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

86 downloads 1309 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 b