maintain payload temperatures at acceptable levels during cold swings to compensate for the radiator heat loss to dark space. Traditional dynamic thermal-control systems have employed various forms of mechanical louvers to regulate heat dissipation into deep space. However, these are bulky (3.3 kg·m−2 ), subject to wear, require additional power, and are costly, exceeding $100K US m−2 [3]. With the trend towards smaller, highly functional satellites; there is a need for cost-effective smart radiator devices (SRD) that are lightweight and reliable. 1.1. OPTICALLY-ACTIVE SMART MATERIALS

Several material systems, mainly involving oxides of transition metals such as V, W, Mn and La, have been found to exhibit a change from metallic to insulator behavior (metal-insulator transition) in response to composition (doping), applied electric field, temperature, or the application of pressure. These transition metals are characterized by partially-filled d-orbitals that contribute to the metallic bonding and electrical conduction. They can readily form a variety of complexes involving the d-orbitals [4]. Chemical bonding of the transition metal can produce an energy splitting, o , of the d-orbitals due to electrostatic ionic and electronelectron interactions. Depending on the electron configuration in the resulting d-related states of the transition metal complex, the system can exhibit metallic or insulating characteristics [4].

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

279

The main smart material systems that have been considered for thermal control applications include V1−x−y Mx N y On [5], where M and N are dopants such as W, Ni, Cr, Ti, and Mo; ceramic tiles based La1−x Mx MnOn perovskite materials (Mx can be Ca, Sr, or Ba [6]); and electrochromic active devices based on WO3 [7, 8]. The alloying or inclusion of M and N dopants is used to modify the characteristics of the material. In this paper, the authors focus mainly on the VOn based smart material system. VOn exhibits one of the largest observed variations in electrical and optical characteristics due to the metal-insulator transition. The transition temperature increases with the oxygen content; varying from 126 K for VO, to 140K for V2 O3 , and to 341 K for VO2 [9, 10]. The metal-insulator transition in Vx On is associated with a change in structure from a tetragonal rutile structure with metallic characteristics above the transition temperature, to a monoclinic structure with insulator-like characteristics below the transition temperature. It is also feasible to induce the metal-insulator transition in VO2 actively using an applied electric field. Measurements by Stefanovich et al. [11] of the switching characteristics in the electrical current using the VO2 in a FET configuration indicate that the metalinsulator transition in VO2 is related to a critical electron density that induces the transition to a metallic state, similar to the classical Mott metal-insulator transition. The Mott criteria for the transition is given by (n c )1/3 αH ≈ 0.25, where n c is the critical electron density and α H is the Bohr radius. The value of the critical density was found experimentally [11] to be about 3 × 1018 cm−3 , in good agreement with theoretical predictions. Since the driving mechanism is an electric field effect, the power requirements to induce the switching are moderate, comparable to the operation of a CMOS device. The broad-band optical transmittance characteristics of a 250 nm thick VO2 coating deposited on c-Si, as measured by FT-IR [5], are shown in figure 1 for the metallic (95◦ C) and insulating (29◦ C) states. The VO2 was prepared by reactive laser ablation [5]. This provided high transmittance in the insulating state that extends from below 2 μm in the NIR to beyond 15 μm in the infrared. The relatively low optical absorption by the VO2 in this insulating state suggests a relatively clean band-gap, free of defect states. In the metallic state, the optical transmittance of the VO2 /Si decreased close to 0%, remaining below 0.5% into the FIR. The nominal transition temperature for undoped crystalline VO2 is about 68◦ C. The corresponding switching in the electrical characteristics of the crystalline VO2 is shown in figure 2. There is a three order of magnitude decrease in the resistivity of the VO2 due to the transition to a metallic state above 70◦ C over a transition region that is about 10◦ wide. La1−x Mx MnOn is a class of double-exchange (DE) ferromagnets [6]. The double-exchange refers to a double electron exchange between neighbouring Mn

280 Transmittance (%)

ROMAN V. KRUZELECKY ET AL. 100 90 80 70 60 50 40 30 20 10 0

29°C

95°C 2

4

6

8 10 12 14 16 Wavelength (micrometers)

18

20

Figure 1. Optical transmittance of undoped VO2 deposited onto Si in the insulating (29◦ C) and metallic (95◦ C) states (near T = 0%)

Electrical resistance (Ω)

and O. This affects the Mn-O-Mn bond angle, and hence, the crystal structure. The DE ferromagnets differ from the vanadium oxide family in that the DE class of materials is metallic at lower temperatures and semiconducting above the transition temperature, over the appropriate compositions. The large change in properties occurs near x = 0.3 in composition with LaMnO3 behaving as an insulator at one extreme of the compositional variation. Bulk samples are typically prepared by sintering a powder mixture of the appropriate composition at high temperatures above 1200◦ C for about 60 h. This can be employed to produce tiles of the material. Electrochromic structures based on WO3 [7, 8] employ an active switching mechanism involving the electric-field induced motion of group 1 ions, such as H+ or Li+ , from a storage reservoir through an ion conducting layer to the WO3 layer. The resulting reactions convert the WO3 , which is IR transparent, to HWO3 (or LiWO3 ) complexes that are metallic and reflective in the IR. This entails a multilayer structure; consisting of a polymer that acts as a storage reservoir for the

105

104

103 20

40

60

Temperature (°C)

80

100

Figure 2. Variation of the electrical resistance of VO2 deposited onto quartz at 500◦ C as measured using the four-point probe method

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

281

H+ or Li+ -ionic colorant, an intermediate solid or liquid electrolyte to facilitate field-induced migration of the colorants, and the active WO3 layer. The layers are sandwiched between two electrodes such as ITO. The transition from an insulating to a metallic state is controlled by the application of a voltage. There is a current flow during switching, typically about 4 mA·cm−2 . 2. Experimental Results This work focused on the growth of VOn on Kapton and Al for various thermalcontrol applications in space. VOn films were also prepared on crystalline Si (c-Si) and quartz to enable various thin-film characterization. Growth on Si facilitated IR characterization of the film thermo-optical characteristics; while growth on quartz enabled characterization of the electrical characteristics and the optical transmittance in the VIS/NIR spectral range. Film characterization included 1. 2. 3. 4. 5. 6.

X-ray diffraction (XRD) to study the film crystal structure, VIS transmittance/reflectance, IR emittance versus temperature, VUV and AO irradiation, SEM observation of the film morphology, Four-point probe resistivity versus temperature.

2.1. FILM PREPARATION

Undoped and doped VO2 films were prepared by reactive pulsed laserablation using a special target and a XeCl excimer laser, as shown in figure 3. The depositions were performed in a controlled gas-phase background consisting of an O2 .Ar−1 gas mixture at total pressures of about 100 mTorr. The substrate temperature was systematically varied from 200 to 350◦ C for growth on Kapton, and from 300 to 500◦ C for growth on Al. The film thickness ranged from 0.1 to about 1.0 μm. The trial test coupons were 25 mm O.D. in size for the coated area. X-ray diffraction (XRD) was employed to study the structure and stoichiometry of the resulting coatings. Deposition conditions were established for the formation of stoichiometric VO2 . At the higher concentrations of O2 in the gas mixture and higher gas-phase pressures, a mixed-phase structure containing VO2 and V2 O5 was observed by XRD. As the total pressure was reduced to below 100 mTorr, and the concentration of O2 in the mixture was reduced to about 2.5%, films that were mainly VO2 were obtained. For film growth on crystalline substrates such as quartz and c-Si, a distinguishable diffraction peak associated with a crystalline

282

ROMAN V. KRUZELECKY ET AL.

Figure 3. Schematic of laser ablation deposition apparatus

VO2/Si VO2

TS=520°C

400000

Intensity (Counts)

200000 0 400000 10

20

30

40

50

60

TS=420°C

200000

0 10

20

30

40

50

400000

60

TS=330°C

200000 0 10

20

30

40

50

60

2θ (degree)

Figure 4. XRD analysis of the deposition temperature effect on the vanadium oxide structure and stoichiometry, as deposited onto c-Si

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

(a)

VO2/Aluminum

Al

700

283

(b) 250

kapton substrate 600

400

Intensity

Al

500

Intensity

200

Al

300

150

100

Al

200

50

VO2

100

0

0 10

20

30

40

50

2θ (degree)

60

70

80

10

20

30

40

50

60

Figure 5. X-ray diffraction spectra of VOn deposited onto (a) polished Al at 520◦ C and (b) Kapton at 300◦ C

VO2 structure could be observed by XRD down to deposition temperatures of about 330◦ C. For film growth on Al, a distinguishable diffraction peak characterizing the crystalline VO2 stoichoimetry (>90%) could be observed by XRD at the higher deposition temperatures near 500◦ C, as shown in figure 5(a), as well as the background Al peaks from the substrate. For film growth on Kapton, typically at deposition temperatures below 350◦ C, the XRD measurements, as shown in figure 5(b), exhibit broad peaks indicative of an amorphous structure. The VOn films exhibited good adhesion to both Al and to Kapton with no evidence of flaking or cracking under microscopic examination.

Figure 6. Photograph of VOn deposited onto a Kapton test coupon

284

ROMAN V. KRUZELECKY ET AL. 1

Transmittance

0.9 0.8 0.7 0.6 0.5 0.4 500

550

600

650

700

750

800

850

Wavelength (nm)

Figure 7. VIS Transmittance of a 100 nm thick VOn deposited onto Glass at 150◦ C

2.2. OPTICAL CHARACTERISTICS

The VIS optical transmittance characteristics of preliminary VOn films deposited onto glass at lower deposition temperatures were examined using a tungsten light source and an Ocean-Optics VIS spectrometer. Figure 7 shows the single pass optical transmittance of a 100 nm thick VOn film relative to the glass substrate. As shown, the VIS transmittance is relatively good, typically above 80% in the VIS, indicating a solar absorptance below 0.2. In most applications, what is required is a high VIS reflectance for external Sun-Shield applications. This can be addressed by using the VOn in a multilayer structure to provide both high VIS reflectance and high IR thermal emittance. These structures are currently being developed on Kapton. Figure 8 shows the (a)

(b) AM0 Spectra

Spectral Irradiance W*m−2*micron−1

2200 2000 1800 1600 1400 1200 1000 800 600 400 200 0 0

0.5

1

1.5

2

2.5

3

3.5

4

Wavelength micrometers ASTM E-490

Wehrli WMO

Figure 8. (a) AM0 spectrum (after NASA [12]) and (b) Optical simulation of VOn -based VIS selective reflector on Kapton

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

285

Sheet Resistance (Ohm/ )

1.00E+07

1.00E+06

1.00E+05

30

40

50

60

70

80

90

100

110

120

Temperature (C) SSH3 150 nm 300 C

SSH4 150 nm 250C

SSH5 150 nm 350C

Figure 9. Sheet resistance (Ohm/) versus temperature for 150 nm thick VOn films deposited onto 5 mil Kapton at various deposition temperatures (250◦ C—SSH04; 300◦ C—SSH03; 350◦ C— SSH05)

attainable VIS reflectance for a relatively simple four-layer structure, as simulated using Zeemax optical simulation software, based on the measured VIS optical characteristics of the VOn on Kapton. A peak reflectance of about 0.9 should be attainable at 0.5 μm, near the peak in the AM0 solar radiation spectrum [12]. The preliminary results suggest that the net effective solar absorptance (α) on Kapton can be reduced below 0.2. 2.3. ELECTRICAL CHARACTERISTICS

The sheet resistivity characteristics of 150 nm thick VOn films deposited onto Kapton were examined using a commercial 4-point probe. Figure 9 shows the variation of the sheet resistivity with temperature for VOn films deposited between 250 (K02) and 350◦ C (K05) on Kapton. Relative to the resistivity characteristics of the crystalline VO2 deposited on quartz, as shown in figure 2, the electrical characteristics of the VOn /Kapton did not exhibit a significant transition near 70◦ C. Lowering the deposition temperature tended to increase the resultant sample resistivity (VO2-SSH-K04), as shown by figure 9. This may be due to a more amorphous structure as the deposition temperature is lowered, as suggested by the XRD measurements, that results in greater carrier scattering and trapping. Typically, the room-temperature resistivity of the amorphous VOn deposited near 300◦ C on Kapton was about an order of magnitude larger than that of crystalline VO2. Since background scattering by the Kapton obscures the XRD measurements,

286

ROMAN V. KRUZELECKY ET AL.

additional XPS measurements are planned to clarify the stoichiometry of the VOn deposited at lower substrate temperatures onto the Kapton. 2.4. THERMAL EMITTANCE

Broad-band sample emissivity data was provided using the measurement set-up shown in figure 10. The samples were placed on a polished Al pad that could be heated or cooled using a proportional-integral controller. The thermally-induced IR emission from the sample was collected and focused onto a pyroelectric detector using a system of ZnSe lenses. An optical chopper and signal lock-in amplifier were used to extend the signal measurement range and to reduce background effects. An f = 25 mm lens allowed signal collection from a small section of the sample surface to reduce background effects. For the initial measurement setup, the pyroelectric detector was used without a window, providing full spectral measurements of the total emitted thermal signal. The thermal emissivity of VOn on 5 mil Kapton HN was estimated using a polished Al backing to compensate for the IR transmission of the Kapton. Figure 11 shows a graph of the measured signal detected by the pyroelectric detector versus the surface temperature, as determined using a miniature thermister, for a 150 nm thick VOn sample deposited at 300◦ C. Relative to a blackbody reference standard, the measurements suggest a net emissivity of about 0.9 for the VOn /Kapton combination. The metal-insulator transition expected near 68◦ C in undoped crystalline VO2 was not resolvable. Figure 12 shows the measured pyroelectric detector signal for an engineered VO2 sample on an Al substrate (ISRD-010), that was structured for tuneable

Figure 10. Preliminary measurement set-up using a pyroelectric detector

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

287

Pyroelectric Detector Signal (V)

7 6 5 4 3 2 1 0

20

30

40

50

60

70

80

90

100

Calibrated Temperature (C) blackbody

Polished AL

VO2_SSH-K-01

Figure 11. Thermal emission of a 150 nm VOn film on Kapton HN (a type of Kapton provided by Dupont)

emittance applications [6], as well as the corresponding signals for a calibrated blackbody reference (ε = 0.88) and polished Al reference (ε = 0.06). At a given sample temperature, the pyroelectric detector signal is proportional to the thermal signal emitted by sample. Figure 13 shows the corresponding sample emissivity of ISRD-10 between 30 and 100◦ C. The emissivity was estimated, at a given measurement temperature T , by the linear interpolation of the measured detector signal relative to the measured signals for the two reference samples; the high-emissivity (blackbody) and

Pyroelectric Detector Signal (V)

5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

Calibrated Temp (C) ISRD-10-Repeat

Blackbody

ISRD-10

Polished Al

Figure 12. Pyroelectric detector signal versus sample temperature

288

ROMAN V. KRUZELECKY ET AL. 0.7

Calibrated IR Emissivity

0.6

0.5

0.4

0.3

0.2

0.1

0

20

30

40

50

60

70

80

90

100

Surface Temperature (C)

Figure 13. Esimated emissivity for VOn /Al sample (ISRD-10) versus temperature

low-emissivity (polished Al) reference standards. The advantage of this technique is that the emissivity of the sample at any temperature can be determined. The accuracy depends on the known emissivities of the reference samples. A commercial Devices and Services Co. (D&S) emissometer was employed to calibrate the emissivity readings near 30◦ C. The sample emissivity varies from about 0.2 at lower temperatures to about 0.65 in the metallic state at higher temperatures, providing a net emissivity tunability of about ε = 0.45, with a relatively gradual variation in the emissivity that is attractive for spacecraft thermal control applications. The rise in the emittance between 40 and 60◦ C corresponds to the formation of the critical electron density to induce the metal-insulator transition. The sharp increase in the emittance between 70 and 80◦ C corresponds to the formation of free electrons in the metallic state, as indicated by the R(T ) data in figure 2. The transition characteristic can be shifted in temperature by doping the VOn as discussed in section 2.5. Moreover, the low-temperature emissivity can be biased upwards by the addition of a suitable high-emissivity layer between the VOn and the Al. as discussed in [6]. 2.5. DOPING

The metal/insulator transition temperature of VO2 can be controlled by the incorporation of suitable donor-like dopant such as W, and/or acceptor-like dopants such as Ti. Table 1 summarizes some of the experimental results for the doping of VO2 prepared at 500◦ C by reactive laser ablation. As indicated, the transition

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

289

TABLE 1. Effect of W and Ti doping on metal-insulator transition temperature of VO2 deposited by reactive Laser Ablation Deposition (LAD) W-doping Ti Doping Transition Temperature

0.0

1.5%

3%

3%

0.0 68◦ C

0% 30◦ C

0% −20◦ C

1% About 0◦ C

temperature can be shifted over a wide range, from above 70◦ C to below −20◦ C, as may be required for a given application. 2.6. PRELIMINARY AO AND RADIATION TESTING

AO testing was performed at the Canadian Space Agency (CSA) by exposing the samples (bare and coated substrate) in the afterglow area of a RF plasma in oxygen. Samples were tested one at a time for a duration of about 1 h each. A piece of Kapton R was used as a reference. Each sample had an area that was masked from exposure to the plasma using a Kapton mask. The oxygen plasma operating parameters are given in the following table 2. Several VO2 samples as deposited on Al were irradiated with atomic oxygen at the CSA test facility. The fluxes, fluences and LEO equivalent time estimations are based on NASA and ESA Standard [ECSS-E-10-04A, 2000]. Examination of the VO2 samples after AO exposure at CSA indicates that there was no resolvable change in mass or the surface morphology. The measurement of the thermo-optical characteristics after AO exposure indicates that the VO2 samples still exhibit a good metal-insulator transition with similar characteristics to those measured prior to the exposure. Recently an independent entity (CNES—the French Space Agency) demonstrated the stability of the thermo-optical properties coating prepared by MPB against Atomic Oxygen, thermal vacuum and space radiation for the equivalent of three years GEO satellite. The results are summarized in table 3. The preliminary results suggest that the VOn formulation could provide good AO protection for space structures, while providing stable thermo-optic characteristics. TABLE 2. Parameters of the atomic oxygen irradiation Atomic oxygen parameter

Value

Fluence Flux Equivalent LEO satellite time to receive same amount of AO. Atoms energy

1017 particles·cm−2 1015 particles·cm−2 ·min−1 6 Months 0.1–0.2 eV

290

ROMAN V. KRUZELECKY ET AL.

TABLE 3. Summary of preliminary space environment testing for VOn on Al Factor

Test facility

Results

Atomic Oxygen

CSA: Equivalent to 6 months LEO exposure to atomic oxygen.

No change in morphology. No change in emissivity tunability. No measurable mass loss.

Radiation

CNES: 6 month long-term radiation exposure at CNES for samples equivalent to 3 years GEO radiation.

ε < 0.002 α < 0.002

3. Discussion Two space applications are currently being developed by MPB for the VOn material system: 1. multifunction protective coating on polymers such as Kapton. 2. thin-film tunable emittance coating on Al for thermal radiator tiles, As a coating on polymers such as Kapton, the VOn structure can offer several functionalities:

r r r r

solar absorptance below 0.2, with high VIS reflectance, IR thermal emittance exceeding 0.8, Rs(T ) < 109 ·/ for ESD protection to below –100◦ C, AO protection.

Work is currently ongoing to optimize the low-temperature deposition process for VOn on Kapton to also provide some tunability of the net thermal emissivity. The goal is a tunability in the net emissivity of about 0.2–0.3 between the high and low-temperature values to facilitate dynamic temperature control of membrane structures. Spacecraft employing passive thermal radiators can require substantial heating at lower temperatures to compensate for the high thermal exchange to dark space. Active systems using mechanical louvers are subject to wear and bear a considerable mass penalty. MPB is developing a thin-film smart radiator device (SRD) based on its VOn technology that is applied directly to the spacecraft Al radiator, resulting in minimal added mass (<100 gm·m−2 ), long-term reliable operation with a target of 15 years, and robust integration with the spacecraft structure. Advanced nano-engineering concepts are being employed to structure the VOn on Al in order to optimize the emissivity tunability. The design optimization and testing is currently ongoing. As previously shown in figure 12, broadband emissivity tunability (4–20 μm) exceeding 0.45 has been achieved.

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

291

(a) Low-Temperature Case:

(b) High-Temperature Case:

Figure 14. Operation of MPB passive thin-film SRD

The MPB thin-film SRD functions to control the differential heat exchange between the spacecraft and vacuum by dynamically controlling the heat return to the spacecraft, the heat transfer to the external radiating surface, and the heat exchange to dark space. Figure 14 provides a schematic of the operation of the SRD below (low-temperature case), and above (high-temperature case) a preselected thermal-switch point. When the spacecraft is cold, heat loss into space is minimized (ε < 0.3) to stabilize the spacecraft temperature. Otherwise, this heat loss needs to be compensated using internal heaters with an additional power penalty. When the spacecraft is hot, maximal heat is dynamically transferred to the radiator and the heat radiation to dark space is maximized (ε > 0.65), while heat return to the spaceship is minimized. The proposed dynamic SRD significantly reduces the heat loss at lower temperatures to enable spacecraft designers to avoid or reduce internal heating requirements. Nominally, the performance of space systems must be derated to account for the effects of the large temperature swings. The SRD can also assist to reduce

292

ROMAN V. KRUZELECKY ET AL.

the peak-to-peak temperature swings of the spacecraft. This would enable existing space systems to operate in temperature ranges of maximum performance.

4. Conclusion High-quality VO2 films have been achieved on Si, quartz and Al substrates by hybrid reactive pulsed-laser ablation. The metal-insulator transition temperature can be tailored from below −20 to above 70◦ C through doping. Below this transition temperature, the VO2 films are insulating in nature and display high optical transmittance over a broad spectral range that extends from the visible to the far infrared. Above the transition temperature, the films exhibit a sharp drop in optical transmission and a corresponding sharp increase in their electrical conductivity and reflectivity. The VO2 -based smart coatings can be employed in a variety of ways to provide very flexible thermal management of spacecraft. The deposition of VOn onto Kapton at relatively low temperatures of 150 and 300◦ C results in an amorphous structure with a weakened metal/insulator transition. Adhesion to Kapton is relatively good, with no observable flaking or cracking. The low-temperature deposition (<350◦ C) results in an amorphous structure that exhibits higher resistivities and a lower conduction activation energy than the crystalline VO2 deposited at 520◦ C, as used for the thin-film SRD applications. VOn -based structures are being developed for Kapton that provide an emissivity exceeding 0.85 and an ε/α > 2 for the thermal control of membrane structures. The VOn coating also exhibits excellent stability under exposure to AO and radiation as encountered in LEO and GEO environments. The thin-film structure can provide several functionalities for polymers including AO and ESD protection, thermal control and selective rejection of incident solar radiation. Work is ongoing to facilitate dynamic variation of the ε/α passively in response to the temperature of the membrane. Both active and passive SRD devices using the VOn on Al are being developed for the dynamic thermal control of spacecraft. Passively, the VO2 can be used as a combined variable-conductance switch/variable-emittance radiator. This enhances the net dynamic variation in emissivity that is possible for a passive SRD device. Combining these effects has provided experimentally measured net variations in IR emissivity from below 0.25 at lower temperatures to above 0.65 at higher temperatures for a passive device on Al. Still greater high-temperature emissivities may be possible by further optimizing the structure and doping of the VO2 to enhance IR optical absorption in the metallic state. The proposed thin-film thermal-control methodology has significant advantages over competitive technologies in terms of weight, cost, structural simplicity with no mechanical components, integration with the space structure,

MULTIFUNCTION SMART COATINGS FOR SPACE APPLICATIONS

293

compatibility with the space environment and dynamic thermal emittance performance. The proposed passive thin-film SRD employs vacuum-compatible layers that have low solar absorptance and good AO resistance. Stability in the presence of ions and protons should also be good due to the strong stoichiometric bonding. An added benefit is the ability to deposit the thin-film SRD on various qualified space materials such as Al and Kapton to minimize the overall mass of the spacecraft thermal control system.

Acknowledgments This work was partially funded by the Canadian Space Agency.

References 1. Whitaker, A. F. and Gregory, J. (1993) LDEF Materials Results for Spacecraft Applications, Reports: N-94-31012; NASA-CP–3257; M–742; NAS–1.55:3257; CONF-9210467. 2. Silverman, E. M. (1995) Space Environmental Effects on Spacecraft: LEO Materials Selection Guide, NASA CR-4661. 3. Tennyson, R. C. and Morison, W. D. (1990) In Proceedings TMS Symposium on Space Environmental Effects on Materials, Anaheim, CA, February 1990. 4. Gilmore, D. G. (1994) Satellite Thermal Control Handbook, The Aerospace Corporation Press, El Segundo, CA. 5. Mahan, B. A. (1975) University Chemistry, 3rd ed., Addison-Wesley Publishing, Reading, MA, USA, pp. 685–743. 6. Kruzelecky, R. V., Haddad, E., Jamroz, W., Soltani, M., Chaker, M., Nikanpour, D., and Jiang, X. X. (2003) In the 33nd International Conference on Environmental System 33-IES, Paper 03ICES-242, Vancouver, Canada, July 2003. 7. Shimakawa, Y., Yoshitake, T., Kubo, Y., Machida, T., Shinagawa, K., Okamoto, A., Nakamura, Y., Ochi, A., Tachikawa, S., and Ohnishi, A. (2002) Applied Physics Letters 80, 4864–4866. 8. Franke, E., Neumann, H., Schubert, M., Trimble, C. L., Yan, L., Woollam, J. A. (2002) Surface and Coating Technology 151–152, 285–288. 9. Douglas, D. T., Swanso, R., Osiander, J., Champion, J., Darrin, A. G., Biter, W., and Chandrasekha, P. (2002) In M. El-Glenk (ed.) Space Technology and Applications International Forum (STAIF)-2002, CP 608, American Institute of Physics, pp. 204–210. 10. Griffiths, C. H. and Eastwood, H. K. (1974) Journal of Applied Physics 45, 2201–2206. 11. Egorov, F. A., Yu. Sh. Temirov, Dvoryankin, V. F., Potapov, V. T., and Sokolovskii, A. A. (1991) Sov. Tech. Phys. Lett. 17, 295–296. 12. Stefanovich, G., Pergament, A., and Stefanovich, D. (2000) Journal of Physics. Condensed Matter 12, 8837–8845. 13. ASTM-E-490. Version SOLAR2000, Model AM0 (Solar irradiance, Air mass zero, distance of 1 A.U.) rredc.nrel.gov/solar/standards/am0/wehrli1985.new.html.

EFFECTS OF SPACE ENVIRONMENT EXPOSURE ON THE BLOCKING FORCE OF SILICONE ADHESIVE PAUL BOEDER,1 RON MIKATARIAN,1 MARY J. LORENZ,1 STEVE KOONTZ,2 KEITH ALBYN,3 AND MIRIA FINCKENOR3 1 The Boeing Company, 13100 Space Center Blvd., Houston, Texas, 77059, U.S.A. 2 NASA Johnson Space Center, M.S.ES4/NASA/JSC, 2101 NASA Road 1, Houston, TX 77058, U.S.A. 3 NASA Marshall Space Flight Center, Mail Code ED31, Bldg. 4711 Room 100B, AL 35812, U.S.A.

Abstract. An issue has been identified with the diode tape used on the International Space Station (ISS) solar arrays to provide a high emittance surface for the underlying diodes. The diode tape consists of silicone pressure sensitive adhesive (Dow Corning QC-7725) with a protective Kapton over-layer to prevent sticking prior to solar array deployment. On-orbit, the Kapton over-layer will erode under exposure to atomic oxygen (AO). Under additional AO exposure, the underlying exposed silicone adhesive is expected to convert to a glass like silicate. Conventional thinking suggests that at some point, the Kapton will be eroded, exposing pure silicone adhesive. The silicone will lose its adhesive characteristics after some time as it converts to a silicate due to AO exposure. This presents a problem because the current operational plan to retract ISS solar array P6 and leave it stored under load for a long duration (6 months or more) may be incompatible with the timing of blocking minimization for the silicone adhesive. Previous testing by Lockheed-Martin Space Systems (LMSS) characterized silicone blocking following exposure to atomic oxygen (AO) in an electron cyclotron resonance (ECR) source facility. LMSS blocking test results, combined with analysis results of predicted ISS on-orbit AO exposure levels indicate the silicone blocking forces resulting from retraction and storage of the P6 4B solar array for an extended period of time may reach unacceptable levels. That is, the silicone may not be sufficiently converted to a silicate. The LMSS ECR source test results and their application to ISS on-orbit conditions are believed to be conservative. To address this conservatism, the Environment Effects Group at Marshall Space Flight Center, under direction from the ISS Program Office Environments Team, performed environments exposure testing with a high energy 5 eV AO beam along with near ultraviolet (NUV) 295 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 295–307. C 2006 Springer. Printed in the Netherlands.

296

PAUL BOEDER ET AL.

radiation and ionizing radiation to more accurately simulate the actual on-orbit environment. Diode tape samples were exposed to each environment constituent individually, put under preload for seven days and then the resulting blocking force was measured using a tensile machine. Additional samples were exposed to AO, NUV and electrons in series and then put under long term (three to ten months) preload to determine the effect of preload duration on the resulting blocking force of the silicone-to-silicone bond. In this paper, we present the results of laboratory tests to determine the effect of space environment exposure (5 eV atomic oxygen, ultraviolet radiation, and electron radiation) on the blocking force of silicone adhesive in a silicone-to-silicone bond. Test results indicate that high energy AO, ultraviolet radiation and electron ionizing radiation exposure all reduce the blocking force for a silicone-to-silicone bond. AO exposure produces the most significant reduction in blocking force. Key words: atomic oxygen, ultraviolet radiation, electron radiation, silicone adhesive, blocking

1. Introduction In this paper, we present the results of laboratory tests to determine the effect of space environment exposure on the blocking force of silicone adhesive in a silicone-to-silicone bond. Environment exposure includes both individual and combined exposure to 5 eV atomic oxygen, ultraviolet radiation, and energetic electron ionizing radiation. Test results available to date and presented herein show that long-term exposure to each of the environment constituents reduces the adhesive properties of a silicone adhesive. Section 2 describes the background and reasoning for performing the testing. Section 3 describes the testing program. Section 4 summarizes the test results and their implications. Conclusions are given in section 5.

2. Background MD-944 diode tape used on the International Space Station (ISS) U.S. solar array panels provides a high emittance surface for the underlying diodes in the solar array panel circuit. MD-944 diode tape consists of a pressure sensitive silicone adhesive (Dow Corning QC-7725) with a protective Kapton over-layer to prevent sticking prior to solar array deployment. On-orbit, the Kapton over-layer will erode under exposure to atomic oxygen (AO). Once exposed to AO, the underlying silicone adhesive is expected to ultimately convert to a silicate with no residual adhesive properties.

EFFECTS OF SPACE ENVIRONMENT EXPOSURE

297

Figure 1. ISS P6 Solar Array Configuration [1]

The P6 solar array is currently mounted on the Z1 Truss Segment on the zenith of ISS as shown in figure 1. The current operational plan is to store the P6 solar array for approximately six months after the P4 solar array is installed and becomes operational. The P6 solar array will eventually be re-deployed outboard of the P4 solar array. The long duration on-orbit storage of the P6 solar array presents a potential problem. When the solar array is folded up for storage, there will be contact between the diode tape on opposing panel faces. If diode tape exposure conditions are just right, it is possible that the diode tape on opposing panels could stick together. This would be possible if the solar array is stored after the Kapton is eroded, but before the silicone adhesive converts to a silicate. If this occurs, then blocking may develop at the diode tape to diode tape interface. If sufficient blocking develops, the diode tape and underlying diodes may be damaged upon redeployment of the solar array. In a worst-case scenario, the force required to break the silicone-to-silicone bond may exceed the solar array deployment mechanisms mechanical capability. Previous testing by Lockheed-Martin Space Systems (LMSS) characterized silicone blocking following exposure to low energy atomic oxygen (AO) in an electron cyclotron resonance (ECR) directional plasma source facility (unpublished). Initial interpretations of LMSS blocking test results, combined with analysis results of predicted ISS on-orbit AO exposure levels, indicate the silicone blocking forces resulting from retraction and storage of the P6 solar array for an extended period of time may reach unacceptable levels. However, we believe the LMSS ECR source test results and their application to on-orbit conditions are conservative. Previously published results indicate that very little AO (on the order of 5 × 1020 AO·cm−2 ) is required to convert a silicone elastomer to a silicate [2]. By similarity, we have reasoned that a similar, minimal amount of AO would be required to eliminate silicone blocking. This hypothesis is also supported by recent laboratory work performed at NASA Glenn Research Center and reported elsewhere in this volume

298

PAUL BOEDER ET AL.

[3]. To explore this question, the Environment Effects Group at Marshall Space Flight Center, under direction from the ISS Program Office Environments Team, performed environmental exposure testing on samples of MD-944 diode tape.

3. Combined Effects Test Program Description The MD-944 diode tape Combined Effects Test Program is being conducted by the Environment Effects Group at the Marshall Space Flight Center [4]. The test program simulates exposure to the various components of the space environment, singly and in combination. The test program calls for samples of the MD-944 diode tape to be exposed to the following: 1. 2. 3. 4. 5.

5 eV atomic oxygen with concurrent vacuum UV radiation; Near UV radiation (NUV); Electron radiation (250 keV energy); Thermal energy atomic oxygen (<0.1 eV); 5 eV atomic oxygen, NUV radiation, and 250 keV electrons combined in series.

Test items 1–3 have been completed while items 4 and 5 are still in progress at the time this paper was prepared. A number of control samples were also prepared to serve as a baseline for testing results. All exposure testing involves exposing the silicone adhesive component of the diode tape to the environment constituent of interest. This was accomplished by mounting the diode tape on an aluminum substrate (block) so that the Kapton over-layer of the tape was bonded to the aluminum substrate. A pressure sensitive adhesive was used to form the Kapton to aluminum bond. Figure 2 shows the sample configuration for the exposure portion of the test. Figure 3 shows a representative tape sample prior to exposure. Test exposure levels are based upon our best estimate of environmental exposure levels on ISS at the time the test program was put in place (July 2003).

Figure 2. MD-944 diode tape sample exeposure configuration (not to scale)

EFFECTS OF SPACE ENVIRONMENT EXPOSURE

299

Figure 3. MD-944 diode tape representative sample, preexposure

The on-orbit exposure period covers the time span from P6 array deployment in December of 2000 to April 2003. The estimated total AO fluence incident on the P6 solar array port panel diode tape is ∼2 × 1021 AO·cm−2 . The total ultraviolet radiation exposure is estimated at >12,000 ESH. The total surface electron dose is estimated at 3.3 × 106 rads. The thermal cycle range for the deployed solar array is −174–+286◦ F. Ultraviolet and electron radiation exposure levels will have continued to increase between April 2003 and the present. However, only a minor increase (10– 20%) in the total AO fluence exposure is expected for the P6 solar array diode tape. This is because ISS has been flying in a minimum drag configuration in order to minimize propellant consumption following the Columbia accident. The minimum drag configuration places the solar array panels edge on to the velocity vector and minimizes the MD-944 diode tape exposure to AO. 3.1. 5eV ATOMIC OXYGEN EXPOSURE TEST

The MSFC Atomic Oxygen Beam Facility (AOBF) produces a flux of 5 eV atomic oxygen atoms of approximately 5 × 1015 atoms·cm−2 ·s−1 . AO fluence is determined from mass loss of Kapton, with corroboration to beam current, duty cycle, and time of exposure. During AO production, vacuum UV radiation is produced, primarily at 130 nm. The VUV dose is dependent on the time in chamber more than the fluence. A sample receiving 1 × 1021 atoms·cm−2 would get 820–900 equivalent sun hours of VUV. Samples of the diode tape with the silicone adhesive side up were exposed to AO fluences ranging from 1.3 × 1018 to 1.5 × 1021 atoms·cm−2 . 3.2. ULTRAVIOLET RADIATION EXPOSURE TEST

A Mercury-Xenon NUV lamp was used as the source of the NUV radiation. The samples were mounted in a vacuum chamber and illuminated from outside of the

300

PAUL BOEDER ET AL.

chamber through a port in the side of the chamber. The plate to which the sample holder was mounted was water cooled and the system operating temperature was 22–25◦ C. During the exposure, the chamber was evacuated to a pressure of 10−6 Torr or less. The lamp was characterized with a radiometer prior to sample exposure, and the output was monitored throughout the test. Samples of diode tape were exposed to 500 equivalent sun-hours (ESH), 1000 ESH, and 2200 ESH. 3.3. ELECTRON IONIZING RADIATION EXPOSURE TEST

The electron radiation energy and fluences for this test were based on radiation environment predictions from the MSFC Environments Group and dose-depth profiling for a 2-mil adhesive layer. The electrons were produced by a PelletronR accelerator capable of generating 200 keV to 2.5 MeV energy electrons. Samples of diode tape were exposed to 250 keV electrons to a fluence of 1.75 × 1013 , 3.90 × 1013 , and 1.15 × 1014 e-cm−2 . 250 keV electrons produce uniform dose-depth profiles as they are energetic enough to penetrate all the way through 2 mils of adhesive. 1.75 × 1013 e-cm−2 produces a calculated dose of 0.6 × 106 rads at a depth of 2 mils in the silicone adhesive but under-doses the front of the adhesive (0–2 mils). 3.9 × 1013 e-cm−2 produces a calculated dose of 1.3 × 106 rads at a depth of 1 mil in the silicone but under-doses the adhesive at lesser depths and overdoses the adhesive at greater depth. 1.15 × 1014 e-cm−2 produces a calculated dose of 3.3 × 106 rads for the surface of the adhesive but overdoses the adhesive at depth. 3.4. THERMAL ENERGY ATOMIC OXYGEN (< 0.1 EV) EXPOSURE TEST

The MSFC Atomic Oxygen Drift Tube Beam Facility produces a flux of <0.1 eV atomic oxygen atoms of approximately 2 × 1016 atoms·cm−2 ·s−1 . During AO production, a minimal amount of vacuum UV radiation is produced. This test had not been completed at the time this paper was prepared. Diode tape samples will be exposed to AO fluences comparable to the fluence levels used in the 5 eV AOBF tests. 3.5. COMBINED EXPOSURE TEST

The combined exposure test involves exposing samples to 5 eV AO, NUV radiation and 250 keV electron radiation in the order indicated. This test was still in progress at the time this paper was prepared. 3.6. THE BLOCKING TEST

The degradation of the MD-944 silicone adhesive as a function of environment exposure is evaluated by a blocking test. The blocking test measures the force (tensile loading) required to fail a bond formed by the silicone adhesive layer of

EFFECTS OF SPACE ENVIRONMENT EXPOSURE

301

Figure 4. Blocking force tensile load test configuration

the tape. After processing the samples, two substrates are pressed together using the exposed silicone adhesive layer of the MD-944 tape (silicone to silicone bond); and placed under 2.5 psi “preload” for a predetermined interval. The majority of the test samples were placed under preload for 7 days prior to performing blocking tests. Some of the combined effects test samples have been placed under longer duration preloads of 3, 6, and 10 months to simulate the effects of long term solar array storage. After preloading, the bond between the substrates is failed by a tensile loading of the bond to quantify the degradation of the silicone adhesive. Figure 4 shows the test set up for the blocking force tensile loading test. Figure 5 shows two samples bonded together and mounted in the tensile test machine prior to initiating a tensile test. 3.7. ANALYTICAL ANALYSES

A small number of samples were set aside so that x-ray photo-electron spectroscopy (XPS) and fourier transform infrared (FTIR) spectroscopy analyses could

Figure 5. Two samples mounted in the tensile test machine prior to test

302

PAUL BOEDER ET AL.

be performed. FTIR analyses were performed using the Horizontal Attenuated Total Reflectance (HATR) technique. The analysis of the tape samples was done by applying the silicone adhesive layer of the tape to the surface of a BBBB ATR crystal, installed in an FTIR spectrometer. One hundred scans, at 4-cm−1 resolution, were made to obtain the spectra reported herein. The spectra collected were evaluated against spectra of the silicone adhesive and the Kapton backing of the tape taken as reference spectra or control samples. The depth of penetration into the sample is typically one to two microns or less. Variables such as sample composition, angle of penetration into the sample, and amount of surface contact with the sample have a large influence on the depth of sample penetration and the internal reflectance “sampling beam.” XPS studies had not yet been completed at the time this paper was prepared. 4. Test Results and Discussion Test results are presented herein for NUV, energetic electron ionizing radiation, and 5 eV AO exposures. Test results are presented as a function of the tensile stress force (pounds per square inch—psi) required to separate two silicone samples bonded together after exposure to various levels of each environment constituent. All samples for which results are presented herein were placed under a nominal 2.5 psi preload for seven days prior to the tensile test being performed. Figure 6 140 7-day preload at room temperature & 2.5 psi Bonding area = 1" x 0.5"

Individual Data Points Mean

120

Random Variation of Mean (95%)

Stress at Max. Load (psi)

100 3 pts 3 pts 80

8 pts

60

40 3 pts 20

0 0

500

1000

1500

2000

NUV Exposure (ESH)

Figure 6. NUV radiation exposure tensile test results

2500

303

EFFECTS OF SPACE ENVIRONMENT EXPOSURE 80 Individual Data Points 70

8 pts

Mean Random Variation of Mean (95%)

3 pts

Stress at Max. Load (psi)

60 3 pts 50

6 pts

40

30

20

10

7-day preload at room temperature & 2.5 psi Bonding area = 1" x 0.5"

0 0.0E+00

2.0E+13

4.0E +13

6.0E +13

8.0E+13

1.0E+14

1.2E+14

250 KeV Electron Exposure (e/cm^2)

Figure 7. 250 keV electron exposure tensile test results

shows the change in tensile stress required to separate two samples exposed to various levels of NUV. Figure 7 shows the change in tensile stress as a function of energetic electron ionizing radiation exposure, while figure 8 shows similar data as a function of 5 eV AO fluence exposure levels. Figure 6 indicates a 67% mean reduction in blocking force after 2200 NUV equivalent sun hours (ESH). We do not know why there is such a large spread in the NUV exposure tensile test results shown in figure 6. Lamp calibration measurements made at the NUV sample fixture location showed a standard deviation of 0.03 for nine measurements. Any systematic errors with the preloading or tensile test phase of the test should also be apparent in the other data sets for electron radiation and AO. Figure 7 shows a 30% mean reduction in blocking force after exposure to a 250 keV electron fluence of 1.15 × 1014 electrons·cm−2 . We speculate that the reason why NUV produced a greater reduction in blocking force as compared to electron radiation is that the NUV radiation is concentrated near the surface of the sample while the electron radiation is spread throughout the material. Figure 8 shows a 100% reduction in blocking force after exposure to a 5 eV AO fluence of 1021 atoms·cm−2 . The samples exposed to this AO fluence did not exhibit any blocking force following a 7-day preload at 2.5 psi. Test results are also presented for selected FTIR analyses. FTIR analyses were performed on control samples as well as samples exposed to various levels of environment constituents. Figure 9 shows the absorbance profile of an unexposed

304

PAUL BOEDER ET AL.

80

3 pts

Individual Data Points Mean Random Variation of Mean (95%)

8 pts

70

Stress at Max. Load (psi)

60

7-day preload at room temperature & 2.5 psi Bonding area = 1" x 0.5"

50

40

3 pts

30

20

10

3 pts

0 0 1.E+17

1.E+18

1.E+19

3 pts

1.E+20

1.E+21

1.E+22

5 eV Atomic Oxygen Kapton Effective Fluence (ao/cm^2)

Figure 8. 5 eV atomic oxygen exposure tensile test results

control sample. Figure 10 shows the absorbance profile of a sample exposed 5 eV AO atoms to a Kapton effective fluence of ∼9.7 × 1019 atoms·cm−2 , while figure 11 shows the absorbance profile of a sample exposed to 5 eV AO atoms to a Kapton effective fluence of ∼1.5 × 1021 atoms/cm2 followed by 2200-ESH NUV. 2.8 2.6 2.4 2.2

Absorbance

2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 4000

3500

3000

2500

2000

1500

1000

Wavenumbers (cm-1)

Figure 9. Sample # 133 (ClC), MD-944 control sample. No environmental exposure

305

EFFECTS OF SPACE ENVIRONMENT EXPOSURE 0.9 0.8

Absorbance

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 4000

3500

3000

2500

2000

1500

1000

Wavenumbers (cm-1)

Figure 10. Sample # 95 (BF4a) exposed to 5 eV AO atoms to a fluence of ∼9.7 × 1019 atoms·cm−2

The absorbance peak at 1250 cm−1 in figures 9 and 10 indicates the presence of the silicone adhesive. This peak is significantly reduced in figure 11 indicating the silicone adhesive has been altered by the AO and near-UV exposure. None of the FTIR analyses detected a significant absorbance band at 1400 cm−1 , which 0.055 0.050 0.045

Absorbance

0.040 0.035 0.030 0.025 0.020 0.015 0.010 0.005 0.000 −0.005 3500

3000

2500

2000

1500

1000

Wavenumbers (cm-1)

Figure 11. Sample # 163 (BF1a) exposed to 5 eV AO atoms to a fluence of ∼1.5 × 1021 atoms·cm−2 followed by 2200-ESH near-UV

306

PAUL BOEDER ET AL.

Figure 12. Diode tape sample following exposure to a total 5 eV AO Kapton effective fluence of 1.48 × 1021 atoms·cm−2

would indicate the presence of a silicate or aliphatic silicone (Si-O-R). The broad absorbance band at 1110–1100 cm−1 in figures 9 through 11 also does not provide a clear indication of the conversion of the silicone adhesive to a silicate (SiOx ). We do not know the depth of the expected silicone/silicate interface so it is difficult to say whether we are looking at the AO effected layer or that we have penetrated into the bulk silicone material with our FTIR analysis. We expect planned XPS analysis to resolve any uncertainty in our FTIR analysis results. The peak at 3000 cm−1 in figures 9–11 is a carbon to hydrogen bond (C–H). Figures 12 shows a sample that has been exposed to 5 eV AO to a total Kapton effective fluence of 1.48 × 1021 atoms·cm−2 . This exposed sample may be compared to the unexposed sample shown in figure 3. No obvious surface features are apparent on the unexposed sample in figure 3, while the post AO exposure sample in figure 12 appears to have formed a glass-like surface coating with fine cracks. FTIR analyses of exposed samples do not show evidence of SiOx formation, but it is unclear whether the FTIR analysis is measuring the properties of the AO affected layer or of the bulk silicone material. XPS depth profile studies are planned to characterize the elemental composition of the diode tape materials before and after exposure to AO. As previously mentioned, the LMSS atomic oxygen blocking test results initially indicated the silicone blocking forces resulting from retraction and storage of the P6 solar array for an extended period of time may reach unacceptable levels. However, more recent analyses conducted at NASA Glenn Research Center to determine the silicone effective fluence correlation for the LMSS tests indicate that the Kapton effective fluence in the ECR facility needs to be 2.64 times higher than in low Earth orbit to replicate the equivalent damage induced in silicone [3] The silicone effective fluence correlation is based on changes in surface hardness of silicone, whereas Kapton effective fluence is based on mass loss. These new results imply that the LMSS atomic oxygen tests based on Kapton effective

EFFECTS OF SPACE ENVIRONMENT EXPOSURE

307

fluence were conservative, and therefore the LMSS tests are more consistent with the results reported here.

5. Conclusions Testing has been performed by the MSFC Space Environment Effects Group to characterize the effect of environmental exposure on the blocking force of silicone adhesive. Testing indicates that long-term exposure to each of the environment constituents (AO, NUV, and electron ionizing radiation) will reduce the adhesion properties of Dow Corning QC-7725 silicone adhesive. AO is the most efficient in this respect, while NUV and electron ionizing radiation will also reduce the adhesive properties of the silicone to a lesser extent. Contrary to expectations, FTIR analyses do not show any evidence of silicate formation following exposure to AO. Additional test plans are in place to 1. Perform XPS scans of control and exposed samples to determine elemental composition as a function of depth in the sample, 2. Complete combined effects exposure testing (5 eV AO, NUV, and electrons in series) followed by long term preloads of 3, 6, and 10 months in duration. 3. Complete sample exposures to low energy AO (<0.1 eV) in the MSFC Atomic Oxygen Drift Tube facility.

Acknowledgments The authors gratefully acknowledge the contributions of Ed Watts and Chuck Semmel of Qualis Corporation and Scott Taylor of MRC.

References 1. Imaging Science and Analysis Group. NASA-JSC, http://sx-isag.jsc.nasa.gov/, SX/. 2. Alred, J., Visentine, J., and Albyn, K. (2001) In AIAA 39th Aerospace Science Meeting, AIAA2001-0097, Reno, NV, January 2001. 3. de Groh, K. K., Banks, B. A., and Ma, D. (2006) in Proceedings of the 7th International Space Conference on Space Materials, Toronto, Canada, 2004, Springer, 401–416. 4. Finckenor, M. M., Edwards, D. L., Vaughn, J. A., Schneider, T. A., Hovater, M. A., and Hoppe, D. T. (2002) Test and Analysis Capabilities of the Space Environment Effects Team at Marshall Space Flight Center, NASA/TP-2002-212076, November 2002.

DRY SLIDING WEAR OF Ti-6Al-4V ALLOY AT LOW TEMPERATURE IN VACUUM

YONG LIU, DEZHUANG YANG, SHIYU HE, AND ZHUYU YE Harbin Institute of Technology, Harbin, 150001, P. R. China

Abstract. Tribological properties of Ti-6Al-4V alloy against GCr15 steel at low temperature in vacuum were measured. The worn surfaces of Ti-6Al-4V alloy specimens were examined by a Hitachi S-570 type scanning electron microscope. The debris were analyzed using a D/max-γB XRD spectrometer. The results showed that the wear rate of Ti-6Al-4V alloy in vacuum ranged within 1.3– 12.0 mg·km−1 at low temperature. The wear rate of Ti-6Al-4V alloy increased with increasing the sliding velocity at low temperature in vacuum. The surface temperature increased to about 32◦ C for 300 m of sliding distance. It was found that plough characteristics on the worn surface provide information of the adhesion wear. Under the low sliding velocity, microcracks formed on the worn surface indicating damage due to low temperature, while no such microcracks were formed under the higher sliding velocity. The XRD analysis showed that the debris contain Fe, implying that the steel disc was wearing off although the original hardness of the steel was higher than that of the Ti-6Al-4V specimen. Key words: titanium alloy, wear, vacuum, adhesion, low temperature

1. Introduction Titanium alloys have been widely used in aerospace industries because of their high strength-to-weight ratio and excellent corrosion resistance [1]. However, their poor wear resistance prevents the applications of titanium alloy as structural materials [2]. The poor tribological properties of titanium alloy can be attributed to low resistance to plastic shearing and low protection exerted by the surface oxide [3]. For these reasons, many surface modification treatments are under continuous development to improve their tribological behavior. Various surface treatments have been widely studied such as, anodizing and MoS2 codeposition [3], ion implantation [4], plasma nitriding [5], laser nitriding alloying [6–9], diamond coating [10], plasma spray [11], and physical vapor deposition [12]. However, in order to optimize the surface treatment it is necessary to understand the 309 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 309–316. C 2006 Springer. Printed in the Netherlands.

310

YONG LIU ET AL.

mechanisms of failure of the untreated alloy under certain conditions as well as the metallurgical phenomena that have to be counteracted by the modified surface microstructure. On the other hand, the low temperature vacuum environment must be considered in space applications where wear resistance is required. The absorption properties, stains, and oxides may be totally or partially removed from the solid surface in low temperature vacuum [13]. Thus, a relatively clean surface can be formed, which might result in increased adhesion or even cold welding due to the strong attractive forces between molecules of the contacting surfaces [14]. This may lead to faulty spread of solar arrays with their pins made of Ti-6Al-4V alloy, a decrease in the life of bearings, and poor contact of electric brushes [15]. Little systematic study has been carried out on Ti alloy in this field. The aim of this study was to explore the dry sliding behavior of Ti-6Al-4V alloy at low temperature in vacuum.

2. Experimental The material used in this study was Ti-6Al-4V alloy (chemical composition in weight percent, 5.5–6.5% Al, 3.5–4.5% V, 0.03% Fe, 0.15% O, balance Ti). The material received in the form of hot rolled rod was annealed at 760◦ C for 60 minutes and cooled in the furnace to obtain an average hardness of 345 HV (35 HRC). The counter material was steel GCr15 (around 1% carbon, 1.5% chromium in weight percent) quenched in water at 840◦ C and then followed by tempering at 150◦ C to have the hardness of 62 HRC. Wear tests were carried out using a pin-on-disc scheme, in which the wear couple was composed of a Ti-6Al-4V alloy pin and a steel disc. The pin was cylindrical in shape, 9 mm in diameter, and 20 mm long, with a spherical end with a radius of 40 mm. The pin was sliding against the steel disc (70 mm in diameter and 10 mm thick). The surfaces of the pin and the disc were both grinded to obtain roughness Ra of about 0.42 and 3.20 μm, respectively. The wear tests were performed at loads of 30 N and within a sliding velocity range of 0.2–1.2 m·s−1 . The sliding distance was 1000 m. The wear results were characterized by the mass loss of the pin using an electronic balance with the precision of ±10−5 g. 3. Results and Discussion The wear teats were carried out under the environmental pressure of 10–5 Pa using a UTI TV 100 type apparatus for vacuum friction and wear (made in Ukraine). The worn surface was prepared for SEM observation by ultrasonic washing in acetone and drying. A Hitachi S-570 type scanning electron microscope was used.

311

DRY SLIDING WEAR OF Ti-6Al-4V ALLOY TABLE 1. Friction coefficient of Ti-6Al-4V under 30 N and various sliding velocities at low temperature in vacuum Sliding velocity, m·s−1

0.2

0.4

0.6

0.8

1.2

Friction coefficient at low temperature Friction coefficient at ambient temperature

0.45 0.75

0.46 0.81

0.33 0.76

0.56 0.68

0.48 0.84

3.1. FRICTION COEFFICIENT

Table 1 lists friction coefficient for Ti-6Al-4V alloy against GCr15 steel under 30 N and different sliding velocities in vacuum. The friction coefficient at low temperature ranged from 0.33 to 0.56, which is lower than that at ambient temperature under the same friction condition (within 0.68–0.84). The measurement results indicate that the friction coefficients of Ti-6Al-4V against steel at ambient temperature and low temperature in vacuum show a nonlinear variation with sliding velocity, implying the complexity of frictional contact and difficulty in describing the friction using an uniform process. In addition, the friction coefficient at low temperature is obviously lower than that at ambient temperature, suggesting that the decrease in temperature will reduce the friction coefficient for Ti-6Al-4V. 3.2. WEAR RATE

Figure 1 shows the variation of wear rate of Ti-6Al-4V with different sliding velocities under 30 N at 203 and 300 K in vacuum. The wear rate of Ti-6Al-4V

Figure 1. Wear rate of Ti-6Al-4V alloy as a function of the sliding velocities under 30 N at low temperature in vacuum

312

YONG LIU ET AL. (a) cracks

(b) edges

(c) adhesion

(d) lamellar tongues

Figure 2. SEM micrographs showing morphologies of the worn surfaces of Ti-6Al-4V alloy under 0.2 m·s−1 and 30 N in vacuum at low temperature

is increased with increasing sliding velocity both at low and high temperature. At ambient temperature, wear rate shows a near-linear variation with sliding velocity, while it is increased nonlinearly with sliding velocity at low temperature. The rate of wear at low temperature is different within the different sliding velocity ranges, i.e., it is slow between 0.2 and 0.6 m·s−1 , fast within 0.6–0.8 m·s−1 , and little changed between 0.8 and 1.2 m·s−1 . The calculation based on figure 1 indicates that the wear rate is higher at ambient temperature than that at low temperature. Figure 2 shows the surface morphologies of Ti-6Al-4V worn at low temperature in vacuum. The main characteristics of the worn surface are the plough grooves and sliding traces that are similar to those at ambient temperature in vacuum. However, there are differences in characteristics of surfaces worn at lower sliding velocities under low temperature and vacuum condition comparing

313

DRY SLIDING WEAR OF Ti-6Al-4V ALLOY TABLE 2. Average microhardness of the worn surfaces of Ti-6Al-4V under 30 N and various sliding velocities in vacuum at low temperature (HV) Sliding velocity, m·s−1 Average microhardness

0.2 650

0.4 648

0.6 652

0.8 655

1.2 689

to these in high temperature vacuum environment. Cracks and black foreign adhered residue can be found in figure 2(a). Figure 2(b) shows the edge morphology due to plastic deformation caused by squeezing of surface during friction. Some adhesion traces and layered steps caused by squeezing deformation can be observed in figures 2(c) and 2(d), respectively. The analysis of figure 2 implies that the worn surface of Ti-6Al-4V after wear at lower sliding velocities under low temperature vacuum condition is characterized by plastic deformation, fracture, and adhesion. No cracks can be observed on the surface worn at higher sliding velocities comparing to that at lower velocities. Adhesion traces and layered steps also can be found on the surface worn at 1.2 m·s−1 under low temperature vacuum condition. It is proposed that the worn surface of Ti-6Al-4V after wear at higher sliding velocities under low temperature vacuum condition is characterized by plastic deformation and adhesion. Table 2 lists the average microhardness of Ti-6Al-4V surface worn under 30 N and different sliding velocities at low temperature in vacuum. The average microhardness of the surface worn at low temperature is higher than that of the original surface (345 HV). Figure 3 shows XRD spectrum for Ti-6Al-4V wear under 0.2 m·s−1 and 30 N at low temperature in vacuum. The worn surface layer of Ti-6Al-4V contains α-Ti and α-Fe phase. It implies that wear took place on

Figure 3. XRD spectrum for the worn surface of Ti-6Al-4V alloy in vacuum at low temperature

314

YONG LIU ET AL.

Figure 4. Temperature of the worn surface of Ti-6Al-4V alloy specimens as a function of the sliding time under 30 N and various sliding velocities in vacuum at low temperature

GCr15 steel during friction. The debris of GCr15 steel will remain on Ti-6Al4V surface. The existence of α-Fe in the worn surface layer of Ti-6Al-4V will result in increase in microhardness. In addition, the hardening caused by plastic deformation during friction also affects the increase in hardness. Figure 4 shows the variation of surface temperature of Ti-6Al-4V with time during sliding at different sliding velocities at low temperature in vacuum. The surface temperature is increased linearly with increasing time. Increase in sliding velocity will lead to an increase in heating rate of the surface. For example, the surface temperature under 1.2 m·s−1 rises from original 203 to 235 K only after 3 min of sliding. The above analysis on morphology and XRD of the worn surface of Ti-6Al-4V against GCr15 steel shows that, the wear of Ti-6Al-4V at low temperature in vacuum is characterized by typical adhesion and plough with material transfer. This implies that the friction process is mainly carried out between Ti-6Al-4V and the transfer material on GCr15 steel. XRD spectrum indicates that there is mutual material transfer between Ti-6Al-4V and GCr15 steel. At low temperature in vacuum, the surface hardness of Ti-6Al-4V is increased leading to the wear of GCr15 steel. The formed Fe debris embedded in Ti-6Al-4V surface results in an increase in hardness of Ti-6Al-4V surface. Comparing the wear results at ambient temperature in vacuum, the difference in contacting condition between sliding surfaces will change the trend of variation in wear rate with sliding velocity, leading to a nonlinear increase of wear rate with increasing sliding velocity at low temperature in vacuum. With the surface temperature increasing with increasing sliding velocity [9], the extent of temperature rise under lower velocities is lower than that under higher velocities. On the other hand, the brittleness of Ti-6Al-4V is decreased with

DRY SLIDING WEAR OF Ti-6Al-4V ALLOY

315

decreasing temperature. Thus, cracks are easily formed at lower sliding velocities. Increase in velocity will reduce the trend of fracture with fewer cracks existing on the worn surface.

4. Conclusion 1. The friction coefficient for Ti-6Al-4V alloy against GCr15 steel in vacuum ranged within 0.68–0.84 at ambient temperature and 0.35–0.55 at low temperature. 2. The wear rate of Ti-6Al-4V alloy increased with increasing the sliding velocity in vacuum, with the trend being higher at ambient temperature than that at low temperature. 3. The surface of Ti-6Al-4V alloy worn at low temperature in vacuum was characterized by increased adhesion. Under the low sliding velocity, microcracks formed on the worn surface indicating of fracture due to low temperature, while no such microcracks were observed under the higher sliding velocity. 4. The XRD analysis showed that the debris contain Fe, implying that the wear took place for the steel disc although the original hardness of the steel was higher than that of the Ti-6Al-4V specimen. 5. The surface temperature increased linearly with increasing time. A higher sliding velocity would create an increased rate in temperature rise.

Acknowledgment The authors are grateful to professor G. D. Gumulya for his interest in and encouragement of this work. This research is supported by National Basis Research Foundation of China.

References 1. Wan, J. Y., Ge, A. M., and Zhou, Y. B. (1987) Titanium Alloys in the Aeronautic Application, Science and Techology Press, Shanghai, China, pp. 5–45. 2. Budinski, K. G. (1991) Wear 151, 203–217. 3. Molinari, A., Straffelini, G., Tesi, B., and Bacci, T. (1997) Wear 208, 105–112. 4. Weisheit, A. and Mordike, B. L. (1995) Surface Engineering 11(2), 160–161. 5. Ignatiev, M., Kovalev, E., Melekhin, I., Smurov, I. Y., and Sturlese, S. (1993) Wear 166, 233–236. 6. Yerramsreddy, S. and Bahadur, S. (1992) Wear 157, 245–262. 7. Ji, H. B., Xia, L. F., Ma, X. X., Sun, Y., and Sun, M. (1999) Tribology International 32, 265. 8. Wilson, A., Matthews, A., Housden, J., Turne, R., and Garside, B. (1993) Surface and Coatings Technology 62, 600–602.

316

YONG LIU ET AL.

9. Herr, W., Matthes, B., Broszeit, E., Meyer, M., and Suchentrunk, R. (1993) Surface and Coatings Technology 57, 77–80. 10. Bloyce, A., Qi, P.-Y., Dong, H., and Bell, T. (1998) Surface and Coatings Technology 107, 125–132. 11. Dong, H. and Li, X. Y. (2000) Materials Science and Engineering A280, 303–310. 12. Kragelski, I. V., Lyubarski, I. M., and Guslyakov, M. I. (1973) Friction and Wear in Vacuum, Mashinostroenie, Moscow. 13. Vlandman, U., Luedtke, W. D., and Ringer, M. (1992) Wear 153(1), 3–30. 14. Miyoshi, K., and Pepper, S. V. (1992) Properties Data for Opening Galileos Partially Unfurled Antenna, NASA, Cleveland, OH, NASA-TM-105355. 15. Sliney, H. E., Lukaszewicz, V., and DellaCorte, C. (1994) The Tribology of PS212 Coatings and PM212 Composites for the Lubrication of Titanium 6Al-4V Components of a Stirling Engine Space Power System, NASA-TM-106462, NASA, Cleveland, OH.

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN: DEPENDENCE ON O-ATOM FLUENCE AND SURFACE TEMPERATURE

DEANNA M. BUCZALA AND TIMOTHY K. MINTON Department of Chemistry and Biochemistry, 108 Gaines Hall, Montana State University, Bozeman, MT 59717

Abstract. Organic polymers are susceptible to erosion from reaction with ambient atomic oxygen in low Earth orbit. We have investigated the linearity of the O-atom fluence dependence of Kapton H erosion and the dependence of Kapton H erosion yield on surface temperature. Sample exposures were performed with a pulsed beam containing hyperthermal O atoms that were generated with a laser detonation source. After exposure, samples were removed from the chamber in which the exposures were done, and postexposure analyses were performed: etch depth (profilometry) and surface topography (atomic force microscopy). A systematic set of exposures, which eroded room-temperature Kapton H from 1.4 to 25 μm, showed that the erosion yield of Kapton H is linearly dependent on O-atom fluence. This result helps validate the use of Kapton H mass loss (or erosion depth) as a linear measure of the O-atom fluence of a materials exposure. The erosion of Kapton H was strongly temperature dependent. At lower temperatures (<100◦ C), the erosion yield appeared to be independent of sample temperature. But above 100◦ C, the erosion yield exhibited an Arrhenius-like temperature dependence, with an apparent activation energy of 0.31 eV. These observations suggest that O-atom-induced erosion of Kapton H proceeds through direct, nonthermal, gassurface reactions, and through reactions that depend on the surface temperature. Key words: atomic oxygen, Kapton, erosion, hyperthermal beam

1. Introduction In the low-Earth orbital environment, spacecraft materials are subject to erosion and oxidation from ambient atomic oxygen. The high relative velocity between a spacecraft and the ambient O atoms results in O-atom reactivity that may be orders of magnitude higher than that of thermal (300 K) O atoms. Organic polymers are especially susceptible to erosion from reactions with O atoms in low Earth orbit, yet they are commonly used on spacecraft as thermal control and structural 317 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 317–329. C 2006 Springer. Printed in the Netherlands.

318

DEANNA M. BUCZALA AND TIMOTHY K. MINTON

materials. The phenomenology of polymer erosion has been investigated extensively in space- and ground-based exposure environments [1–3]. A Kapton H standard witness sample is commonly used to identify the effective O-atom fluence of an exposure [3, 4]. Kapton H is assumed to erode linearly with O-atom fluence and, within the accuracy of this assumption, this material serves as a convenient reference material for determining the fluence of a test. The detailed mechanisms of polymer erosion resulting from reactions of hyperthermal atomic oxygen are still poorly understood, as controlled O-atom environments are difficult to achieve and the mechanisms are complex. Even while research into the mechanisms by which organic polymers degrade under hyperthermal O-atom attack continues to reveal the rich chemistry that occurs in the extreme environment of low Earth orbit [3, 5–7], new insight into the erosion mechanisms may nevertheless be gained from erosion phenomenology. The dependence of the erosion yield on sample temperature can provide general information about the erosion mechanisms. If the rate limiting processes occur in thermal equilibrium with the surface, then a temperature-dependent study of the erosion rate will yield an effective activation energy for the rate-limiting processes. On the other hand, if the rate-limiting processes do not occur in thermal equilibrium with the surface, then these processes would not be expected to depend on surface temperature, but rather on the collision energy between incident O atoms and the surface. Koontz et al. [8] have studied the temperature dependence of the erosion of Kapton H when this material was exposed to different thermal O-containing environments, and an Arrhenius form [Re = A exp(E a /RT )] was used to describe the dependence of Kapton erosion yield, Re , on the surface temperature, T . Activation energies, E a , of less than 0.29 eV were obtained. However, activation energies derived from temperature-dependent sample exposures in low Earth orbit are considerably smaller [9]. In fact, no temperature dependence was observed for Kapton erosion in low Earth orbit [10–12]. Tagawa and coworkers [13] recently reported a very careful study of the temperature-dependent reactivity of a Kapton-like polyimide upon exposure to two different O-atom beams, one containing O atoms with an average translational energy of 5.0 eV and the other containing O atoms with an average translational energy of 1.1 eV. By varying the sample temperature from 253 to 353 K and measuring mass loss in situ with a quartz crystal microbalance, they derived small activation energies of 5.7 × 10−4 eV for the 5.0 eV exposure and 4.5 × 10−2 eV for the 1.1 eV exposure. The low or nonexistent activation energies (or minimal temperature dependence) observed for hyperthermal O-atom exposures and the substantial activation energies observed for thermal O-atom exposures suggest that high-energy O-atom-surface collisions, characteristic of those occurring on spacecraft surfaces at LEO altitudes, overcome any reaction barriers that would limit the erosion rate when oxygen atoms bombard a surface at low incident energies. As the erosion yield of a polymer under hyperthermal O-atom attack does not appear to depend on

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN

319

surface temperature, it may be concluded that the erosion yield is determined by gas–surface reactions that do not occur in thermal equilibrium with the surface. The experiments on the temperature dependence of erosion have thus far not systematically probed the temperature dependence of polymer erosion at temperatures above the highest temperature used by Tagawa and coworkers (i.e., 353 K). And the studies by Tagawa and coworkers utilized relatively low fluence exposures (<1018 O atoms·cm−2 ), so the surfaces were fairly smooth during the exposures. After a significant fluence of directed hyperthermal O atoms (>1019 O atoms·cm−2 ), polymer surfaces become very rough [3]. Multiple collisions on such rough surfaces can drive incident O atoms toward thermal equilibrium with the surface. Therefore, reactions in thermal equilibrium with the surface might be more important for higher fluence exposures that result in many micrometers of surface erosion. We have conducted a study of the dependence of the erosion yield of Kapton H on surface temperature over the temperature range 298–573 K. The surface roughness of the eroded samples was characterized by atomic force microscopy. Accompanying every exposure for this study and for other (unrelated) studies in our laboratory was a Kapton H reference specimen that was held at 296 K. Analysis of the erosion depth of all these reference samples as a function of exposure duration has provided a measure of the linearity of the dependence of the Kapton H erosion yield on O-atom fluence. 2. Experimental Details The sample exposures were conducted in the source region of a molecular beam apparatus [3, 5]. A schematic diagram of the experimental arrangement is shown in figure 1. The pulsed, hyperthermal, O-containing beam was produced with a laser detonation source that is similar to the original laser-detonation source SOURCE PULSED

NOZZLE

SAMPLE MOUNT

CHAMBER

VALVE

COLLIMATING APERTURE

MIRROR SKIMMER

IONIZER

QUADRUPOLE MASS FILTER

CO2 LASER MAIN SCATTERING CHAMBER

Figure 1. Schematic diagram of the beam source, sample mount location, main scattering chamber, and rotatable mass spectrometer detector

320

DEANNA M. BUCZALA AND TIMOTHY K. MINTON

Figure 2. Sample mount and sample position numbering scheme

designed by Physical Sciences Inc. [14]. The source produces neutral beams (ion content 1%) containing 60–70% O atoms and 30–40% O2 molecules traveling at velocities in the vicinity of 8 km·s−1 . The average O-atom translational energy used in the experiments reported here was approximately 5 eV, and the energy width (full width at half maximum) was approximately 2.5 eV. Characterization of the beam has been done to determine that the hyperthermal O atoms are in their ground electronic state, O(3P) [15]. In another similar study, we have concluded that the O2 molecules in the beam are in the ground O2 (3 g− ) state [16]. These results establish that the exposure environment used in the laboratory subjects materials to ground-state O and O2 , as do space exposures. The samples were exposed to the hyperthermal beam at a distance of 40 cm from the apex of the conical nozzle source. A small portion of the O-atom beam passed through several apertures and regions of differential pumping before entering a mass spectrometer, which was used to characterize the ratio of O to O2 in the beam and the velocity distributions of these species. The Kapton-equivalent O-atom fluence of an exposure was determined by measuring the erosion depth of a Kapton H reference sample that is assumed to have an erosion yield of 3.00 × 10−24 cm3 ·O-atom−1 . The Kapton-equivalent fluence measurements led to the conclusion that the average O-atom flux of one beam pulse at the samples was 1.67 × 1015 O atoms·cm−2 ·pulse−1 . The samples were placed in a sample mount (see figure 2) such that the beam impingement angle was a few degrees off normal. The entire mount was exposed to the beam at a repetition rate of 2 pulses per second. Kapton H samples 1.22 cm diameter and 125-μm thick were placed in the sample mount, with a stainless steel wire mesh placed over them. The mesh allowed the erosion depth between

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN

321

the exposed and unexposed areas of the samples to be measured. The sample mount contains nine locations for samples. The Kapton H reference sample was placed in position 5. The reference sample was thermally isolated from the rest of the sample mount and was maintained at 296 K. The temperature of the samples at the other eight positions can be elevated from room temperature to 573 K by resistive heaters embedded in the sample mount. The specific temperatures used were 298, 333, 373, 423, 498, and 573 K. Kapton H samples were always placed in positions 4 and 7 of the sample mount. In order to ensure uniform temperature of these thermally insulating samples, thin films of Kapton H were bonded to silicon substrates with a thermally conductive, silver-filled polyimide adhesive (Ablebond 70–1). This adhesive was cured in vacuum, with a maximum curing temperature of 573 K. When mounted in the sample mount, the back sides of the silicon substrates were in intimate contact with metal surfaces that were at the equilibrium temperature of the heated mount. All samples used for studying the temperature dependence of the erosion yield were exposed to 100,000 pulses of the hyperthermal beam. Many additional exposures (for other purposes) were carried out over a two-year period, and the Kapton H reference samples for all the exposures during this period were used to study the dependence of Kapton H erosion depth on exposure duration. After exposure, samples were removed from the chamber and examined by profilometry and atomic force microscopy (AFM). The wire mesh allowed the measurement of many steps on exposed samples with the use of a Dektak-3 (Veeco Metrology Group) surface profiler. The average erosion depth and corresponding standard deviation for a given sample were calculated from measurements of 40–45 different step heights. It has previously been shown that the determination of the erosion yield of Kapton may be accomplished with equivalent results by mass loss with a microbalance or by step height measurements with a profilometer [17]. AFM measurements of surface roughness of the exposed regions of sample surfaces were obtained with a nanoscope IIIa tapping mode atomic force microscope (Digital Instruments).

3. Results and Analysis The erosion depth of the Kapton H reference samples is plotted in figure 3 as a function of exposure duration, which varied from 28,000 (∼4 h) to 250,000 pulses (∼35 h). These exposure durations correspond to Kapton-equivalent O-atom fluences of 3.68 × 1019 and 4.11 × 1020 atoms·cm−2 . These data were collected over a two-year period, totalling 2,800,227 beam pulses and approximately 389 h of exposures. There is variation in the per-pulse flux from exposure to exposure, but within the fluence uncertainties of the various exposures, the erosion depth appears to increase linearly with the O-atom fluence (exposure duration). The linear regression fit to the data goes to zero at a fluence greater than zero, which is

322

DEANNA M. BUCZALA AND TIMOTHY K. MINTON 14

Erosion Depth/µm

12 10 8 6 4 2 0 0

50

100

150

200

250

300

Exposure Duration/1000 beam pulses

Figure 3. Erosion of Kapton H as a function of exposure duration

consistent with earlier reports of an induction period before erosion commences [18]. However, the uncertainty in the fit would allow for the line to pass through the origin. The erosion depths of Kapton H are plotted as a function of sample temperature in figure 4. Clearly, the erosion depth has a strong temperature dependence, especially at temperatures greater than 400 K. The erosion depths of the 20 18

Erosion Depth/µm

16 14 12 10 8 6 4 2 250

300

350

400

450

Temperature/Kelvin

500

550

600

Figure 4. Erosion depth of Kapton H as a function of temperature. The r points represent sample position #7, the ◦ points represent sample position #4, and the points represent the Kapton H reference sample (maintained at 296 K)

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN

323

Ratio of Kapton/Kapton Ref

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5

300

400

500

600

Temperature/Kelvin

Figure 5. Kapton erosion ratio as a function of temperature. The points r represent sample position #7, and the ◦ points represent sample position #4. The dashed curve is the best fit through all the points with a two-term Arrhenius-like function (see text)

variable-temperature samples were divided by the erosion depths of the respective Kapton H reference samples, and the ratios are plotted as a function of temperature in figure 5. These data were fit very well by an Arrhenius-like equation that includes a constant term: erosion ratio = A exp(−E a /RT ) + C. The values of the constants, A, E a , and C, are 1137, 0.31 eV, and 0.9114, respectively. The constant, C, must be added to the Arrhenius term, because the erosion ratio tends toward a constant value of approximately one at very low temperatures. AFM images of unexposed and exposed Kapton H surfaces are shown in figure 6. All images were collected with the same area, 2.5 μm × 2.5 μm, and they are presented with a z-scale of 500 nm. Root-mean-square (rms) surface roughnesses were calculated with the Nanoscope IIIa software and are presented in figure 7. Before exposure, the Kapton samples had smooth surfaces, with rms surface roughnesses of approximately 7 nm. The exposed surfaces exhibited the familiar “shag-carpet” topography [3]. Surfaces that were eroded at the lowest temperature exhibited the largest roughness, and as the sample temperature was increased, the roughness decreased significantly, reaching a minimum and then increased again. The roughness measurements are consistent with the visual changes in surface topography (see figure 6). 4. Discussion Kapton H is often used as a reference standard to determine the (Kaptonequivalent) O-atom fluence of an exposure. This fluence is interpreted as being

324

DEANNA M. BUCZALA AND TIMOTHY K. MINTON

Figure 6. AFM images of unexposed (A) and exposed (B-G) Kapton H. All images are 2.5 × 2.5 mm and have a z scale of 500 nm

equivalent to a dose of normal–incidence, 5 eV O(3P) atoms that would erode a Kapton H surface a specified amount (i.e., the erosion depth of the Kapton H witness sample of an exposure). Although rarely discussed, the flux of the Kapton equivalent dose must be assumed to be low enough to avoid effects from gasphase collisions near the surface. A safe upper limit to avoid such effects would be a continuous exposure at a flux of 1017 O atoms cm−2 ·s−1 or pulsed exposure (of ∼10 μs duration) where the peak flux is less than approximately 1021 atoms 70

Roughness/nm

60 50 40 30 20 10

unexposed sample

0 0

298

333

373

423

493

573

Temperature/Kelvin

Figure 7. Surface roughness as a function of sample exposure temperature

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN

325

cm−2 ·s−1 [19]. To calculate the Kapton-equivalent O-atom fluence, the value of 3.00 × 10−24 cm3 ·O-atom−1 is taken to be the erosion yield of Kapton H when it is eroded by 5 eV O(3P) atoms at normal incidence. Note that Kapton-equivalent fluences are often reported for a variety of exposure environments, including environments that contain O atoms with translational energies much less than 5 eV and environments that are not well characterized, such as oxygen or air plasmas. Presumably, the validity of the reported Kapton-equivalent fluence improves as the exposure environment approaches that of a continuous stream of 5 eV O(3P) striking a surface at normal incidence, with fluxes not greater than those given above. The beam source used in the studies reported here contains mostly O(3P) atoms, and the impingement angle on the surface is near normal. But there are also O2 molecules present in the beam, and the beam is pulsed, not continuous. Assuming a pulse width at the sample surface of approximately 10 μs and a per-pulse flux of 1.67 × 1015 O atoms·cm−2 ·pulse−1 , the peak flux is approximately 1.67 × 1020 O atoms·cm−2 ·pulse−1 . Despite this high peak flux, the number of atoms/molecules in the pulse is not sufficiently high that gas-phase collision effects are very important above the sample surface [19]. Therefore, the O2 in the beam is expected to be the most significant deviation from the ideal exposure situation, as the O2 could be involved in collisional processes [6, 7] and in reactions with radical sites generated at the surface [3]. In spite of the complication of O2 in the hyperthermal beam, the main uncertainty in the study of the dependence of Kapton H erosion yield on fluence likely comes from the variation in the beam flux for exposures conducted on different days. The data in figure 3 provide a view of this variation, particularly in the erosion depths measured for exposure durations of 100,000 beam pulses. Even though we attempted to operate the beam reproducibly, the flux apparently varied by roughly ±10%. This level of reproducibility of the beam flux represents the current state of the art for our source. As we did not have an independent measure of the flux for the different exposures, the relative fluence for the data reported in figure 3 is taken to be given by the number of beam pulses. Within the uncertainty in the fluence, our data show that the erosion yield of Kapton H varies linearly with fluence. While it is possible that the erosion yield may be slightly nonlinear with fluence, this nonlinearity must be small. Therefore, our data support the assumption that the erosion yield of Kapton H varies linearly with 5 eV O(3P) fluence. This result further validates the use of Kapton H mass loss or erosion depth as a linear measure of the O-atom fluence of a materials exposure. The erosion depth of Kapton H showed a significant dependence on surface temperature. As the surface temperature was increased from room temperature to 573 K, the erosion depth increased by a factor of 3.3. This observation may seem to contradict the result of Tagawa and coworkers [13] which showed that the erosion yield of a Kapton-like polymer was almost independent of temperature. However,

326

DEANNA M. BUCZALA AND TIMOTHY K. MINTON

they studied the mass-loss rate over a narrower temperature range. They concluded that the erosion rate of the polyimide had an Arrhenius temperature dependence, with an activation energy of 5.7 × 10−4 eV for the temperature range 253–353 K. This measurement employed a quartz crystal microbalance, and mass-loss data were obtained during the initial erosion period of the polymer. We also observed a negligible temperature dependence in the similar temperature range 298–373 K. In fact, our results agree with those of Tagawa and coworkers in this temperature range, within the uncertainty of our data. The obvious temperature dependence that we observed became apparent at temperatures above those used by Tagawa and coworkers. The two-term fit to the data in figure 5 suggests that the loss of material from a Kapton H surface is the result of temperature-dependent and temperatureindependent mechanisms. The constant term, C, results from a mechanism which is temperature-independent. This mechanism likely involves the direct reaction of O atoms with the surface on a time scale too short for thermal equilibrium to be achieved. Direct reactions of 5 eV O atoms with the surface should be able to overcome even significant reaction barriers, so reaction would proceed independently of surface temperature and yield an apparent activation energy of zero. For example, hydrogen abstraction has a barrier of aaproximately 0.25 eV [20], while hydrogen elimination and C–C bond-breaking reactions have barriers near 2 eV [21, 22]. Any of these reactions may occur, given the available collision energy. Our results do not permit a conclusion about the nature of the temperature-independent mechanism. If the initial reactions of hyperthermal O atoms with a hydrocarbon polymer are similar to those that occur with gas-phase hydrocarbons [15, 20–22], then perhaps all these reactions, as well as others, are involved. A temperature-dependent reaction process becomes important as the sample temperature increases. An Arrhenius functional form describes well the temperature dependence of Kapton H erosion depth at the higher temperatures used. This dependence suggests a mechanism that may take place in thermal equilibrium with the surface, implying that in order for this mechanism to occur, O atoms must transfer their energy to the surface and become trapped (either through physisorption or through reaction) before the rate limiting erosion process(es) proceed(s). The trapping probability increases with surface roughness, which allows for multiple bounces at the surface that drive incident atoms toward thermal equilibrium and increase the likelihood that a reaction will occur. In the studies reported here, and in any situation where macroscopic amounts of material are removed by highly directional O atoms, the surface reaches a steady-state roughness. Therefore, the conclusion of a mechanism that depends on thermal accommodation (or trapping) of incident O atoms is consistent with a steady-state erosion of rough surfaces. Trapping would be expected to be less probable on smooth surfaces, possibly

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN

327

reducing the importance of the thermal mechanism and thus the temperature dependence of erosion. The temperature dependence of the Kapton H erosion yield is undoubtedly linked to the dependence of surface roughness on temperature, but the relationship is not clear. At lower temperatures, where the temperature-independent mechanism dominates, a very rough surface is produced. But as the temperature is increased and the temperature-dependent mechanism starts to dominate, the eroded surface does not become as rough. The thermal mechanism retains no memory of the incident beam direction, so O atoms that become trapped may react with equal probability anywhere on the surface. In addition, the reactivity of trapped atoms increases at higher temperatures. Both these factors may contribute to the smoothing of the surface. But why the roughness reaches a minimum and begins to rise again with temperature is difficult to explain. The ultimate dependence of surface roughness on temperature represents a complicated interplay between the temperature-dependent and temperature-independent mechanisms.

5. Conclusion The erosion depth of Kapton H as a function of exposure duration in a hyperthermal O-atom beam was investigated for a broad range of exposure durations covering almost two orders of magnitude. Within the reproducibility of the exposure flux provided by our laser-detonation source, the erosion depth of Kapton H was found to be linearly dependent on O-atom fluence. This result is consistent with use of Kapton H as a linear erosion standard for materials testing in atomic oxygen environments. The dependence of Kapton H erosion by hyperthermal atomic oxygen on surface temperature was studied by directing a hyperthermal O-containing beam at Kapton H samples held at various temperatures and then conducting postexposure analyses of these samples by profilometry and by atomic force microscopy. Although the detailed erosion mechanisms are not known, it appears that a temperature-independent process dominates the erosion at temperatures lower than approximately 400 K and that both this temperature-independent process and an additional temperature-dependent process contribute to the erosion at temperatures above 400 K. At the highest temperature used (573 K), the erosion yield resulting from the temperature-dependent process was 3.3 times that resulting from the temperature-independent process. The temperature-independent process is believed to involve direct reactions of hyperthermal O atoms with the surface, while the temperature-dependent process is believed to proceed after incident O atoms have become trapped on the surface, perhaps through physisorption or reaction. The activation energy for the temperature-dependent process was

328

DEANNA M. BUCZALA AND TIMOTHY K. MINTON

determined to be 0.31 eV, which is similar to the barrier for an H-atom abstraction reaction.

Ackowledgments This research was supported by AFOSR MURI Grant No. F49620-01-1-0335 and by a DEPSCoR Grant administered by AFOSR (No. F49620-01-1-0212). The authors wish to thank Prof. John Tully (Yale University) for helpful discussions.

References 1. Tennyson, R. C. (1991) Canadian Journal of Physics 69, 1190. 2. In L. A. Teichman and B. A. Stein (eds.), NASA/SDIO Space Environmental Effects on Materials Workshop, Hampton, VA, 28 June–1 July 1988, NASA conference Pulication 3035 (NASA, Washington, D.C., 1989). 3. Minton, T. K. and Garton, D. J. (2001) In R. A. Dressler (ed.), Chemical Dynamics in Extreme Environment, World Scientific, Singapore, pp. 420–489. 4. Minton, T. K. (1995) Protocol for Atomic Oxygen Testing of Materials in Ground-Based Facilities, Version Number 2, Jet Propulsion Laboratory, Publication 95–17, Pasadena, CA. 5. Zhang, J., Garton, D. J., and Minton, T. K. (2000) Journal of Chemical Physics 117, 6239. 6. Minton, T. K., Zhang, J., Garton, D. J., and Seale, J. W. (2000) Journal of High Performance Polymers 12, 27. 7. Zhang, J. and Minton, T. K. (2001) Journal of High Performance Polymers 13, S467. 8. Koontz, S. L., Albyn, K., and Leger, L. J. (1991) Journal of Spacecraft and Rockets 28, 315. 9. Gregory, J. C. and Peters, P. N. (1988) In J. Visentine (ed.), Atomic-Oxygen Effects Measurements for Shuttle Missions STS-8 and 41-G, Vol. 2, NASA Technical Memorandum 100459, NASA, Houston, TX, pp. 4.1–4.5. 10. Brinza, D. E., Chung, S. Y., Minton, T. K., and Liang, R. H. (1994) Final Report on the NASA/JPL Evaluation of Oxygen Interactions with Materials–3 (EOIM-3), NASA Contactor Report 198865, JPL Publication 94–31, NASA, Pasadena, CA. 11. Peters, P. N., Gregory, J. C., and Swann, J. T. (1986) Applied Optics 25, 1290. 12. Meshishnek, M. J., Stuckey, W. K., Evangelides, J. S., Feldman, L. A., Peterson, R. V., Arnold, G. A., and Peplinski, D. R. (1987) Effects on Advanced Materials: Results of the STS-8 EOIM Experiment, Report No. SD-TR-87-34, Aerospace Corporation, Los Angeles, CA. 13. K. Yokota, M. Tagawa, and N. Ohmae, Journal of Spacecraft and Rockets 1, 143. 14. Caledonia, G. E., Krech, R. H., and Green, B. D. (1987) AIAA Journal 25, 59. 15. Garton, D. J., Minton, T. K., Maiti, B., Troya, D., and Schatz, G. C. (2003) Journal of Chemical Physics 118, 1585. 16. Troya, D., Schatz, G. C., Garton, D. J., Brunsvold, A. L., and Minton, T. K. (2004) Journal of Chemical Physics 120, 731. 17. Unpublished results from T. K. Minton and R. A. Krech. See [4] for details on both types of measurements. 18. Kinoshita, H., Tagawa, M., Yokota, K., and Ohmae, N. (2001) Journal of High Performance Polymers 13, 225.

EROSION OF KAPTON H BY HYPERTHERMAL ATOMIC OXYGEN

329

19. Cline, J. A., Minton, T. K., and Braunstein, M. (2004) In Proceedings of the AIAA 37th Thermophysics Conference, Paper No. 2004-2685, Portland, OR, 28 June–1 July 2004. 20. Andresen, P. and Luntz, A. C. (1980) Journal of Chemical Physics 72, 5842. 21. Gindulyte, A., Massa, L., Banks, B. A., and Rutledge, S. K. (2000) Journal of Physical Chemistry A 104, 9976. 22. Troya, D., Pascual, R. Z., Garton, D. J., Minton, T. K., and Schatz, G. C. (2003) Journal of Physical Chemistry A 107, 7161.

TRANSPARENT ARCPROOF PROTECTIVE COATINGS: PERFORMANCE AND MANUFACTURABILITY ISSUES

JOHN GRIFFIN, NISCHALA UPPALA, JYOTHI VEMULAPALLI, AND PAUL D. HAMBOURGER Cleveland State University, Cleveland, OH 44115

Abstract. Highly transparent thin films with moderately high sheet resistivity approximately 108 ohms·square−1 ( · −1 ) are needed for protection of photovoltaic arrays and other spacecraft surfaces from static charging in space. They may also be useful for dust control on Mars and the Moon. Indium tin oxide (ITO) codeposited with the transparent insulator MgF2 is promising for these applications, but it is difficult to deposit films with reproducible sheet resistivity. We report experiments on a small dual RF magnetron sputter coater, using plasma emission monitoring (PEM) to improve control of film composition. Results show that PEM is useful for composition control but must be supplemented by periodic in situ measurements of coatings’ optical or electrical properties. We have successfully coated both rigid (glass, quartz) and flexible substrates. Key words: arcproof coatings, spacecraft charging, dust management

1. Introduction Charge buildup on nonconductive spacecraft surfaces, due to solar proton and electron emissions, can cause damaging arcing. The optimum protective coating for these surfaces would have a sheet resistivity approximately 108 ohms·square−1 ( · −1 ) and must be highly transparent spectral range if used on photovoltaic arrays or optical windows. An excessively conductive surface is undesirable in low Earth orbit since it may lead to large current flow between the spacecraft power system and the conductive space plasma. Application of these coatings will require production coating of a variety of substrates ranging from flat glass to the complex, flexible polymeric structures of inflatable satellites [1]. Similar coatings may have dust control applications on the Moon and Mars, since the dust is held on surfaces by electrostatic charge. Previous work [2] has shown that thin films of codeposited indium tin oxide (ITO) and MgF2 can be made with the desired sheet resistivity and have the 331 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 331–340. C 2006 Springer. Printed in the Netherlands.

332

JOHN GRIFFIN ET AL. 1e+11 1e+10

Sheet Resistivity (Ω/ )

1e+9 1e+8 1e+7 1e+6 1e+5 1e+4 1e+3 1e+2 1e+1 1e+0

Avg. film thickness = 65 nm

0

5

10

15

Calculated Wt. % MgF2

Total Solar Transmittance

0.90

0.88

0.86

0.84

0.82

0.80

Avg. film thickness = 65 nm Transmittance of SiO2 substrate = 0.93

0

2

4

6

8

10

12

14

16

Calculated Wt. % MgF2 Figure 1. Sheet resistivity and solar transmittance of ITO–MgF2 vs. composition [2]

extremely high solar optical transmittance needed for high-efficiency photovoltaic arrays, as shown in figure 1. Unfortunately, the sheet resistivity of ITO–MgF2 is rather sensitive to the MgF2 /ITO ratio, leading to irreproducible results when depositing coatings in laboratory-scale equipment [3]. Thus this material cannot be used unless a method is found for reliable industrial-scale coating deposition. Industrial deposition of transparent oxide coatings is frequently done by sputtering, using magnetron sputter guns driven by RF or medium-frequency current sources. To investigate the possibility of depositing ITO–MgF2 by this technique, we have prepared films

TRANSPARENT ARCPROOF PROTECTIVE COATINGS

333

using two independently powered RF magnetrons with, respectively, ITO and MgF2 targets, facilitating adjustment of film composition. Since plasma emission monitoring (PEM) of sputter discharges is a technique well known to the vacuum coating industry, we investigated its use to control the composition of our films. We find that monitoring the intensity of ITO and MgF2 plasma emission lines improves sample reproducibility but probably will have to be supplemented by periodic in-situ resistance or optical measurements for reliable production. It would clearly be helpful if sheet resistivity were less strongly dependent on film composition. We find that the addition of high-purity air (N2 /O2 mixture) during deposition appears to accomplish this.

2. Experimental Techniques A schematic view of the deposition chamber is shown in figure 2. Films were deposited by simultaneous operation of two 5.1 cm diameter 13.56 MHz magnetron sputter guns. RF power (generally <100 W) to each gun was controlled independently to adjust film composition. Sputtering was carried out in argon gas at approximately 6 mTorr pressure. Samples normally were made without addition of oxygen or air because this system produces highly transparent, conductive ITO without it. The background pressure with argon turned off and pump throttle valve set as for deposition was <2 × 10−5 Torr. (Pressure with argon off and throttle wide open was typically ∼1 × 10−6 Torr.) Sample thickness was determined by readings of a single quartz crystal monitor (QCM) located near the sample. The QCM had been calibrated separately for MgF2 and ITO by measuring films of each, deposited on optically flat quartz, with PLAN VIEW

Optical Fiber

Optical Fiber

Sample QCM below sample

Pinhole shield

Pinhole shield

ITO Gun

MgF2 Gun Shield

Figure 2. Layout of deposition chamber

334

JOHN GRIFFIN ET AL.

a profilometer [4]. MgF2 /ITO composition ratios were estimated from deposition rate measurements made on each sputter gun at least once during each deposition run. (We had previously found the deposition rate to be approximately a linear function of RF power.) Light from the plasma in front of each gun was collected by an optical fiber oriented approximately as shown in figure 2. Each fiber terminated in a short tube with a small hole at its outer end. This reduced the number of scattered particles reaching the fiber [5]. Emissions were analyzed by a two-channel grating spectrometer with wavelength resolution approximately 0.4 nm. Experiments showed that each fiber received a negligible amount of light from the other sputter gun. Substrate temperature during deposition, estimated from thermocouple measurements, was <40◦ C. Thus, it is likely that our samples were highly disordered or amorphous. However, this demonstrates the feasibility of coating flexible polymeric substrates. Most samples discussed in this paper were deposited on borosilicate glass. However, we have successfully coated Mylar, Kynar, and Upilex. Each substrate was covered by an aluminum mask to produce a sample measuring 0.3 × 1.9 cm2 with electrical contact arms along the edge. Electrical resistance measurements were made at room temperature in ambient atmosphere by four-terminal methods to eliminate the effect of contact resistance, using appropriate guarded cabling and high-input resistance electrometers.

3. Results and Discussion 3.1. PLASMA EMISSION

A typical PEM spectrum for each target is shown in figure 3. The spectra clearly are very different. We chose to use ITO and MgF2 lines at 453 and 384 nm, respectively, because they showed the best correlation with sheet resistivity. Higher resolution plots of these lines are shown in figure 4. The possible benefits of composition control by plasma emission monitoring are shown in figure 5, where we plot sheet resistivity for the same samples vs. the MgF2 /ITO plasma intensity ratio and vs. the MgF2 concentration determined by the quartz crystal monitor. Note the closer correlation of sheet resistivity to the intensity ratio than to the estimated MgF2 concentration. On the other hand, we sometimes found large shifts of sheet resistivity relative to plasma intensity ratio and QCM data, usually after opening the chamber to remove samples. This is shown in figure 6. Based on these results, it appears that PEM can facilitate production of ITO– MgF2 but must be supplemented by periodic in-situ measurements on coated products or witness coupons.

ITO

Intensity

100

50

0 200

300

400 Wavelength (nm)

500

600

500

600

MgF2 800

Intensity

600 400 200 0 200

300

400 Wavelength (nm)

Figure 3. Broadband optical spectrum of each target’s discharge. Intensity is measured in arbitary units ITO

Intensity

100 ITO Indium Tin Oxygen Argon

50

0 448

450

452 454 Wavelength (nm)

456

458

MgF2

Intensity

400

MgF2 Magnesium Fluorine Argon

200

0 382

383 384 385 Wavelength (nm)

386

Figure 4. Higher-resolution spectra of the monitored plasma emissions, together with data on relevant elemental emission lines

336

Sheet Resistivity (Ω Ω// )

JOHN GRIFFIN ET AL. Sheet Resistivity vs Optical Intensity Ratio 1.E+12 1.E+11 1.E+10

Run 272

1.E+09 1.E+08

Run 274

Run 273

1.E+07 1.E+06 1.E+05 1.E+04 1.0

1.5 2.0 2.5 3.0 3.5 4.0 MgF2/ITO Optical Intensity Ratio

Sheet Resistivity vs Vol% MgF2 Sheet Resistivity (Ω Ω// )

1.E+12 1.E+11

Run 272 Run 273

1.E+10 1.E+09

Run 274

1.E+08 1.E+07 1.E+06 1.E+05 1.E+04 15

20

25

Vol% MgF2 (from QCM) Figure 5. Sheet resistivity vs. MgF2 /ITO intensity ratio (upper graph) and vs. estimated MgF2 ˚ on glass concentration (lower graph). Sample thickness 400 A

3.2. COATING STABILITY

The sheet resistivity of several films measured at room temperature is plotted vs. time since deposition in figure 7 (glass substrates) and figure 8 (Upilex substrates). All samples were stored in ambient air. As noted by the authors of [2], sheet resistivity generally increases over time, probably due to absorption of atmospheric oxygen. However, the stability is similar to that of [2] even though our films are somewhat thinner. In addition, we note that stability of our films on Upilex substrates appears to be similar to that of films on glass. Some films in figures 7 and 8 show unusually large resistivity increases. The reason for this is unknown, and the films’ microstructure has not been examined.

TRANSPARENT ARCPROOF PROTECTIVE COATINGS

337

Sheet Resistivity vs Optical Intensity Ratio

Sheet Resistivity (Ω Ω// )

1.E+11 1.E+10 1.E+09 1.E+08

Run 325

1.E+07

Run 326

1.E+06 1.E+05 1.E+04 1.E+03 1.E+02

0

0.5

1

1.5

2

Sheet Resistivity (Ω Ω// )

MgF2/ITO Optical Intensity Ratio

1.E+12 1.E+11

Sheet Resistivity vs Vol% MgF2

1.E+10 1.E+09

Run 325

1.E+08

Run 326

1.E+07 1.E+06 1.E+05 1.E+04 1.E+03 1.E+02

0

10

20

30

Vol% MgF2 Figure 6. Data similar to figure 5, but showing poor correlaton of sheet resistivity with PEM and QCM data

3.3. EFFECT OF AIR INJECTION

The effect of injecting a small amount of high-purity air during deposition is shown in figure 9, where we plot sheet resistivity vs. MgF2 concentration with and without air injection. Argon flow rate was 0.9 SCCM in both cases, so the partial pressure of air was approximately 10% of the total pressure. Although there is considerable scatter in the data, the addition of air appears to make sheet resistivity less dependent on composition. Note that figure 9 shows a considerably higher MgF2 concentration than do figures 5 and 6. We believe this is due to a calibration error in the QCM data of figure 9, which were taken early in

338

JOHN GRIFFIN ET AL. 400 Å ITO-MgF2 on glass Run 272 & 282 1.E+12 1.E+11

Sheet Resistivity (Ω Ω// )

1.E+10 1.E+09 1.E+08 1.E+07 1.E+06 1.E+05 1.E+04 1.E+03

0

50

100

150

Age (days)

Figure 7. Sheet resistivity vs. time since deposition (glass substrates). Samples stored in ambient air

Sheet Resistivity (Ω Ω// )

the project, that does not affect the remainder of this paper. It appears impossible to correct for this error in a reliable manner, so we have not attempted to do so. The data of figure 9 for zero air flow suggest an abrupt increase in sheet resistivity at a “critical” MgF2 concentration, the origin of which is unknown. It

1.E+12 1.E+11 1.E+10 1.E+09 1.E+08 1.E+07 1.E+06 1.E+05 1.E+04 1.E+03

500 Å ITO-MgF2 on Upilex Run #115

0

100 200 300 400 500 600 Age (days)

Figure 8. Sheet resistivity vs. time since deposition (Upilex substrates). Samples stored in ambient air

TRANSPARENT ARCPROOF PROTECTIVE COATINGS

339

Air Flow = 0 Air Flow = 0.1 SCCM 1.E+12 1.E+11

Sheet Resistivity (Ω Ω// )

1.E+10 1.E+09 1.E+08 1.E+07 1.E+06 1.E+05 1.E+04 1.E+03

0

5

10 15 20 25 30 35 40 45 50 Vol. % MgF2 (from QCM)

Figure 9. Sheet resistivity vs. MgF2 concentration with and without addition of air. Argon flow 0.9 SCCM in both cases

might indicate a compositional metal–insulator transition or the onset of electron percolation between conducting and insulating granules.

4. Conclusion ITO–MgF2 coatings have been successfully deposited by sputtering from dual RF magnetrons. Control of film composition using intensity measurements of plasma emission lines improves reproducibility of sheet resistivity. These techniques are familiar to the vacuum coating industry. However, plasma emission monitoring will probably have to be supplemented by in situ measurements of coating properties. Since high-resistance measurements would be difficult in the vicinity of the sputter discharge, measurement of optical properties would be preferable. We find that the reflectance of ITO–MgF2 diminishes as sheet resistivity increases, suggesting that simple reflectance measurements might suffice. The dual-magnetron method has been used to deposit ITO–MgF2 on flexible polymeric substrates that are increasingly of interest for space applications.

340

JOHN GRIFFIN ET AL.

Stability of the coating on polymers appears to be similar to that on glass provided flexure is limited. Preliminary data suggest that injecting high-purity air during deposition may make sheet resistivity less strongly dependent on film composition. This could be very helpful in production. In the next several months we will investigate the durability of ITO–MgF2 under vacuum ultraviolet exposure, in situ optical and electrical properties measurement methods, and the benefits of air injection.

Acknowledgments The authors gratefully acknowledge the support of NASA Glenn Research Center Cooperative Agreements NCC3-740, NCC3-1023, NCC3-1033, and NCC3-1065. We thank Bruce A. Banks, Joyce A. Dever, Thomas W. Kerslake, Craig H. Marshall, and Deborah L. Waters for many helpful discussions.

References 1. Kerslake, T. W., Waters, D. L., Scheiman, D. A., and Hambourger, P. D. (2003) In 1st International Energy Conversion Conference, Paper AIAA 2003-5919, Portsmouth, VA. 2. Dever, J. A., Rutledge, S. K., Hambourger, P. D., Bruckner, E., Ferrante, R., Pal, A. M., Mayer, K., and Pietromica, A. J. (1998) NASA Technical Memorandum 1998-208499, August 1998. 3. Cashman, T., Kaur, J., Muhieddine, L. K., Shanbhag, M., Ubaid, S. H., Welch, B., Vemulapalli, J., and Hambourger, P. D. (2002), In Proceedings of ICPMSE-6. Toronto, Canada, 1–3 May, 2002, Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 73–80. 4. Cashman, T., Demko, R., Uppala, N., Vemulapalli, J., Welch, B., and Hambourger, P. D. (2003) Vacuum Technology and Coating, September 2003, p. 38. 5. C. H. Marshall (private communication).

THE STUDY OF THE EFFECTS OF ATOMIC OXYGEN EROSION ON THE MICROSTRUCTURE AND PROPERTY OF VO2 THERMOCHROMIC COATING USING CSA’S SPACE SIMULATION APPARATUS XIN XIANG JIANG,1 DARIUS NIKANPOUR,1 MOHAMED SOLTANI,2 MOHAMED CHAKER,2 ROMAN V. KRUZELECKY,3 AND EMILE HADDAD3 1 Advanced Materials and Thermal Group, Spacecraft Engineering Space Technologies, Canadian Space Agency, 6767 route de l’Aeroport, St-Huber, Quebec, Canada 2 ´ INRS Energie-Mat´ eriaux and Telecommunication, University of Quebec, Quebec, Canada 3 MPB Communications In, Pointe-Claire, Quebec, Canada

Abstract. In this study, a thermochromic VO2 coating, which has been studied for spacecraft smart thermal radiator application, was exposed to a ground-based atomic oxygen flux to determine the potential effects of space environment on its performance. The coating sample was prepared using laser ablation deposition technique. Atomic oxygen exposure experiment was conducted on the samples for equivalent 6 months and 3 years in typical LEO environment. Mass loss of the coating samples due to atomic oxygen exposure was measured. Characterization of total hemispherical emittance of the coating before and after atomic oxygen exposure indicates that the atomic oxygen erosion affects the thermal-optical performance of the coating to an extent. X-ray photoelectron spectroscope (XPS) analysis of the coating samples was performed and an increase in the oxygen concentration in outermost layer of the coating due to the atomic oxygen exposure was identified. Possible mechanism for the change in thermo-optical property of the coating was discussed. Key words: smart material, thin-film coating, atomic oxygen, space environment effects, thermochromic

1. Introduction Because of the extreme and cycling temperature environment on Earth orbits, thermal control for spacecraft is always needed to provide a regulated temperature 341 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 341–350. C 2006 Springer. Printed in the Netherlands.

342

XIN XIANG JIANG ET AL.

environment for onboard instruments and electronic devices to function properly during all phases of the spacecraft’s mission. Radiator (a thermal louver to eject heat into deep space), heater, and associated control unit constitute a conventional thermal control subsystem. However, such thermal control subsystem can be heavy and expensive [1], particularly for high power consumption spacecraft. As the density of instruments on the spacecraft tends to be high and the mission tends to be complex, thermal control subsystem of higher performance in terms of mass, design simplicity, and cost than the conventional technique would be needed. To meet these requirements, two key technologies are currently under study and development, which are the MEMS (micro electro-mechanical systems) based micro-mechanical louver [2] and the smart radiator using either thermochromic or electrochromic materials [3]. Of particular interest is the thermochromic coating based smart thermal radiator because of its passive mechanism and no need for sensor, control unit, and power consumption. VO2 is a typical thermochromic material, which can change its thermal emittance, also the transmittance and reflectance, due to a metal-insulator phase transition upon the change of environment temperature [4]. Such property can be effectively used to regulate heat rejection by the spacecraft into deep space so as to achieve a passive thermal control for the spacecraft. Through doping, the transition temperature can be modulated to be suitable for the spacecraft thermal control [5]. Coating can be deposited on different substrates using various methods including chemical vapor deposition [6], electron beam evaporation [7], RF sputtering [8], sol-gel technique [9], and pulsed laser vacuum deposition [10]. A recent works by Kruzelecky et al. [11] and Soltani et al. [4] demonstrated that a tuneability of ∼60% in emissivity of VO2 coating can be achieved by proper control of the stoichiometry and microstructure of the coating. However, there is a concern if space environment, particularly the atomic oxygen, would degrade the performance of the coating. Such concern is based on two considerations that firstly atomic oxygen can generally degrade the performance of optical coating; and secondly several phases, such as VO2 and V2 O5 etc., exist in V–O family [12] and atomic oxygen bombardment of the coating may alter its surface stoichiometry composition and, consequently, its thermo-optical performance. It is for this reason that the current study is conducted to understand the potential effects of such space environment on the coating as to provide data for the design of VO2 -based spacecraft thermal radiator to meet the mission requirements.

2. Experimental Procedure 2.1. SAMPLE PREPARATION

In this study, two samples of VO2 coating were deposited on aluminum alloy substrates using laser ablation deposition technic [4]. Metallic vanadium target

EFFECTS OF ATOMIC OXYGEN EROSION

343

was used. A pulsed laser beam generated by a XeCl Excimer laser at wavelength of 308 nm and pulse duration of 20 ns is introduced into the deposition chamber through a quartz window and focused by quartz lens on the target. The deposition was performed in a controlled gas phase consisting of O2 and Ar gas mixture at a pressure of 90 mTorr so that reaction between metallic vanadium atom vapor and oxygen produce desired VO2 instead of other forms of vanadium oxide, such as V2 O5 . The substrate is heated and maintained at 520◦ C to facilitate the thin film crystalline growth. The final thickness of the coating after deposition is about 150 nm. 2.2. CHARACTERIZATION

X-ray diffraction (XRD) analysis of the coating sample was firstly performed using Philips diffractometer to determine its crystalline structure and particularly if other form of vanadium oxide other than VO2 was formed during the deposition process. A 6300 F scanning electron microscope (SEM) was used to observe the microstructure of coating surface, such as the surface smoothness and crystallization of the coating. An important objective of this study is to understand the mechanism of the potential effects of atomic oxygen exposure on the microstructure of the VO2 thermochromic coating. For this purpose, the surfaces of the coating samples before and after atomic oxygen exposure were specifically analyzed using ESCALAB 220I-XL X-ray photoelectron spectrometer to determine if there would be changes in surface chemistry. Total hemispherical emittance was measured using AZ Technology’s reflectometer (Temp 2000) on the coating samples before and after atomic oxygen exposure. A series of such tests was performed at different temperatures ranging from 25 to 60◦ C to determine the thermochromic behavior of the coatings. 2.3. ATOMIC OXYGEN EXPOSURE

The atomic oxygen exposure experiment on the coating samples was performed using Canadian Space Agency’s space simulation apparatus, as shown in figure 1. It is a RF plasma based atomic oxygen generator and is also equipped with barometers, thermocouples, a flashlamp, an oxygen lamp, Krypton and Xenon continuum lamps for generation of UV/VUV radiation, and an irradiance profiler. However, the UV/VUV lamps were turned off during the exposure experiment due to the reason to explicitly study the interaction between atomic oxygen and the coating. Vacuum is generated using a Danielson Tribodyn 200/57 three-module molecular drag pump, with typical operating pressure at ∼100 mTorr inside the test chamber for atomic oxygen flux to be produced, although before the introduction of oxygen for producing atomic oxygen, pressure on the order of 10–4 Torr was achieved. Oxygen flow rate is set to 15 c·c·m−1 . Previous experiment on the

344

XIN XIANG JIANG ET AL.

Figure 1. CSA’s space simulation apparatus

characterization of the facility indicated that the atomic oxygen source is of energy level of 0.1 eV. Two coating samples were used in atomic oxygen exposure experiment. The exposure times were 2 and 12 h for the first and second samples respectively. To determine the effective atomic oxygen fluence so as to determine the equivalent LEO (low Earth orbit) exposure time, Kapton thin film was used as witness sample during the atomic oxygen exposure experiment in accordance with ASTM Standard E 2089-00. The change in thickness of the Kapton witness sample was used to calculate the effective fluence. The masses of the VO2 coating samples before and after atomic oxygen exposure were also measured in accordance to ASTM standard in anaerobic chamber using a microbalance with an accuracy of 0.002 mg.

3. Experimental Results and Discussion Figure 2 shows the XRD pattern for VO2 coating sample deposited on aluminum substrate. Because of the penetration of X-ray through the VO2 coating, the diffraction peaks from aluminum substrate are picked up and are relatively strong. VO2 diffraction peak is easily distinguished although it is relatively weak, indicating probably imperfect crystallization. Concerning the degree of surface degradation, the mass losses of the coating samples after the atomic oxygen exposures were firstly measured. Table 1 lists the parameters of atomic oxygen exposure experiment and the mass losses occurred on two samples. Mass loss of the coating sample after equivalent 6 months LEO exposure was found to be too small to be accurately measured (within the error of microbalance). However, when the equivalent LEO exposure time increases to

345

EFFECTS OF ATOMIC OXYGEN EROSION TABLE 1. Parameters of atomic oxygen exposure experiment on VO2 coating samples Sample no. Exposure time (h) Effective fluence (atoms·cm−2 ) Equivalent LEO exposure time* Mass loss (mg) Reaction efficiency (×10−27 cm3 ·atom−1 ) ∗

#1

#2

2 1.56 × 1021 ∼6 months negligible —

12 9.5 × 1021 ∼3 years 0.018 0.42

Calculation is based on typical LEO/AO flux of 1 × 1014 atoms·cm−2 · s−1 [14]

3 years, mass loss of 0.018 mg was measured. Giving the density of VO2 of 4.34 g·cm−3 [13], the atomic oxygen reaction efficiency (Re ) of the VO2 coating can be calculated according to the eq. (1) [14]: Re =

(M/ρ) Volume of material lost = F F

(1)

where M—Mass loss of the sample (g) F—Fluence of atomic oxygen (atoms·cm−2 ) ρ—Density of the material (g·cm−3 ) It can be seen from the results that atomic oxygen reaction efficiency of VO2 coating is very low, in the same order of magnitude as that of high quality SiO2 VO2/Aluminum

Al

700 600

Al

Intensity

500 400

Al

300

Al

200

VO2

100 0 10

20

30

40

50

60

70

80

2θ (degree)

Figure 2. XRD pattern of VO2 coating deposited on aluminum substrate

346

XIN XIANG JIANG ET AL.

Semi-hemisphere emittance

0.5 0.4 0.3 0.2 Non AO exposed

0.1

Equivalent 3 years LEO AO Exposure

0 20

30

40

50

60

70

Temperature (C)

Figure 3. The effects of AO exposure on the total hemispherical emittance

coating (<0.8 × 10–27 cm3 ·atom−1 ), which is commonly used for protecting material in atomic oxygen environment. This means the surface erosion of VO2 coating is rather limited from the mass loss point of view. However, degradation or change in the thermo-optical properties of the coating can’t be ruled out because several other factors, such as the change in surface texture and stoichiometry of VO2 coating, could still affect the thermo-optical properties of the coating. A significant feature of the thermochromic characteristic of the VO2 coating is that its thermo-optical properties change drastically within a relative narrow temperature range, as confirmed by previous study [4, 11]. Figure 3 shows the thermochromic behavior (change of thermal emittance with respect to the change of temperature) of VO2 coating sample before and after equivalent 3 years LEO atomic oxygen exposure. It can be seen that although the atomic oxygen exposure does not alter the basic thermochromic behavior of the coating, it elevates overall total hemispherical emittance and also broadens the transition temperature range slightly. In a study by Dillon et al. [15] on VO2 coating deposited on single crystal Si substrate using sputtering process, it was also found that plasma oxygen bombardment tends to increase the thermal emittance of the coating although specific causes could not be identified. It is noted the VO2 samples used in this study, which are deposited on Al substrate by laser ablation process, are somewhat different from those used by Dillon et al. in terms of thermo-optical property, probably due to the difference in coating/substrate interfacial structure and deposition conditions. However, one conclusion may be drawn from this experiment and the study done by Dillon et al. [15] is that atomic oxygen can generally affect the thermo-optical property of VO2 coating to an extent. Characterization of the microstructure of the coating before and after atomic oxygen exposure was performed. Figure 4 shows the SEM observations of the surface of the coating samples before and after atomic oxygen exposure. By

347

EFFECTS OF ATOMIC OXYGEN EROSION

(a)

(b)

Figure 4. SEM images of VO2 coating surface (a) before and (b) after equivalent 3 year LEO atomic oxygen exposure

a comparison, it is evident that a small number of surface defects (holes) are created after the atomic oxygen exposure. However, the creation of the defects of this size (up to a micrometer) is somewhat puzzling because stoichiometric VO2 is not a volatile material. A possible explanation is that, on one hand, VO2 coating has relatively high density of crystalline structure defects as indicated by weak XRD peak; on another hand, the atomic oxygen is quite reactive and can effectively introduce structure defects in oxide crystalline material through mechanisms of substitution, vacancy, and interstitial [16]. Upon atomic oxygen exposure, density of structure defects in the VO2 coating is likely to increase substantially. This may result in possible concentration of defects at certain sites, forming small holes on the surface, as observed by the SEM. The formation of microscale surface defects can increase the microsurface roughness of the coating. This may be a contributing factor to the overall increase in the thermal emittance of the coating because it generally increases with the increase of surface microroughness. Characterization of the surface chemistry (stoichiometry composition) of the coating before and after atomic oxygen exposure was further performed. Figure 5 shows the XPS analysis of the coating before and after equivalent 6 months and 3 years LEO/AO exposures. It can be seen that, firstly, the intensity of O1s peak increases and, secondly, the peak for V2p1/2 becomes more prominent after equivalent 3 years LEO/AO exposure. The increase in intensity of O1s peak is an ˚ indication of the increase in oxygen concentration in the outermost layer (<100 A) of the VO2 coating. This is further confirmed by the XPS quantitative analysis that the ratio of oxygen (O) to vanadium (V) for the coating has increased from stoichiometrically balanced 2.0–2.1 after equivalent 3 year LEO atomic oxygen exposure. Such increase in O/V ratio certainly would create a stoichiometrical imbalance between O and V in terms of VO2 coordination and consequently, a large number of vacancies (defects) for vanadium atoms would be generated in the

348

XIN XIANG JIANG ET AL. 1,5 × 106

O1e

VO2/Al after equivalent 3 years LEO AO exposure

1,0 × 106

V2p3/2 V2p1/2

Intensity

5,0× 104 1,0 × 106

VO2/Al after equivalent 6 months LEO AO exposure

5,0× 104 1,0× 106 VO2/ Al Before AO irradiation 5,0 × 104

0,0 540

535

530

525

520

515

510

505

Binding energy (eV) Figure 5. XPS analysis of the coating before and after equivalent 6 months and 3 years LEO/AO exposures

crystalline structure. The defects reduce the energy bandgap for photon emission, this is probably the main reason why thermochromic behavior (thermal emittance with respect to temperature) of the coating is affected by atomic oxygen to an extent. 4. Conclusion In this study, a thermochromic VO2 coating sample, which is being developed for smart spacecraft thermal radiator application, is subjected to a test in groundbased atomic oxygen flux to determine the potential effects of space environment on its thermo-optical property. Equivalent 6 months and 3 years LEO/AO exposure experiments were performed on two samples respectively. The experiment results show that VO2 coating tends to have very good resistance to the atomic oxygen erosion as judged from the mass loss results. However, characterization of thermal emittance of the coating before and after atomic oxygen exposure indicates that the atomic oxygen exposure can affect the thermal-optical performance of the coating to an extent, increasing overall total hemispherical emittance and slightly

EFFECTS OF ATOMIC OXYGEN EROSION

349

broadening the transition temperature range. SEM and XPS analysis of coating samples before and after the atomic oxygen exposure were performed to explore the causes for such change. The analysis suggests that an increased density of surface and crystalline structure defects, caused by atomic oxygen bombardment, is probably the main reason responsible for the changes in thermo-optical property of the coating.

References 1. Gilmore, D. G (Ed), (1994) Satellite Thermal Control Handbook, The Aerospace Corporation Press, El Segundo, California 2. Darrin, A. G., Osiander, R., Champion, J., Swanson, T., and Douglas, D. (2002) Variable emissivity through MEMS technology, In Proceedings of Space Technology and Application International Forum 2000, M. S El-Genk (ed.), AIP Conference Proceedings, American Institute of Physics Vol. 504, January 19, pp. 803–808. 3. Tachikawa, S., Shimazaki, K., Ohnishi, A., Hirosawa, H., Shimakawa, Y., Ochi, A., Okamoto, A., and Nakamura, Y. (16–20 June 2003) In Proceedings of 9th International Symposium on Materials in a Space Environment, Noordwijk, The Netherlands, European Space Agency, pp. 34–47. 4. Soltani, M., Chaker, M., Haddad, E., Kruzelecky, V., and Nikanpour, D. (May/June 2004). Optical switching of vanadium dioxide thin films deposited by reactive pulsed laser deposition, Journal of Vacuum Science and Technology. A22 (3), pp. 859–864. 5. Takahashi, I., Hibino, M., and Kudo, T. (October 1999) Thermochromic properties of doubledoped VO2 thin films fabricated from polyvanadate-based solutions, In Proceedings of SPIE, Vol. 3788, pp. 26–32. 6. Sahana, M. B., Subbanna, G. N., and Shivashankar, S. A. (2002) Phase transformation and semiconductor-metal transition in thin films of VO2 deposited by lowpressure metalorganic chemical vapor deposition, Journal of Applied Physics 92, 6495– 6504. 7. Lee, M. H., and Kim, M. G. (1996) RTA and stoichiometry effect on the thermochromism of VO2 thin films, Thin solid films, 286, 219–222. 8. Jerominek, H., Picard, F., and Vincent, D. (1993) Vanadium oxide films for optical switching and detection. Optical Engineering 32, 2092p. 9. Partlow, D. P., Gurkovich, S. R., Radford, K. C., and Denes, L. J. (1991) Switchable vanadium oxide films by a sol-gel process, Journal of Applied Physics 70, 443p 10. Kim, D. H., and Kwok, H. S. (1994) Pulsed laser deposition of VO2 thin films. Applied Physics Letters 65, 3188p 11. Kruzelecky, R. V., Hadda, Jamroz, E., Soltani, M., Chaker, M., Nikanpour, D., and Jiang, X. (July 2003) Passive Dynamically Variable Thin-Film Smart Radiator Device, In Proceeding of 2003 SAE Conference, 2003-01-2472, Vancouver, BC, Canada. 12. Parker, S. P. (1992) Encyclopedia of Chemistry, S. P. Parker (ed), 2nd ed., McGraw-Hill, 1152p. 13. Chudnovskiy, F., Luryi, S. (2002) Switching Device Based on a First-Order Metal-Insulator Transition Induced by an External Electric Field, in Future Trends in Microelectronics: the Nano Millennium, S. Luryi, J. M. Xu, and A. Zaslavsky, (ed), Wiley Interscience, pp. 148–155.

350

XIN XIANG JIANG ET AL.

14. Silverman, E. M. (August 1995), Space environmental Effects on Spacecraft: LEO Materials Selection Guide, NASA contract report 4661, E. M. Silverman (ed.), part 1 15. Dillon, R. O., Le, K., Ianno, N. (2001) Thermochromic VO2 sputtered by control of a vanadiumoxygen emission ratio, Thin Solid Films, 398–399, pp. 10–16 16. Kim, H. S., Lee, C. H., Lee, C. E., Kim, K. M., Noh, S. J., Hong, C. S., Hur, N. H., Shim, S. Y., and Ri, H. C. (2001) Oxygen-plasma effects of a La0.7 Ca0.3 MnO3 single crystal, Applied Physics Letters 79(25), 4177–4179.

DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION

Q. WEI, S. Y. HE, AND D. Z. YANG Space Materials and Environment Engineering Laboratory, Harbin Institute of Technology, P. O. Box 432, Harbin 15001, P. R. China

Abstract. The change in optical transmittance of JGS3 optical quartz glass was studied using ground simulation for the space environment conditions of high vacuum, heat sink, and 140 keV low energy protons and the kinetic model for the evolution of color centers in the process of radiation damage was explored. The experimental results show that radiation damage occurs in the surface layer of quartz glass under large flux and low-energy proton radiation. The optical density change increases rapidly and then a saturation trend appears with increasing absorption dose. A kinetic model for the evolution of color centers in quartz glass irradiated with protons is proposed based on experimental data, from which the change in optical density can be given. The model fitted curve is similar to experimental ones. It is believed that the proposed kinetic model can be used in the quantitative description for the change in optical property of quartz glass with increasing absorption dose under proton radiation with low energy. Key words: quartz glass, protons radiation, color center, optical density

1. Introduction Quartz glass has a number of excellent properties, and is extensively used as optical windows and optical lens in spacecraft [1, 2]. The properties of optical elements or materials would be degraded [3, 4] as the spacecraft flies in orbits subjected to the radiation by protons and electrons with a wide energy spectrum of the Van Allen radiation belts. In the past years, the study concerning radiation effects of space charged particles on optical materials focused on the particles with energies over 1 MeV. However, in the Van Allen radiation belt the flux of particles increases with their energy decreasing. The protons and electrons with energies below 200 keV have large fluxes up to 108 particles·cm−2 ·s−1 . The lower the particle energy is, the relatively shorter is the penetration depth of the particle into materials. The absorbed energy and the particles mainly concentrate in the surface layers of the materials. Consequently, the effects of proton radiation with such low energies on 351 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 351–357. C 2006 Springer. Printed in the Netherlands.

352

Q. WEI ET AL.

optical materials are important, especially for long-term mission spacecraft. The present study is aimed at investigation of the change of spectrum properties of a JGS3 quartz glass induced by irradiation with protons with energy of 140 keV in vacuum environment and heat sink, and analyze the color center evolution kinetics. 2. Experimental The experimental material is JGS3 optical quartz glass with the impurity level of less than 5 × 10−3 %, in which OH groups are removed. The sample dimensions are 20 × 20 × 2 mm, and its surfaces are polished. The average transmittance of the samples in 200–3200 nm region is 93%. The radiation experiments were conducted in a system that can simulate the radiation of protons with 30–200 keV under 10−4 Pa vacuum and 77 K heat sink environment. The protons energy of 140 keV is chosen, the beam density is 0.2 μA·cm−2 , and the highest protons fluence is 2 × 1016 cm−2 . The transmittance of the samples before and after irradiation was analyzed using a spectrophotometer of UV-3101PC type made by SHIMADZU, Japan. 3. Results and Discussion The change in transmittance spectrum of the JGS3 optical quartz glass irradiated with the protons is shown in figure 1. With increasing fluence, the transmittance of the quartz glass decreases, mainly in the near-ultraviolet to the visible region, with little changes in the infrared region, as shown in figure 1(a). Figure 1(b) shows that the obvious transmittance change occurs in the region of 220–250 nm. E centers are the primary radiation defects in the quartz glass [5, 6]. The absorption peaks of the quartz glass at 220–250 nm can be related to the E

Figure 1. The spectral transmittance (a) and transmittance change for JGS3 quartz glass before and after protons irradiation with 140 ke V

DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION

353

Figure 2. The change in optical density at 230 nm with the absorption dose for quartz glass after 140 keV protons radiation

centers and their variants [7, 8]. The optical density change of quartz glass at 230 nm is discussed in this paper. Generally the degree of optical absorption of a material is indicated by the optical density, expressed as the logarithm of the ratio of the transmitted light intensity T to the incident light intensity T0 , i.e., T A = lg (1) T0 The absorption dose and the radiation fluence are correlated with each other by the following formula: = 1.6 × 106 Eψ/Rρ

(2)

Here is the absorption dose in Gy, E the particle energy in MeV, the radiation fluence in protons·cm−2 , R the incidence particle penetration depth in material in cm, and ρ the material density in g·cm−2 . Figure 2 shows the increase in optical density at 230 nm with the absorption dose for the samples irradiated with 140 keV protons. When the absorption dose reaches a critical value, the change in optical density tends to level off. This trend might be related to a dynamic balance between the production and annihilation of color centers induced by the proton radiation.

4. The Role of Kinetics of Color Center Formation in Radiation Damage 4.1. HYPOTHESES

Radiation damage is a complicated physical and chemical process, which depends not only on the radiation dose , dose rate J , and radiation energy E, but also

354

Q. WEI ET AL.

on material category and defect distribution. The following hypotheses can be proposed for discussing the radiation damage kinetics: 1. The concentration of potential trap defects Ni (x) and distribution of defects along incidence direction do not change during irradiation. 2. During irradiation, a constant fraction of potential trap defects f i is converted to color centers per unit time. The dn i is the number of potential trap defects converted to color centers per unit time dt and qi is the annihilation fraction of color center during radiation. 3. The concentration of type i color centers n i0 = 0, when irradiation begins. Before irradiation, the absorption spectrum of quartz glass is regarded as the intrinsic or background absorption. 4. The inelastic ionization effect predominates for low-energy radiation. All the secondary electrons or vacancies originated from the ionization are trapped by potential trap defects. Assume that σ is average electron trapped section within the concentrated layer that color centers occur in. 4.2. THE KINETIC MODEL

Let J be the absorption dose rate, and n i the concentration of color centers at time t. According to the hypotheses above, the growth rate of color centers is given by dn i = σ J [ f i (Ni (x) − n i ) − n i qi ] dt

(3)

If the concentration of defect traps Ni is constant, the equation can be solved as follows: f i Ni (x) n i = n i0 + − n i0 1 − e−( fi +qi )σ J t f + qi i f i Ni (x) = n i0 + − n i0 1 − e−( fi +qi )σ (4) f i + qi Based on assumption (3) above, n i = n i0 = 0 at t = 0, the concentration of color centers can be given as ni =

f i Ni (x) f i Ni (x) 1 − e−( fi +qi )σ J t = 1 − e−( fi +qi )σ f i + qi f i + qi

(5)

The eq. (5) presents a relation between the absorption dose and the average concentration of color centers per unit area of surface, and reflects the accumulation kinetic of color centers.

DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION

355

4.3. THE EXPRESSION FOR OPTICAL PROPERTY

It was found that the relation between the integral areas of absorption curve and the oscillator strength and concentration of color centers could be expressed as [9] nf =

9mc 1017 n Z nZ α W = const × αm W m 2e2 h (n 2Z + 2)2 (n 2Z + 2)2

(6)

where n is volume concentration of color centers; e the unit charge; f the oscillator strength; n Z the refractive index of media; m ∗ the electronic reduced mass of color center; α(E ∗ ) the absorption coefficient, and E ∗ the photon energy. α m the absorption coefficient of absorption peak, W the full width at half height. According to the above equation, when W is constant, the concentration of color centers is proportional to the absorption coefficient, i.e., n = ξ α. Meanwhile, there is a relation of α = 2.303 A between the absorption coefficient and optical d density, and then the relation between optical density and concentration of color centers can be given by A = 0.434dα =

0.434 dn = gdn ξ

(7)

where ξ and g are scale factors. Because the effects of proton radiation with low energy mainly focus on the materials surface, the color centers concentrate in a very thin layer on the surface, and d is thickness of this layer. According to assumption (3), n io = 0. The concentration of color centers is proportional to the optical density change induced by radiation, i.e. Ai = Ai − Ai0 = dg (n i − n i0 ) = dgn i = Ai

(8)

Then, according to eqs. (5), (7), and (8), the optical density change can be given by f i Ni (x) 1 − e−( fi +qi )σ J t f i + qi f i Ni (x) 1 − e−( fi +qi )σ = Ais (1 − e−β ) = dg f i + qi

Ai = dg

(9)

i (x) where, Ais = dg fifiN+q ; β = ( f i + qi )σ . i Because d is the thickness of the concentrated layer of color centers, eq. (9) reflects the change in average optical density per unit area of sample surface. According to the above analysis, under radiation with 140 keV protons, the optical density change of quartz glass can be given by following formula:

A = a(1 − e−bx )

(10)

356

Q. WEI ET AL. TABLE 1. Parameters fitted for the change in optical density of fused quartz at 230 nm under proton radiation Parameters Proton energy (keV)

A

b

140

0.16031

3.9385E-9

Figure 3 shows the results of mathematical fit to the experimental curve. The parameters are listed in table 1. Comparing the experimental curve with the fitted one, it can be suggested that the functions fitted from the kinetic model of color centers can approximately describe the optical property change for the quartz glass under proton radiation with low energy.

5. Conclusion Space proton irradiation with low energy is characterized by a high flux and results in obvious radiation damage effects on the surface layer of quartz glass. Initially, the change in optical density increases rapidly, and then a saturation trend appears with increasing absorption dose. Based on the analysis of the experimental results, a kinetic model for the evolution of color centers in quartz glass irradiated with protons is proposed. The change in average optical density per unit area of surface can be given by Ai = Ais 1 − e−β . The mathematic fitted curve is similar to the experimental one. We believe that the proposed kinetic model can be used in the

Figure 3. The fitted curves for the change in optical density of quartz glass at 230 nm under proton radiation

DAMAGE KINETICS OF QUARTZ GLASS BY PROTON RADIATION

357

quantitative description of changes in optical property of quartz glass with increasing absorption dose under protons radiation with low energy.

References 1. Englisch, W. (1989) In Proceedings of Society of Photo-Optical Instrumentation Engineer, Orlando, Florida, America, Vol. 1118, pp. 42–48. 2. Wei, Q., Liu, H., He, S. Y. (2004) Radiation Effects & Defects in Solids 159, 195–201. 3. Garrett, H. B., and Hastings, D. (1994) The space radiation environment, In Proceedings of American Institute of Aeronautics and astronautics, AIAA94-0590, pp. 1–12. 4. Blue, M. B. (1994) Degradation of Optical components in Space, National Aeronautics and Space Administration (NASA), N94-31029, pp. 217–225. 5. Marshall, C. D., Speth, J. A., and Payne, S. A., (1997) Journal of Non-Crystalline Solids 212, pp. 59–73. 6. Levy, P. W. (1985) In Proceedings of Society of Photo-Optical Instrumentation Engineer, Bellingham, WA, Vol. 541, pp. 1–13. 7. Griscom, D. L. (1980) Journal of Non-Crystalline Solids 40, pp. 241–245. 8. Leach, R. D. (1995) Journal of AIAA 3564, pp. 1–17. 9. Fang, S. G., and Zhang, Q. R. (1989) Physics of Color Center in Crystal, Shanghai Jiao Tong University Press, Shanghai, China, pp. 36–38.

MICROSCOPIC MECHANISMS AND DYNAMICS SIMULATIONS OF O+ (4 S3/2 ) REACTING WITH METHANE

LIPENG SUN AND GEORGE C. SCHATZ Department of Chemistry, Northwestern University Evanston, IL

Abstract. The reaction O+ (4 S3/2 ) + methane is studied as a benchmark for developing the theory of polymer erosion by O+ under LEO conditions. Ab initio electronic structure calculations show that the interaction of O+ with CH4 can lead to a large number of reaction products such as charge transfer, hydride abstraction, H elimination, etc. Based on the information obtained from these quantum chemistry calculations, a direct dynamics classical trajectory simulation was carried out at 5eV relative translation energy and the chemical reaction channels predicted by the ab initio calculations are confirmed. Key words: ab initio calculation, direct dynamics, reaction mechanism, O+ + CH4 1. Introduction When space vehicles travel in low Earth orbit (LEO) or geosynchronous Earth orbit (GEO), their surfaces are under constant bombardment by energetic atoms/molecules, ions, electrons, and various sources of electromagnetic radiation. Among the collision processes, atomic/molecular ions are responsible for the space vehicle charging and atmospheric drag. Under the LEO and GEO environments, ions may have kinetic energies ranging from 1 to 106 eV [1, 2]. As a consequence, various chemical processes that carry mass away from the surfaces can be induced and enhanced due to ion/surface collisions that cause chemical reaction, energy transfer, and surface ionization. Therefore, it is important to understand the ion/surface erosion mechanisms at the microscopic level and to evaluate the relative importance of ions compared to neutrals (such as atomic oxygen) in producing erosion. Although there have been extensive studies on the O(3 P) erosion mechanisms of the surface materials of spacecraft [3], surprisingly, the microscopic reaction mechanisms of the O+ (4 S) ion, the most abundant ionic species in LEO, are mostly unknown [4]. Recent experimental work has been carried out for O+ reaction with small alkane molecules in the gas phase using the guided-ion beam time-of-flight (TOF) technique[5] and for O+ reaction 359 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 359–364. C 2006 Springer. Printed in the Netherlands.

360

LIPENG SUN AND GEORGE C. SCHATZ

with decanethiolate/Au(111) self-assembled monolayers [6]. However, theoretical investigation for these reaction systems has not been done. As a benchmark system, the reaction of O+ with CH4 can provide detailed information about the formation/cleavage of C–O, O–H, C–H, and H–H bonds which should be useful for understanding the closely related polymer erosion mechanisms in O+ /surface collisions. In addition, this is a useful system for developing theoretical methods that enable the direct simulation of O+ /surface collisions in the future. Therefore, in the work reported here, both ab initio electronic structure theory and classical trajectory simulations are employed to study the O+ + CH4 reaction mechanisms and dynamics.

2. Results 2.1. Ab initio CALCULATIONS

Under the LEO environment the kinetic energy of O+ is about 5eV relative to the traveling space vehicle. With this energy, a large number of reactions may be energetically accessible. To evaluate the reaction energies, the ab initio electronic structure calculations were performed by using the quantum chemistry software package GAMESS [7]. It has been shown from previous studies that for ionmolecule reactions the MP2/6-31G∗∗ level of calculation gives reasonable energetics and geometries. The calculated structures and energies are shown in figure 1. Charge transfer can be viewed as a simple type of chemical reaction in which one or several electrons move from one chemical species to another. For the [O rCH4 ]+ system, charge transfer, O+ + CH4 → CH4 + + O, is exothermic and can be described by a near-resonant electron transfer mechanism. In this scenario, charge transfer occurs at primarily large impact parameters with negligible momentum transfer between the colliding species. The recently measured experimental charge transfer cross section supports this picture [8]. The large charge ˚ 2 ) allows us to model the subsequent chemical transfer cross section (about 80A processes by studying the O + CH4 + collision system. The equilibrium structures of the charge transfer intermediate CH4 + have been extensively studied in the past twenty years [9–12]. Its symmetry, due to John–Teller effects, is reduced from Td to C2v , D2d and C3v with C2v as its most stable structure as depicted in figure 1. After charge transfer, many processes can take place on the several accessible electronic states. On the lowest quartet electronic state, the primary products that are thermodynamically favorable are found to be CH3 + OH+ , H2 CO+ +2H, H2 O+ +CH2 , and CH3 O+ + H. All of these reactions are exothermic. In addition, when O approaches the CH4 + ion there is a reaction intermediate formed, and the stable structure due to the ion-molecule electrostatic interaction, as shown in figure 1, is the configuration formed when O approaches the vertex of CH4 + . The

MICROSCOPIC MECHANISMS AND DYNAMICS SIMULATIONS Reactant

Charge Transfer

Reaction Intermediates

2.0

0.0

E (eV)

Products CH2+ + H2 + O CH3+ + H + O

+

O+ + CH4

OH+ + CH3 + O+

−2.0

361

H2CO+ + 2H

CH4+

H2O+ + CH2 +

CH3O+ + H

+

O

CH2+ + H2O CH3+ + OH

−4.0

H2CO+ + H2

−6.0 OH

+

−8.0

Figure 1. Energy diagram of O+ + CH4

CH4 moiety of the reaction intermediate is very close to the C3v symmetry of CH4 + with the charge located on the CH4 moiety. It is interesting to notice that there is no reaction barrier found from [H3 CH–O]+ → CH3 + OH+ which indicates that direct dissociation should readily occur. With such a strong electrostatic interaction (−1.57 eV) between CH4 + and O, a reaction mechanism is expected in which O and CH4 + first form the [H3 CH–O]+ complex and then this is trapped in a deep potential energy well for a certain period of time and then this either directly dissociates to OH+ + CH3 or undergoes further chemical reactions such as OH+ + CH3 → CH2 + H2 O+ . This reaction mechanism should be especially important for low collision energies where all the above exothermic reactions are expected to occur. With increase in the colliding energy, however, this complex mediated mechanism becomes less and less important. Instead, a direct reaction mechanism becomes more important. For example, for the [H3 CH–O]+ → CH3 + OH+ reaction, the incoming O atom can directly grab one H atom from CH4 + without exchanging much of its kinetic energy. This mechanism is usually referred as a stripping mechanism. There are also endothermic reactions, i.e., CH3 + + H + O and CH2 + + H2 +O possible on the quartet potential energy surface. Since the available energy for the reaction (5eV) is much larger than the dissociation

362

LIPENG SUN AND GEORGE C. SCHATZ

threshold (0.8 eV for CH4 + → CH3 + + H and 1.7 eV for CH4 + → CH2 + + H2 ), these reactions occur with a collision induced dissociation mechanism in which the CH+ 4 cation received enough internal energy from the collision with O to dissociate into CH3 + + H or CH2 + + H2 . The reactions on the electronic doublet potential energy surface are highly exothermic. Also, a reaction intermediate, formed by O atom insertion to the C–H bond, is found. The major product ions are CH2 + , CH3 + , and H2 CO+ . It is worth noting that at low collision energies CH3 + can only be produced on the doublet surface which implies that intersystem crossing to the doublet state is required. Although spin-orbit coupling is usually weak, at low collision energies, when the system is trapped as a reaction intermediate, the crossing seam can be accessed multiple times before dissociation, so that the transition from quartet to doublet states is enhanced. Since a majority of the energetically accessible reaction channels are exothermic, at high impact energies, the primary products such as CH3 O+ and H2 CO+ can further isomerize to H2 C–OH+ and HC–OH+ or dissociate directly to HCO+ . These products are also shown in the figure 1. 2.2. DIRECT DYNAMICS SIMULATION

The energetics calculations in the previous section provide qualitative information about which reactions can occur, but quantitative information about the reaction dynamics can only be revealed by performing dynamics simulations. To make a further understanding of the [O·CH4 ]+ reaction dynamics, direct dynamics classical trajectory simulations in which the energy gradient is obtained directly from a semiempirical Hamiltonian (PM3) were performed on the quartet potential energy surface, i.e., only spin-allowed products are included. The trajectory initial conditions were chosen by a quasi-classical sampling method for the CH4 + . The colliding O atom has an energy of 5 eV relative to CH4 + . The products found from the simulation and their branching fractions are shown in table 1. Many reaction products, especially the oxygenated species, are found. The CH3 + + H + O product is the result of a collision induced dissociation reaction mechanism. The appearance of CH3 + OH+ is in agreement with the TABLE 1. The reaction products and their branching fractions, based on PM3 calculations on the lowest quartet potential surface OH+ CH+ CH+ HCO+ H2 CO+ CH+ H2 O+ CO+ 3 3 2 + CH3 + H + O + OH + 3H + 2H + H2 + O + CH2 + H2 + 2H other Branching fraction 12.4%

7.0%

12.4% 29.5%

2.3%

24.8%

6.2%

5.4%

<1%

MICROSCOPIC MECHANISMS AND DYNAMICS SIMULATIONS

363

expected reaction mechanism on the quartet potential energy surface. HCO+ is a secondary product that comes directly from H2 CO+ . The small fractional population of the H2 CO+ cation radical indicates that when it is formed the energy transferred to H2 CO+ is sufficient for its dissociation. The CH2 + product supports a CH4 + collision induced dissociation mechanism. Interestingly, a CO+ product was also found with a non-negligible branching fraction. It is worth noting that + the total fraction of CH+ 3 (19%), which is much smaller than that of HCO , is not in agreement with experiment [8]. As discussed in the previous section, this may partially due to the fact that the current calculations are based on the electronic quartet potential energy surface and do not treat intersystem crossing. Besides the analysis above, we also noticed that PM3 gives only a qualitatively correct potential energy surface. A detailed quantitative picture of the reaction branching ratio needs to be made on a more precise potential energy surface. However, this is outside the scope of this paper.

3. Conclusion Overall, the quantum chemistry calculations and classical trajectory simulations have found that a large number of chemical reactions occur on both quartet and doublet potential energy surfaces in O+ + CH4 collisions. This is in sharp contrast to what is found for O(3 P) collisions with methane, where the dominant products at 5 eV involve either OH + CH3 or H + CH3 O formation, along with the dissociation channels associated with CH3 O unimolecular decay. The charge transfer happens via a quasi-resonant mechanism at large impact parameter and is the dominant process. Chemical reactions including the formation of C–O and O–H bonds occur at smaller impact parameters, and we have interpreted these results by an intermediate-mediated reaction mechanism at low collision energy and a CID/direct reaction mechanism at high impact energy. The subsequent secondary reactions involve unimolecular dissociation mechanisms. Intersystem crossing is expected to play some role in the reaction dynamics, especially at low energies. This is an important conclusion since when O+ collides with the surface covered by polymer materials, the O+ may be trapped on the surface which provides enough time for intersystem crossing to happen. With this understanding, our future work will focus on O+ reactions with alkyl thiol selfassembled monolayer surfaces and other types of material surfaces.

Acknowledgment This research was supported by AFOSR MURI Grant F49620-01-1-0335.

364

LIPENG SUN AND GEORGE C. SCHATZ

References 1. Dressler, R. A. and Murad, E. (2001) In Chemical Dynamics in Extreme Environments, R. A. Dressler (Ed.), World Scientific, Singapore. 2. Jacobs, D. C. (2001) In Chemical Dynamics in Extreme Environments, R. A. Dressler (Ed.), World Scientific, Singapore. 3. Troya, D. and Schatz, G. C. (2001) International Reviews in Physical Chemistry T. K. Minton, R. A. Dressler, and D. J. Garton (Eds.), in Advanced Series of in Physical Chemistry: Chemical Dynamics in Extreme Environments, World Scientific, Singapore, and references therein. 4. Lyubimova, G. V. and Shestakov, A. F. (1994) Kinetics and Catalysis 35, 208. 5. Levandier, D. J., Chiu, Y. H., and Dressler, R. A. (2004) J. Chem. Phys.120, 6999. 6. Qin, X., Tzvetkov, T., and Jacobs, D. C. (2003) Nuclear Instruments and Methods in Physical Research B, 203, 130. 7. Schmidt, M. W., Baldridge, K. K., Boatz, J. A., Elbert, S. T., Gordon, M. S., Jensen, J. H., Koseki, S., Matsunaga, N., Nguyen, K. A., Su, S. J., Windus, T. L., Dupuis, M., and Montgomery, J. A. (1993) Journal of Computational Chemistry 14, 1347. 8. Levandier, D. J., Chiu, Y.-h., Dressler, R. A., Sun, L., and Schatz, G. C. (2004) Journal of Physical Chemistry A, A108, 9794–9804. 9. Paddonrow, M. N., Fox, D. J., Pople, J. A., Houk, K. N., and Pratt, D. W. (1985) Journal of the American Chemical Society, 107, 7696. 10. Frey R. F. and Davison E. R. (1988) Journal of Physical Chemistry, 88, 1775. 11. Wetmore, S. D., Boyd, R. J., Eriksson, L. A., and Laaksonen, A. (1999) Journal of Physical Chemistry, 110, 12059. 12. Signorell, R. and Merkt, F. (1999) Journal of Physical Chemistry, 110, 2309.

THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) WITH PERFLUORINATED HYDROCARBONS DIEGO TROYA∗ AND GEORGE C. SCHATZ Department of Chemistry, Northwestern University, 2145 Sheridan Rd, Evanston, IL

Abstract. We have studied the reactions of hyperthermal atomic oxygen with perfluorinated hydrocarbons using quantum-mechanical and molecular dynamics methods. Electronic structure calculations reveal that the reaction barriers for all of the primary reaction channels in the O(3 P) + CF4 and C2 F6 systems are larger than those in the analogous reactions with methane and ethane. F abstraction to give OF is not favored due to the large electronegativity of both F and O. The F elimination process, O(3 P) + Cn F2n+2 → F + OCn F2n+1 , is possible through a barrier of about 3 eV. This barrier is ∼1.2 eV larger than in the case of O(3 P) + Cn H2n+2 reactions. Direct C–C breakage in perfluorinated hydrocarbons proceeds through a barrier of about 2.5 eV, 1 eV larger than that in unsubstituted alkanes. Molecular dynamics calculations indicate the presence of additional dynamical barriers for reaction. This makes the reactivity of perfluorinated saturated alkanes to be very small even at 5 eV collision energy, in agreement with experiments. Key words: perfluorinated alkanes, hyperthermal atomic oxygen, C–C breakage, low Earth orbit erosion, molecular dynamics, quantum chemistry, fluoropolymer degradation 1. Introduction The mechanisms whereby erosion of fluoropolymers in low Earth orbit (LEO) takes place are currently not well understood. On-orbit experiments have detected degradation of thermal control materials composed of metallized TEFLONR FEP (fluorinated ethylene propylene) in the Hubble Space Telescope [1–4]. Groundstate atomic oxygen (O(3 P)) is the most abundant species in the LEO atmosphere [5], and is assumed to be the dominant factor in erosion of materials in LEO. The LEO reaction efficiency (volume of material removed per incident O atom) ∗ Current address: Department of Chemistry, Virginia Tech. 107 Davidson Hall, Blacksburg, VA, 24061-0212 USA

365 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 365–375. C 2006 Springer. Printed in the Netherlands.

366

DIEGO TROYA AND GEORGE C. SCHATZ

of TEFLONR FEP measured in the Evaluation of Oxygen Interactions with Materials III (EOIM-III) flight experiment (0.05 × 10−24 cm3 .atom−1 ) is ∼60 and ∼90 times smaller than the reaction efficiencies of Kapton and polyethylene, respectively [6]. These data are consistent with those of earlier missions [7]. The presence of ultraviolet (UV) and X-ray radiation, proton and electron irradiation, and thermal cycling likely contributes to the deleterious effects of hyperthermal O(3 P) on polymers in LEO, and this complicates the elucidation of the microscopic mechanisms responsible for LEO erosion. A number of ground-based experiments have tried to separate the contributions of atomic oxygen from the rest of environmental factors that might enhance TEFLONR FEP degradation in LEO. Studies of TEFLONR FEP erosion with thermal oxygen atoms (collision energy, E coll = 0.04 eV) revealed a reaction efficiency about 40 times smaller than that of polyethylene or Kapton [8]. Cazaubon et al. measured a reaction efficiency of 0.4 × 10−24 cm3 .atom−1 in an experiment that uses a beam of hyperthermal (E coll = 5 eV) atomic oxygen that contains some UV radiation [9]. The fact that this measured erosion rate is about one order of magnitude larger than that in LEO experiments implies that the dose of UV radiation might play an important role in the erosion process. Very recent experiments of near-thermal (E coll = 0.11 eV) O(3 P) collisions with semi-fluorinated alkanethiolate self-assembled monolayers (SAM’s) indicate that perfluorinated alkanes do not react with low energy ground-state atomic oxygen [10]. This is in sharp contrast to the large degradation observed with unsubstituted alkanethiolate SAM’s under identical initial conditions. Minton and coworkers have crossed a hyperthermal ground state atomic oxygen beam [11] with a beam of gas-phase perfluoropropane (C3 F8 ), and no signals associated with oxidation of the perfluoroalkane were detected at E coll = 4.6 eV [12]. On the other hand, oxidation and C–C breakage reactions of propane (C3 H8 ) with that hyperthermal O(3 P) beam have been recently described in detail [13]. FEP polymers were simultaneously exposed to UV light and hyperthermal atomic oxygen by Rasoul et al. [14]. Postirradiation surface analysis by X-ray photoelectron spectroscopy (XPS) revealed the presence of oxygen coordinated to the surface of the polymer. In addition, the F : C ratio slightly decreased (by about 15%) with respect to that of the pristine polymer. Silvered fluorinated propylene ethylene films were independently irradiated with 5 eV atomic oxygen, 10 MeV protons and UV light of wavelengths below 400 nm by Nakayama et al. [15]. Surface analysis of the exposed films indicated that whereas the effects of proton or UV light are negligible, 5 eV atomic oxygen degrades the fluorinated polymer. These experiments also revealed a decrease in the F : C ratio of the O(3 P)-exposed polymer with respect to the original one, and the presence of oxygen at the polymer surface. A series of linear hydrocarbon polymers, with F : C ratios ranging from 0 (polyethylene) to 2 (TEFLONR ) have been exposed in the absence of UV light to hyperthermal atomic oxygen (of undetermined collision energy) by Gonzalez et al. [16]. XPS analysis of the exposed polymers reveals that in the case of TEFLONR

THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 367

the F : C ratio of the eroded sample decreases ∼20% with respect to that of the pristine film, in agreement with the set of experiments by Rasoul et al. [14] and Nakayama et al. [15]. However, this last set of experiments disagrees with the other two in the fact that oxygen was not detected in the surface of the irradiated polymers [16]. Irrespective of this, the small decrease of the F : C ratio in all of the experiments suggests that the mechanisms of mass loss are not dominated by processes in which single F atoms are abstracted or knocked-out from the polymer, as these processes would give rise to more noticeable decreases in the F : C ratio. The unclear origins of the superior performance in LEO of fluorinated alkane polymer with respect to unsubstituted alkanes and the disagreement between different experiments indicate the need for further studies. In this paper, we present a theoretical study of the reactions of O(3 P) with fluorinated alkanes based on quantum-mechanical (QM) calculations and molecular dynamics simulations. First-principles QM calculations are developed to identify all of the relevant pathways that can lead to degradation of fluorinated alkanes in LEO environment. Classical trajectory calculations that use a high-accuracy semiempirical QM Hamiltonian are carried out to simulate the collision of O(3 P) with fluorinated species in LEO and derive information about the relative importance of the different open reaction channels and their dynamics properties. In order to gain additional insight, the results of the simulations developed here are compared with earlier calculations of the reactions of hyperthermal O(3 P) with unsubstituted alkanes [13, 17–21].

2. Electronic Structure Calculations First-principles molecular orbital calculations have been used to characterize reagents, products and connections between them in the reactions of O(3 P) with perfluoromethane (CF4 ) and perfluoroethane (C2 F6 ). We have used the density functional theory (DFT) hybrid functional B3LYP with the 6-31+G* basis set for these calculations. There are three main product channels in the reactions of O(3 P) with a short-chain fluorinated alkane: O(3 P) + Cn F2n+2 → OF + Cn F2n+1 , F abstraction O(3 P) + Cn F2n+2 → F + OCn F2n+1 , F elimination O(3 P) + Cn F2n+2 → OCm F2m+1 + Cn−m F2(n−m)+1 , C–C breakage Table 1 shows the energetics of these reactive processes. Structures of the saddle points involved in the different O(3 P) + C2 F6 reaction pathways are plotted in figure 1. The main differences between O(3 P) reactions with short-chain alkanes and their perfluorinated analogues is the large endothermicity and barrier associated with the abstraction of a fluorine atom. Whereas H abstraction by O(3 P) to give OH

368

DIEGO TROYA AND GEORGE C. SCHATZ

TABLE 1. Calculated reaction energies and barriers (in eV) for the O(3 P) + CF4 , CH4 , C2 F6 , and C2 H6 reactionsa Reaction Energy O(3 P) + CX4 (X = F, H) OF + CF3 2.966

OH + CH3 0.161

F + OCF3 1.168

H + OCH3 0.548

O(3 P) + C2 X6 (X = F, H) OF + C2 F5 2.549

OH + C2 H5 −0.061

F + OC2 F5 0.831

H + OC2 H5 0.368

OCF3 + CF3 −0.534

OCH3 + CH3 −0.234

O(3 P) + C2 F6 CF3

O(3 P) + C2 H6 CH3

2.523

1.577

Reaction Barrier O( P) + CX4 (X = F, H) 3

O(3 P) + CF4 →OF + CF3

O(3 P) + CH4 →OH + CH3

O(3 P) + CF4 →F + OCF3

3.205

0.240

3.035

O(3 P) + CH4 →H + OCH3 1.778

O(3 P) + C2 X6 (X = F, H) O(3 P) + C2 F6 →OF + C2 F5 2.795 a

O(3 P) + C2 H6 →OH + C2 H5 0.069

O(3 P) + C2 F6 →F + OC2 F5 3.149

O(3 P) + C2 H6 →H + OC2 H5 1.813

B3LYP/6-31+G* calculations. The energies are zero-point corrected.

Figure 1. Schematic representation of the structures and energy profile of the primary reaction pathways involved in O(3 P) + C2 F6 collisions at hyperthermal energies

THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 369

is approximately thermoneutral and proceeds through a barrier well below 0.5 eV, more than 2.5 eV are needed to generate OF in O(3 P) collisions with fluorinated alkanes. O and F are very electronegative species, and the formation of a bond between them is not favored. It can be also observed that the F elimination reaction (which yields F + OCn F2n+1 ) is about 0.5 eV more endoergic than for the corresponding H elimination reaction. Also, the barriers for F elimination are 1.2 eV larger than those for H elimination. The products generated in the C–C breakage reaction of O(3 P) with C2 F6 are thermodynamically more stable than those of the analogous reaction with C2 H6 . However, perfluorination leads to a 1 eV increase in the C–C breakage barrier height. The fact that C–F bonds (dissociation energy, De = 4.93 eV at the B3LYP/631+G* level for C2 F6 ) are slightly stronger than C–H bonds (De = 4.64 eV at the same level of theory for C2 H6 ) does not fully account for the large differences in the barrier heights between fluorinated and unsubstituted alkanes noted in table 1 for all of the primary processes. We have just discussed the reasons for the unfavourable formation of OF. As it can be seen in the saddle point structures in figure 1, F elimination or C–C breakage are produced when the incoming O(3 P) atom forms a bond with the carbon atoms of the alkane backbone. In the case of unsubstituted alkanes, H atoms bonded to the hydrocarbon chain do not hinder significantly the approach of O(3 P) to the C atoms of the chain. However, the larger size of F atoms induces repulsive interactions with the electronic cloud of the incoming O(3 P). This shielding effect makes it more difficult for O(3 P) to reach the C atoms in fluorinated alkanes and leads to an increase in the reaction barriers for F elimination and C–C breakage that is larger than what would be expected based on bond-strength arguments alone. The C–C breakage saddle points for O(3 P) + perfluorinated alkane reactions have been studied before by Massa and coworkers [22]. However, it should be noted that the saddle point reported in that work for the O(3 P) + C2 F6 reaction is not the lowest energy one. Instead, it corresponds to a higher-energy reaction pathway.

3. Reaction Dynamics Calculations We have performed direct-dynamics classical trajectory calculations of collisions of hyperthermal O(3 P) with CF4 and C2 F6 to gain deeper insight into the dynamics of the degradation processes of fluoropolymers in LEO. In our direct-dynamics calculations, the energy gradients needed at each integration step to solve the Hamiltonian equations of motion are obtained from quantum-mechanical calculations. Ab initio or DFT calculations of the energy gradients are unwieldy at this time, and we have resorted to use the MSINDO semiempirical Hamiltonian

370

DIEGO TROYA AND GEORGE C. SCHATZ

[23]. This QM technique has proven useful in earlier calculations of O(3 P) + short-chain saturated alkanes [13, 17–19] and self-assembled monolayers [20, 21]. For the reactions under study, restricted open-shell (ROHF) calculations using MSINDO adequately reproduce the B3LYP/6-31+G* data. For instance, the ROHF/MSINDO barrier for the O(3 P) + CF4 → F + OCF3 reaction (2.936 eV), is in excellent agreement with B3LYP/6-31G* calculations (3.035 eV). In the case of reactions with C2 F6 , ROHF/MSINDO underestimates the B3LYP/6-31+G* F elimination and C–C breakage reaction barriers by about 0.5 eV. Batches of 10,000 trajectories were integrated for the O(3 P) + CF4 and O(3 P) + C2 F6 systems with E coll = 4.5, 5.0, 5.5, and 6.0 eV. We used a time integration step of 10 a.u. The maximum sampling impact parameter was 3.5 a.u for O(3 P) + CF4 and 4.0 a.u. for O(3 P) + C2 F6 . Trajectories are started (stopped) at ∼12 a.u. separation distance between reagents (products). The initial coordinates and momenta of the reagent molecules are sampled from zero-point motion. The calculations reveal that the abstraction reaction pathways (which yield OF + Cn F2n+1 products) are negligible in both O(3 P) + CF4 and O(3 P) + C2 F6 . The number of reactive trajectories giving OF found in O(3 P) + C2 F6 collisions at E coll = 4.5, 5.0, 5.5, and 6.0 eV is 0, 1, 1 and 3, respectively. For O(3 P) + CF4 only 1 reactive trajectory to OF + CF3 was found, and this was at the largest collision energy explored, E coll = 6.0 eV. Whereas the H abstraction reaction to generate OH is the dominant product channel in collisions of hyperthermal O(3 P) with C2 H6 , perfluorination almost completely suppresses the formation of alkyl radicals by F abstraction. Notwithstanding that from an energetic point of view F abstraction by O(3 P) only requires ∼3 eV, we see that even at E coll = 6 eV this channel is negligible. The reasons for this are tied to the shape of the potential energy surface. As it can be seen in figure 1, the saddle point connecting reagents and products has a ‘late’ character (i.e., the transition state is very similar ˚ from the C atom to products, with the F atom that is abstracted being at 2.1 A that it is bonded to in reagents). It is well-known that in reactions with ‘late’ transition states the reagent translational energy is not effective in surmounting the barrier, such as we see here for the O(3 P) + CF4 → OF + CF3 and O(3 P) + C2 F6 → OF + C2 F5 reactions. Stretching of C–F bonds will on the other hand promote F abstraction, as this motion is strongly coupled to the abstraction reaction coordinate. This suggests a strong increase in the cross section for F abstraction for vibrationally excited fluoroalkanes. Therefore, in the future it will be interesting to investigate the effect of the substrate temperature on the cross section. Figures 2 and 3 display the cross sections for the F elimination reactions, 3 O( P) + CF4 → F + OCF3 and O(3 P) + C2 F6 → F + OC2 F5 , respectively, in comparison with the corresponding data for hydrogenated hydrocarbons. In the case of the reaction involving CF4 , F elimination is the only primary reaction channel other than F abstraction. Figure 2 shows that although F elimination is more important than F abstraction (which is negligible), the cross sections for F

THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 371 2 3

O( P) + CF4 -> F + OCF3 (F elimination) 3

O( P) + CH4 -> H + OCH3 (H elimination)×0.1

cross section / a.u.

1.5

×0.1

1

0.5

0

4.5

5

Ecoll / eV

5.5

6

Figure 2. Excitation functions (cross sections vs Ecoll ) for the O(3 P) + CX4 → X + OCX3 (X = F, H) reactions. The cross sections of the H elimination reaction are divided by 10

elimination are very small. We carried out additional calculations to locate the threshold of reactivity for this channel, and 1 reactive trajectory was found out of 10,000 trajectories calculated at E coll = 4.0 eV. This suggests that the lowest collision energy for F elimination is about 1 eV larger than the energy of the saddle 1 3

O( P) + C2F6 -> OCF3 + CF3 (C-C breakage) 3

O( P) + C2F6 -> F + OC2F5 (F elimination)

0.8

3

cross section / a.u.

O( P) + C2H6 -> OCH3 + CH3 (C-C breakage)×0.1 3

O( P) + C2H6 -> H + OC2H5 (H elimination)×0.01

0.6

0.4

×0.1

0.2

×0.01

0

4.5

5

Ecoll / eV

5.5

6

Figure 3. Excitation functions (cross sections vs Ecoll ) for the O(3 P) + C2 X6 → X + OC2 X5 (solid lines) and the O(3 P) + C2 X6 → OCX3 + CX3 (dotted lines) (X = F,H) reactions. The cross sections of the X = H reactions have been divided by 100 and 10, respectively

372

DIEGO TROYA AND GEORGE C. SCHATZ

point for this process (∼3 eV, see table 1). Although the ROHF/MSINDO barrier for F elimination in the O(3 P) + C2 F6 reaction is about 0.5 eV smaller than that in O(3 P) + CF4 , the size of the F elimination cross section in O(3 P) + C2 F6 remains quite small. In fact, the F elimination cross section in O(3 P) + C2 F6 is much smaller than that in O(3 P) + CF4 (a factor of ∼6 at E coll = 6 eV). The threshold for reactivity is also well above the minimum energy saddle point located for this reaction pathway. Thus, much as it happens for the F abstraction reaction, the features of the potential energy surface are such that there is a dynamical barrier to reaction induced by the inefficacy of collision energy to explore geometries close to those of the minimum energy saddle point. F abstraction or elimination reactions require a large elongation of the C–F bond that breaks. Whereas vibrational excitation of the C–F bond would promote reactivity, a fast O(3 P) atom is not able to efficiently promote the C–F stretching needed for reaction. In addition, in collisions in which a substantial fraction of the momentum of the incoming O(3 P) is transferred to C–F stretching, motions other than those leading to reaction are strongly coupled and can inhibit reaction, relaxing the instantaneous excitation of the C–F stretch to other modes. This is known as “frustrated collisions” and has been described before in studies of H with vibrationally excited HF and H2 O [24–26]. The cross sections for the C–C breakage reaction, O(3 P) + C2 F6 → OCF3 + CF3 , are displayed in figure 3 in comparison with the results for O(3 P) + C2 H6 . No attempts to accurately locate the threshold for reactivity were made in this case due to the large computational expense involved. However, it can be inferred from the small cross section found at E coll = 4.5 eV, that the minimum collision energy leading to C–C breakage is well above the energy of the minimum-energy saddle point located for this reaction. Therefore, dynamical barriers seem to be present in all of the reactions of O(3 P) with perfluorinated alkanes investigated in this work. It is remarkable that the C–C breakage cross sections in the O(3 P) + C2 F6 reaction is larger than the F elimination cross section. This behavior is opposite to that of unsubstituted alkanes [13, 18, 20, 21]. For instance, in the O(3 P) + C2 H6 reaction at E coll = 5.75 eV, the H elimination: C–C breakage cross-section ratio is 4.00 (16.64 : 4.16 a.u.). However, in the O(3 P) + C2 F6 reaction the F elimination: C–C breakage cross section ratio is 0.12 at E coll = 5.5 eV and 0.24 at E coll = 6.0 eV. The H elimination and C–C breakage reaction barrier heights in O(3 P) + C2 H6 are very similar. Thus the preference for H elimination over C–C breakage comes from the fact that when O(3 P) strikes one of the carbon atoms of the ethane molecule, 3 C–H bonds can break leading to H elimination, but only 1 C–C bond can break leading to C–C breakage. In O(3 P) + C2 F6 the barrier height for C–C breakage is slightly smaller than that of F elimination. Also, the C–C bond is in this case the most labile bond in C2 F6 (De = 3.66 eV, compared to De = 4.93 eV for a C–F bond at the B3LYP/6-31+G* level) and this is why C–C breakage is

THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 373

more important than F elimination. It is interesting to notice that the C–C bond in C2 F6 is 0.4 eV weaker than the C–C bond in C2 H6 (De = 4.07 eV at the B3LYP/631+G* level). This is a consequence of electronic density removal from the C–C bond by the very electronegative fluorine atoms. Notwithstanding the larger importance of C–C breakage over F elimination in O(3 P) + C2 F6 collisions at hyperthermal energies, the size of the C–C breakage cross-section at E coll = 5.0 eV is one order of magnitude smaller than that in O(3 P) + C2 H6 . The F elimination cross section is about two orders of magnitude smaller than the H elimination one at the same collision energy. Furthermore, the total cross section for reaction in O(3 P) + C2 F6 is 0.14 a.u. at E coll = 5.0 eV. This is to be compared with a 39.3 a.u. cross-section for all of the reactive processes in O(3 P) + C2 H6 (including, H abstraction, H elimination and C–C breakage).

4. Concluding Remarks Our calculations indicate that fluorinated alkanes are much less susceptible to reaction than hydrogenated alkanes upon collisions with hyperthermal O(3 P), with the O(3 P) + C2 H6 / O(3 P) + C2 F6 ratio of total reactive cross sections at 5.0 eV being 39.3/0.14 = 270. This is to be compared with the EOIM-III on-orbit experimental measurements that reported a reaction efficiency for polyethylene 88 times larger than that of TEFLONR FEP [6]. Our calculations are also in agreement with the gas-phase crossed-beams experiments of the Minton group that indicate no oxygen-containing products in the O(3 P) + C3 F8 reaction at E coll = 4.6 eV [12], but substantial reactivity for hyperthermal collisions of O(3 P) with C3 H8 [13]. The reasons for a decrease in the reactivity of alkanes with respect to hyperthermal O(3 P) upon fluorination are manifold —C–F bonds in fluorinated saturated alkanes are 10% stronger than C–H bonds in unsubstituted species. —The barriers for C–C breakage or F elimination in fluorinated alkanes are about 1 eV larger than the corresponding barriers in unsubstituted alkanes as a consequence of the larger size of F atoms, which inhibits the approach of O(3 P) to the carbon backbone, —H atom abstraction by OH is dominant in O(3 P) + alkane reactions, but is negligible in O(3 P) + fluorinated alkane reactions. —Additional dynamical barriers arise due to the inefficacy of collision energy to provide geometries close to those of the first-order saddle points for reaction. The results of our calculations can also contribute to the understanding of XPS experiments of surface composition after hyperthermal atomic oxygen exposure [14–16]. In all of these experiments it was found that the F : C ratio of the

374

DIEGO TROYA AND GEORGE C. SCHATZ

irradiated polymer was only slightly smaller than that of the pristine polymer. This suggests that mass removal could involve ejection of CF2 units into the gas phase. This is in contrast with the microscopic mechanisms characterized in erosion experiments of unsubstituted hydrocarbon polymers by O(3 P), in which CO, CO2 and H2 O are the volatile species [27]. Our calculations show that O addition to fluoropolymers is only minor and the dominant erosion process is C–C breakage. Therefore we can tentatively infer that erosion of fluorinated polymers mainly involves C–C breakage with formation of volatile OCF2 products that scatter into the gas phase. Although OCF species have not been found in O(3 P) + C3 F8 collisions at E coll = 4.6 eV [12], larger collision energies in the beams that have been used to study TEFLONR erosion, together with the presence of UV light or ions and defects in the exposed fluoropolymers would facilitate OFC species formation. Future calculations involving condensed-phase fluorinated alkanes, such as self-assembled monolayers, will help to verify this. Time-of-flight detection of the scattered species in a hyperthermal O(3 P) steady-state bombardment experiment of TEFLONR would directly provide information about the microscopic mechanisms of fluoropolymer erosion in LEO.

Acknowledgments This research was supported by AFOSR MURI GrantNo. F49620-01-1-0335 and NSF Grant No. CHE-0131998. The authors wish to thank Timothy K. Minton (Montana State University) for sharing experimental data prior to publication and for his continuing encouragement.

References 1. de Groh, K. K., Gaier, J. R., Hall, R. L., Espe, M. P., Cato, D. R., Sutter, J. K., and Scheiman, D. A. (2000) Journal of High Performance Polymers 12 (1), 83. 2. Dever, J. A., de Groh, K. K., Banks, B. A., Townsend, J. A., Barth, J. L., Thomson, S., Gregory, T., and Savage, W. (2000) Journal of High Performance Polymers 12 (1), 125. 3. de Groh, K. K., Dever, J. A., Sutter, J. K., Gaier, J. R., Gummow, J. D., Scheiman, D. A., and He, C. (2001) Journal of High Performance Polymers 13, S401. 4. Dever, J. A., de Groh, K. K., Messer, R. K., McClendon, M. W., Viens, M., Wang, L. L., and Gummow, J. D. (2001) Journal of High Performance Polymers 13, S373. 5. Jursa, A. (1976) U. S. Standard Atmosphere. (U.S. Government Printing Office, Washington D. C., 1976). 6. Koontz, S. L., Leger, L. J., Visentine, J. T., Hunton, D. E., Cross, J. B., and Hakes, C. L. (1992) Journal of Spacecraft and Rockets 32(3), 483. 7. Leger, L. J. and Visentine, J. T. (1986) Journal of Spacecraft and Rockets 23(5), 505. 8. Koontz, S. L., Albyn, K., and Leger, L. J. (1991) Journal of Spacecraft and Rockets 28(3), 315. 9. Cazaubon, B., Paillous, A., and Siffre, J. (1998) Journal of Spacecraft and Rockets 35(6), 797.

THEORETICAL STUDY OF REACTIONS OF HYPERTHERMAL O(3 P) 375 10. Wagner, A. J., Wolfe, G. M., and Fairbrother, D. H. (2004) Journal of Chemical Physics 120(8), 3799. 11. Garton, D. J., Minton, T. K., Maiti, B., Troya, D., and Schatz, G. C. (2003) Journal of Chemical Physics 118, 1585. 12. Minton, T. K., private communication (2004). 13. Garton, D. J., Minton, T. K., Troya, D., Pascual, R., and Schatz, G. C. (2003) Journal of Physical Chemistry A 107(23), 4583. 14. Rasoul, F. A., Hill, D. J. T., George, G. A., and O’Donnell, J. H. (1998) Polymers for Advanced Technologies 9, 24. 15. Nakayama, Y., Imagawa, K., Tagashira, M., Nakai, M., Kudoh, H., Sugimoto, M., Kasai, N., and Seguchi, T. (2001) Journal of High Performance Polymers 13, 2001. 16. Gonzalez, R. I., Phillips, S. H., and Hoflund, G. B. (2003) Journal of Applied Polymer Science 92 (3), 1977. 17. Troya, D., Pascual, R. Z., and Schatz, G. C. (2003) Journal of Physical Chemistry A 107 (49), 10497. 18. Troya, D., Pascual, R. Z., Garton, D. J., Minton, T. K., and Schatz, G. C., Journal of Physical Chemistry A 107(37), 7161. 19. Troya, D., Schatz, G. C., Garton, D. J., Brunsvold, A. L., and Minton, T. K. (2004) Journal of Chemical Physics 120(2), 731. 20. Troya, D. and Schatz, G. C. (2003) European Space Agency Special Publication SP-540, 121. 21. Troya, D. and Schatz, G. C. (2004) Journal of Chemical Physics 120, 7696 . 22. Gindulyte, A., Massa, L., Banks, B. A., and Miller, S. K. R. (2002) Journal of Physical Chemistry A 106, 5463. 23. Ahlswede, B. and Jug, K. (1999) Journal of Computational Chemistry 20(6), 563. 24. Schatz, G. C. and Kuppermann, A. (1980) Journal of Chemical Physics 72(4), 2737. 25. Lendvay, G., Bradley, K. S., and Schatz, G. C. (1999) Journal of Chemical Physics 110, 2963. 26. Barnes, P. W., Sims, I. R., Smith, I. W. M., Lendvay, G., and Schatz, G. C. (2001) Journal of Chemical Physics 115, 4586. 27. Minton, T. K. and Garton, D. J. (2001) In Advances Series in Physical Chemistry Vol 11: Chemical Dynamics in Extreme Environments, R. A. Dressler (ed.) (World Scientific, Singapore, 2001), pp. 420.

SIMULATION OF UV INFLUENCE ON OUTGASSING OF POLYMER COMPOSITES

R. H. KHASSANCHINE, A. N. GALYGIN, A. V. GRIGOREVSKIY, AND A. N. TIMOFEEV Joint-stock company “Kompozit,” 4, Pionerskaya str., Korolev, Moscow region, Russia

Abstract. Aspects concerning influence of UV radiation on outgassing processes in vacuum for acrylic copolymer-based composites are discussed. Mathematical models describing the outgassing processes in materials being subjected to UV radiation as well as deposition of emerged volatile products are given. Comparative analysis of experimental results of UV action on outgassing kinetics for materials with different filler-binder ratios at different temperatures is carried out. Key words: outgassing, polymeric composite, UV radiation, desorption, diffusion, remission

1. Introduction Outgassing of a polymeric composite subjected to ultraviolet (UV) radiation in vacuum involves the following basic processes: desorption of volatile products (VP) adsorbed or generated on the surface of the material; diffusion and desorption of gaseous substances adsorbed by material or generated in it due to UV radiation; evaporation of material being the result of UV action. The influence of UV radiation on outgassing processes in polymeric composites due to photodecomposition of polymeric components may drastically exceed those induced by exposure to solar radiation. This is stipulated by a number of features of the photochemical processes that occur in materials under exposure to UV, and that the major part of VUV-radiation is absorbed by the thin near-surface layer of material. The latter can cause the “heavy” in mass but easily surfacecondensable VP to be moved from the thin layer of a polymeric composite into the gaseous phase.

377 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 377–383. C 2006 Springer. Printed in the Netherlands.

378

R. H. KHASSANCHINE ET AL.

2. Mathematical Model To describe mathematically the influence of UV on physical and chemical processes that occur in a material and on its surface, we have made use of principal postulates accepted in the thermal-vacuum outgassing model [1, 2]. The concentration change of outgassing components in a hermetically enclosed substrate is stipulated by the following processes: surface desorption at the material-vacuum interface; photodestruction of the material; chemical reactions; evaporation of the material, and diffusion resulting from the aforesaid processes. Mathematical models of outgassing processes in composites use effective coefficients, i.e., parameters describing the processes that take place in laboratory and on-board experiments. With these assumptions, the change of concentration of i-type outgassing component Ci (x, t) (i = 1, 2, 3 . . . N ) in a material being in vacuum under exposure to UV-radiation can be described by the following differential equations: M ∂ 2 Ci (x, t) ∂Ci (x, t) photo = Di − σi→m + χi Ci (x, t) ∂t ∂x2 m=1 + Aoi exp[α photo (x + v · t − h)], x ∈ [0, h − v · t] , t > 0,

(1)

that satisfy the initial and boundary conditions: Ci (x, t) |t=0 = Ri , Di

x ∈ [0, h]

(2)

∂Ci (x, t) ∂Ci (x, t) |x=h−v·t + (ki + a0 ) Ci (x, t)|x=h−v·t = |x=0 = 0, t > 0, ∂x ∂x (3) photo

where σi→m —weighting coefficient of photodestruction of i-type component through j-channel, sec−1 ; χ i —chemical reaction rates with involvement of i-type component, sec−1 ; Di, ki —effective diffusion and desorption coefficients of i-type component respectively; Ri —concentration of i-type component in material at initial moment, gμ−1 m−3 × 10−12 ; h—thickness of material, m; v—evaporation rate of material, m·s−1 ; A0i —parameter that depends on composition of material under test and UV source; α photo —effective coefficient of linear reduction of UV radiation, a0 —coefficient that allows for the effect of radiation on processes in the near-surface layer with involvement of i-type component, m·s−1 The “heavy” outgassing components are not able to diffuse through the material. Therefore it is sufficient to examine the dynamics of change of their concentrations in the near-surface layer (x = h − v · t) that could be described with the

SIMULATION OF UV INFLUENCE

379

help of the equation dCi∗ (h − v · t, t) = −ki∗ Ci∗ (h − v · t, t) + Si∗ (t), dt

Ci∗ t=0 = Ri∗ ,

(1 )

where ki∗ —effective desorption coefficient of i-type “heavy” component; Si∗ (t), Ri∗ —source function and initial concentration of i-type “heavy” component in the near-surface layer of material. photo Functions Ci (x, t) obtained while solving the eqs. (1)–(3) with σi→m weighting coefficients may enter the equations describing the changes of concentration of m-type outgassing component (m = 1, 2, 3 . . . M) that could be generated in the course of destruction of proper i-type component. Searching in such a way all feasible outgassing components one may create a system of equations describing changes of their concentrations in the material. Having found the distribution Ci (x, t) in a sample from eqs.(1) to (3), the dependencies of VP mass in the material and the flux through the unit surface of the material-vacuum boundary on time t are respectively determined from the following expressions h−v·t

Msi (t) = S0

Ci (x, t) d x, 0

t Fi (t) =

(v + ki + a0 ) Ci (h − v · τ, τ )dτ, 0

(4) where Ci (h − v · t,t)—concentration of i-type VP in the near-surface layer; S0 — surface area of a sample of material (source). To study outgassing in vacuum, quartz microbalances are used that convert the changes of mass added onto the quartz piezoresonator surface into alteration of output frequency of self-oscillator. Thus, the outgassing process is monitored by observing the deposition of VP onto the surface of the sensitive mass loss sensor. Therefore, to interpret the experimental data we have to introduce the model that binds the mass loss rate dMsi (t)/dt and the deposition rate of part of this mass dMci (t)/dt on the condensation surface. The latter can cause the “heavy” in mass but easily surface-condensable VP to be moved from the thin layer of a polymeric composite into the gaseous phase d Msi (t)/dt = −(v + ki + a0 )S0 Ci (h − v · t, t) d Mci (t)/dt = −αcs d Msi (t)/dt − kci Mci (t) − χci Mci (t),

(5) (6)

where kci —effective coefficient of re-emission of i-type VP from the condensation surface; χ ci —chemical reaction rates with involvement of i-type component on the condensation surface; α cs geometrical factor that depends on arrangement of VP source with respect to the unit of the condensation surface.

380

R. H. KHASSANCHINE ET AL.

As for the laboratory experiments on VP outgassing and deposition where alterations in thickness of VP source may be neglected, solution of the system (5)–(6) for Mci (t) has the following form: ∞ f n + gn Mci (t) = αcs S0 (ki + a0 ) [exp (−bi t) − exp(−ηci t)] ηci − bi n=1 gn (1 − exp (−ηci t)) , (7) + kci 2Ri k i + a0

; tgλn h = ; where f n = 2 λn Di 1 + h (ki + a0 )/Di + λn Di /(ki + a0 ) λn sin 2λn h + 2α photo · cos λn h [cos λn h − exp (−α photo · h)] ; h · (α photo2 + λ2n ) · bi M photo bi = λ2n Di + βi ; βi = σi→m + χi ; ηci = kci + χci . (8)

gn = A0

m=0

The outgassing rate of VP source and deposition and re-emission rates on/from the condensation surface respectively are defined mainly by the constants in eqs. (7) and (8). The total mass of VP deposited on the condensation surface can be determined by summing masses of individual components: Mc total (t) =

N

[Mcn (t) + n (t, χcn )],

(9)

n=1

where n (t, χcn )—n-type part of VP mass that has chemically reacted on the condensation surface by time t; N —number of VP components.

3. Results of Numerical Analysis Figure 1 shows plots of space-time dependences of VP concentrations in polymeric composites for the following case. Prior to t0 = 400 h outgassing in vacuum occurs with equal effective diffusion and desorption coefficients, initial concentrations of VP in material and β—generalized first-order rate of reactions involving the specified type of VP. At the moment t0 , as shown in figure 1(a)—temperature of material increases stepwise, 1(b)—UV source activates. With temperature growth the outgassing rate increases due to growth of effective diffusion and desorption coefficients that leads to quicker reduction of VP concentration in material. Increase of outgassing rate under the UV action accompanying

381

SIMULATION OF UV INFLUENCE

Figure 1. Space-time dependences of VP concentrations in material: (a)—stepwise increase of temperature of material, (b)—actuation of UV source at t = 400 h

with photodestruction of material is tied with increase of VP concentration in it. 4. Experimental Set-Up Figure 2 shows a schematic diagram of the “Vesy” equipment that was used to study influence of UV radiation on outgassing kinetics at different sample temperatures. 1

2 3 7 170 mm

70 mm

4 5

6

Figure 2. Schematic diagram of the “Vesy” equipment: (1)—reservoir that is filled with liquid nitrogen to control thermostatically the mass flow sensor; (2)—vacuum chamber body; (3)—cryogenic shield; (4)—mass flow sensor; (5)—heating table with sample; (6)—UV sources; (7)—quartz window

382

R. H. KHASSANCHINE ET AL.

Figure 3. Spectral density of the source of UV radiation

To determine the outgassing rates of materials in vacuum, we have used the quartz microbalance with the 10 MHz AT-cut mass-sensitive piezoresonator thermally stabilized at (82 ± 5)◦ K. A sample was attached to the table in the vacuum chamber. Experiments were carried on at pressure no greater than (2 ± 0.5) × 10−4 Pa. The DDS-30 deuterium lamp with spectral density shown in figure 3 was used as the source of UV radiation. This lamp was mounted inside the vacuum chamber so that the whole sample surface of the VP source material was irradiated uniformly.

5. Results and Discussion Experimental results of influence of UV radiation on outgassing of polymeric composites are given in figure 4. Data processing shows that outgassing rate in polymeric materials under exposure to UV radiation depends on both the temperature and the polymeric binder/filler (ZnO in our case) volume ratio. Plots (a) and (b) in figure 4 present polymeric binder/filler volume ratios of 1:5.5 and 1:1 respectively. Plot 4(c) shows experimental data of ECOM-1 thermal coating3 mass loss due to outgassing process. Both samples prior to t = 8 h are subjected to thermal-vacuum action. Then one of them was subjected to UV-radiation (curve2) that lead to a significant increase of mass loss in the outgassing process as compared to mass loss (curve-1) of the sample that was not exposed to UV radiation. Finally, plot 4(d) gives a computational curve showing change of outgassing rate when activating UV radiation at the moment t = 24 h. The results of numerical calculations of the outgassing process when polymeric composites are subjected to UV radiation show a significant dependence of the amount of VP emerged from the exposed materials under study both on

SIMULATION OF UV INFLUENCE

383

Figure 4. Experimental curves of outgassing kinetics: (a; c)—ECOM-1; (b)—model material with equal binder/filling agent volumes; (d)—influence of actuation of UV source at t = 24 h on outgassing rate

the basic parameters characterizing the material itself and on the emission source spectrum. Experimental data show that the outgassing rate depends on the temperature of the material and the volume ratio of the polymeric binder/filling agent in the case when polymeric materials are subjected to UV radiation.

References 1. Khassanchine, R. H., Grigorevskiy, A. V., and Galygin, A. N. (2002) Some aspects of simulation of outgassing processes under thermal vacuum exposure to coatings applied to space vehicles, Protection of Materials and Structures from Space Environment, ICPMSE-6, Toronto-Canada, May 2002. 2. Kostiuk, V. I. and Khassanchine, R. H. (2002) Journal of Cosmonautics and Rocket Engineering, 28, 155–163. 3. Grigorievski, A., Gordeev, J., Gurov, A., Kiseleva, L., and Shuiski, M. (2000) In Proceedings 8th International Symposium, Material in a Space Enviroment and 5th International Conference, Protection of Materials and Structures from Space Environment, Arcachon, France, 5–9 June, 2000 pp. 81–87.

THE IMPACT OF HIGH-VELOCITY PARTICLES ON THERMAL PIPELINES IN SPACECRAFT

SEMKIN N. D., VORONOV K. E., AND NOVIKOV L. S. Samara State Aerospace University, Samara

Abstract. During long-duration space flights spacecraft components are exposed to space environment factors. The performance of Thermal Pipelines (TP) and their diagnostics (control of functional integrity and TP emission power) and simulating high-velocity particle interaction with MIMS-structure, as part of a TP wall, are considered. The simulation experiments of high-velocity particles interaction with TP using various types of accelerators are described.

1. Introduction During long-duration space flights spacecraft components and systems are exposed to the deep space factors. High-velocity particles (space debris, micrometeorites) present the greatest danger to high-temperature thermal pipelines (TP), which are important elements of the on-board installation. Small particles lead to changes in optical characteristics, while large ones cause depressurization of TP with a resulting release of the heat-carrier into open space, with the contamination of the space environment becoming one of dangerous consequences. The performance of Thermal Pipelines and their diagnostics (control of functional integrity and TP emission power) are discussed in this paper together with experiments simulating high-velocity particle interaction with Metal-InsulatorMetal-Semiconductor (MIMS) structures being a part of the TP wall design. Experiments simulating the high-velocity particle interaction with TP using various types of accelerators were conducted and are described in this paper.

2. Electric Simulation of Thermal Pipes Interaction of high-velocity particles with a TP wall may result in perforation of MIMS-structure (figure 1) that results in an increased flow of heat-carrier during the perforation. The voltage applied to the structure restores it, if necessary, and can serve as a diagnostic parameter for detection of such particles. As a result of the perforation of the TP wall, the TP system is losing its hermeticity, which 385 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 385–392. C 2006 Springer. Printed in the Netherlands.

386

SEMKIN N. D. ET AL.

Figure 1. Design of thermal pipe (TP). 1—TP wall (thickness 0.3–0.5 mm, material: Nb) 2—dielectric (Al2 O3 )—thickness 0.1 mm 3—black cover (lanthanum manganate), thickness 0.1 mm 4—metallic film (metal-dielectric structure coat), thickness 0.05 mm 5—radiation power meter 6—sensor of permeability control 1–4—MIMS-structure (general view).

in turn, is causing the TP wall temperature as well as the temperature of the TP system to change. Taking into account both the dependence of the current through a dielectric (for example, A12 O3 ) on temperature and its time dependence, it is possible to establish the type of damaging interaction and to obtain information on particle parameters. According to the equivalent electrical diagram of TP (figure 2), the expression for the channel resistance of a MIM-structure to penetration ( through layers 1-2-4, figure 1) is obtained RX =

3U a R L Rd 1 U b R L Rd − − R bc 2 U Rd − (Rd + R L )U a U Rd − (Rd + R L )U b

Figure 2. Electrical model of MIM-structure perforation

THE IMPACT OF HIGH-VELOCITY PARTICLES

387

where U is the basic voltage, Rd is the dielectric resistance, R L is the limiting resistance that limits the current of the basic voltage, Rbc —resistance of black cover, Ua , Ub —voltages that can be determined in time of penetration from two measuring points (a and b as marked in figure 1 and 2), R1, R2—resistance of the black cover layer (figure 2). In case of TP cooling, as a result of its depressurization and escape of the heat-carrier into space, the change of resistance of the black cover layer may help to determine the average power on the TP. The emissive power is calculated according to the Stephan–Boltzmann formula. 3. Models of TP Perforation Caused by High-Velocity Particles Various versions of interaction of a high-speed particle with a MIM-structure as part of a TP wall may be considered: 1. Crater formation on the external surface of the TP wall (black cover) without damaging the dielectric and internal metal wall. 2. Crater formation in the black cover and the dielectric without mechanical locking in a MIM-structure. 3. Crater formation in a MIM-structure with mechanical locking of the latter, without depressurization of TP. 4. Penetration through a TP wall, forming a perforation. The first three cases do not change the operational conditions of TP. The first two cases are less important than the case of mechanical locking of a TP wall during full-scale operation conditions. In case of a penetration through TP wall, the heat-carrier (liquid and gaseous) escapes. Let us assume, that after penetration through a TP wall in time t (time of particle passing through a TP wall) the closed MIM-structure will be restored. Because of the pressure difference the heat-carrier in the form of sodium vapour (or liquid) can escape through the formed perforation. In this case we assume that: 1. Vapor escapes without condensation on the formed channel walls, so that the MIM-structure is not short-circuited. 2. Vapour escapes until the heat-carrier mass is used up completely. 3. The thermal pipe cools down after the escape of the heat-carrier due to a decreased thermo-emission from the heat source. According to this model it is possible to obtain the equation connecting time of the heat-carrier escape (figure 3) and the cross-section of the perforation S0 . te =

M , j · S0

(1)

388

SEMKIN N. D. ET AL.

Figure 3. TP Cooling model

where M is the mass of the heat-carrier, j is the density of the vapour stream. Assuming the section of the hole is S0 = 10−2 cm2 , then te ≥ 2 hours. Let us estimate the time of TP cooling after the escape of the heat-carrier, assuming the stream from the heat source q = 0. Then the equation describing TP cooling will look like: d CmT = −δT 4 S, dt

(2)

where Cm is an average thermal capacity of a niobium TP wall, m is the mass of Niobium, S is the area of the lateral surface. The initial condition for eq. (2) is: at t = 0, T = 1150 0 K. Integrating eq. (2), we shall obtain the dependence of

THE IMPACT OF HIGH-VELOCITY PARTICLES

389

Figure 4. Current vs. Temperature dependence for TP with various black covers

temperature of TP wall cooling: T0 T (t) = , 3t 3 1+ t0

(3)

0 where t0 = δCmT is a typical time of cooling. T0 4 S During TP cooling, the resistance Rbc of the black cover and Rd of the dielectric (figure 4) are increased. Figure 4 shows experimental dependence of the dielectric and black cover currents Ibc and IL of a thermal pipe model on temperature. Since the temperature dynamics during the vapour escape (T = T0 = const) and after the escape are different and te t0 , it is possible to determine the time of depletion of the heat-carrier mass based on the change of the conductivity current of MIM-structure in time, ddtId . The beginning of the escape process is determined by the fact of closurerestoring MIM-structure, and the value Sh is evaluated from formula (1). Given the size of the hole, and using models of open penetration of a MIMstructure, the parameters of a high-velocity particle are determined. The process of vapor escape with its condensation on the walls of the hole channel, can be controlled by burning of a conducting jumper. The moment of full depletion of the heat-carrier is determined by the moment when the jumper in the created penetration hole disappears. In this model a probability remains of an incomplete depletion of the heatcarrier mass.

390

SEMKIN N. D. ET AL.

The third model assumes that liquid sodium flows from a TP wick into the hole. In this case it is likely that higher breakthrough voltage will be necessary for restoring a MIM-structure. Knowing the value of the breakthrough voltage and taking into account the model of formation of the conducting jumper, it is possible to identify the loss of hermeticity in the TP, as a result of its perforation, as well as to restore the information about the penetration hole size. Let us calculate the resistance of a MIM-structure dielectric in an operating mode (T = 900◦ C) from knowing the dependencies between TP temperature along the length Z of a pipe T = T (Z), and the specific conductivity on temperature δ = δ(T ). The combined conductivity of the whole TP is defined by the following expression: 2π r G= h

h δ (Z ) d Z ,

(4)

0

where r is the radius of the pipe, h is the thickness of dielectric (h T). The eq. (4) with allowance for the above dependencies will be transformed to: 2π r G= h

T

δ (Z ) d Z dT . dT

To

1 2πr lδ0 , or R = . h G Let us evaluate the electrical energy power necessary to burn the short-circuit jumper: 2 E P= Rcc . Rcc + Rl If δ = const = δ0 , then G =

where Rl is load resistance, Rcc is short-circuited jumper resistance. The energy necessary for evaporation of the heat-carrier mass can be expressed as: Ee = Cm(T e − T ) + qm where Cm is mass thermal capacity, qm is the heat of vaporization, and Te is the temperature of evaporation.

4. Experimental Results The experiments of perforating a TP wall having a MIM-structure were conducted by using an electromagnetic accelerator with particles in the form of conductors 0.2 mm in diameter and l = 3 mm. For a thermal pipe with a niobium wall having thickness of 0.3 mm and a layer of dielectric (Al2 O3 ) having a thickness of 0.1

THE IMPACT OF HIGH-VELOCITY PARTICLES

391

Figure 5. Cases of crater formation (left) and full perforation (right)

mm and a black-cover (spinel h = 0.1 mm), the perforating velocity for a particle of this size is estimated as ∼5.3 km·s−1 . Figure 5 (the left view) shows a crater formed by a particle with velocity 2.5 km·s−1 and a full penetration ( the right view). Perforation of the black cover is observed on a large area as a result of an unloading wave on the TP surface. For a TP wall, heated up to T = 900◦ C, penetration velocity of a particle decreases to 4.8 km·s−1 . The influence of fine particle fluxes on optical properties of TP’s was investigated experimentally using an explosion type accelerator [1], and an electrostatic accelerator over the range of masses 5 × 10−12 –10−15 kg [2] with particle materials chosen from: C (graphite, black soot), Al2 O3 , Al, TiC, Fe, W. The velocity range of particles was 1.5–8 km·s−1 . The particle density in all experiments was within 0.1–0.65 g·m−2 . Experiments demonstrated that the absorption coefficient As is considerably increased from 0.15 up to 0.43–0.6, and blackness factor E is changed slightly from 0.9 up to 0.92. Lanthanum manganate, lanthanum chromide and spinel were used as black covers for TP.

5. Conclusions 1. The absorption coefficient of the black surfaces of the investigated TP’s was found to increase substantially from 0.15 to 0.60 upon interaction with a flux of particles 0.1–0.65 g·m−2 that is equivalent to interaction with particles having a mass of 10−15 –10−12 kg and velocity 1–30 km·s−1 in space over a period of 2–2.5 years [3]. 2. The conducted experiments demonstrated that hermeticity and emissivity control of TP’s is possible by presenting their structure as a distributed

392

SEMKIN N. D. ET AL.

RC-equivalent circuit and estimating the conductivity currents of the black coatings and MIM-structures. The results also suggest that the discussed approach and method could be used for monitoring in flight conditions.

References 1. Novikov, L. S., Akishin, A. I., Semkin, N. D., and Voronov, K. E. (1997) In Proceedings of the 7th International Symposium on Materials in Space Environmen,t Toulouse, France, 16–20 June, 1997 (SP-399, August 1997), pp. 493–496. 2. Emma, A. T., Mark, K. H., Kay, L. (1977) In Proceedings of the Second European Conference on Space Debris, ESOC, Darmstadt, Germany, 17–19 March, 1997, (ESA SP-393, May 1977), pp. 429–434. 3. Sch¨afer, F. and Schneider, E. (1997) In Proceedings of the Second European Conference on Space Debris, ESOC, Darmstadt, Germany, 17–19 March, 1997, (ESA SP-393, May 1977), pp. 435–443.

PHYSICAL MECHANISM OF SOLAR CELL SHUNTING UNDER HIGH VELOCITY IMPACT OF SOLID PARTICLES V. A. LETIN,1 A. B. NADIRADZE,2 AND L. S. NOVIKOV3 1 FSUE RPE “KVANT,” 3-d Mytischinskay 16, 129626, Moscow, Russia 2 Moscow Aviation Institute, Volokolamskoe shosse 4, 125993, Moscow, Russia 3 Skobeltsyn Institute of Nuclear Physics Moscow State University, 119992 Moscow, Russia

Abstract. The mechanism of spacecraft solar array (SA) degradation caused by shunting of individual elements under impact of high-velocity solid particles has been analyzed. It was shown that the shunting effect appears as a result of irreversible structural changes in the semiconductor structure and it may be observed at an impact of rather large (over 50–100 μm), high-velocity (over 5–7 km·s−1 ) particles. A shunting model is presented, and critical impact parameters at which the effect manifests itself are identified. Key words: solar cells, solid particles, high-velocity impact, shunting

1. Introduction It is known that solar array degradation in space is a multicomponent function of environment, depending mainly on deterioration of the solar array parameters under the corpuscular radiation effect, thermal cycling, charging, etc. [1]. Insufficiently studied factors causing the degradation include effects of micrometeorites, comet and artificial-origin particles. It is assumed that the coverglasses of solar cells (SC) erode and fracture under impacts of meteorite bodies. In this case, power drop of a LEO solar array does not exceed 0.2–0.25% a year [2, 3]. However, under certain circumstances, the impacting solid particles could cause the solar cells shunting and substantial degradation of their power. Shunting occurs under particular structural changes in the semiconductor structure and may be observed at high-velocity (over 5–7 km·s−1 ) impact of rather large (over 50–100 μm) particles [4, 5]. 393 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 393–400. C 2006 Springer. Printed in the Netherlands.

394

V. A. LETIN, A. B. NADIRADZE, AND L. S. NOVIKOV

2. Shunting Model 2.1. PHYSICAL MECHANISM OF SHUNTING

From the theory of high-velocity collisions [6] it is known that immediately after collision of a particle with a solid body surface the impacted material is compressed, at first, in the shock wave and then expands in relief waves coming from the target free edges. The thermodynamic state of the substance changes within very wide limits. At collision velocities of 5–15 km·s−1 the initial pressure behind the shock-wave front reaches several million atmospheres, whereas the energy density exceeds several times the sublimation energy. Under certain conditions of shock-wave loading, the semiconductor starts melting. During the melting-crystallization process, precipitates (impurity clusters) having a high electric conductivity may originate. The disturbance of the structure with impurities may in turn bring partial or complete loss of the p–n junction rectifying properties, increase in leakage current, and substantial deterioration of the solar cell characteristics even when the impact zone area is relatively small. 2.2. SHUNTING PARAMETERS

To evaluate energy losses associated with high-velocity impacts, it is necessary to assess the shunting parameters. Leakage currents passing through the shunting region would be defined by the shunting resistance value Rsh . This value may be expressed through the length L sh = π Dsh and specific resistance ρsh of the shunting region: Rsh = ρsh /L sh (figure 1). The L sh value characterizes the damage geometry and is associated with the collision parameters, i.e., with particle sizes, velocity, density, with the thickness of a solar cell coverglass, and so on. The specific resistance ρsh characterizes the physical properties of the shunting structure and depends on the peculiar features of the latter (on the concentration and distribution of defects, impurity concentration profile, etc.), as well as on the conducting layer thickness.

a)

Dsh

Dk

δ

b)

Iph

Rs Id Ish

n-region Ish p-region. Strike zone.

Rsh

IL

I′sh R′sh UL

RL

Figure 1. (a) Impact zone and (b) equivalent circuit diagram of a SC under high-velocity impact by a solid particle: Rsh —initial shunting resistance; Rsh —shunting resistance of region, originated at impact

PHYSICAL MECHANISM OF SOLAR CELL SHUNTING

395

In case of a single-layer target, it is possible to apply the known semiempirical relation, connecting the impact crater dimensions Dk with the impact parameters [5]. Dk = 2ak K σ (ρk /ρt )1/3 (ρk Vk /σ∗ )1/μ

(1)

where ak , Vk —particle radius and velocity, respectively; ρk , ρt —striker and target densities, accordingly; μ—self-modulation parameter; K σ = 2.38; μ = 3.45. This expression is in good correlation with the experiment and allows evaluation parameters of impact craters within the velocity range of 5 through 80 km·s−1 . The parameter σ∗ characterizes the qualities of striker and target materials under shock-wave loads and demonstrates at what shock-wave amplitude the crater formation process stops (for fused silica and aluminum striker σ∗ = 6.6 GPa). In case of a two-layer target the crater dimensions may be calculated using expressions (2) and (3). Assuming that the shock-wave propagation in the first layer of a target is subject to the law P ∼ r −μ , whereas the crater formation in the second layer stops when P ≤ σ2 , it is possible to obtain the following expression for the crater diameter in the second layer of the target: Dsh = 2 r∗2 − δ 2 , (2) where r∗ = 0.5 Dk (σ∗1 /σ∗21 )1/μ , σ∗1 —primary shock-wave pressure; σ∗21 — primary shock-wave pressure at which the refracted component would have the amplitude σ∗2 after the shock wave splits at the two layer interface. To evaluate the σ∗21 value it is necessary to solve the problem of shock wave refraction at a two medium boundary [6]. For the glass–silicon pair σ∗21 = 5.2 GPa. It is impossible to calculate the ρsh value. However, analyzing the processes of shunting structure formation, it was found that the ρsh value should not depend strongly on the impact parameters. This hypothesis would be confirmed experimentally. 2.3. CRITICAL PARAMETERS OF THE IMPACT

It was demonstrated above, that the particle impact must be accompanied by semiconductor melting in the p-n junction region for the shunting effect to manifest itself. Therefore, there must be some critical values of the impact parameters at which this condition is satisfied. The impact heating effects are associated with the substance entropy growth when passing through the shock-wave front. Then at the subsequent pressure drop to the initial level, the accumulated entropy remains unchanged, since the relief process runs almost isoentropically. The accumulated entropy manifests itself as an internal energy of the substance i.e. the material temperature increases. If the

396

V. A. LETIN, A. B. NADIRADZE, AND L. S. NOVIKOV

accumulated internal energy turns to be higher than the target substance melting point then the target substance transforms into the liquid state. Entropy increment, due to the substance impact compression, was found applying the Rankine–Hugoniot’s equations [6]. The pressure value at the substance melting initiation was determined by the intensity of shock wave within which the entropy increase will be equal to the entropy difference of the substance in its initial state and at its melting point. For silicon, this value is estimated at S = 1.08 kJ.(kg.K)−1 . In this case, the calculated value of shock-wave pressure σsh obtained according to reference [6], is equal to 68 GPa. At a known critical amplitude of the shock wave σsh , the critical velocity of particles would be equal to: ck 2 σsh c1 2 σsh ck c1 Vsh = + + + − + (3) 2sk ρk sk 2s1 ρt st 2sk 2s1 where c1 , s1 —constants of Hugoniot’s adiabat, ρ—mass density of the material, k, t—all terms with subscripts k, t refer to the materials of the particle and target accordingly. Applying the expression describing the crater diameter as a function of impact parameters (1), (2), it is easy to find the critical radius of particles ash assuming that the crater diameter Dk = Dsh : ash = (δ/K σ )(ρm /ρk )1/3 (σsh,21 /ρm Vk2 )1/μ

(4)

The obtained relations define the collision critical parameters (critical velocity and particle radius) corresponding to the semiconductor melting initiation, depending on the striker and target properties as well as on the thickness of the solar cell protective shielding. Outcome of calculating the particle critical radius depending on the collision velocity and the particle material density is shown in figure 2. In the calculations it was assumed that σsh,21 = 54.8 GPa. The curves shown in figure 2 are restricted from the left side (marked by black circles) and correspond to critical velocity values calculated by formula (3).

3. Electrostatic Accelerator Tests The tests were conducted on an electrostatic accelerator fitted with a Van-de-Graff generator. The aluminum particles of 0.1–10 μm were accelerated to velocities from 0.5 to 10–15 km·s−1 . The particle flux hitting a sample was 0.8–1.2 s−1 [7]. SC of the n+ -p-p+ type with the thickness of about 200 μm and the depth of the p-n junction of about 0.5 μm were used as samples.

PHYSICAL MECHANISM OF SOLAR CELL SHUNTING

397

Critical radius, ash/δ

2

Al Ti Fe

1.5

1

0.5

0

1

10

100

Impact Velocity, km/s

Figure 2. Dependence of critical radius of particles on the velocity of impact

In this experiment, changes of the solar cell parameters at each particle impact were recorded. The solar cell reverse current at the fixed bias voltage value (around 2.5 V) was recorded. This enabled to control reverse voltage changes of the cells at each particle impact. The solar cell typical I–V characteristics are given in figure 3 and inner resistance and back bias current of SC during irradiation—in figure 4. The critical velocity of collisions was determined as Vsh = 7.5–12.5 km·s−1 , specific resistance of the shunting region ρsh = 0.068 ·cm, and critical pressure value corresponding to the obtained Vsh value is equal to σsh = 74.5 GPa.

Figure 3. I–V characteristic of solar cell before and after bombardment

398

V. A. LETIN, A. B. NADIRADZE, AND L. S. NOVIKOV

Figure 4. Inner resistance (marked open squares) and back bias current of SC (marked black circles) during bombardment

4. Investigation of Solar Array Retrieved from “MIR” Orbital Station According to the program MIR-NASA of Russian—USA cooperation in space, a part (eight panels with 0.8 m2 each) of a mounted solar battery (MSB), which operated in the space for 10.5 years, was returned to the Earth on board the Space Shuttle (STS-89) in January 1998. Ground investigations of the MSB gave a unique opportunity to obtain new information about the MSB’s materials’ behavior in space, power degradation, and also permitted to understand the nature and mechanism of this degradation [8, 9]. The experimental investigations of the MSB provided the opportunity to evaluate the influence of the micrometeorites and debris on the solar batteries and to specify existing models describing substance space-time distribution and to estimate typical peculiarities of high-speed collisions [10]. The carried out investigation showed that solar batteries (SB) coverglass damage was caused by high-speed collision and, in the majority of cases particle’s velocity was about 5–7 km/s and the SB damage was of erosive nature with the overall damaged area constituting ∼34% from the total SB’s area. The degree of power degradation across the MSB surface was very heterogeneous and a special investigation program concerning the anomalous segments of the MSB surface was conducted. For this purpose two panels of the MSB were partially disassembled. MSB elements with signs of micrometeoroid impacts were disassembled and analyzed. Single solar cells were found exhibiting short circuit behavior that was due to the high-speed collisions (figure 5). As can be seen from figure 5, the effect of SC shunting due to high-velocity collision with micrometeoroid particles and space debris was confirmed. For the curve 7 (Vx x = 29 mV, Isc = 107 mA) the estimation of shunting parameters

PHYSICAL MECHANISM OF SOLAR CELL SHUNTING

399

Figure 5. I–V characteristic of the solar element from returned solar panel: 1, 2 standard I–V characteristics; 3, 4—anomalous I–V characteristics of the solar element with increased series resistance; 5, 6, 7—anomalous I–V characteristics of the solar element with shunting p-n junction

Rsh = 0.271 , Dsh = 0.075 cm, L sh = 0.235 cm, ρsh = 0.064 ·cm was in good agreement with the results obtained for the spacecraft “VEGA” and using the electrostatic accelerator.

5. Conclusions Solar arrays could be shunted exhibiting short-circuit behavior under impacts of high-velocity solid particles, the shunting being the aftereffect of structural changes in a semiconductor. The presence of shunting effect was directly confirmed while conducting ground tests and investigating a solar array fragment recovered from the “MIR” Space Station. The critical velocity of particles causing the shunting is around Vsh = 7–12 km·s−1 . The specific resistance of shunting regions found to be about 0.065 ·cm for silicon solar cells was evaluated on the basis of the data obtained onboard “MIR” Space Station, and laboratory modeling results.

Acknowledgments The authors with to express their gratitude to colleagues Y. L. Lisovsky, N. D. Sizova, and V. P. Sidorenko for their assistance in editing and computer design.

400

V. A. LETIN, A. B. NADIRADZE, AND L. S. NOVIKOV

References 1. Letin, V. A., Zayavlin, V. R., and Gubanova, I. A. (1988) Journal of Electrotechnical Industry V. 22, N. 13, pp. 1–44. 2. Rauschenbach, H. S. (1980) Solar Array. Handbook. Litton Educational Publishing Inc., New York, pp. 1–549. 3. Letin, V. A., Holeva, M. N., and Zayavlin, V. R. (2002) Journal of Uzbekistan Academy of Science, 1, pp. 7–16. 4. Rijov., Y. A., Burgasov, M. P., Nadiradze, A. B., and Svirschevsky S. B. (1990), In 17th International Symposium on Rarefied Gas Dynamic, Aachen, FRG. 5. Burgasov, M. P., and Nadiradze, A. B. (1993) In Proceedings of ESPC-93, Austria, 23–27 August 1993, p. 767. 6. Kinslow, R. (1970) High Velocity Impact Phenomena, (ed.) Academic Press, New York and London, pp. 1–511. 7. Novikov, L.S. (2003) High Velocity Collisions in Space, Scobeltsyn Institute of Nuclear Physics, Moscow, 2003, pp. 1–11. 8. Letin, V.A. (2002) In Proceedings of 6th European Space Conference, Porto, Portugal, 6–10 May, 2002, pp. 713–718. 9. Letin, V. A., and Babayevsky, P. G. (2001) Journal of High Performance Polymers 13 S453–S460. 10. Letin, V. A., Deev, I. S., Scurat, V. E., and Tsetlin, V. V. (2002) II Stage Report on after Flight Studies of a Retrieved Fragment of Mountable Solar Array 17KS5810-0, RSC Energia-SPRE “Kvant”, Korolev, Moscow.

DETERMINATION OF GROUND-LABORATORY TO IN-SPACE EFFECTIVE ATOMIC OXYGEN FLUENCE FOR DC 93-500 SILICONE KIM K. DE GROH,1 BRUCE A. BANKS,1 AND DAVID MA2 1 NASA John H. Glenn Research Center, 21000 Brookpark Rd., M.S. 309-2, Cleveland, Ohio 44135 U.S.A. 2 Lockheed Martin Space Systems. 1111 Lockheed Martin Way, M/S 7LRS/157, Sunnyvale, CA 94089, U.S.A.

Abstract. The objective of this research was to calibrate the ground-to-space effective atomic oxygen fluence for DC 93-500 silicone in a thermal energy electron cyclotron resonance (ECR) oxygen plasma facility. Silicones, a commonly used spacecraft material, do not chemically erode with atomic oxygen attack like other organic materials. Silicones react with atomic oxygen and form an oxidized hardened silicate surface layer. Therefore, the effective atomic oxygen fluence in a ground-test facility should not be determined based on mass loss measurements, as they are with organic polymers, such as Kapton, a polyimide. A technique has been developed at the Glenn Research Center to determine the equivalent amount of atomic oxygen exposure in an ECR ground-test facility to produce the same degree of atomic oxygen damage as in space. The approach used was to compare changes in the surface hardness of ground-test (ECR) exposed DC 93-500 silicone with DC 93-500 exposed to low Earth orbit (LEO) atomic oxygen as part of a shuttle flight experiment. The ground to in-space effective atomic oxygen fluence correlation was determined based on the fluence in the ECR source that produced the same hardness for the fluence in-space. Because microhardness measurements need to be obtained on the very surface layer of a rubber substrate (with primarily elastic deformation) traditional techniques for microhardness that apply large forces and indentations based on plastic deformation, could not be used. Therefore, a nanomechanical measurement system operated in conjunction with an atomic force microscope (AFM) was used to determine the surface hardness of the silicones. The nanomechanical system can provide ultralight load indentations and can continuously measure force and displacement as an indent is made. Hardness versus contact depth measurements were obtained for five ECR exposed DC 93-500 samples (ECR exposed for 18–40 h, corresponding to Kapton effective fluences of (4.2–9.4) ×1020 atoms·cm−2 , respectively) and for a space exposed DC 93-500 from the evaluation of oxygen interactions with materials III (EOIM III) shuttle flight experiment, exposed to LEO atomic oxygen for 2.3 × 1020 atoms·cm−2 . Pristine controls for the ECR tests and for the EOIM III flight 401 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 401–416. C 2006 Springer. Printed in the Netherlands.

402

KIM K. DE GROH ET AL.

sample were also evaluated. A ground-to-space correlation value was determined based on correlation values for four contact depths (150, 200, 250, and 300 nm), which represent the near surface depth data. The results indicate that the Kapton effective atomic oxygen fluence in the ECR facility needs to be 2.64 times higher than in LEO to replicate equivalent exposure damage in the ground-test silicone as occurred in the space exposed silicone. Key words: space environment, environmental durability, low Earth orbit, atomic oxygen, oxidation, silicones, nanomechanical hardness, ground testing, space flight experiment, calibration

1. Introduction Silicones, a family of commonly used spacecraft materials, do not chemically erode away with atomic oxygen (AO) attack like other organic materials that have volatile oxidation products. Silicones react with AO and form an oxidized hardened silicate surface layer. Often loss of methyl groups causes shrinkage of the surface skin and “mud-tile” crazing degradation. But, silicones often do not lose mass [1], and some silicones actually gain mass during AO exposure. Therefore, the effective AO fluence for silicones in a ground-test facility should not be determined based on mass loss measurements, as they typically are with polymers such as Kapton [2]. Another method for determining effective fluence needs to be employed. A technique based on changes in surface hardness has been developed at the NASA Glenn Research Center. Specifically, this technique has been used to determine the equivalent amount of atomic oxygen oxidation via oxide hardness measurement. The approach was to compare changes in the surface hardness of ground-laboratory exposed DC 93-500 silicone with DC 93-500 exposed to LEO AO as part of a shuttle flight experiment. The ground to in-space effective atomic oxygen fluence was determined based on the Kapton effective fluence in the ground-laboratory facility that produced the same hardness for the fluence in-space. Because microhardness measurements need to be obtained on the very surface layer of a rubber substrate (with primarily elastic deformation) traditional techniques for microhardness characterization that apply relatively large forces and characterize hardness based on plastic deformation, could not be used. Therefore, nanomechanical testing using ultralight load indentations and continuous load-displacement monitoring was used to determine the surface hardness of the silicones. Hardness versus contact depth were obtained for five DC 93-500 samples exposed to AO in an electron cyclotron resonance (ECR) thermal energy source facility and for space exposed DC 93-500 from the evaluation of oxygen interactions with materials III (EOIM III) shuttle flight experiment. Ground-to-space correlation values were determined based on the near-surface depth data.

DETERMINATION OF GROUND-LABORATORY

403

2. Experimental Procedures 2.1. MATERIALS

2.1.1. DC 93-500 silicone The product name for DC 93-500 is Dow CorningR 93-500 Space Grade Encapsulant. It is supplied as a two-part liquid component kit. When the liquid components are thoroughly mixed, the mixture cures to a clear, flexible, low volatility space grade elastomer, which is suited for the protection of electrical/electronic assemblies, and has numerous other spacecraft applications. The service temperature range of DC 93-500 is −45–200◦ C (−49–392◦ F). 2.1.2. Ground-laboratory ECR AO exposed DC 93-500 The ground-laboratory test samples were exposed to AO in an ECR facility. These samples were approximately 1/4'' × 1/4'' (0.635 × 0.635 cm) in size, and approximately 0.020–0.025'' (0.051–0.064 cm) thick. These samples were made to have the same thickness as the space flight sample, as thickness variations of soft materials could affect hardness values. All samples (pristine controls, ground-laboratory ECR exposed and the space flight sample) were mounted using the same type and thickness sample holder for hardness characterization. 2.1.3. In-space exposed DC 93-500 The in-space exposed DC 93-500 sample was exposed to low Earth orbit (LEO) AO as part of the evaluation of oxygen interactions with materials III (EOIM III) shuttle flight experiment flown on STS-46. This sample was exposed to directed ram AO from within the shuttle bay and received a LEO atomic oxygen fluence of 2.3 ± 0.3 × 1020 atoms·cm−2 [3]. The flight sample was 2.54 cm in diameter and was approximately 0.020–0.025'' (0.051–0.064 cm) thick. The silicone received enough AO while in LEO to cause significant microcrazing of the surface. Preand postflight micrographs of the flight sample surface are shown in figure 1. A pristine control sample fabricated at the same time as the flight sample was also available for testing. 2.2. ELECTRON CYCLOTRON RESONANCE SOURCE AO EXPOSURE

The ground-laboratory test samples were exposed to AO in an electron cyclotron resonance (ECR) source facility at the Lockheed Martin Space System Company (LMSSC) Advanced Technology Center (ATC). This directional plasma system provides approximately 90% thermal AO (0.1 eV) and approximately 10% ionized species (15–20 eV). The effective exposed area is 6–8'' diameter with approximately 20% drop-off at 4'' off center. The Kapton effective atomic oxygen flux is approximately 6 × 1015 atoms·cm−2 ·s−1 based on mass loss measurements. Five

404

KIM K. DE GROH ET AL.

Figure 1. Preflight and postflight images of the EOIM III DC 93-500 sample: (a) preflight, and (b) postflight showing surface crazing due to AO oxidation

DC 93-500 samples were exposed to ECR AO from 18 to 40 h. The sample temperature during ECR exposure was 80◦ C. The ground-laboratory exposed samples and the corresponding Kapton mass loss effective fluence values are provided in table 1. TABLE 1. ECR AO exposed samples # AO (h)

AO Kapton effective fluence (× 1020 atoms·cm−2 )

0 18 20 24 30 40

0 4.2 4.7 5.6 7.0 9.4

DETERMINATION OF GROUND-LABORATORY

405

2.3. NANOMECHANICAL HARDNESS TESTING

2.3.1. Nanoindentation introduction The hardness of a material can be defined as the resistance to penetration [4], or the resistance to plastic deformation, and is dependent on the type of test used to determine its value. Traditional hardness testers force an indenter with known geometry into the surface of a material with a known applied load. Depending on the type of test, the hardness (H ) is expressed by a number that is either inversely proportional to the depth of indentation for a specified load and indenter, or equal to the load (P) over the area of indentation (A), H = P/A [5]. It should be kept in mind that the hardness value obtained in a particular test serves only as a comparison between materials or treatments [5]. Nanoindentation arose from the realization that an indentation test with a sharp indenter applied with a low force is an excellent way to measure very small volumes of materials and thin films. But, a need existed to be able to determine the indentation area without a high-magnification microscope. Therefore, depthsensing indentation (DSI) techniques were developed. Nanoindentation refers to depth-sensing indentation testing in the submicrometer range. In nanoindentation techniques, the load and displacement of the indenter are recorded during the indentation process and these data are analyzed to obtain the contact area, and thereby mechanical properties, without having to see the indentations. DSI techniques provide a means for studying elastic and time dependent plastic properties of materials. In the TriboScope system, the procedure used to calculate the hardness (H ) from the load-displacement data is presented by Oliver and Pharr [6], and is known as the Oliver and Pharr method. This method accounts for the curvature in the unloading data (nonlinear unloading) and uses a procedure for determining contact area (A) at peak load based on the indenter shape and depth of penetration. The hardness is defined as the mean pressure the material will support under load and is computed from the peak indentation load (Pmax ) and the projected area of contact at peak load, H = Pmax /A [6]. This hardness is different from the conventional definition of hardness because the nanomechanical hardness is calculated utilizing the contact area at maximum load whereas in conventional tests the area of the residual indent after unloading is used. 2.3.2. TriboScope nanomechanical test system A Hysitron Inc. TriboScopeR Nanomechanical Test System operated in conjunction with a Park Scientific AutoProbe atomic force microscope (AFM) was used to determine the surface hardness of the silicone samples. The TriboScopeR nanomechanical test instrument is a quantitative depth-sensing nanoindentation system that uses a three-plate capacitive force/displacement transducer. The Hysitron nanomechanical system can provide ultralight load indentations (less

406

KIM K. DE GROH ET AL.

than 25 μN) and continuously measures force and displacement as an indent is made. It should be noted that the maximum force experienced by these samples (the sample force, Fsample ) is not the same as the prescribed applied force (the force applied to the center plate of the transducer, Fapplied ), primarily due to the large displacement of these samples with soft substrates. 2.3.3. Indentation procedures The transducer tip was brought into contact with the surface of the sample using the AFM electronics and feedback system. Once in contact with the surface the tip was left at an ultralight preload and the system was left to settle for 2 min prior to starting the indent, to minimize drift. Indentations in the test samples were taken using a trapezoid loading curve with a 20 s hold period (5 s ramp up, 20 s hold, 5 s ramp down) using a Berkovich indenter. A Berkovich tip is a three-sided pyramid tip with an area-to-depth function which is the same as that of a Vickers indenter [7]. The total included angle of a Berkovich tip is 142.3◦ with a half angle of 65.35◦ . This tip geometry has been used as the standard for nanoindentation. The average radius of curvature is typically between 100 and 200 nm. Indents were made with applied forces ranging from 75 to 900 μN, and up to 1500 μN when possible. The maximum applied load was determined by the limit of the transducer displacement for each sample (≈4700 nm). Each indentation was taken at a new sample location. Approximately 20–25 individual indentations were obtained for each sample. Preloads of 0.112 mg (1.10 μN) were used for less than 200 μN applied indent force indents. Preloads of 0.125 mg (1.23 μN) were typically used for higher than 200 μN load forces. In some of the earlier test sets preloads of 0.111 mg (1.09 μN) were used for higher forces, but were determined not to affect the hardness values for the higher loads.

3. Results and Discussion 3.1. PRISTINE DC 93-500 (ECR CONTROL)

During indent testing of the pristine DC 93-500, negative unloading force values were observed in the real time force versus displacement data during the second set of indents, possibly indicating tip contamination. Tests were conducted that verified that the sample tip was contaminated from the silicone (i.e., repeated indents were consistently “softer” than the original data taken at the same location prior to tip contamination). After tip cleaning, additional data was taken to verify that the tip was not changed. The results of the noncontaminated pristine data obtained during two separate indent sessions at two different locations are provided in figure 2 along with a curve fit for the data. An image of the sample surface is also

407

DETERMINATION OF GROUND-LABORATORY 25

Hardness (MPa)

20

15

0.2 mm

10

5

Y = 5175/(0.01+1.0X1.5)+0.45 0

0

500

1000

1500

2000

2500

3000

3500

Contact Depth (nm)

Figure 2. Hardness versus contact depth for the pristine DC 93-500 ECR control sample

provided in figure 2. As can be seen, the hardness of the pristine silicone is greater at the near surface than deeper in the bulk of the material. This is commonly found for both metal and polymer materials using nanoindentation techniques. Polymers typically display greater hardness values at the surface because of air or light induced cross-linking and ambient oxidation of the sample. 3.2. 18 H AO EXPOSED DC 93-500

The 18 h ECR exposed sample received a Kapton effective fluence of 4.2 × 1020 atoms·cm−2 . Hardness data was acquired on the 18 h AO exposed sample during two indentation sessions at different locations. The hardness versus contact depth data is provided in figure 3 along with a curve fit for the data. An image of the sample surface is also provided in figure 3 and shows that very fine microcracks have developed. An interesting effect was observed in the 18 h sample data at higher applied indent forces (400, 600, 800, and 900 μN). An apparent “film breakthrough” was displayed in the loading curve. It appears that the tip was broken through a hardened surface layer and down into a softer underlying layer. An example is provided in figure 4 for the 900 μN applied force indent in the 18 h AO exposed sample. This type of breakthrough was observed for most of the oxidized films after some loading force is applied. It should be noted that the screen image shown in figure 4, provides the displacement at breakthrough, not the actual contact depth.

408

KIM K. DE GROH ET AL. 25

Hardness (MPa)

20

15

0.2 mm

10

5

0

1.5

Y = 0.86/(0.01+0.00005X 0

500

1000

1500

2000

2500

3000

)+0.3 3500

Contact Depth (nm)

Figure 3. Hardness versus contact depth for the 18 h ECR AO exposed DC 93-500

3.3. 20 H AO EXPOSED DC 93-500

The 20 h ECR exposed sample received a Kapton effective fluence of 4.7 × 1020 atoms·cm−2 . Hardness data was acquired on the 20 h AO exposed sample during two separate indentation sessions at two different locations. The hardness

Figure 4. TriboScope software sample force versus displacement curve for 900 μN applied force indent in the 18 h AO exposed DC 93-500 sample

409

DETERMINATION OF GROUND-LABORATORY 25

Hardness (MPa)

20

15

0.2 mm 10

5

0

Y = 6950/(0.01+0.5X1.5)+0.5 0

500

1000

1500

2000

2500

3000

Contact Depth (nm)

Figure 5. Hardness versus contact depth for the 20 h ECR AO exposed DC 93-500

versus contact depth data is provided in figure 5 along with a curve fit for the data. There was a fair amount of scatter in the data for contact depths between approximately 200–500 nm. Film breakthrough was observed when using the 500, 700, and 800 μN applied indent forces. An image of the sample surface is also provided in figure 5 and shows that the surface has finer craze line than the 18 h exposed sample. 3.4. 24 H AO EXPOSED DC 93-500

The 24 h ECR exposed sample received a Kapton effective fluence of 5.6 × 1020 atoms·cm−2 . The 24 h sample was indented during one session at two different regions. The hardness versus contact depth data is provided in figure 6 along with a curve fit for the data. Film breakthrough was observed for all indents at or above 200 μN applied indent forces. An image of the sample surface is also provided in figure 6 and shows that the surface has a well-developed crazed texture. 3.5. 30 H AO EXPOSED DC 93-500

The 30 h ECR exposed sample received a Kapton effective fluence of 7.0 × 1020 atoms·cm−2 . The hardness versus contact depth data is provided in figure 7 along with a curve fit for the data. Film breakthrough was observed for all indents at or above 300 μN applied indent forces. An image of the sample surface is also provided in figure 7 and shows that the surface has a well-developed crazed texture like the 24 h exposed sample.

410

KIM K. DE GROH ET AL. 25

Hardness (MPa)

20

15

0.2 mm

10

5

0

Y = 4350/(0.01+0.25X1.5)+0.5 0

500

1000

1500

2000

2500

3000

3500

Contact Depth (nm)

Figure 6. Hardness versus contact depth for the 24 h ECR AO exposed DC 93-500

3.6. 40 H AO EXPOSED DC 93-500

The 40 h ECR exposed sample received a Kapton effective fluence of 9.4 × 1020 atoms·cm−2 . The hardness versus contact depth data was obtained at two separate regions of the sample. The hardness versus contact depth data is provided in figure 8 along with a curve fit for the data. Film breakthrough was observed for indents at or above 625 μN applied indent forces. An image of the sample surface 25

Hardness (MPa)

20 15 0.2 mm

10 5 0

Y = 4825/(0.01+0.23X1.5)+0.5 0

500

1000 1500 Contact Depth (nm)

2000

2500

Figure 7. Hardness versus contact depth for the 30 h ECR AO exposed DC 93-500

411

DETERMINATION OF GROUND-LABORATORY 25

Hardness (MPa)

20

15

0.2 mm

10

5

0

Y = 0.46/(0.01+0.000015X1.5)+0.5 0

500

1000

1500

2000

2500

3000

Contact Depth (nm)

Figure 8. Hardness versus contact depth for the 40 h ECR AO exposed DC 93-500

is also provided in figure 8 and like the 24 and 30 h exposed samples, shows a well-developed crazed texture. 3.7. EOIM III CONTROL DC 93-500

The hardness versus contact depth data for the EOIM III control sample is provided in figure 9 along with a curve fit for the data. 25

Hardness (MPa)

20 15 10 5

Y = 630/(0.01+0.065X1.57)+0.45 0

0

500

1000

1500

2000

2500

3000

3500

Contact Depth (nm) Figure 9. Hardness versus contact depth for the EOIM III control sample fabricated at the same time as the flight sample

412

KIM K. DE GROH ET AL. 25

Hardness (MPa)

20

15

0.2 mm

10

5

0

Y = 169.5/(0.01+0.01X1.5)+0.9

0

500

1000

1500

2000

2500

3000

Contact Depth (nm)

Figure 10. Hardness versus contact depth for the EOIM III flight sample

3.8. EOIM III DC 93-500 FLIGHT SAMPLE

The EOIM III flight sample was exposed to a LEO AO fluence of 2.3 × 1020 atoms·cm−2 . The hardness versus contact depth data is provided in figure 10 along with a curve fit for the data. Film breakthrough was observed for indents at 600 and 700 μN applied indent forces. An image of the sample surface is also provided in figure 10 and shows that the surface has a well-developed crazed texture. 3.9. GROUND-LABORATORY TO IN-SPACE EFFECTIVE FLUENCE FOR THE ECR FACILITY

The AO exposure does cause surface hardening, as expected, as the silicone is oxidized. The nanomechanical hardness data indicate that the AO exposed samples are substantially harder at the near surface than the pristine silicone. Another observation is that the hardened layer decreases with depth, as one would expect, with the hardness of the AO exposed samples becoming the same as pristine DC 93-500 at a depth of approximately 1500 nm. The consistency of the hardness for all samples below this depth was excellent, and shows excellent repeatability of this nanomechanical technique. There was a fair amount of scatter and data overlap for some samples at the low applied indent forces. Surface roughness, such as that which occurs with the “mud-tile” cracking of the oxidized silicate layer, and inhomogeneities in the surface oxide could contribute to hardness variations, particular for low-force indents. Although film breakthrough was observed at various applied loads for all oxidized films (e.g., the ECR and flight samples), specific trends were not investigated.

413

DETERMINATION OF GROUND-LABORATORY 16

Hardness (MPa)

y=1.054E-20x + 3.693E+00

Ground Data (ECR Source)

14

Space Data (EOIM-III)

12 10

6.10 × 1020 atoms/cm2 in ECR facility equivalent to 2.3 × 1020 atoms/cm2 in LEO

8 6 4 2 0

0

2E+20 4E+20 6E+20 8E+20 Kapton Effective AO Fluence (atoms/cm2)

1E+21

a.

Hardness (MPa)

16 14

Ground Data (ECR Source)

12

Space Data (EOIM-III)

10 8

6.16 × 1020 atoms/cm2 in ECR facility equivalent to 2.3 × 1020 atoms/cm2 in LEO

y = 4.521E-21x + 1.376E+00

6 4 2 0

0

2E+20

4E+20

6E+20

8E+20

1E+21

Kapton Effective AO Fluence (atoms/cm2)

b. Figure 11. Hardness versus Kapton effective AO effective fluence: (a) at a contact depth of 150 nm, and (b) at a contact depth of 300 nm

This would require additional analyses of the force versus distance curves to determine the specific depth of contact and applied load for which breakthrough occurred. This information was not readily available. As previously mentioned, the observation of film breakthrough confirms that an embrittled oxidized layer has developed over the soft underlying silicone layer. Using values obtained from the curve fit equations for each data set (the curve fit equations are given in the hardness vs. depth charts); plots of hardness versus Kapton effective fluence at contact depths of 150, 200, 250, and 300 nm were produced. These depths represent the “near-surface” data. The 150 nm contact depth and 300 nm contact depth plots are provided in figures 11(a) and 11(b),

414

KIM K. DE GROH ET AL.

respectively. The hardness for the ECR pristine sample and the EOIM III control samples were very similar. The EOIM III control was slightly harder at a contact depth of 150 nm, which is not surprising. Even if the mix ratios and degrees of cure were the same for both controls, because the EOIM III control was made years earlier there would be a difference in the quantity of residual outgassing products and oxidation that could increase the hardness slightly over time. These aging differences may also affect the flight sample hardness, but are considered negligible, as the hardness difference of the two controls was determined to be insignificant at a contact depth of 150 nm (and they were essentially the same at 300 nm). Based on the linear fits for the hardness versus exposure time data, the ECR Kapton effective fluence that provided the same hardness as the EOIM III sample for 150, 200, 250, and 350 nm depth was 6.10, 6.00, 6.04, and 6.16 × 1020 atoms·cm−2 , respectively. Averaging these values provides 6.08 × 1020 atoms·cm−2 , therefore the Kapton effective atomic oxygen fluence in the ECR facility needs to be 2.64 times higher than in LEO to replicate equivalent exposure damage in the ground-test silicone as occurred in the space exposed silicone. When comparing the surface crazing, the extent of crazing in the 18 and 20 h ECR exposed samples appears to be less developed than the flight sample, while the 24 h ECR exposed sample seems more substantially crazed like the flight sample. This is a qualitative assessment only, and may be contributed to through sample handling procedures. For comparison purposes, the hardness versus contact depth curves for the EOIM III flight sample and the 24 h ECR sample, which was closest for the necessary ECR fluence for equivalent damage, were overlaid along with the pristine ECR sample data and are provided in figure 12. It should be noted that consistency in the sample thickness, sample holder, and nanoindentation technique for obtaining nanomechanical hardness data is very important. This is particularly true when the hardness changes substantially with depth such as with these samples. It is believed that these results will apply to chemically similar silicones in thermal energy AO systems. Further testing should be conducted to verify this. Similar tests should be conducted to get a ground-laboratory to in-space effective fluence correlation for silicones in higher energy AO facilities.

4. Relevance to ISS This method is relevant to silicone pressure sensitive adhesive (PSA) samples that were ECR exposed for International Space Station (ISS) solar array diode tape blocking assessment. The DC 93-500 samples as described herein were ECR exposed simultaneously with silicone PSA samples. On-orbit, the silicone PSA surface will eventually convert to a hardened silicate glass layer, converting a

DETERMINATION OF GROUND-LABORATORY

415

20

Pristine Hardness (MPa)

15

24 hrs ECR AO (5.6 E 20 at/cm2) EOIM III (2.3 E 20 at/cm2)

10

5

0

0

500

1000

1500

2000

2500

3000

Contact Depth (nm)

Figure 12. Hardness versus contact depth for pristine, 24 h ECR exposed and LEO space exposed DC 93-500 silicone

sticky surface to a glassy nonblock surface. By using the correlation of ground and space DC 93-500 data, the ground-test effective fluence of the silicone PSA can be deduced and the degree of blocking versus AO exposure that can be expected on-orbit for the ISS diode tape can be more accurately determined.

5. Summary and Conclusions Traditional techniques for determining effective AO fluence exposures in groundlaboratory facilities based on mass loss measurements should not be used for silicones. Therefore, a technique has been developed for ground-laboratory to inspace effective AO fluence determination for silicones based on changes in surface hardness, which occurs as the silicone is oxidized. Specifically, nanoindentation hardness measurements were determined for ground-test (ECR) exposed DC 93500 silicone and DC 93-500 exposed to LEO AO as part of the EOIM III shuttle flight experiment. The ground to in-space effective AO fluence was determined based on the Kapton effective fluence in the ECR source that produced the same hardness for the fluence in-space. Preliminary calibration and optimization testing of the nanomechanical system with silicone samples were conducted and critical issues reviewed. Hardness versus contact depth measurements were obtained for five ECR exposed DC 93-500 samples (ECR exposed for 18 h–40 h, corresponding to Kapton mass loss effective fluences of (4.2–9.4) × 1020 atoms·cm−2 , respectively) and for the EOIM III LEO exposed DC 93-500 silicone. Pristine

416

KIM K. DE GROH ET AL.

controls for the ECR tests and for the EOIM III flight sample were also measured. The ground-to-space correlation value was determined based on correlation values for four contact depths (150, 200, 250, and 300 nm), which represent the “near-surface” depth data. The results indicate that the Kapton effective AO fluence in the ECR facility needs to be 2.64 times higher than in LEO to replicate equivalent exposure damage in the ground-test silicone as occurred in the space exposed silicone.

Acknowledgments The authors would like to thank Phil Abel of NASA Glenn Research Center for his consistent help with hardware setup and introductions to the Park Scientific AFM and Hysitron systems. They also greatly appreciate all the many helpful insights and advice on calibration and testing procedures received from Lance Kuhn and Richard Nay of Hysitron.

References 1. Hung, C. and Cantrell, G. (1994) Reaction and Protective of Electrical Wire Insulators in AtomicOxygen Environment, NASA TM 106767. 2. ASTM E 2089-00. (2000) Standard Practices for Ground Laboratory Atomic Oxygen Interaction Evaluation of Materials for Space Applications, American Society for Testing and Materials, Philadelphia, USA, June 2000. 3. Knootz, S. L., Leger, L. J., Rickman, S. L., Hakes, C. L., Bui, D. T., Hunton, D. E., and Cross, J. B. (1995) Journal of Spacecraft and Rockets 32(3), 475–482. 4. Smallman, R. E. (1985) Modern Physical Metallurgy, Butterworths and Co. (Publishers) Ltd, England. 5. Avner, S. H. (1974) Introduction to Physical Metallurgy, 2nd ed., McGraw-Hill Inc., New York, 1974. 6. Oliver, W. C. and Pharr, G. M. (1992) Journal of Material Research 7(6), 1564–1583.

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS MASAHITO TAGAWA,∗ KOSHI MATSUMOTO, HIROAKI DOI, KUMIKO YOKOTA,∗ AND NOBUO OHMAE Department of Mechanical Engineering, Faculty of Engineering, Kobe University, Rokko-dai 1-1, Nada, Kobe 657-8501, JAPAN

Abstract. Feasibility of atomic oxygen concentration using reflecting mirrors was studied. Pyrex was selected as a material of reflecting surface due to its low recombination yield of atomic oxygen. Efficiency of reflection at Pyrex surface was measured to be 87.1%. The focal point of atomic oxygen reflected by a concave mirror was predicted by a computer simulation based on Hard-Cube model. It was estimated from the computer-simulated results that the difference between incident and exit angles of atomic oxygen due to energy loss during inelastic scattering event was evaluated to be at most 5 degrees depending on the incident angle. The computational results on trajectories of atoms scattered at the Pyrex concave mirror were compared to experimental results, and good correlation was observed. A high-efficiency atomic oxygen concentrator is designed based on the experimental results reported herein. Key words: atomic oxygen, low Earth orbit, concentrator, space environmental utilization

1. Introduction One of the problems in material testing in a flight experiment is a low flexibility of fluence control of atomic oxygen. If atomic oxygen flux could be controlled in a flight experiment, a multiple fluence exposure testing may be achieved in one flight experiment. Such multiple fluence test capability may reduce cost and time for a material testing in low Earth orbit (LEO) and provide wider knowledge of materials’ reaction in space environment. Also an accelerated test capability in LEO would be important to verify the survivability of materials in atomic oxygen space environment especially for atomic oxygen resistant materials. This ∗

Corresponding authors.

417 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 417–429. C 2006 Springer. Printed in the Netherlands.

418

MASAHITO TAGAWA ET AL.

is because atomic oxygen resistant materials require greater atomic oxygen fluences to see the effect of atomic oxygen exposure. Moreover, high-atomic oxygen flux reduces a required exposure time in LEO and may eliminate contamination problems in material tests. Therefore, use of atomic oxygen concentrator will drastically increase the degree of freedom in material tests conducted in LEO. One of the ways to achieve higher atomic oxygen flux in LEO is lowering the altitude of space vehicle. However, it is practically impossible to change the altitude of International Space Station (ISS) only for a material exposure test. The other way to achieve higher atomic oxygen flux in a given altitude is to concentrate (or to focus) atomic oxygen in a small targeting area. For this purpose, a highefficiency atomic oxygen concentrator needs to be developed. An attempt has already been tried by STS-85 in 1997 (see section 2 for detail). Three parabolic horns with different size have been tested to collect atomic oxygen. However, it has been reported that the concentration factor was not proportional to the size of the parabolic horn due to gas build-up [1]. Not only a material testing in LEO, but also a new thruster system requires to concentrate atomic oxygen. The new ion engine system uses atomic oxygen in LEO as a propellant (a conventional ion engine carries Xenon as a propellant) [2]. If the air breathing ion engine (ABIE) were successfully developed, it would be applied to compensate a drag of satellite at relatively low altitude (100–200 km) [3]. However, ABIE requires a high-efficiency atomic oxygen concentration device to maintain plasma in its reactor. Since the efficiency of the concentrator limits the operating altitude of ABIE, the development of high-efficiency atomic oxygen concentration device is a key for realizing this new ion engine system. In this paper, an attempt to develop a high-efficiency atomic oxygen concentration device is reported. Basic properties of atomic oxygen scattering at solid surfaces were investigated to develop atomic oxygen concentrator. Some key points to achieve high-efficiency in atomic oxygen concentration will be addressed based on the computational and experimental results obtained in atomic oxygen beam facilities at Kobe University [4].

2. Atomic Oxygen Focusing Device Aboard STS-85 The Effects of Space Environment on Materials Experiment (ESEM) had flown in-space on the STS-85 mission, returned to Earth and analyzed as one element of collaboration between the former National Space Development Agency of Japan (NASDA) and the National Aeronautics and Space Agency (NASA). The primary objectives of the ESEM experiments were to investigate the effects of atomic oxygen cosmic dust and man-made debris, and shuttle induced contamination on materials.

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 419

Figure 1. Atomic oxygen concentrators used in ESEM experiment flown on STS-85 [1]

In one of the experiments specimens were exposed in specially designed concentrators that were designed to increase the atomic oxygen flux onto the specimens and thus to provide accelerated time testing. The atomic oxygen concentrator is a horn-type aluminum structure that was expected to have ×4, ×9 and ×16 fold accelerations (see figure 1). STS-85 was launched at 10:41 am EDT on August 7, 1997. The ESEM pallet faced the ram direction for a total of 77.00 h, 52.48 h while at an altitude of 287– 296 km, and 24.52 h while at an altitude of 256 km. There were some exposures to the ram direction every day of the mission. The normal vector to the ESEM surface was perpendicular to the velocity vector for a large portion of the mission. The orbiter touched down at 7:08 am EDT on August 19, 1997. Atomic oxygen fluence was estimated from the mass change of Kapton witness sample. By assuming the erosion rate of Kapton to be 3.0 × 10−24 cm3 .atom−1 , atomic oxygen fluence during the flight was estimated to be 8.6 × 1019 atoms.cm−2 . In contrast, a detailed computer model predicted the total fluence of atomic oxygen during the mission to be 1.0 × 1020 atoms.cm−2 . After retrieval, laser profilometry was used to measure the recession of specimens flown under the atomic oxygen concentrators. Table 1 is a summary of the data for the average thickness changes of Kapton specimens placed under the focusing concentrators during the ESEM experiment flight. From the mass loss data of the Kapton witness sample, the recession of 2.6 μm is expected on the ×1 Kapton specimen. As is shown in table 1, the reaction rates are increased over ambient rates, but the erosion depths do not correlate with acceleration factors at ×9 and ×16. It was suggested that the pressure build-up in each of the focusing concentrators, provides an opportunity for recombination of atomic oxygen.

420

MASAHITO TAGAWA ET AL. TABLE 1. Recession of the polyimide films exposed to atomic oxygen under concentrators flown in ESEM experiment [1]. Acceleration factor Material

×4

×9

×16

Kapton

9 μm

18 μm

15 μm

3. Experimental Details 3.1. ATOMIC OXYGEN BEAM SOURCE

The atomic oxygen source used in this study was a laser detonation atomic oxygen source. This type of source was originally developed by Physical Sciences Incorporation (PSI) [5]. It uses laser-induced breakdown phenomenon of the oxygen gas. Principle of operation in the atomic oxygen source can be described as follows; pure oxygen gas is introduced into the nozzle throat through a PZT pulsed supersonic valve. A powerful laser pulse of TEA CO2 laser is focused on to the oxygen gas cloud at the nozzle throat. By absorbing the laser energy, high-density and high-temperature oxygen plasma is formed at the nozzle throat. Once plasma is formed, it propagates forward absorbing the laser energy that is included in the tail of the laser pulse. The plasma propagation is along the incident laser axis, and oxygen molecules are decomposed and accelerated in the front of the propagating plasma wave. Thus, an intense hyperthermal atomic oxygen beam pulse is formed. The PSI-type atomic oxygen source is attached to the space environment simulation facility at Kobe University (figure 2) [4]. The atomic oxygen beam is

Figure 2. Laser detonation (PSI-type) atomic oxygen beam source at Kobe University used in this study

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 421

Intensity (arb. unit)

1.2 1.0 0.8 0.6 0.4 0.2 0.0

0

2

4

6

8

10

12

14

16

18

20

Translational Energy (eV)

Figure 3. Typical translational energy distribution of the atomic oxygen beam used in this experiment. Mean translational energy is 5 eV

always monitored by the time-of-flight (TOF) measurement system consisting of a quadrupole mass spectrometer (QMS) and a multichannel scalar. A typical TOF spectrum of the atomic oxygen in the beam is shown in figure 3. We used the relationship of P(E) ∝ t 2 N (t) to calculate the translational energy distribution. In this spectrum, the component arising from the cracking of molecular oxygen in the ionizer of QMS was subtracted. Mean translational energy of the atomic oxygen was calculated to be 5.0 eV, which corresponds to the orbital impact velocity of atomic oxygen. The atomic oxygen flux in the beam was measured by an Ag-coated quartz crystal microbalance (QCM) with an accommodation coefficient of 0.62 [6]. The principle of measurement is explained in following section. Since the reaction of atomic oxygen with Ag is a non-linear phenomenon, only the initial reaction, which gave a linear mass gain, was used to calculate atomic oxygen flux [7]. A typical atomic oxygen flux at the sample position is 1.4 × 1014 atoms.cm−2 /.s−1 . 3.2. ATOMIC OXYGEN DETECTION BY AG-QCM

The flux of concentrated atomic oxygen was also measured using QCM with silver electrodes [8]. The resonant frequency of a QCM decreases with time when atomic oxygen oxidizes the silver electrode of QCM. Since silver is not oxidized by molecular oxygen but by atomic oxygen at room temperatures, formation of silver oxide signifies atomic oxygen reaction at the silver surface. A change in mass of the silver film was recorded by measuring the resonance frequency of QCM with a resolution of 0.1 Hz during atomic oxygen exposure. Resonance frequency of QCM is expressed by the following formula; f = −f20 W/NAρ.

(1)

422

MASAHITO TAGAWA ET AL.

where, N is the frequency constant, A is the electrode area, ρ is the density of quartz, and f0 is the resonant frequency. Since N , A, ρ, f 0 are known factors, one can calculate the mass change (W) of the sensor crystal from the frequency shift (f) of the QCM. It is natural to consider that reaction yields of materials with atomic oxygen depend on its translational energy. Evidently, reaction yield of polyimide depends strongly on the translational energy of atomic oxygen. It has been reported that the polyimide erosion by 1.1 eV atomic oxygen is much lower than that by 5.0 eV beam [9]. The translational energy of the reflected atomic oxygen may be decreased due to inelastic scattering and the reaction yield of silver would also be affected. Since the effect of translational energy on silver oxidation is unknown, the reaction yield of atomic oxygen with silver surface was assumed to be 1.0, instead of 0.62 for 5 eV case. Therefore, the calculated flux in this study represents the lower limit of the actual atomic oxygen flux.

4. Results and Discussion 4.1. MATERIAL SELECTION OF REFLECTING SURFACE

Atomic oxygen is electrically neutral, thus no electromagnetic devices, such as electrostatic lenses for instance, can be applied to concentrate atomic oxygen. The only practical way to change the direction of atomic oxygen is to use reflecting surfaces. Reflection of atoms/molecules at solid surfaces has been investigated as gas-surface scattering phenomena. It has been recognized that atomic oxygen would react with solid surfaces and form an oxide, or recombine to molecular oxygen when scattered at solid surfaces. Therefore, material for reflecting surface, which gave small cross-section of oxidation and recombination, need to be selected. Table 2 lists the recombination cross-section of atomic oxygen at various solid surfaces [10]. It is obvious that Pyrex gives the lowest recombination TABLE 2. Recombination cross-section of atomic oxygen at solid surfaces [10]. Metals

Oxides

Material

Recombination cross-section

Material

Recombination cross-section

Ag Cu Fe Ni Au Mg

2.4 × 10−1 1.7 × 10−1 3.6 × 10−2 2.8 × 10−2 5.2 × 10−3 2.6 × 10−3

CuO Mn2 O3 Fe2 O3 CrO3 ZnO Pyrex

4.3 × 10−2 1.3 × 10−2 5.2 × 10−3 2.5 × 10−4 4.4 × 10−4 3.1–4.5 ×10−5

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 423

Figure 4. Experimental configuration of the reflectivity measurements of the atomic oxygen reflecting mirror with QCMs

cross-section of atomic oxygen. In contrast, Au is known to be atomic oxygen resistant metal but has relatively large recombination cross-section. Thus, these two materials were chosen as candidates for reflecting surface materials to be used for atomic oxygen concentrators. An experiment was conducted to see the effect of recombination cross-section of reflecting surfaces on atomic oxygen concentration. The configuration of the experiment is shown in figure 4. Two flat surfaces were prepared: Pyrex and Au. These are optical mirrors 7.62 cm in diameter. These reflecting surfaces were placed downstream at 46.5 cm from the nozzle. Incident angle and exit angles of atomic oxygen were fixed at 45◦ . Flux of concentrated atomic oxygen was measured by Ag-QCM located 10.0 cm from the mirror surface. Figure 5 indicates the mass gain of QCM during the test. Atomic oxygen fluxes concentrated by Au and Pyrex reflecting surfaces were 1.13 × 1014 , and 1.48 × 1014 atoms.cm−2 .s−1 , whereas those at 56.5 cm away from the nozzle without mirrors was estimated to be 2.5

Mass gain (μg)

2.0

Pyrex

1.5

Gold

1.0 0.5 0.0 0

500

1000

1500

2000

2500

3000

Elapsed time (s)

Figure 5. Mass gain of the Ag-QCM due to oxidation by the atomic oxygen reflected by flat mirrors

424

MASAHITO TAGAWA ET AL. 2.5

Mass gain (μg)

2.0

Concave mirror

1.5 1.0

Flat mirror

0.5 0.0 0

500

1000

1500

2000

2500

Elapsed time (s)

Figure 6. Mass gain of the Ag-QCM due to oxidation by the atomic oxygen reflected by Pyrex mirrors

1.7 × 1014 atoms.cm−2 .s−1 . By comparing these values, the efficiency of reflection at Au and Pyrex surfaces were evaluated to be 66.5 and 87.1%, respectively. It was concluded that the recombination cross-section is important as a material property for a reflecting surface. Pyrex was the material to be selected as a reflecting surface in the following experiments. 4.2. SHAPE EFFECT OF REFLECTING MIRROR

The effect of the shape of the reflecting surface was examined. Two Pyrex reflecting surfaces were prepared: flat and concave mirrors. The concave mirror had a 15.24 cm radius of curvature of the reflecting surface (7.62 cm in focal length). Configuration of the experiment was the same as that described in previous section. Figure 6 indicates the frequency shift of QCM during the test. Atomic oxygen fluxes concentrated by flat and concave reflecting surfaces were 1.60 × 1014 , and 3.18 × 1014 atoms.cm−2 · s−1 . By comparing these values, it was obvious that the concave Pyrex mirror provided an atomic oxygen flux twice as high as the flat mirror. It was, thus, concluded that the use of Pyrex concave mirror is effective to concentrate atomic oxygen. 4.3. COMPUTER SIMULATION BASED ON HARD-CUBE MODEL

Usefulness of a Pyrex concave mirror to concentrate atomic oxygen was demonstrated in the previous section. The remaining problem is the determination of focal point of atomic oxygen. The major problem is that, unlike light scattering, atoms and molecules lose their energy in the scattering event and usually do not exit in the specular direction. In order to maximize the concentration of atomic oxygen the estimation of exit direction of the scattered atom is important. For this purpose computer simulation of a scattering event of atomic oxygen at Pyrex surface was carried out using the Hard-Cube model [11]. The Hard-Cube model

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 425

mg ui u in

θi

θr

ut u rn

u rt

u it

ms

(urt=u it)

vs

Figure 7. Principle of the Hard-Cube model

is one of the primitive methods to simulate scattering events. It assumes elastic collision of hard-atom (ball) and hard-surface moiety (cube) that vibrates in the direction of surface normal. The principle of Hard-Cube model is shown in figure 7. In this model, the average exit angle of oxygen atom is expressed in the following formula: μ kT 1 1 − μ −1 θ¯r = cot + cot θi (2) 1 + μ 1 + μ E i cos2 θi where θi is incident angle, E i is incident energy, k is Boltzmann constant, T is surface temperature, θr is average exit angle, and μ is ratio of the mass of incident atom and surface moiety (m g /m s ). With eq. (2), average exit angle can be computed with parameters shown above. The only unknown value in eq. (2), the effective surface mass (m s ), was determined to be 200 amu by comparing with the experimental data of hyperthermal F atom scattering at Si (001) surface reported by Minton et al. [12]. Differences between exit and incident angles are plotted as a function of incident angle in figure 8. It is obvious that the 5 deg. off 6

qr - qi (deg.)

4 2

Specular reflection

0 −2 −4 0

20

40

60

80

Incident angle (deg.)

Figure 8. Computational results on difference between incident and exit angles of 5 eV atomic oxygen reflected at the solid surface

426

MASAHITO TAGAWA ET AL.

Figure 9. Computational results on the trajectories of 5 eV atomic oxygen reflected at a concave mirror surface

maximum difference in exit and incident angles is calculated to be 5◦ at the incident angle 42◦ . The focal point of the atomic oxygen reflected at the Pyrex concave mirror in the experimental configuration shown in figure 4 was computer-simulated. The trajectories of oxygen atoms reflected at the Pyrex concave mirror are shown in figure 9. It is observed that atomic oxygen is focused near the position x = 44 cm, y = −6 cm in the figure even though aberration is fairly large. The focal position of (44, −6) corresponds to the average exit angle of 55◦ (incident angle is 45◦ ). This computational result was compared with those measured in an experiment in the same configuration. Flux measurements of scattered atomic oxygen were carried out with the exit angle between 20 and 90◦ . Detection of atomic oxygen was made by Ag-QCM with the mirror-detector distance of 8.0 cm. The results are shown in figure 10. As is clearly seen in figure 10, atomic oxygen flux shows its peak at the exit angle between 45 and 60◦ . The Hard-Cube model used in this study is known as a simple scattering model and sometimes it does not describe the hyperthermal atom-surface scattering phenomena precisely. It was, however, confirmed that results of the simple computer simulation based on Hard-Cube model agreed well with those of the scattering experiment. This finding suggests that even for the simple scattering model it is useful to design the atomic oxygen concentrator. 4.4. EFFECT OF MULTIPLE MIRRORS

Three concave mirrors were used to increase the atomic oxygen flux concentration. The experimental setup is shown in figure 11. Incident angles of atomic oxygen

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 427 6

Flux (×1014 a atoms/cm2/s))

(a)

(b)

5 4 3 2 1 0

0

20

40

60

80

100

Exit angle (deg.)

Figure 10. (a): Experimental setup of the angular distribution of reflected atomic oxygen, and (b): its results with incident angle of 45◦ . The atomic oxygen flux peaks at exit angles between 45–60◦

for upper and middle mirrors were 45◦ , whereas that for the lower mirror was 55◦ . These mirrors were installed to point their exit directions toward the QCM detector. Atomic oxygen flux detected by Ag-QCM is shown in figure 12. Atomic oxygen fluxes concentrated with a single mirror are distributed between 0.6 and 3.8 × 1014 atoms.cm−2 /.s−1 depending on the mirror-QCM distances (Average flux: 1.92 × 1014 atoms.cm2 .s−1 ). With three mirrors, atomic oxygen flux reaches as high as 1.31 × 1015 atoms.cm−2 .s−1 . Note that the atomic oxygen flux detected without reflecting mirror was subtracted as background, the fluxes mentioned above are the net atomic oxygen fluxes reflected by the reflecting mirrors. From the experiment described herein, the possibility for concentration of atomic oxygen by reflecting surfaces was clearly demonstrated.

upper mirror

center mirror

Ag-QCM detector lower mirror

Figure 11. Experimental configuration of the atomic oxygen concentration with multiple concave mirrors

428

MASAHITO TAGAWA ET AL.

Flux (×1014 atoms/cm2/s)

14 12 10 8 6 4 2 0

1

2

3

Number of mirrors

Figure 12. Average atomic oxygen fluxes amplified by the multiple concave (1–3 mirrors)

5. Conclusion Use of Pyrex reflecting mirrors to concentrate atomic oxygen in LEO was proposed in this study. The reflection efficiency of 5 eV atomic oxygen at Pyrex surface was measured to be as high as 87.1%. The focal point of reflected atomic oxygen was predicted by the simple computer simulation based on the Hard-Cube model. It was calculated that the difference between incident and exit angles of atomic oxygen due to energy loss during inelastic scattering event was at most 5◦ depending on the incident angle. The computational results on trajectories of atoms scattered at the Pyrex concave mirror were compared to experimental results, and good correlation was observed. As a result, the atomic oxygen flux at the focal point increased one order of magnitude higher than the initial exposure condition by using three Pyrex mirrors. Even though the directionality and energy distribution of the incoming atomic oxygen beam may be partially lost by the reflection at the concentrator wall, this new technology allows us to see early previews of atomic oxygen effects that won’t be seen until later on ISS materials and components, especially for atomic oxygen resistant materials. It is also a key technology for the development of ABIE and for the other applications using atomic oxygen as a natural resource for space environmental utilization. Use of the concentrator opens the new possibility in atomic oxygen research in LEO.

Acknowledgments A part of this work was supported by the grant-in-aid of scientific research contract No. 14350511 and 15560686 from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Kurata Memorial Hitachi Science Foundation.

ATOMIC OXYGEN CONCENTRATION USING REFLECTING MIRRORS 429

The author would like to thank H. Kinoshita of Kobe University for stimulating discussion.

References 1. United States Developed ESEM Experiments. Evaluation of Space Environment and Effects on Materials, Final Report, available at the website http://setas-www.larc.nasa.gov/esem/AOE.html 2. Stuhlinger, E. (1967) Ion Propulsion for Space Flight, McGraw-Hill, New York. 3. Nishiyama, K. (2003) In Proceedings of 54th International Astronautical Congress, Bremen, Germany, 29 September–3 October, 2003. 4. Yokota, K., Seikyu, S., Tagawa, M., and Ohmae, N. (2003) In Proceedings of the 9th International Symposium on Materials in a Space, Noordwijk, The Netherlands, ESA SP-540, 16–20 June, 2003, pp. 265–272. 5. Caledonia, G. E., Krech, R. H., Upschulte, B. L., Sonnenfroh, D. M., Oakes, D., and Holtzclaw, K. W. (1992) Fast Oxygen Atom Facility for Studies Related to Low Earth Orbit Activity, AIAA Paper 92-3974. 6. Tagawa M., Yokota K., Kinoshita H., and Ohmae, N. (2003) In Proceedings of the 9th International Symposium on Materials in a Space, Noordwijk, The Netherlands, 16–20 June, 2003, pp. 247–252. 7. Kinoshita, H., Tagawa, M., Yokota, K., and Ohmae, N. (2001) High Performance Polymers 13(4), 225–234. 8. Matijasevic, V., Garwin, E. L., and Hammond, R. H. (1990) Review of Scientific Instruments 61, 1747–1749. 9. Yokota, K., Tagawa, M., and Ohmae, N. (2003) Journal of Spacecraft and Rockets 40(1), 143–144. 10. Greaves, J. C. and Linnet, J. W. (1959) Transactions of the Faraday Society 55, 1323–1330. 11. Logan R. M. and Stickney R. E. (1966) The Journal of Chemical Physics 44, 195–201. 12. Minton T. K., Giapis K. P., and Moore T. A. (1997) Journal of Physics and Chemistry A 101, 6549–6555.

ATOMIC OXYGEN SOURCE CALIBRATION ISSUES: A UNIVERSAL APPROACH

CARL WHITE, JUAN CARLOS VALER, ALAN CHAMBERS, AND GRAHAM ROBERTS The University of Southampton, School of Engineering Sciences, Astronautics Research GroupSouthampton, Hants, United Kingdom

Abstract. There are a number of different ground-based facilities capable of producing atomic oxygen with fluxes similar to those found in orbit. The problem of using these sources to study and measure the resistance of a material to atomic oxygen is that the atomic oxygen beams typically have different kinetic energies, oxidation and quantum levels, as well as different fluxes. There may also be other species present such as VUV which may affect a material’s response to AO. Consequently for a given material, it is not surprising that different erosion yields are reported for exposures in different types of sources. In order to understand and compare the results from the different sources there is a need to calibrate each source. This paper briefly describes the different types of AO source and sensors for measuring AO flux. It also identifies the other species which need to be measured in a “round robin” experiment to provide a calibration of each source which is essential if our understanding of AO erosion mechanisms is to be enhanced. 1. Introduction It is known that many institutions in the academia, military, and government sectors have facilities that simulate one or more aspects of the environment present in Low Earth Orbit (LEO), and that the facilities available to reproduce the atomic oxygen (AO) in LEO do so with different degrees of success. The objective of this paper is to highlight the need for a “round robin” characterization of the major AO facilities in order to provide a better understanding of the factors affecting the erosion mechanism and rate. Due to the synergistic nature of the effects of the space environment on spacecraft materials, measurements of the UV dose and contamination must be included in a “round robin” experimental package. This paper includes an overview of some methods to produce LEO-like AO as well as a brief discussion on the major methods to evaluate AO sources. 431 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 431–441. C 2006 Springer. Printed in the Netherlands.

432

CARL WHITE ET AL.

2. Atomic Oxygen Sources It is widely accepted, that the atomic oxygen in LEO has the following characteristics:

r r r r

Kinetic Energy of approximately 5 eV Flux of approximately 5 · 1015 atoms · cm−2 · s−1 Ground Quantum Level Very low level of ionized species

The AO flux in LEO varies with altitude and time (the latter related to the solar cycle), and this variation can be of the orders of magnitude level. Experience with ground-based simulation facilities has shown that different sources produce AO with characteristics different from those mentioned above. Hence it is difficult to compare the results produced in different facilities. The situation is further complicated by the lack of a reliable and exact standard technique to measure AO flux. The following discussion will only address those methods that are successful in producing an AO flux and kinetic energy similar to those found in orbit. 2.1. ION SOURCES [1]

This type of source has enjoyed great interest during the last fifteen years. A beam of positive or negative ions that is later electrostatically accelerated and focused is created by either RF excitation or electron bombardment; an energy of around 5 eV can be achieved. In a later process, the ions are neutralized by either charge exchange or surface neutralization. Charge exchange can be achieved by the injection of fast ions into the beam that would also result in an overall increase of the kinetic energy of the beam. However, this may not always be possible due to divergence of the beam induced by the space charge that in turn results in a reduction of the kinetic energy of the species present in the beam. Gaseous charge exchange is inefficient due to low cross sections. Surface neutralization presents the complication that the reflected beam would not have a uniform kinetic energy. Unfortunately, and in addition to the charge exchange limitations of this method, the Columbian repulsion of the ions in the beam before neutralization, prevents this method from producing very high fluxes, although in theory LEO-like densities can be achieved. 2.2. LASER DISCHARGE

The production of atomic oxygen atoms is achieved by introducing a molecular oxygen gas pulse into a conical nozzle and firing a laser light pulse. The formed

ATOMIC OXYGEN SOURCE CALIBRATION ISSUES

433

plasma expands into the conical nozzle, creating a detonation/blast wave and generating a beam comprising a mixture of mainly neutral AO but with some ions. The flux that can be achieved in this kind of source is similar to LEO, with energies (that can be controlled) in the range of 3–8 eV [1]. A difficulty of this method is that it generates UV emission together with the AO atoms. It has been reported [2] that in the spectral range of 115–180 nm, a laser pulsed detonation source produced an intensity of 104 equivalent suns. It should be noted that the production of plasma can also be achieved with a continuous source of laser light, but the achieved beam intensities tend to be lower than with a pulsed laser. This approach also produces ionized species and UV radiation. 2.3. MICROWAVE OR RF DISCHARGE

This method is not very different from the laser discharge methods described above. This technique creates a plasma by means of a microwave or RF discharge, which is later accelerated in a nozzle. A kinetic energy of about 5 eV can be accomplished with this technique. As with the laser, microwave and RF discharges ionized species and UV radiation in addition to AO. 2.4. ELECTRON-STIMULATED DESORPTION

One side of a specially designed ceramic is placed under a high pressure of molecular oxygen, while the other side is exposed to vacuum; the latter side is usually coated with a thin film of silver. The molecular oxygen on the high-pressure side dissociates into AO and migrates through the ceramic towards the low pressure side. Upon reaching the other surface, and before it recombines into molecular oxygen, a flux of electrons impinges it causing the atomic oxygen atoms to leave with a kinetic energy similar to LEO conditions. This method has provided [3] fluxes similar to LEO for a short amount of time, without a significant production of ions or any UV emission. Further research is needed to increase the flux of oxygen through the ceramic membrane so as to allow the production of higher AO beam densities. The energy distribution in the beam under high flux conditions still needs to be determined. 3. Overview of Atomic Oxygen Flux Measuring Techniques A review of a variety of AO flux measuring techniques for space-based applications is given in [4]. Much of the discussion in that paper is also valid for the groundbased application discussed here. The following section is a brief description of the most important methods for measuring AO fluxes.

434

CARL WHITE ET AL.

3.1. KAPTON

Kapton-H is a polymeric material commonly used in spacecraft thermal insulation. Impinging AO erodes the material by producing volatile reaction products. The resultant mass loss can be used to measure the AO total flux; so far Kapton mass loss is regarded as the standard measure of AO flux. The erosion yield for Kapton has been well established as 3 × 10−24 cm3 · atom−1 [5, 6]. However, it is important to mention that the erosion yield of Kapton-H is believed to be dependant on the energy of the impinging AO, which is a problem when using it to compare results from facilities that produce beams with different energies [4]. There are some practical difficulties with the use of Kapton-H witness samples. It is not possible to have real time measurements, since the sample has to be taken out from the vacuum chamber to be weighed. Hence the method is not sensitive to local changes of flux. There is also the risk of Kapton being contaminated and even reacting with the atmosphere and thus changing its mass. This material is hygroscopic and due care should be taken on weighting it. Wolan and Hoflund [7] have described the limitations of Kapton for accurately measuring AO fluxes. Nevertheless, the use of Kapton-H witness samples provides a simple method for determining AO fluence with order of magnitude accuracy. It is also important to consider that Kapton-H is in the process of being discontinued, and it has been replaced by Kapton-HN. The latter is a material with a slightly different erosion yield [8]. 3.2. QUARTZ CRYSTAL MICROBALANCES [4]

Quartz Crystal Microbalances (QCM) have successfully been used to measure AO fluxes both in ground-based simulation facilities and in LEO. The use of crystals covered with either silver or carbon is considered to be a mature technology. QCMs provide an in situ measurement of the flux of atomic oxygen that is very specific to atomic oxygen. If the crystal is coated with silver, the AO will chemisorb and therefore its mass will increase; if coated with carbon, the AO flux will erode it with a corresponding decrease in mass. A specially cut quartz crystal is excited by the control unit and its frequency is measured; a change of frequency can be accurately related to a change of mass of the crystal. With the use of a suitable calibration curve, the rate of change of frequency can be related to an AO flux. QCMs can be very small, have high accuracy and repeatability, can have low mass, allow remote operation, and are relatively inexpensive. They have the limitation though that the material that is used to coat the crystal has a limited useful lifetime, after which the instrument is no longer capable of detecting AO. A limitation on silver-coated QCMs is that AO can recombine on the silver surface,

ATOMIC OXYGEN SOURCE CALIBRATION ISSUES

435

thus preventing it from being chemisorbed into the silver film; these recombined atoms would not be detected as part of the flux by the QCM. Further research on silver-coated QCMs needs to be performed to improve their accuracy, since the sticking coefficient between the AO and the silver is not accurately known, and there are some uncertainties related to the chemistry between those two species and the diffusion rate of AO on silver [3]. QCMs can also be used to monitor the deposition of contaminants in a vacuum chamber. For this application, it is enough to use an uncoated crystal. Provided that this crystal is kept at about the same temperature as the chamber, contaminants should deposit on the crystal as the same pace as on the rest of the chamber. The corresponding change in mass would in turn change the oscillation frequency of the crystal, what can be used to accurately determine the deposition rate of contaminants. 3.3. CATALYTIC PROBE

This method is simple, reproducible, and very sensitive. The probe consists of a wire, coil, or foil and it may be movable in the test area. Its highly catalytic surface is heated by the recombination of reactive species. A method very appropriate to absolute measurements of atom concentrations was applied to the O-atom reactions by Elias, Ogryzlo, and Schiff [9]. The probe is a silver-coated platinum wire coil, large enough so that the temperature rise due to recombination is not large (T < 100◦ C). It is operated isothermally at an elevated temperature and the difference in electrical power to produce the same wire temperature in the presence and absence of O-atoms is measured. The obvious disadvantage of this method is the lack of specificity to detect AO. It has been reported that oxygen molecules in the metastable state 92.2 kJ.mole−1 above the ground state will produce an erroneously large heat release in the detector. This error is avoided with the complete absence of O2 in the system. In this scenario, measurements using this method are in excellent agreement with values obtained by NO2 titration. 3.4. ELECTRON SPIN RESONANCE

This method is highly specific for the detection of atomic oxygen recombination. The resonance lines are easily identified and other information such as temperature can be obtained for any parametric excited state. Among its principal disadvantages are cost and complexity. Also, the fact that it gives only space averages of concentration over a considerable length of the flow tube. Absolute or even relative concentrations are calculated with some difficulty from experimental data. The signal depends on the average collision frequency of the atom and this complicates an experiment with added gases.

436

CARL WHITE ET AL.

The only rate data so far reported by this method are neither qualitatively nor quantitatively reliable. 3.5. NO2 TITRATION [10]

This technique is based on the following reactions: NO2 + O −→ NO + O2 NO + O −→ NO2 + hυ The titration involves the careful measure of the flux of NO2 in a fast flow tube. As the first reaction is about five orders of magnitude faster than the second, no energy emission would take place from the second reaction when the number density of O is less or equal than the number density of NO2 . When the flux of O is higher than that of NO2 , the second reaction begins to take place and light emission would be present. This emission is a greenishyellow afterglow that has been well determined by spectral analysis. When the number density of O doubles that of NO2 , the maximum light emission occurs; the concentration of NO remains constant during the peak emission. This glow can be detected by means of a photomultiplier, so that the endpoint of the titration can de determined with precision. This method is highly sensitive and specific. It has been successfully applied to pressures below one torr, and the presence of neutral gases such as Argon does not interfere with the measurements. Unfortunately, the implementation of this method is not easy, since the equipment required is not only bulky, but it is not commercially available and has high demands in terms of vacuum requirements. 3.6. MASS SPECTROMETERS [4]

The main principle behind this kind of instrumentation is the ionization of all the species in the beam so as to later divert and detect them. This technique detects mass-to-charge ratios, and as such, would give the same reading for both O+ and O+2 2 . Mass spectrometers have a virtually indefinite useful detection life and can also be used to detect other species present in the flux. The most common type is the quadruple mass spectrometer, which has been successfully used on numerous occasions for both flight and ground AO measurements. A disadvantage of this technique is that the equipment required may be expensive, consume a large amount of power, and it is generally bulky. Other difficulties may include the recombination of AO before detection and that its inability to detect kinetic energies (unless supplemented by additional equipment for a timeof-flight measurement or other technique).

ATOMIC OXYGEN SOURCE CALIBRATION ISSUES

437

3.7. ACTINOMETERS [4, 11]

These methods have the significant advantages of allowing in-situ measurement, high specificity and sensitivity, low complexity, low cost, remote operation, and high potential for miniaturization. For the purposes stated in this paper, suitable in-situ actinometers candidates are silver, carbon, and zinc oxide (ZnO). All give a resistance increase when AO is absorbed and/or reacts with the sensor material. Resistance is easily measured in-situ and hence, unlike Kapton, real-time determination is possible. Silver actinometers use the change in resistance of thin silver films as they are exposed to atomic oxygen. The progressive conversion of the silver into nonconducting silver oxides causes the film resistance to increase, the extent of which depends on the total fluence (integrated flux) of atomic oxygen to which the film is exposed. This method assumes uniform oxidation. The relation between the resistance, R, of the silver sensing film and its dimensions is simply given by R = ρo (1 + 3λ/(8τ ))L/W τ where ρo is the resistivity of the conductor, L is the length, W is the width, τ is the thickness of the sensing element, and λ is the mean free path of the conducting ˚ for silver films between 200 and 4000 A ˚ of electrons, value that is around 530 A thickness. The silver oxidation process apparently involves three stages. The first stage is ˚ This is typical of a process controlled by linear up to a depth of 340 +/− 100 A. surface reaction. The second stage appears to be parabolic, a typical response of a process controlled by diffusion. This suggests that the oxide layer does not inhibit ˚ of oxide has the transport of oxygen atoms until a depth of approximately 340 A formed. The final stage involves a rapid increase in the rate of change of the resistance, ˚ This later stage is believed to be an end-effect caused up to a depth of about 400 A. by the breakdown of the very thin conduction films into discrete islands, and the data collected on this range in not reliable. One disadvantage of this method is that it requires calibration with a known source of atomic oxygen, to obtain a plot of film resistance vs. flux for a particular film thickness and time dose. Once a calibration has been performed (or one of the published ones is used), it presents a reliable, accurate, and reproducible method to determine the flux of atomic oxygen. This method has a limited useful life as the silver is consumed in the oxidation process. Unlike silver, carbon releasing volatile oxidation products and hence carbon actinometers are not diffusion limited. For carbon actinometers, resistance (R) can be calculated by [12]: R = τo Ro /(τ o − F Y)

438

CARL WHITE ET AL.

where τ o is initial thickness in cm, Ro is initial resistance, F is fluence in atoms · cm−2 , and Y is erosion yield in cm3 · atom−1 . In common with silver, carbon actinometers are consumed by the oxidation process and hence have a limited life. Osborne [13] has demonstrated that thin, sputtered films of ZnO are also sensitive to AO flux. They have the advantage over silver and carbon actinometers that they can be regenerated by heating to moderate temperatures which, in principle, allows their useful lifetimes to be extended indefinitely. Unfortunately, recent experience with these sensors suggests that their response is variable and dependent on film deposition conditions. Currently, there is research at the University of Southampton on thick film actinometers that may have a more reproducible response to AO flux. It should be emphasized that actinometers like the ones discussed here have the great advantage of being suitable for miniaturization. It is even conceivable to produce a small integrated circuit with more than one actinometer in it, that would require minimum power, and that would have all the electronics required for it. This small “chip” could be produced in large quantities and used as the test sample for the proposed “round robin” test.

4. Atomic Oxygen Beam Energy Measurment A good and reliable way to measure the energy of a beam of AO is the “Time-ofFlight” technique. Its basic principle is to measure the time oxygen atoms in the beam require to travel a known distance. By dividing the distance by the time we would know the velocity. The kinetic energy can be calculated by means of the following equation: E = mv2/2

where E is the kinetic energy, v is the speed, and m is the mass of an atom of oxygen.To provide an accurate speed measurement, beam modulation methods are typically used. In many instances, mechanical modulation is achieved by means of a chopper. A good overview of this method and of some of the instruments required to implement it can be found in [14].

5. A Sensor to Monitor UV Radiation As mentioned before, it is known that UV radiation and atomic oxygen have synergistic effects. Given the fact that the production of AO can lead to the emission of UV in the vacuum chamber, this parameter should be monitored so as

ATOMIC OXYGEN SOURCE CALIBRATION ISSUES

439

to make a correct evaluation of a space environment simulator. Needless to say, this characterization of the UV environment in the simulator should also take into consideration other sources like lamps. This could be achieved by solar-visible blind aluminium gallium nitride photodetector. This kind of sensor has been proposed [15] to monitor the environment in the proximities of the International Space Station. This detector is sensible to light in the range 120–285 nm. Due to the importance [15] of detecting with accuracy of the Lyman-alpha emission, a sensor of this type can be covered with a sapphire window so as to cut out the UV spectrum below approximately 130 nm. Therefore, the emission between 120 and 130 nm could be inferred by the difference in the measurement on the covered and the uncovered sensors. It is important to mention that such sensors are not very costly, require little volume and power to operate. They are also very suitable for automatic control. They required though that temperature be monitored since their response is sensitive to temperature.

6. Proposed “Round Robin” Experiment to Determine Atomic Oxygen Fluxes We can conclude from the discussion presented above, that different AO sources produce fluxes with different excited states and energies and that sometimes these differences can be significant. There may also be differences in the radiation level and contaminant levels which could either accelerate or retard an AO reaction. With such variations a full source calibration is required if the behaviour of exposed materials is to be accurately quantified and understood. The rationale behind a “round robin” is to define a standard set of experiments and measurements which if conducted in a source would produce a universally accepted calibration of that source. Thus, an “ideal” method or equipment to compare and evaluate AO sources should have the following qualities:

r Ease of manufacture and operation. r Low cost. r Relatively small size and ease of integration with major AO simulator facilities.

r Able to measure the flux density. r Ability to measure the beam energy. r Capable of providing a qualitative and quantitative chemical analysis of the species present in the beam.

440

CARL WHITE ET AL.

For absolute confidence in the outcome, the calibration should as far as possible use a standard experimental package and be carried out under the supervision of the same team. It would also be beneficial if this package is suitable of AO exposure in LEO, so as to allow the comparison of orbital and laboratory fluxes. Based on the review on AO flux measurement techniques presented in this publication, it seems reasonable to propose the design of a small integrated circuit with a carbon-based actinometer and a silver-coated or carbon-coated QCM crystal. If problems of reproducibility can be overcome, a ZnO actinometer could also be included since it is potentially very sensitive and reusable. To measure the deposition of contaminants on the internal surfaces of the space simulator, a QCM crystal without coating can be used. It can be easily added to the circuit and share some of the electronics from the other QCMs on it. Kapton witness samples must also form part of the experimental package since it is currently regarded as being the “industry” standard and the majority of quoted erosion yields are based on Kapton reference samples. The proposed experiment should also include two solar-visible blind aluminium gallium nitride photodetectors; one of them should be covered with a sapphire window. This sensor arrangement would allow the UV radiation to be monitored in the chamber, with particular attention to the 120–130 nm spectrum range. A small temperature monitor should be included. To overcome the practical difficulties associated with working in different facilities, the development of a miniaturized integrated package is seen as being an essential development. 7. Challenges The discussion above presents the need for, and the method of, conducting a series of “round-robin” tests of laboratory-based AO simulation facilities worldwide in order to obtain a benchmark comparison. Although feasible in the nearterm, such a program nevertheless presents some additional challenges. These include: 1. The chemical nature of the oxygen atoms (e.g., ionic, neutral) is harder to determine using the proposed techniques. Mass spectrometers should be used where available, but this may be hard to implement in all facilities. 2. The proposed “round robin” experiment does not measure AO kinetic energies. Suitable techniques like time-of-flight mass spectroscopy should be used when possible. 3. The proposed techniques should first be integrated and tested in a “standard facility”. This process would ease the calibration of the instrumentation and allow a more reliable comparison among facilities.

ATOMIC OXYGEN SOURCE CALIBRATION ISSUES

441

References 1. Kleiman, J., Iskanderova, Z., Gudimenko, Y., and Horodetsky, S. (2003) In Proceedings of the 9th International Symposium on Materials in a Space Environment, The Netherlands, ESA Publication SP-540, 2003. 2. Grossman, E., Guzman, I., Viel-Inguimbert, V., and Dinguirard, M. (2003) Journal of Spacecraft and Rockets 40(1), 110–113. 3. Valer, J. C. (2000) Thesis for the Masters in Aerospace Engineering Degree presented at the University of Tennessee Space Institute, August 2000. 4. Osborne, J. J., Harris, I. L., Roberts, G. T., and Chambers, A. R. (2001) Review of Scientific Instruments 72(11), pp. 4025–4041. 5. Reddy, M. R. and Mater, J. (1995) Science 30, 281. 6. Leger, L. J. and Visentine, J. T. (1986) Aerospace America 24, 32. 7. Wolan, J. T. and Hoflund, G. B. (1999) The Journal of Vacuum Science And Technology A 17(2), 662–664. 8. Tim Minton, personal communication. 9. Elias, L., Ogryzlo, F. A., and Schiff, H. I. (1959) Canadian Journal of Chemistry 37, 1690. 10. Thrush, B. A. (1967) Science 156, 470–473. 11. Harris, I. L., Chambers, A. R., and Roberts, G. T. (1997) Review of Scientific Instruments 68(8), 3220–3238. 12. White, C. B., Rao, J., Roberts, G. T., Chambers, A. R., Lawson, K. J., and Nichols, J. R. (2003) In Proceedings of the 9th International Symposium on Materials in a Space Environment, The Netherlands, ESA Publication SP-540, 2003. 13. Osborne, J. J., Roberts, G. T., Chambers A. R., and Gabriel, S.B. (1999) Review of Scientific Instruments 70(5), 2500–2506. 14. Auerback, D. J. (1988) Atomic and Molecular Beam Methods, Vol. 1, Oxford University Press, pp. 362–379. 15. Dinguirard, M., Mandeville, J. C., Van Eesbeek, M., Tighe, A. P., Durin, C., Chambers, A., Gabriel, S., Goulty, D., and Roberts, G. (2001) American Institute of Aeronautics and Astronautics Publication, # 2001-5070.

LOW-COST SPACE MISSIONS FOR SCIENTIFIC AND TECHNOLOGICAL INVESTIGATIONS

DANIEL RANKIN, DR. ROBERT E. ZEE, FREDDY M. PRANAJAYA, DANIEL G. FOISY, AND ALEXANDER M. BEATTIE Space Flight Laboratory, University of Toronto Institute for Aerospace Studies 4925 Dufferin Street, Toronto, Ontario

Abstract. The Space Flight Laboratory at the University of Toronto Institute for Aerospace Studies (UTIAS/SFL) is pioneering the use of commercial off-the-shelf technologies in space to support education and rapid access to space for Canadian researchers. The Canadian Advanced Nanospace eXperiment (CanX) Program empowers researchers across Canada by providing spacecraft under 10 kg that facilitate inexpensive research in low Earth orbit. The limited resources available to these “nanosatellites” force innovation in the way systems are built, how experiments are defined, and in the instrumentation that is used to support international class science and technology missions. Graduate students at UTIAS/SFL build nanosatellites under the supervision and mentoring of UTIAS/SFL staff while payloads are defined and developed by other universities and businesses across Canada. CanX-1, the program’s first nanosatellite, was launched on 30 June 2003. The objectives of CanX-1 were to qualify critical system technologies in support of future nanosatellite missions, such as CanX-2. Currently under development, CanX-2 is planned to be the first nanosatellite in the CanX program to support a variety of experiments for researchers across Canada. CanX nanosatellites are an excellent means to acquire fast heritage for new technologies—technologies that will be critical in supporting the demanding satellite missions of tomorrow. Highperformance radiation-robust computer systems, high-speed radio transceivers, and electric propulsion systems, are a few examples of microspace technologies under development at UTIAS/SFL that can be tested and proven through the CanX Program. Key words: Low-Cost space missions, nanosatellites 1. Introduction The Canadian Advanced Nanospace eXperiment (CanX) series of satellites was started in September of 2001 at the University of Toronto Institute for Aerospace 443 J.I. Kleiman (ed.), Protection of Materials and Structures from Space Environment, 443–454. C 2006 Springer. Printed in the Netherlands.

444

DANIEL RANKIN ET AL.

Figure 1. Student training at UTIAS/SFL

Studies’ Space Flight Laboratory (UTIAS/SFL). The CanX program began with the intention of providing an opportunity for Canadian graduate engineering students to learn about the field of microsatellite engineering, while at the same time providing a low-cost orbital platform for Canadian scientists. CanX missions use the CubeSat standard developed by Stanford and CalPoly universities, with the aim of lowering satellite launch costs through standardization that allows small cube shaped satellites to be launched inside a compatible deployment system. A single CubeSat has a mass of less than 1 kg and a side length of 10 cm, and is cube shaped as the name suggests. The CanX program also takes advantage of the latest advances in technologies that are applicable to space, by using a relatively short design cycle. As an integral part of the CanX program, student participation is essential for the success of the program. The expected design cycle of a CanX satellite, lasting approximately 18–24 months, nicely coincides with the length of time it typically takes for students to complete a Master’s degree. In this way, the students are able to experience a complete satellite development cycle, and leave the CanX program with training in all phases of satellite design, construction, testing, and operations (see figure 1). In addition to the student team responsible for much of the work on a CanX satellite, there is also a team of UTIAS/SFL staff members who may design some subsystems of the satellite. The range of fields covered by the SFL staff members includes computer engineering, power systems engineering, radio frequency communications, systems engineering, propulsion design, and satellite testing. The staff members have previous experience in microsatellite design, and are able to mentor the students and share the lessons learned through previous SFL satellite experiences to help ensure the success of the current CanX mission.

LOW-COST SPACE MISSIONS

445

2. Facilities The Space Flight Laboratory (SFL) at the University of Toronto Institute for Aerospace Studies (UTIAS) is a modern satellite engineering facility built within the confines of a world-recognized centre for space research. The laboratory has incorporated significant facilities to allow most of the design, assembly, and testing of SFL satellites to be accomplished in-house. For satellite design, there are significant computing facilities along with many of the latest software packages for aiding the development mechanical, electrical, and software designs. There are also facilities for constructing basic mechanical and electrical prototypes of flight systems. After prototypes are constructed, SFL has facilities to conduct tests on the prototype systems. SFL has two thermal cycling chambers, which can test items ◦ ◦ within a temperature range of −70 C to +180 C. SFL also possesses equipment to operate a small vacuum chamber which can be used within a thermal chamber, thus allowing in-house thermal-vacuum testing of spacecraft components, or in the case of CanX series satellites the entire spacecraft. For radio testing, SFL possesses a small anechoic chamber. There are also facilities available for spacecraft testing, as well as instruments such as oscilloscopes, spectrum analyzers, and signal generators. For vibration testing and EMI/EMC testing, UTIAS/SFL has relationships with other departments at the University of Toronto, some Canadian government agencies, and industry partners to provide support for testing that cannot be conducted in-house. For final spacecraft integration and assembly, there is a Class 10,000 clean room located at UTIAS/SFL. The clean room facility at SFL allows for integrated functional testing in-house, saving the cost of having to move the established test support equipment to an off site clean room facility (see figure 2). UTIAS/SFL also contains ground station facilities for communicating with and tracking SFL satellites. There are two separate ground stations at the present

Figure 2. Clean room facility at UTIAS/SFL

446

DANIEL RANKIN ET AL.

Figure 3. Mission control at UTIAS/SFL

time: the MOST ground station, and the CanX ground station. The MOST ground station has facilities for VHF/UHF/S-band satellite communications using Yagi antennas, as well as S-Band downlink capability with a 2.3 m parabolic dish antenna. At present, the CanX ground station works in half-duplex with UHF uplink and downlink using a dual Yagi antenna array. The ground station control area at SFL also has a large screen projection system, so that major spacecraft events can be watched easily by larger groups (see figure 3).

3. CANX-1 The CanX-1 project commenced in September 2001 as the first in a series of CanX satellites to be designed and built at UTIAS/SFL. The program goals of CanX-1 were to provide education for students and to set up laboratory infrastructure for future CanX use. CanX-1 is a nanosatellite (satellite < 10 kg). At 2 kg and in the shape of a 10 cm cube (figure 4), the mission goals for CanX-1 are to demonstrate systems for use in future CanX missions, as well as to demonstrate several technologies in space. The systems included in CanX-1 are

r r r r r r

A custom designed on board computer (OBC) using an ARM7 processor. A custom designed UHF radio operating in Amateur radio frequencies. A magnetic attitude control system including an on board magnetometer. Two CMOS imagers (colour and monochrome) on a custom designed board. A CMC electronics GPS receiver. A Xiphos Q4 board.

LOW-COST SPACE MISSIONS

447

Figure 4. CanX-1 (without antennas)

The CanX-1 OBC is responsible for managing all of the satellite systems including power, payloads, attitude control, and the UHF radio. The OBC operates at 40 MHz, with application programs running in 512 kB of error detection and correction (EDAC) controlled SRAM memory. For bulk data storage, there is also 32 MB of flash memory on board. The satellite bootloader is written on a 128 kB EPROM, in order to increase the reliability of this critical system. Power is supplied to CanX-1 using an array of Emcore triple junction Gallium Arsenide solar cells. These cells provide approximately 2 W of power to the system when the satellite is in sunlight, with the exception of the +Z face of the satellite that only has one cell instead of the nominal two. The UHF radio system is the only means of communication between the ground and CanX-1. It operates in the amateur band of frequencies with a downlink frequency of 437.88 MHz, and was custom-designed for CanX-1 by students at SFL. The radio system is half-duplex, working at a data rate of 1200 bps. The magnetic Attitude Control System (ACS) for CanX-1 utilizes three airwound magnetic torquer coils. These torquers, orthogonally mounted underneath three of the CanX-1 solar panels, provide the torques needed to implement a “B-dot” control law. This controller is used to reduce the angular momentum of the satellite, i.e., detumble the satellite. Through the use of a 3-axis magnetometer supplied by Honeywell, the rate of change of the magnetic field is sensed and current is applied to the magnetic torquer coils in order to slow the rotation of the spacecraft. The software needed to accomplish these tasks is run by the OBC on the EDAC protected SRAM. The CMOS imagers are included on the CanX-1 mission for two purposes. The monochrome imager, with a field of view of approximately 5◦ , is intended mainly to conduct ground based star-tracking experiments on the returned images. The colour imager, with a field of view of approximately 40◦ , is to be used to

448

DANIEL RANKIN ET AL.

Figure 5. From left to right, CanX-1, DTUSat, AAU CubeSat

take pictures of Earth and its moon. Both imagers are manufactured by Agilent Technologies and have a resolution of 640 × 480 pixels. CanX-1 also contains two other payloads. The Xiphos Q4 board was provided by Xiphos Technologies of Montreal, Quebec. The Q4 board contains an array of sensors for on-orbit evaluation, including angular rate sensors, temperature sensors, and magnetometers. The other payload is a CMC Electronics GPS receiver. This GPS receiver acquires signals through two small patch antennas on the Y panels of the spacecraft. The GPS receiver is able to determine spacecraft position and velocity when multiple GPS satellites are in view. The CanX-1 design was completed in October 2002, with acceptance testing taking place through January–April of 2003. CanX-1 was launched from Plesetsk Cosmodrome in northern Russia on June 30th, 2003, along with the MOST microsatellite also assembled at UTIAS/SFL. CanX-1 was launched in a deployment tube along with two other CubeSats (DTUSat and AAU CubeSat) developed by Danish university student teams (see figure 5 and 6). Unfortunately, no contact was ever made with CanX-1 after launch. In the first couple of months following launch, none of the three satellites, CanX-1, DTUSat, or AAU CubeSat was contacted. Various large dish antennas were commissioned in an attempt to make contact. AAU CubeSat was finally contacted after securing the use of a 8 m dish antenna in Denmark. However, they were only able to download minimal telemetry and were not able to complete their mission. To this day neither CanX-1 nor DTUSat have been contacted. After going through numerous contingency scenarios, and after obtaining the help of the Algonquin Radio Observatory and a 9 m ground station at Defence Research and Development Canada (Ottawa) to try and make contact with CanX-1, the UTIAS/SFL student team was unable to do so. The absence of data makes it difficult to draw definitive conclusions. A complete failure report has been written based on the format used for recent NASA failures, and the most plausible causes of failure have been identified. All that

LOW-COST SPACE MISSIONS

449

Figure 6. NLS-1 launch tube containing CanX-1, DTUSat, AAU CubeSat mounted on rocket upper stage

can be done is to extrapolate from the extensive environmental and functional testing performed on the satellite prior to launch. In addition to the satellite failure modes, the possibility of damage from the launch tube or interaction with adjacent satellites leading to damage have been considered. The lessons learned from CanX-1 and the accompanying recommendations are being applied to the CanX-2 project now underway. Among the major programmatic improvements, some funding from NSERC has allowed the UTIAS/SFL staff to be more heavily involved in supervision, to develop critical components (radios and power systems), and to work side by side with students in an integrated satellite team. In addition, UTIAS/SFL is now working with nanosatellite developers at the University of Tokyo to develop a gentler and more reliable separation system based on a space proven design. This separation system will also allow each nanosatellite to be ejected independently from the launch vehicle. CanX-1 achieved its goal of establishing the essential elements of the CanX program and has produced valuable lessons that will only serve to increase the reliability of future CanX missions.

4. CANX-2 Initiated in September 2003, CanX-2 is the second in the series of CanX satellites. At present, CanX-2 is in the detailed design phase. CanX-2 is planned to be one

450

DANIEL RANKIN ET AL.

Figure 7. Solid model of CanX-2

of the first operational science nanosatellites for Canadian researchers. The size of the satellite is roughly double that of CanX-1, allowing for greater available volume and surface area for power generation (figure 7). CanX-2 uses a UTIAS/SFL developed OBC, this time based on an ARM9 processor at 18 MHz. The OBC will use 512 kB of EDAC protected SRAM and 32 MB of flash memory for mass data storage (see figure 8). Power is provided by triple junction GaAs solar cells, and with more surface area compared to CanX-1, the maximum power generation is increased to 4 W. Primary radio communications is accomplished using half-duplex UHF, and the bit rate for CanX-2 is planned to be 4800 bps. CanX-2 also includes an extensive suite of science and engineering payloads. These payloads consist of:

r r r r r r r

A three-axis momentum-bias coarse pointing attitude control system. A high-data-rate S-Band transmitter. Two CMOS imagers (colour and monochrome). A Nanocalorimeter. An Atmospheric Spectrometer. A GPS receiver for GPS radio occultation studies. A Surface Materials experiment.

Like CanX-1, the CanX-2 Attitude Control System (ACS) utilizes three magnetorquer coils. In addition, CanX-2 will include a momentum wheel provided by Dynacon, Inc. The goal for the system is to have both attitude determination and pointing within one degree of accuracy. The ACS is included to both test the

LOW-COST SPACE MISSIONS

451

Figure 8. CanX on-board computer

momentum wheel technology, and to support the other science and engineering payloads. The S-Band transmitter is an SFL designed product to test the performance of miniaturized S-Band technology for nanosatellites, while also increasing the amount of science data that can be received by the ground. The maximum designed information rate for the S-Band transmitter is 256 kbps, and this can be scaled on the fly; this allows higher data rates when the satellite is at higher elevations in the sky. The two CMOS imagers that are being planned for CanX-2 are very similar to the CanX-1 imagers. These imagers, manufactured by National Semiconductor, are slightly larger than the CanX-1 imagers, and both have planned fields of view of about 30◦ with a resolution of 1200 × 1024. Again, the monochrome imager provides the option of doing ground-based star tracking experiments, while the colour imager will be used mainly to take pictures of interesting targets like Earth and its moon. Both cameras may be used for on orbit calibration of the ACS performance. The Nanocalorimeter experiment is provided by Dr. Harry Ruda of the University of Toronto. The proposed experiment is aimed at studying the influence of space-based radiation on the properties of semiconductor nanowhiskers with and

452

DANIEL RANKIN ET AL.

without double barrier structures. The end goal of this experiment is to develop a semiconductor-based heat-sensing array with sensors as small as 50–100 nm. The Atmospheric Spectrometer, provided by Dr. Brendan Quine of York University, is an Earth imaging spectrometer. It provides measurements of airborne greenhouse gases to support the goals of the Kyoto protocol. The payload operates in the near infrared band, and features a surface resolution of 1 km, which will enable the identification of local variation and sources of pollution emission. The GPS experiment, designed by Dr. Susan Skone of the University of Calgary, uses a directional GPS antenna mounted on the outer surface of the satellite to take measurements as GPS satellites are occulted by the Earth’s atmosphere. From this data, a detailed profile of tropospheric water vapor can be generated, along with atmospheric electron densities. The surface materials experiment for CanX-2 is provided by Dr. Jacob Kleiman of the University of Toronto in collaboration with Prof. Alan Chambers from the University of Southampton, UK. This experiment uses a specifically developed array of 4 atomic oxygen sensors. One or more sensors will be treated by Dr. Kleiman’s group to enhance their oxidative stability. The behavior of all sensors will be evaluated during the flight. The plan is to monitor the changes in sample thickness as a result of atomic oxygen erosion to evaluate the effectiveness of the special surface treatment. In addition, a network communications experiment involving an innovative satellite communication protocol developed by Dr. Michel Barbeau of Carleton University is included in CanX-2. This experiment will test a satellite networking protocol under the open-source operating system of eCos. With the design of CanX-2 well underway, the flight hardware is planned for completion by the end of 2004. With a projected launch in late-2005, CanX-2 will be the second Canadian nanosatellite launched into space.

5. Conclusion The CanX program serves Canada and international partners by providing exceptional hands-on training to graduate students in the area of space systems engineering. At present, this program is unique in Canada. University researchers, industry and government have opportunities to fly science and technology experiments in space cheaply and rapidly. The CanX program is intended for aggressive experimentation in space. In combination with the training aspect of the program, missions are completed in less than two years for a few hundred thousand dollars. In keeping with the nano/microspace philosophy, redundancy in the satellite is traded for simple, good design, with mission risk distributed over multiple low-cost missions rather than over multiple components in a single mission. The CanX program is pioneering the low-cost exploitation of space and intends to revolutionize

LOW-COST SPACE MISSIONS

453

Canadian space activity in the 21st century. The UTIAS Space Flight Laboratory encourages companies, university groups and government agencies across the world to collaborate on these low-cost missions. For more information please visit www.utias-sfl.net.

Acknowledgements The UTIAS Space Flight Laboratory gratefully acknowledges the following sponsors of the CanX program:

r r r r r r r r

Defense Research and Development Canada (Ottawa) Natural Sciences and Engineering Research Council (NSERC) MacDonald Dettwiler Robotics Dynacon Incorporated Ontario Centers of Excellence, CRESTech Division Canadian Space Agency Xiphos Technologies Inc. Radio Amateur Satellite Corporation (AMSAT)

In addition, the following organizations have made valuable donations to the program: AeroAntenna Technology Inc. Agilent Technologies Altera Altium Analytical Graphics Ansoft

454 ARC International ATI Autodesk @lliance Technologies Cadence CMC Electronics EDC E Jordan Brookes Emcore Encad Honeywell Micrografx National Instruments Natural Resources Canada Raymond EMC Rogers Corporation Stanford University Texas Instruments The Mathworks Wind River

DANIEL RANKIN ET AL.

SUBJECT INDEX

ab initio calculation, 359, 360 absorptance, 75, 124, 171, 236, 239, 240–242, 252, 257, 259, 260, 266, 277, 284, 285, 290, 293 acrylic, 62, 168, 377 actinometers, 437, 438 activation energy, 51, 292, 317, 318, 326, 327 adhesion, 268, 283, 292, 307, 309, 310, 312–315 adhesive, 123, 180, 295–302, 305–307, 321, 414 algorithm, 27, 265 alkanes, 365–367, 369, 370, 372–374 aluminum, 10, 11, 13, 14, 18, 62, 73, 74, 154–156, 158, 162, 175, 177, 180, 183, 190, 217–222, 233, 242, 253, 265, 268, 283, 298, 334, 342, 344, 345, 395, 396, 419 aluminum foil, 158, 253 angle dependence, 144, 146 angular distribution, 96, 97, 100, 427 antenna, 446, 448, 452 AO laboratory simulation, 252 AO protection, 171, 277, 289, 290 arc plumes, 9 arcing, 9, 14, 16, 17, 18, 331 array, 9–18, 21, 71, 75, 77–85, 93, 94, 98, 101–104, 124, 154, 295–297, 299, 301, 306, 393, 398, 399, 414, 446, 447, 448, 452 atmosphere, 38, 54, 127, 194, 218, 249, 254, 334, 365, 434, 452 atomic oxygen (AO), 51–56, 58, 75, 77, 102, 141–151, 154, 167, 168, 247–249, 252, 253, 255, 257–262, 266, 277, 278, 289, 290, 295–300, 304, 306, 317–328, 341–349, 365, 366, 373, 401–403, 414, 417–438, 452 atomic oxygen attack, 248, 257, 401 atomic oxygen-bombardment, 342, 349 atomic oxygen-effective fluence, 252 atomic oxygen-erosion, 52, 102, 144, 249, 341, 452

barrier to reaction, 372 binder, 35, 62, 74, 168, 169, 172, 247, 377, 382, 383 blends, 248 blocking, 17, 295–297, 300, 301, 303, 306, 307, 414, 415 bulk metallic glass, 217 calibration, 237, 303, 337, 402, 415, 416, 431, 434, 437, 439, 440, 451 Canadian Advanced Nanospace eXperiment (CanX), 443–453 Canadian Space Agency (CSA), 277, 289, 290, 293, 341, 343, 344, 453 carbon/epoxy composites, 209 carbon–carbon composite, 326, 372, 373 C-C breakage, 365–374 ceramer, 247–249, 252, 256–262 charged particle(s), 1, 7, 22, 35, 37, 107, 108, 115, 122, 141, 233, 238, 239, 241, 277, 351 charging, 18, 51, 52, 56–58, 115, 331, 359, 393 clean room, 445 coatings (see also protective coatings), 45, 61, 62, 64, 68, 71, 73, 77, 85, 87–91, 93, 94, 104, 167, 168, 171–173, 247, 248, 252, 254, 255, 257, 258, 262, 265, 266, 272, 277, 281, 292, 331, 332, 339, 343, 392 collision energy, 318, 326, 363, 365, 366, 370–373 colour center, 112, 113, 351–355 composite material (see also carbon/epoxy composites), 248, 249 concentrator, 123, 124, 258, 417–419, 426, 428 contaminants, 17, 128, 129, 253, 435, 440 conversion, 31, 247, 248, 253, 265, 269, 270, 273–275, 306, 437 coverglass, 14, 17, 394, 398 crack(s), 153, 160, 163, 164, 217, 220, 222, 223 cross-linking, 213, 214, 254, 407 damage, 1, 2, 5, 7–10, 16, 35, 37, 62, 66, 71, 74, 76–81, 84, 85, 90, 93, 94, 101, 102, 104, 108, 112, 124, 125, 153, 159–161, 175, 176,

455

456

SUBJECT INDEX

damage (cont.) 180, 181, 183, 185, 190, 214, 217–222, 233, 240–245, 248, 254, 266, 306, 309, 351, 353, 354, 356, 394, 398, 401, 402, 414, 416, 449 debonding effect, 209, 212 debris, 72, 73, 141, 153–155, 158, 159, 164, 175–178, 181, 183–186, 189, 190, 218, 223, 235, 309, 314, 315, 385, 398, 418 defect(s), 68, 88, 111, 112, 279, 354 degradation, 1, 3, 5, 21, 22, 25, 27–29, 32, 35, 37, 39–41, 62, 65, 68, 74, 77, 80, 84, 87, 89, 90, 93, 94, 104, 123–125, 127–139, 152–154, 167, 170, 173, 190, 209, 225, 231, 258, 266, 278, 300, 301, 344, 346, 365–367, 369, 393, 398, 402 desorption, 43, 44, 377, 378, 379, 380, 433 detector(s), 116, 127, 128, 286, 287, 319, 426, 427, 435, 439 diffusion, 10, 43, 44, 62, 112, 377, 378, 380, 435, 437 direct dynamics, 359, 362 dose rate, 45, 353, 354 Earth, 1, 9, 51, 85, 94, 107, 115, 141, 153, 167, 175–177, 209, 247, 248, 277, 278, 306, 317, 318, 331, 341, 344, 359, 365, 398, 401–403, 417, 418, 431, 443, 448, 451, 452 electrochromic, 279, 280, 342 electron beam, 2, 52, 54–57, 226, 342 electron cyclotron resonance (ECR), 295, 297, 306, 401–416 electron flux, 44, 49, 115–122, 240 electron radiation, 4, 36, 43–49, 107, 108, 110, 111, 240, 241, 243, 296, 298–300, 303 electron(s), 1, 2, 4–6, 8, 10, 14, 36–40, 43–49, 52–57, 61–65, 73, 107–122, 158, 175, 181, 219, 226, 238–243, 252, 278, 279, 288, 295–303, 307, 309, 310, 331, 339, 342, 343, 354, 360, 366, 401–403, 432, 433, 435, 452 ellipsometry, 126, 133, 265, 269, 274 enamel, 62, 91, 168, 169, 173 environmental durability, 402 erosion effect(s), 85 erosion rate, 51–53, 58, 141–151, 318, 326, 366, 419 erosion yield, 169, 172, 173, 317–321, 325, 327, 434, 438, 440

fabric, 168, 170, 173 FEP, 278, 365, 366, 373 FEP/Teflon, 278 FEP/Teflon-silvered (Ag-FEP), 77 fiber(s), 168, 170, 333, 334 fiberglass, 10 flexibility, 21, 81, 124, 248, 417 fluence, 1–8, 35–37, 49, 51, 52, 56, 58, 64–68, 87–91, 127, 129–132, 138, 142, 145, 146, 148, 150, 151, 167, 169–172, 247, 249, 252, 256–258, 262, 289–300, 303–307, 317–328, 317–327, 344, 345, 352, 353, 401–404, 407–419, 434, 437, 438 fluorinated polymer(s), 366, 374 fluorine, 335, 367, 373 fluoropolymer degradation, 365 FPGA, 21–32 fracture, 37, 38, 153, 160–165, 217, 220, 221, 223, 225, 231, 313, 315, 393 friction coefficient, 311, 315 Fuel Oxidizer Reaction Products (FORP), 193–208 geomagnetic, 115 glass transition temperature, 161, 164, 254 glass-optical, 107–113 ground laboratory facility (see also AO laboratory simulation), 124, 402 ground-based testing, 85 hardness, 182, 254, 306, 309, 310, 314, 315, 401–415 hazards, 141, 175 heating, 72, 84, 94, 210, 211, 265, 277, 290, 291, 314, 381, 395, 438 hermetic, 378, 385, 390, 391 high temperature, 161, 162, 254, 273, 280, 312, 313 high-energy, 45, 53, 54, 124, 141, 150, 226, 236, 248, 249, 255, 295, 296, 318 high-velocity impact, 155, 161, 393, 394 hydrocarbon polymers, 247, 326, 366, 374 hyperthermal-beam, 317, 320, 321, 325 hyperthermal atomic oxygen, 143, 154, 317–328, 365, 366, 373, 420 hyperthermal atomic oxygen beam, 420 hyperthermal-reaction, 326, 365–374 impact, 35, 51, 52, 71–73, 76, 81, 87, 89–93, 96, 98–104, 122, 153–155, 158, 160–169,

SUBJECT INDEX 173, 175, 176, 179–190, 217–219, 221–223, 233, 235, 237, 240, 241, 243–245, 258, 266, 360, 362, 363, 370, 385, 393–397, 421 inclination, 115, 116, 186 inelastic scattering, 422, 428 inorganic, 173, 247, 248 insulator, 277–280, 286, 288, 289, 292, 331, 339, 342, 385 International Space Station (ISS), 18, 71–85, 93–104, 168, 175–183, 186, 187, 190–198, 205–208, 295–299, 414, 415, 418, 428 ionizing radiation, 43, 44, 153, 296, 300, 302, 303, 307 Kapton, 17, 18, 51, 54, 58, 75, 76, 102, 142, 153, 158–164, 247, 252–258, 262, 278, 281–299, 302, 304, 306, 317–328, 344, 366, 401–404, 407–410, 413–416, 419, 420, 434, 437, 440 Kapton H, 247, 252, 253, 255, 256, 262, 286, 287, 317–328 Kapton HN, 286, 287 Kapton polyimide, 278 LEO environment, 155, 341, 360, 367 LEO erosion, 365 Long Duration Exposure Facility (LDEF), 154, 155, 176, 278 loss tangent (tan δ), 39, 209, 213 low Earth orbit (LEO), 9, 51, 85, 141, 153, 167, 175, 176, 247, 248, 277, 278, 306, 317, 318, 331, 344, 359, 365, 401–403, 417, 443 low temperature, 1, 108, 292, 309–315, 323 magnetic field, 447 mass loss, 35, 37, 41, 47–52, 55–58, 142–144, 151, 154, 169–172, 209, 210, 212, 214, 215, 252, 257, 258, 262, 290, 299, 306, 317, 318, 321, 341, 344–348, 367, 379, 382, 401–404, 415, 419, 434 mass spectrometer(s), 52, 56, 319, 320, 421, 436, 440 material-outgassing, 43–49 mechanical properties, 123–125, 129, 133, 136–139, 155, 167, 170, 209, 241, 405 meteoroids, 154, 176, 183 methane, 210, 359, 363, 365 methyl, 35, 40, 41, 247, 249, 254, 255, 261, 402

457

methyl silicone rubber, 35, 40 microcrack(s), 171, 219, 221, 222, 257, 258, 262, 309, 315, 407 microhardness, 248, 313, 314, 401, 402 mitigation, 9, 16, 18, 21, 71, 76–78, 80–85, 93, 101, 102 model(s), 44–49, 62, 76, 84–90, 99–101, 115, 116, 120, 125–127, 153, 158, 161–165, 176, 186, 204, 205, 252, 272–274, 351, 354, 356, 360, 378, 379, 383, 386–390, 393, 394, 417, 419, 424–428, 450 modeling, 71, 76, 84, 101, 239, 267, 271, 399 molecular dynamics, 365, 367 molecular oxygen, 421, 433 Mylar, 116, 155, 233, 238–245, 334 nanomechanical hardness, 402, 405, 412, 414 N -nitrosodimethylamine (NDMA), 193, 194, 198–208 O-atom, 317–328, 435 on-board computing platform, 21–32 on-orbit, 17, 18, 80, 85, 176, 200, 295–299, 365, 373, 414, 415, 448 optical characteristics, 93, 94, 167, 171, 173, 281, 284, 385 optical coatings, 71, 85, 93, 94, 104, 266 optical constants, 126, 265, 269, 271–274 optical degradation, 87–90, 133 optical density, 351, 353, 355, 356 optical materials, 107, 351, 352 optical microscopy, 36 optical properties, 61–63, 66–69, 75, 77, 107, 125, 138, 159, 168, 225, 226, 230, 231, 233, 239, 241, 245, 252, 259, 260, 266, 289, 339, 346, 391 optical quartz glass, 107, 108, 351, 352 orbital debris, 73, 155, 175, 176, 181–186, 189, 190, 235 organic, 43, 124, 128, 129, 247, 248, 317, 318, 401, 402 oscillator, 271, 355, 379 outgassing, 43–49, 124, 212, 377–383, 414 oxidation, 141, 149, 150, 248, 253, 257, 317, 366, 402, 404, 407, 414, 422–424, 431, 437, 438 oxide(s), 74, 154, 171, 249, 258, 265, 280, 282, 309, 331, 332, 343, 347, 402, 412, 421, 422, 437 oxide layer, 249, 258, 437

458

SUBJECT INDEX

oxygen atom(s), 150, 258, 299, 300, 318, 366, 426, 432, 433, 437, 438, 440 oxygen plasma(s), 167, 168, 172, 173, 247, 289, 401, 420 paint, 44, 61, 74, 75, 87, 154, 172, 181, 182, 184 paint-thermal control, 74 paint-white, 44, 61 perfluorinated alkanes, 365, 366, 372 pitting, 71–73, 76–81, 84, 85, 93, 94, 98, 104 planetary, 154 plasma interaction(s), 9, 10 plasma-asher, 253 polyethylene, 141–151, 162, 218, 219, 225, 366, 373 polyimide, 51–58, 141–151, 153, 159, 168–173, 233, 252, 278, 318, 321, 326, 401, 420, 422 polymer materials, 363, 407 polymer(s), 44, 52, 53, 123, 124, 139, 142–144, 146, 153–161, 169, 242, 247, 248, 250, 255, 258, 280, 318, 319, 325, 326, 359, 360, 363, 366, 367, 374, 377, 407 polymeric composite, 47, 377, 379 power, 14, 16, 17, 22, 24, 25, 30, 71, 84, 87, 123, 124, 143, 159, 176, 177, 268, 270, 275, 278, 279, 291, 316, 331, 333, 334, 342, 385–387, 390, 393, 398, 435–439, 444, 447, 449, 450 product-volatile, 43, 44, 254, 377 protective coating, 102, 167, 173, 247, 249, 257, 258, 262, 290, 331 proton (ir)radiation, 4, 6, 7, 87, 88, 110, 113, 356 proton flux(es), 68 purge, 93 Pyrex, 417, 422–424, 426, 428 quantum chemistry, 359, 360, 363, 365 quartz crystal microbalance (QCM), 51–58, 142–151, 333–339, 421–427, 434, 435, 440 quartz glass, 107–113, 351–357 radiation, 1–8, 21–25, 32, 35–37, 40–49, 61–69, 87– 91, 94, 107–116, 119–125, 129, 132, 133, 137, 141, 146, 150, 153–155, 159, 225, 229, 231, 233–249, 252, 265, 266, 270,

277, 278, 285, 289–292, 296, 298–303, 307, 343, 351–359, 366, 377, 378, 381–383, 386, 393, 433, 438–440, 443, 451 radiation protection, 21 radio-frequency plasma, 253 reaction mechanism, 359–363 reaction pathways, 367, 370 reaction products, 51, 58, 193, 194, 359, 362, 434 re-configurable computing, 21 reflectance, 61–69, 125, 171–173, 236, 239–241, 243, 252, 257, 259, 260, 265–267, 270–275, 281, 284, 285, 290, 302, 339, 342 reflectance-spectral, 239 reflecting mirrors, 417, 427, 428 reliability, 25, 35, 107, 108, 225, 447, 449 remission, 43, 46, 48, 377 roughness, 126, 145, 146, 170, 310, 319–324, 326, 327, 347, 412 satellite, 18, 115–120, 153, 154, 289, 418, 443–453 scanning electron microscopy (SEM), 158, 181–185, 190, 217, 219, 221, 222, 252, 281, 310, 312, 343, 346, 347, 349 sensor, 53, 54, 142, 342, 379, 381, 386, 422, 437, 438–440 SEU-mitigation, 21 shielding, 154, 168, 175, 176, 369, 396 shunting, 393–399 silica, 168, 171, 248, 252–262, 395 silicone, 35–41, 123–138, 168, 171, 173, 247–249, 252–258, 262, 295–307, 401–407, 412–416 simulation test, 107, 167 single event upset (SEU), 21–32 smart material, 277, 278, 279, 341 software, 30, 176, 185, 271–275, 285, 323, 360, 408, 445, 447 solar sail, 233, 234, 243, 245 solar-absorptance, 124, 171, 236, 239, 241, 242, 277, 284, 285, 290, 293 solar-array, 9–18, 71, 74–85, 93, 94, 96, 98, 101–104, 154, 159, 262, 295–299, 306, 310, 393, 398, 399, 414 solar-cell, 1, 2, 4, 7, 8, 16, 17, 35, 75, 84, 94, 98, 101, 102, 107, 123, 124, 209, 257, 258, 393–399, 447, 450

SUBJECT INDEX solar-irradiance, 125–128, 266, 267, 270–272, 275 solar-radiation, 61, 63, 66, 124, 285, 292, 377 solid particles, 393, 399 space environment, 1, 7, 9, 21, 22, 35, 43, 51, 52, 61, 69, 71, 87, 93, 107, 108, 113, 115, 123, 124, 141–143, 153, 155, 167, 175, 193, 209, 217, 225, 231, 233, 235, 236, 247, 265, 266, 272, 277, 278, 290, 293–296, 298, 307, 309, 317, 331, 341, 342, 348, 351, 359, 365, 377, 385, 393, 401, 402, 417, 418, 420, 428, 431, 439, 443 Space Station-International, 18, 71, 72, 93, 104, 175, 185, 193, 295 space-applications, 24, 107, 252, 265, 277, 290, 310, 339 space-components, 154 spacecraft materials, 43–49, 154, 167, 173, 248, 317, 402, 431 space-durability, 123 space-environment effects, 307 space-environmental utilization, 428 space-exposure, 107, 130, 139, 320 space-mission, 21, 443 spectral reflectance, 61, 68, 236, 239, 270, 271 spectrometer, 36, 52, 56, 211, 226, 284, 302, 309, 319, 320, 334, 343, 421, 436, 440, 450, 452 spectrum, 6, 7, 38, 40, 44, 45, 49, 61, 63, 68, 87, 89, 90, 108, 109, 123–130, 137, 139, 142, 171, 172, 213, 225, 227, 229, 241, 258, 265– 267, 270–274, 284, 285, 313, 314, 334, 335, 351, 352, 354, 383, 421, 439, 440, 445 surface modification, 309 surface morphology, 36, 37, 41, 161, 209, 212, 213, 215, 289 synergistic effect(s), 35, 52, 54, 55, 107–113, 141–144, 147–149, 151, 248, 438 synergy, 51, 141, 248 Teflon, 77, 278, 366 tensile properties, 231 tensile strength, 38, 41, 126, 136–139, 164, 226, 227, 242 terephthalate, 225 thermal control-coatings, 61, 87, 266 thermal control-materials, 155, 365

459

thermal-analysis, 211, 252 thermal-cycling, 107, 141, 155, 212, 366, 393, 445 thermal-diffusivity, 62 thermal-emittance, 159, 239, 240, 277, 278, 284, 286, 290, 293, 342, 346, 348 thermal-expansion, 209 thermal-pipelines, 385 thermal-stability, 248 thermochromic, 341–343, 346, 348 thin-film coating, 277, 341 titanium alloy, 309, 315 total mass loss (TML), 212 trajectory, 157, 221, 359, 360, 362, 363, 367, 369–371 transition state, 370 transmission, 107–113, 286, 292 transmittance, 110, 113, 123, 125, 128–136, 139, 226, 240, 252, 257–266, 279–281, 284, 292, 332, 342, 351, 352 tribological properties, 309 tuneable emittance, 277 ultimate tensile strength, 126, 139, 164, 242 ultraviolet (UV) energy, 75 UV exposure, 63, 123, 126–130, 132–139, 149, 225, 226, 237, 239, 242, 302–305 UV irradiation, 61, 64, 225, 227–230 UV radiation, 123, 132, 133, 137, 155, 159, 225, 229, 231, 239, 248, 249, 277, 298–303, 343, 366, 377, 378, 381–383, 433, 438, 440 vacuum thermo-cycling, 214 vacuum ultraviolet (VUV) radiation, 123–139, 149, 154, 155, 225–231, 236, 248, 249, 277, 278, 281, 299, 343, 377, 431 Van Allen (belts), 351 vent, 93, 96, 99 volatile products, 43, 44, 254, 377 VUV-degradation, 125 VUV-exposure, 123, 126, 128, 129, 132–139, 225, 226 VUV-radiation, 123, 132, 133, 137, 155, 225, 229, 231, 249, 277, 343, 377 VUV-source, 137, 139, 225, 226 water dump, 93–104 wavelength, 65, 66, 75, 108–110, 123–138, 143, 156, 171–173, 225–227, 236, 239–243,

460

SUBJECT INDEX

wavelength (cont.) 259, 260, 261, 265–267, 271, 280, 284, 334, 335, 343 wear, 217, 254, 278, 290, 309–315 witness sample, 52, 58, 142, 148, 151, 252, 253, 318, 324, 344, 419, 434, 440

X-ray diffraction, 218, 281, 283, 343 X-ray Photoelectron Spectroscopy (XPS), 54, 142, 143, 150, 225–227, 229, 231, 252, 258, 262, 286, 301, 302, 306, 307, 341, 347–349, 366, 373 X-ray spectrum, 313, 314

AUTHOR INDEX

Abraimov, V. V., 455 Albyn, K., 455 Alred, J., 455 Baker, D., 455 Banks, B. A., 455 Beattie, A. M., 443 Bernhard, R. P., 455 Boeder, P., 455 Buchanan, V., 455 Buczala, D. M., 317 Chaker, M., 455 Chambers, A. R., 455 Chernik, V. N., 455 Christiansen, E. L., 455 de Groh, K. K., 455 Dee, L., 455 Dever, J., 455 Dworak, D., 455 Edwards, D. L., 455 Eliaz, N., 153 Eliezer, S., 153 Fan, C. Z., 217 Fraenkel, M., 153 Ferguson, D. C., 455 Finckenor, M., 455 Galofaro, J. T., 455 Galygin, A. N., 455 Gao, Y., 209 Geng, H., 455 Gerasimova, T. I., 455 Geurkov, V., 21 Gouzman, I., 455 Gray, P., 455 Griffin, J., 331 Grigorevskiy, A. V., 455 Grigoryan, O. R., 455 Grossman, E., 455

Haddad, E., 455 Hambourger, P. D., 455 Hovater, M., 455 Hu, Z., 455 Hubbs, W., 455 Hyde, J. L., 455 Jamroz, W., 455 Jia, Y. Z., 217 Jing, Q., 217 Jiang, X. X., 277 Karniotis, C. A., 247 Khassanchine, R. H., 455 Kirischian, L., 455 Kiseleva, L. V., 455 Kleiman, J., 455 Koontz, S., 455 Kostiuk, V. I., 455 Kruzelecky, R. V., 455 Lam, C. W., 193 Letin, V. A., 393, 455 Li, G., 455 Li, Z., 209 Liu, H., 455 Liu, R. P., 217 Liu, Y., 309 Ma, D., 455 Maeda, K., 51 Maman, S., 153, 455 Matsumoto, K., 455 Mikatarian, R., 455 Minton, T. K., 455 Nadiradze, A. B., 455 Naumov, S. F., 455 Nehls, M., 455 Nikanpour, D., 455 Novikov, L. S., 87, 115, 167, 385, 393 Ohmae, N., 455

461

462 Pankop, C., 455 Peng, G., 225 Petrov, V. L., 115 Pranajaya, F. M., 443 Prosvirikov, V. M., 455 Rankin, D., 443 Roberts, G. T., 455 Schatz, G. C., 455 Schmidl, W., 455 Seikyu, S., 455 Semkin, N. D., 455 Semmel, C., 455 Shevelva, V. N., 115 Smirnova, T. N., 455 Sokolova, S. P., 455 Soltani, M., 455 Solovyev, G. G., 87 Soucek, M., 455 Sun, L. L., 217 Tagawa, M., 460 Tchurilo, I. V., 460 Terterian, I., 21 Thompson, D. W., 265 Timofeev, A. N., 43 Troya, D., 460 Tsvelev, V. M., 460 Uppala, N., 460

AUTHOR INDEX Valer, J. C., 460 Vasil’ev, V. N., 87 Vayner, B. V., 460 Vemulapalli, J., 460 Verker, R., 153 Visentine, J. T., 460 Voronov, K. E., 460 Wang, H. Y., 217 Wang, W. K., 217 Wang, X. Y., 217 Wei, Q., 460 Wertz, G., 460 West, B., 193 White, C., 460 Wong, B., 277 Woods, B. W., 265 Woollam, J. A., 460 Yan, L., 460 Yang, C., 460 Yang, C., 217 Yang, D., 1, 35, 107, 209, 225, 309, 351 Ye, Z., 309 Yokota, K., 460 Zee, R. E., 443 Zelenkevich, A. P., 460 Zhang, J., 460 Zhang, L., 460 Zhang, X. Y., 35