Preface
Preface
In November 1776, Alessandro Volta performed his classic experiment disturbing the sediment of a shallow lake, collecting the gas and demonstrating that this gas was flammable. The science of Biomethanation was born and, ever since, scientists and engineers have worked at understanding this complex anaerobic biological process and harvesting the valuable methane gas produced during anaerobic decomposition. Two lines of exploitation have developed mainly during the last century: the use of anaerobic digestion for stabilization of sewage sludge, and biogas production from animal manure and/or household waste. Lately, the emphasis has been on the hygienic benefit of anaerobic treatment and its effect on pathogens or other infectious elements. The importance of producing a safe effluent suitable for recirculation to agricultural land has become a task just as important as producing the maximum yield of biogas from a given type of waste. Therefore, anaerobic digestion at elevated temperatures has become the main area of interest and has been growing during the last few years Anaerobic digestion demands the concerted action of many groups of microbes each performing their special role in the overall degradation process. Both Bacteria and Archaea are involved in the anaerobic process while the importance, if any, of eukaryotic microorganisms outside the rumen environment is still unknown. The basic understanding of the dynamics of the complex microflora was elucidated during the latter part of the last century where the concept of inter-species hydrogen transfer was introduced and tested. The isolation of syntrophic bacteria specialized in oxidation of intermediates such as volatile fatty acids gave strength to the theories. Lately the use of molecular techniques has provided tools for studying the microflora during the biomethanation process in situ. However, until now the main focus has been on probing the dynamic changes of specific groups of microorganisms in anaerobic bioreactors and less emphasis has been devoted to evaluating the specific activities of the different groups of microbes during biomethanation. In the future we can expect that the molecular techniques will be developed to allow more dynamic studies of the action of specific microbes in the over-all process. From the present studies we know that many unknown microbes are found in anaerobic bioreactors. Especially within the domain of Archaea, there are whole phyla of microbes such as the Crenarchaeota, which make up significant fractions of microbes in a reactor but without cultured representatives. Improving the techniques for the isolation of presently unculturable microbes is a major task for the future.
X
Preface
Anaerobic digestion of waste has been implemented throughout the world for treatment of wastewater, manure and solid waste and most countries have scientists, engineers and companies engaged in various aspects of this technology. Even though the implementation of anaerobic digestion has moved out of the experimental phase, there is still plenty of room for improvements. The basic understanding of the granulation process, the basis for the immobilization of anaerobic microbes to each other without support material in UASB reactors, is still lacking. Like any other bioprocess, anaerobic digestion needs further control and regulation for optimization. However, until now suitable sensors for direct evaluation of the biological process have been lacking and anaerobic bioreactors have generally been controlled by indirect measurements of biogas or methane production along with measurements of pH and temperature. The newly development of an on-line monitoring system for volatile fatty acids could be a major step in the right direction and the use of infra-red monitoring systems could bring the price down to a reasonable level. A better performance of large-scale anaerobic bioreactor systems for treatment of complex mixtures of waste can be expected to be based on on-line monitoring of the process in the future along with controlling software for qualified management of these plants. Besides treatment of waste, anaerobic digestion possesses a major potential for adding value to other biomass converting processes such as gasification, bioethanol or hydrogen from ligno-cellulosic materials. Conversion of ligno-cellulosic biomass will often leave a large fraction of the raw material untouched which will be a burden for the over-all economy of the process and will demand further treatment.Anaerobic digestion will on the other hand be capable of converting the residues from the primary conversion into valuable methane, which will decrease the cost and the environmental burden of the primary production. Biomethanation is an area in which both basic and applied research is involved. Major new developments will demand that both disciplines work together closely and take advantage of each other's field of competence. The two volumes on Biomethanation within the series of Advances in Biochemical Engineering and Biotechnology have been constructed with this basic idea in mind and, therefore, both angles have been combined to give a full picture of the area. The first volume is devoted to giving an overview of the more fundamental aspects of anaerobic digestion while the second volume concentrates on some major applications and the potential of using anaerobic processes. The two volumes will therefore be of value for both scientists and practitioners within the field of environmental microbiology, anaerobic biotechnology, and environmental engineering. The general nature of most of the chapters along with the unique combination of new basic knowledge and practical experiences should, in addition, make the books valuable for teaching purposes. The volume editor is indebted to all the authors for their excellent contributions and their devotion and cooperation in preparing these two volumes on Biomethanation. Lyngby, January 2003
Birgitte K. Ahring
CHAPTER 1
Applications of the Anaerobic Digestion Process Irini Angelidaki 1 · Lars Ellegaard 2 · Birgitte Kiœr Ahring 3 1 2 3
Environment & Resources, The Technical University of Denmark, Block 115, 2800 Lyngby, Denmark. E-mail:
[email protected] BWSC, Gydevang 11, 3450 Allerød, Denmark Environmental Microbiology & Biotechnology, Biocentrum, The Technical University of Denmark, Block 227, 2800 Lyngby, Denmark
At the start of the new millennium waste management has become a political priority in many countries. One of the main problems today is to cope with an increasing amount of primary waste in an environmentally acceptable way. Biowastes, i.e., municipal, agricultural or industrial organic waste, as well as contaminated soils etc., have traditionally been deposited in landfills or even dumped into the sea or lakes without much environmental concern. In recent times, environmental standards of waste incineration and controlled land filling have gradually improved, and new methods of waste sorting and resource/energy recovery have been developed. Treatment of biowastes by anaerobic digestion processes is in many cases the optimal way to convert organic waste into useful products such as energy (in the form of biogas) and a fertilizer product. Other waste management options, such as land filling and incineration of organic waste has become less desirable, and legislation, both in Europe and elsewhere, tends to favor biological treatment as a way of recycling minerals and nutrients of organic wastes from society back to the food production and supply chain. Removing the relatively wet organic waste from the general waste streams also results in a better calorific value of the remainder for incineration, and a more stable fraction for land filling. Keywords. Anaerobic digestion, Reactors, Codigestion, Biowastes, Solid waste, Slurries, Manure, Industrial waste
1
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
2
Factors Influencing the Biogas Process . . . . . . . . . . . . . . .
4
2.1 2.2 2.3
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 5 5
3
Digestion of Slurries . . . . . . . . . . . . . . . . . . . . . . . . .
7
3.1 3.2 3.3 3.4 3.5 3.6
General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Process and Plant Configuration . . . . . . . . . . . . . . . Veterinarian Aspects . . . . . . . . . . . . . . . . . . . . . . Combined Digestion of Livestock Waste and Industrial Waste Process Tank Configuration . . . . . . . . . . . . . . . . . . Equipment Characteristics . . . . . . . . . . . . . . . . . .
. . . 7 . . . 8 . . . 9 . . 11 . . . 13 . . . 14
Advances in Biochemical Engineering/ Biotechnology, Vol. 82 Series Editor: T. Scheper © Springer-Verlag Berlin Heidelberg 2003
2
I. Angelidaki et al.
3.6.1 3.6.2 3.6.3 3.6.4 3.6.5 3.6.6 3.6.7 3.7 3.7.1
Mixing Technique . . . . . . . . . Pumping . . . . . . . . . . . . . . Heat Exchanging . . . . . . . . . . Biogas Cleaning . . . . . . . . . . Biogas Transmission . . . . . . . . Fiber Separation . . . . . . . . . . Odor Control Systems . . . . . . . Operational Experience and Results Production Results . . . . . . . . .
4
Digestion of High-Solid Wastes . . . . . . . . . . . . . . . . . . . 23
4.1 4.2 4.3 4.3.1 4.3.2 4.3.3 4.4
Pretreatment of Municipal Solid Waste . . . . . . . . . . . . Post Treatment . . . . . . . . . . . . . . . . . . . . . . . . . Biological Treatment . . . . . . . . . . . . . . . . . . . . . . Wet Digestion Systems . . . . . . . . . . . . . . . . . . . . . Dry digestion Systems . . . . . . . . . . . . . . . . . . . . . Multi-stage Anaerobic Digestion Systems . . . . . . . . . . . Summary of Processes Used for Anaerobic Treatment of Solid Wastes . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
References
. . . . . . .
. . . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . .
. . . . . . . . .
. . . . . .
. . . . . . . . .
. . . . . .
14 15 16 17 20 21 22 22 22
24 25 25 26 27 29
. . . 31
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1 Introduction Anaerobic degradation or digestion is a biological process where organic carbon is converted by subsequent oxidations and reductions to its most oxidized state (CO2), and its most reduced state (CH4).A wide range of microorganisms catalyze the process in the absence of oxygen. The main products of the process are carbon dioxide and methane, but minor quantities of nitrogen, hydrogen, ammonia and hydrogen sulfide (usually less than 1% of the total gas volume) are also generated. The mixture of gaseous products is termed biogas and the anaerobic degradation process is often also termed the biogas process. As the result of the removal of carbon, organic bound minerals and salts are released to their soluble inorganic form. The biogas process is a natural process and occurs in a variety of anaerobic environments. Such environments are, marine and fresh water sediment, sewage sludge, mud, etc. The interest in the process is mainly due to the following two reasons: – A high degree of reduction of organic matter is achieved with a small increase – in comparison to the aerobic process – in the bacterial biomass. – The production of biogas, which can be utilized to generate different forms of energy (heat and electricity) or be processed for automotive fuel. Biogas has a lower calorific value than natural gas, and in specific applications, such as automotive fuel, treatment of the biogas to improve its quality is required.
3
Applications of the Anaerobic Digestion Process Table 1. Calorific value of biogas and natural gas
Gas composition
Biogas 65% CH4
Biogas 55% CH4
Natural gas
Upper calorific value KWh/m3STP a Lower calorific value KWh/m3STP a
7.1 6.5
6.0 5.5
12.0 10.8
a
STP (standard temperature and pressure), i.e. the volume at 0 °C and 1 bar pressure.
In Table 1, the upper and lower calorific value of biogas in comparison to natural gas is shown. The biogas process has been known and utilized for many years, but especially after the rise of energy prices in the 1970s, the process received renewed attention due to the need to find alternative energy sources to reduce the dependency on fossil fuels.Although the price of fossil fuels decreased in 1985, the interest in the biogas process still remains due to the environmental benefits of anaerobic waste degradation. Additionally, the biomass used for biogas production was originally produced by photosynthetic fixation of carbon dioxide from the atmosphere, and combustion of biogas thus does not add extra carbon dioxide to the atmosphere as it does when combusting fossil fuels formed millions of years ago. The anaerobic degradation process has been used for years for energy production and waste treatment. It is used in closed systems where optimal and controlled conditions can be maintained for the microorganisms. The process can be utilized for the fast and efficient degradation of different waste materials. The anaerobic process is today mainly utilized in four sectors of waste treatment. 1. Treatment of primary and secondary sludge produced during aerobic treatment of municipal sewage. The process is utilized to stabilize and reduce the final amount of sludge and at the same time biogas is produced, which can be used to partly cover the need for energy in the sewage treatment plant. This application is widespread in the industrialized world in connection with the establishment of advanced treatment systems for domestic wastewater. 2. Treatment of industrial wastewater produced from biomass-, food-processing or fermentation industries. These wastewater types are often highly loaded and can successfully be treated anaerobically before disposal directly to the environment or sewage system. The produced biogas can often be utilized to cover the need for process energy. With the environmental concerns and cost of alternative disposal this application of the anaerobic process is increasing. 3. Treatment of livestock waste in order to produce energy and improve the fertilizing qualities of manure. Due to more strict rules concerning the usage, distribution, and storage of manure this application is growing especially in countries with a high animal production density. 4. A relatively new sector for use of the anaerobic processes on an industrial scale is treatment of the organic fraction of municipal solid waste (OFMSW). The aim of this process is first of all to reduce the amount of waste in other treatment systems, i.e. landfills and incineration plants, and secondly to recycle the nutrients from this type of waste to the agricultural sector.
4
I. Angelidaki et al.
2 Factors Influencing the Biogas Process The biogas process as a complex biological process is influenced by several environmental factors. The interdependence of the bacteria is a key factor of the biogas process. Under conditions of unstable operation, intermediates such as volatile fatty acids and alcohols accumulate at different rates depending on the substrate and the type of perturbation causing instability. Thus, changes in the concentration of intermediates indicate disturbance of the biogas process. The most important environmental factors that can influence the balance of the system are temperature, pH, substrate composition and toxins. 2.1 Temperature
Temperature is one of the main environmental factors affecting bacterial growth. Anaerobic bacteria are affected in the same way as the aerobic ones. Growth rates often increase with increasing temperature up to a certain limit, while there is a rapid decrease in growth as the temperature approaches the upper limit for survival of the bacterium. Besides the influence on growth rates of bacteria, temperature also influences physical parameters such as viscosity, surface tension and mass transfer properties.Apart from the general steady state dependency on temperature, also temperature stability is important, since even relatively small changes in the temperature result in an efficiency drop until adaptation has occurred. Treatment of waste in anaerobic reactors is normally carried out within two temperature ranges: around 25–40 °C, known as the mesophilic range, and higher than 45 °C, known as the thermophilic range. Several advantages of anaerobic waste digestion at thermophilic temperatures have been reported previously. At higher temperatures: – – – –
the rate of digestion is faster, and thus shorter retention times are required, smaller reactor volumes are required for treating the same amount of waste, higher rate and efficiency of particulate matter hydrolysis, more efficient destruction of pathogens.
Poor stability was previously believed to be associated with thermophilic temperatures. However, many years of experience with full-scale biogas plants operating at thermophilic temperature, have demonstrated that this is not the case. Temperature has a positive effect on the digestion rate, resulting in higher volumetric methane production rates. Casali and Senior found that the methanogenic rate in refuse increased 2.6 times when the temperature was increased from ambient temperature to 30 °C and further 3 times when the temperature was increased from 30 to 40 °C. However, the ultimate methane yield from organic matter is not influenced by temperature in the range 30 to 60 °C. Consequently the effect of temperature on
Applications of the Anaerobic Digestion Process
5
growth rates can only be seen when the loading rates are high or retention times short. Varel et al. found with waste of beef-cattle in semi continuous reactor experiments that the effect of temperature on the rate of methane production was more noticeable at short retention times.As the retention time was decreased, active fermentation could only be achieved at thermophilic temperatures.At 3 days RT, active fermentation could only be achieved at temperatures between 45 and 60 °C and at 2.5 days RT only at 60 °C. Methanogenesis is also possible under psychrophilic conditions (below 25 °C) but at lower process rates. Anaerobic bacteria can adapt quite easily to low temperatures, and high rate anaerobic treatment has been achieved at psychrophilic conditions, when bacteria have been immobilized or otherwise retained in the process. It has been found that when shifting from mesophilic to psychrophilic temperatures the microbial populations still have the same composition as under mesophilic temperatures. This could indicate that at these temperatures psychrotolerant rather than true psychrophilic organisms are active. 2.2 pH
The anaerobic digestion process is limited to a relatively narrow pH interval from approx. 6.0 to 8.5; a pH value outside this range can lead to imbalance. Each of the microbial groups involved in anaerobic degradation has a specific pH optimum and can grow in a specific pH range. The methanogens and acetogens have pH optimum at approx. 7, while acidogens have lower pH optimum around 6. Methanogens at pH lower than 6.6 grow very slowly. In an anaerobic reactor, instability will as a rule lead to accumulation of VFA, which can lead to a drop in pH (acidification). However, accumulation of VFA will not always be expressed as a pH drop due to the buffer capacity of some waste types. In manure there is a surplus of alkalinity, which means that the VFA accumulation shall exceed a certain point before this can be detected as a significant change in pH. This means that when a drop in pH in the reactor is eventually observed, the concentration of volatile fatty acids is most probably very high and the process may already have been affected. There are many factors which influence pH. It is especially organic acids and carbon dioxide which lower pH, while ammonia will contribute to an increase of pH. Other compounds contributing to the buffering capacity are hydrogen sulfide and phosphate. 2.3 Toxicity
A number of compounds are toxic to the anaerobic microorganisms. Methanogens are commonly considered to be the most sensitive to toxicity of the microorganisms in anaerobic digestion. However, the process can acclimatize, and higher concentrations of the toxicant can be tolerated after a period of adaptation. The most common inhibitor for the anaerobic process is ammonia. In anaer-
6
I. Angelidaki et al.
obic digestion ammonia originates from soluble ammonia in the influent, from protein degradation and other compounds such as urea. Many substrates used for anaerobic treatment often contain ammonia in toxic concentrations. Such substrates include pig and poultry manure, slaughterhouse waste, potato, juice, highly proteinaceous sludge, wastewater from shale oil and coal liquefaction processes. The results concerning ammonia-N inhibitory level are conflicting, as they depend on parameters such as pH, temperature and adaptation of the inocula. It is generally accepted that it is the non-ionized form of ammonia that is responsible for inhibition, pH has a significant effect on the level of ammonia inhibition, as the pH value determine the degree of ionization. The free ammonia ratio to total ammonia/ammonium ratio can be calculated from the equilibrium relation as follows: [NH3]/[T – NH3] = 1/(1+[H+]/Ka) where, [NH3] and [T – NH3] are the free ammonia and the total ammonia/ammonium concentrations respectively, and Ka is the dissociation constant, which is temperature dependent. In Fig. 1 this dependency is shown graphically. As pH and temperature increase the free ammonia fraction also increases. Especially pH has a strong influence on the degree of ionization of ammonia. Bhattacharya and Parkin found the maximum tolerable free-ammonia concentration to be only 55 mg-N/l. Most reported inhibitory levels are, however, higher.Angelidaki and Ahring found that the biogas process could be adapted to tolerate free ammonia concentration of 800 mg-N/l. The inhibitory level of free ammonia depends strongly on the degree of acclimatization of inoculum to ammonia. As the free ammonia concentration decreases with decreasing pH it could be expected that even a slight reduction of pH would have a positive effect on ammo-
Fig. 1. pH and temperature influence on NH3 dissociation
Applications of the Anaerobic Digestion Process
7
nia inhibition. It was indeed found by several investigators that a pH decrease reduced ammonia inhibition. Free ammonia inhibition result in VFA accumulation, which in turn lower pH and decrease the ratio of free ammonia, with the result that free ammonia inhibition is relieved. Due to this self-stabilizing mechanism, processes can be maintained in a stable ammonia inhibited state, where a balance between VFA concentration and ammonia loading exist. The carbon/nitrogen (C/N) ratio is also important for process stability. A C/N ratio of 25 to 32 has been reported to have a positive effect on the methane yield. At lower C/N ratios the risk of excess nitrogen not needed for biomass synthesis and therefore becoming inhibitory increases. Opposite, a very high C/N ratio would lead to N deficiency for biomass synthesis.Waste with very high COD concentration and low content of nitrogen such as olive mill effluents has been shown not to be able to be degraded alone. Addition of either nitrogen or codigestion with wastes with a lower C/N ratio was needed in order to digest olive mill effluents successfully. Anaerobic treatment of wastewater containing high sulfate concentrations can cause inhibition as a result of the formation of hydrogen sulfide. It has been reported that total hydrogen sulfide concentrations of 100 to 300 mg/l or free hydrogen sulfide concentrations of 50 to 150 mg/l caused severe inhibition resulting in complete cessation of biogas production. Long chain fatty acids, such as oleate and stearate, have been found to be toxic to the anaerobic process. No adaptation to the fatty-acid toxicity was observed. However, the presence of particulate material can increase the resistance of the process to long-chain fatty acids as LCFA are absorbed on the particulate material and thus not active as inhibitor. For other organic compounds such as phenols, chloroform and formaldehyde a reversible toxicity has been observed. Heavy metals are toxic for anaerobic microorganisms in concentrations in the range 10–3 to 10–4 M. However, experiments have shown that acclimatization to high heavy metal concentrations can occur and often the level of heavy metals would become an environmental problem before affecting the process. In a reactor, the actual concentration of soluble metal ions is normally low due to precipitation of insoluble metal salts, e.g., as sulfides. It has been shown that less than 2% of the metals may be in the soluble form. Although anaerobic bacteria are inhibited by several toxicants they can also tolerate a number of toxicants. Complete anaerobic degradation of pentachlorophenol has been reported. A number of other toxicants have been reported to be detoxified in anaerobic reactors such as nitroaromatic compounds, chlorinated aliphates, N-substituted aromatics and azo dyers.
3 Digestion of Slurries 3.1 General
One type of anaerobic waste treatment plants which have emerged since the mid 1980s in Denmark and elsewhere, are large scale manure treatment plants serv-
8
I. Angelidaki et al.
ing a number of farmers, treating a slurry of manure, often with a fraction of other waste as supplementary substrate. The main objective is to extract energy and to improve the fertilizing quality of manure, resulting in less nutrient leakage to ground and surface waters from the agricultural sector. The description below gives an impression of the most important elements of such a slurry treating biogas plant, also illustrating some of the practical design considerations not directly related to the reactor system itself, as well as illustrating the gradual technical development waste treatment plants often undergo. 3.2 Process and Plant Configuration
The main elements in a typical large-scale manure biogas plant, are shown in Fig. 2. Raw material is transported by special constructed lorries from/to farmers and is kept in pre-storage buffer tanks. It is then pumped through heat exchangers into the reactor at the desired flow rate. The effluent is pumped to an
Fig. 2. Concept of manure biogas plants top Farm biogas plant bottom Centralized biogas plant
Applications of the Anaerobic Digestion Process
9
after-storage tank, before it is brought back to the farmers and finally utilized on fields as fertilizer. A biogas plant is in reality much more complex than described above. Many unit operations are or may be utilized in such a plant i.e.: – – – – – – – –
transport/pumping stirring/mixing macerating/grinding heat exchanging biogas treatment and cleaning biogas compression and transportation biogas storage filtration/separation
Figure 3a shows a picture of a biogas plant (Blaabjerg biogas plant, Denmark), where some of these unit operations are employed, is shown. In Fig. 4 the general lay-out of a full-scale biogas plant (Blaabjerg, Denmark) is shown. Many process and lay-out variations are possible, often dictated by local conditions and needs. 3.3 Veterinarian Aspects
Mixing and redistribution of manure, from several farms, requires adoption of a sufficient level of sanitation, in order to avoid the risk of spreading pathogens. The veterinarian requirements in Denmark concerning manure are to ensure that the manure is kept at a thermophilic temperature (> 50 °C) for a minimum
Fig. 3. A centralized biogas plant (Blaabjerg, Denmark)
10
I. Angelidaki et al.
Fig. 4. Typical lay-out of a centralized biogas plant (Blaabjerg, Denmark). This plant is treat-
ing approx. 100,000 ton of manure and industrial waste per year, and producing approx. 3 million m3 of biogas per year
of 4 hours. The required sanitation can be obtained directly in a thermophilic process by observing special pumping routines, whereas a mesophilic process requires a passive pre- or after sanitation stage (Fig. 5). It has been documented that the above requirements ensure an effective pathogen reduction, reaching a level equal to or better than 4 log-units when using Faecal streptococcus as an indicator organism. Certain types of supplementary waste suitable for codigestion in manure based biogas plants, such as sewage-sludge, require more strict sanitation, due to the potential content of human pathogens. Originally such types of waste required heating to 70 °C for a minimum of 1 hour. A special Danish veterinarian research effort, utilizing high temperature resistant bacteria and vira as indicators, has resulted in the definition of alternative temperature/time combinations, which in conjunction with a biogas process, are considered equivalent to a 70 °C/1 h heating. For a thermophilic biogas process (minimum 52 °C), a guaranteed temperature/retention time of 52 °C/10 hours, 53.5 °C/8 hours or 55 °C/6 hours is considered equivalent to 70 °C/1 hour. The retention can take place in separate sanitation tanks before or after the reactor stage, or directly in the reactor if pumping sequences allow the necessary pauses. For mesophilic biogas plants slightly higher retention times are required and sanitation must take place in a separate tank at a thermophilic temperature level. While inclusion of “problematic” waste in the past required a separate pretreatment step, all the biomass can now be sufficiently sanitized in a single thermophilic process, provided the longer guaranteed retention times is built into the design.
Applications of the Anaerobic Digestion Process
11
Fig. 5. Plant configuration examples: 1 Thermophilic plant with two reactors and heat ex-
changing; 2 Mesophilic plant with two reactors, heat exchanging and thermophilic post sanitation
3.4 Combined Digestion of Livestock Waste and Industrial Waste
Construction of Joint Biogas Plants gives the possibility for combined anaerobic treatment and utilization of livestock waste and several types of organic waste from the food processing industry. Organic industrial waste is usually characterized by high pollution loads and often contains high concentrations of rapidly degradable substrates such as saccharides, starches, lipids and proteins. Manure usually has a rather low solids concentration (typically 3–5% total solids (TS) for pigs and 6–9% for cattle and dairy cows). In addition the manure contain particles and straw/fibers with a content of ligno-cellulose. This fiber fraction is highly recalcitrant to degradation and will often pass through the reactor mainly undigested. The high content of water together with the high fraction of fibers in manure is the reason for the low methane yields of manure, typically ranging from 10–20 m3 CH4/ton of manure treated. However, manure is an excellent basic substrate for co-digestion of industrial waste, which could otherwise be difficult to process alone. The reasons for this are: – the high content of water in manure acts as a solvent for the dryer types of wastes, solving problems of pumping and mechanical treatment of solid wastes.
12
I. Angelidaki et al.
– the high buffering capacity of manure protects the process against failure due to pH drop if the VFA concentration increase. – manure is rich in a wide variety of nutrients necessary for optimal bacterial growth. By combining different types of waste, such as manure, municipal solid waste and organic industrial waste, a much higher gas yield can be obtained from biogas reactors since organic industrial wastes in particular are more easily degraded and have a higher gas potential than manure. As most types of industrial organic waste result in methane yields varying from 30 to 500 m3/ton, these sources constitute a very attractive supplementary substrate for manure based biogas plants. The gas yields from different types of organic industrial waste is shown in Table 2. Besides increasing the yield, addition of easily degradable material has been shown to stabilize the anaerobic digestion process if added in a controlled fashion. This effect could be partly due to a higher active biomass concentration in the reactor, which will be more resistant to inhibitory compounds. Furthermore, the inorganic parts of some organic wastes, such as clays and iron compounds,
Table 2. Methane yield from different types of industrial waste
Type of organic waste Composition of the organic material
Organic content Methane yield (%) (m3/ton)
Stomach and intestine content
Carbohydrates, proteins and lipids
15–20
40–60
Flotation sludge (dewatered)
65–70% proteins, 30–35% lipids
13–18
80–130
Bentonitebound oil
70–75% lipids, 25–30% other organic matter
40–45
350–450
Fish-oil sludge
30–50% lipids and other organic matter
80–85
450–600
20–30
150–240
Source sorted organic Carbohydrates, proteins, and lipids household waste Whey
75–80% lactose and 20–25% protein
Concentrated whey
75–80% lactose and 20–25% protein 18–22
7–10
Size water
70% proteins and 30% lipids
10–15
40–55 100–130 70–100
Marmelade
90% sugar, fruit organic acids
50
300
Soya oil/Margarine
90% vegetable oil
90
800–1000
Methylated spirits
40% alcohol
40
240
Sewage sludge
Carbohydrates, lipids, proteins
Concentrated sewage Carbohydrates, lipids, proteins sludge
3–4
17–22
15–20
85–110
Applications of the Anaerobic Digestion Process
13
have been shown to counteract the inhibitory effect of ammonia and sulfide, respectively. 3.5 Process Tank Configuration
Continuous feeding/pumping is often preferred in order to optimize heat exchangers. In order to obtain the necessary guaranteed retention time for sanitation purposes, configurations with two or three reactors (or sanitation tanks) operated in parallel are often seen, Fig. 6. At any given time, one reactor is being fed, one emptied and one resting, i.e., ensuring the sanitation retention time. As fewer larger tanks are often preferred for economical reasons, various schemes with discontinuous pumping or combinations with smaller buffer tanks can also be seen.
Fig. 6. A biogas plant with two reactors and post sanitation
14
I. Angelidaki et al.
3.6 Equipment Characteristics 3.6.1 Mixing Technique
Initially, submerged electrically driven medium speed mixers, mounted on guiding rails were utilized for most mixing tasks. These mixers soon proved rather costly to operate, especially where access and routine service inspection is difficult as is the case for digesters or other closed process tanks. Slow moving, top mounted central mixers with a freely suspended shaft with two propellers have become the preferred solution for reactor mixing. The mixers usually work continuously with a mixing power input of 3–4 W/m3. Service requirements are very limited, but a correct reactor level relative to the upper impeller is very critical in order to avoid annoying floating layer formation. In storage tanks several types of mixers are in use, including improved submerged models. The mixing energy input is often discontinuous (high power for a short period) and average mixing power input typically varies from 10 W/m3 in prestorage/mixing tanks to 1 W/m3 in after storage tanks. The position and type of mixers in combination with tank geometry have proven to be very critical. Hydrodynamic favorable solutions, allowing the material to flow to the mixers, generally work best in combination with mixing at different depths. Figure 7 schematically shows the most common mixer configurations.
Fig. 7. Common mixer configurations: A Top mounted reactor mixer; B Side mounted mixer
in process tank/storage; C Submerged mixer on guiding rail; D Top mounted mixer; E Top mounted mixer with submerged angle gear
Applications of the Anaerobic Digestion Process
15
Fig. 8. Reactor top, showing centrally mounted top mixer. Mixer shaft extends to the bottom of
the reactor (approx. 17 m) with two propellers (3.2 m) placed at the top (just below liquid surface) and at the bottom
In Fig. 8 a picture of the reactor top with a centrally mounted top mixer is shown. 3.6.2 Pumping
Eccentric worm pumps have been used for continuous feeding/extraction of biomass (low flow, high pressure) while centrifugal pumps with open impellers and cutters have been used for biomass transfer between storage tanks (high flow, low pressure). Feeding pumps, in particular those handling fresh biomass have proven to be rather costly to operate due to high wear rates, but, despite several efforts, no viable alternative has been found so far. Low pump speed (< 300 rpm) and a pressure drop as low as possible can somewhat alleviate the problems.
16
I. Angelidaki et al.
3.6.3 Heat Exchanging
Most biogas plants in Denmark utilize heat exchanging between in-going fresh biomass and outgoing digested biomass, although a few plants have been built with no or limited heat exchanging. The choice depends on the value of heat, i.e., the possibility of utilizing heat for other purposes. Due to the high viscosity of manure and other biomass, especially when fresh, it is seldom possible to obtain a turbulent flow without generating an excessive pressure drop. Laminar or transition flow is the result with rather poor heat transfer coefficients. The most common types of exchangers utilize curved flow channels to break up the thermal boundary layers by secondary flow or multi pass design to break up the flow at regular intervals. Most designs attain heat transmission values of approx. 300 W/m2/°C with flowing velocities up to approx. 1 m/s, when exchanging fresh biomass against digested biomass, or approx. two times as much if exchanging against water. Dimensioning, especially for pressure drop, relies very much on experience, as prediction of biomass viscosity is very uncertain. A good design margin and spare pump capacity is necessary to avoid problems. In designs with rather narrow flow passages (< approx. 30 mm) it is considered wise to macerate the fresh biomass, which is also likely to improve the biogas yield slightly. Different heat exchanger configurations are shown in Fig. 9. Typical dimensioning practice (in Denmark) leads to heat exchanger installations with a resulting temperature difference of 10–15 °C. The remaining heat-
Fig. 9. Different heat exchanger configurations
Applications of the Anaerobic Digestion Process
17
Fig. 10. Typical exchanger used in full scale biogas plants
ing up to process temperature, and for heat loss compensation, is by hot water coils in the reactors or with a final biomass/water heat exchanger plus fewer coils for heat loss compensation. In this connection, it is important to recognize and include the heat loss due to evaporation of the water contained in the humid biogas leaving the reactor. A typical heat exchanger used in full scale biogas plants is shown in Fig. 10. On the input side of heat exchangers (and pumping systems) the risk of clogging is manly due to the possible presence of foreign objects. On the digestate side of heat exchangers, one must reckon with the formation of struvite (MgNH4PO4), which forms due to over-saturation when the biomass is cooled. The scaling occurs mainly in the cold sections and will gradually close flow channels. The struvite needs to be dissolved by flushing with a circulating weak acid at regular intervals. Residual scaling also occurs in cold pipes down stream of heat exchangers, although at a much slower rate. 3.6.4 Biogas Cleaning
Dry biogas from manure generally consist of approx. 60–70% v/v methane (CH4), 30–40% v/v carbon dioxide (CO2), 1–2% v/v nitrogen (N2), 1000– 3000 ppm hydrogen-sulfide (H2S) and 10–30 ppm NH3 . The concentration of hydrogen-sulfide stated is typical when operating on manure alone. Other
18
I. Angelidaki et al.
substrates, even in relatively small amounts, can result in quite different levels, both lower and higher. Raw biogas is a humid gas, which needs to be cooled and carefully drained before utilization. Early attempts to dry biogas using regenerative absorption equipment designed for pressurized air proved difficult due to accelerated absorbent decomposition, and the preferred solution has instead become to simply cool the biogas and to design the gas system with care to avoid undrained low points. In the case where the biogas is to be utilized in a boiler and emission requirements for sulfur-dioxide are not too strict, no further cleaning is necessary. In many cases, however, the biogas is utilized in gas engines for Combined Heat and Power (CHP) generation. In this case the hydrogen-sulfide content needs to be kept below approx. 700 ppm for conventional gas engines in order to avoid excessive corrosion and too rapid (and costly) lubrication oil deterioration. One method of reducing the hydrogen-sulfide content is to add a commercial ferrous solution to the reactor feed. Ferrous compounds bind sulfur as insoluble products in the liquid phase, which reduces the evolution of gaseous hydrogen-sulfide. The method has been utilized in a number of plants, but is rather costly if applied continuously, since the consumption of ferrous material on a stoichiometric basis has proven to be 2–3 times the desired reduction in gaseous hydrogen-sulfide. In some cases, however, ferrous containing waste products can be obtained for codigestion most of the year, with commercial ferrous addition acting as a back up possibility only. Often secondary H2S biogas cleaning is necessary, and in this case a biological oxidation process has become the dominating solution. By injection of a small amount of air (2–8 % v/v) into the raw biogas, the hydrogen-sulfide content can be biologically oxidized either to free (solid) sulfur or (aqueous) sulfurous acid according to following reactions: 2 H2S + O2 Æ H2O + 2 S 2 H2S + 3 O2 Æ 2 H2SO3
(1) (2)
The reaction will occur spontaneously and can take place in the reactor headspace on the floating layer (if any) and reactor walls, if air is injected directly in the headspace. Due to the acidic nature of the products with a risk of corrosion and the dependency of a stable floating layer, it is often preferred to isolate the process in a separate reactor as shown schematically in Fig. 11. The reactor is somewhat similar to a scrubber, consisting of a porous filling (randomly packed plastic elements or similar) where microorganisms can grow, and a sump, pump, and nozzle arrangement allowing regular showering of the filling. Showering has the function of washing out acidic products and supplying nutrients to the microorganisms. The sump must therefore contain a liquid with a high alkalinity and contain essential nutrients, for which digested manure, preferably screened, is the ideal and readily available choice.A reactor loading of approx. 10 m3/h of biogas per m3 of reactor filling and a process temperature around 35 °C has been the normal choice, and the process has proven very efficient, provided that sufficient air is injected (slightly more than stoichiometri-
Applications of the Anaerobic Digestion Process
19
Fig. 11. Schematic diagram of system for biological H2S oxidation
cally needed). Sump pH must be maintained at 6 or higher.A washing procedure, where the filling elements are gurgled through with an air/water mixture, has to be carried out at regular intervals in order to prevent free sulfur deposits from closing the reactor filling. In some cases, where biogas is stored or passing through a digestate after storage, the H2S reactor is omitted and only air injected. Cleaning is then relying on the formation of a floating layer in the after storage, on which the microorganisms can grow and perform the oxidation. A floating layer can usually be maintained with the choice of a low mixing intensity, without too many problems for operating the tank as a buffer storage. This solution is certainly more cost effective, but can also be more unreliable as floating layers can be rather unstable, i.e., sinking overnight perhaps to resurface some days later. At least some periods with reduced cleaning efficiency can be expected. In Fig. 12, a reactor used for biological hydrogen-sulfide reduction is shown.
20
I. Angelidaki et al.
Fig. 12. Hydrogen sulfide reduction reactor, 80 m3 with 50 m3 filling material. H2S is oxidized
by a biological process to acidic products or free sulfur, with injection of a small amount of atmospheric air upstream of the reactor
3.6.5 Biogas Transmission
Biogas Plants are usually situated at some distance from the nearest town, while biogas utilization in an engine CHP installation often takes place inside or in the vicinity of the town in order to connect to a district heating system. As a consequence biogas often has to be transported up to several km between the plants. The first plants utilized high pressure (2–4 bar) dry transmission with piston compressors and absorption drying equipment.As mentioned earlier, absorption drying presented problems, and as piston compressors also proved vulnerable and operationally expensive, which has caused the low pressure wet transmission to be gradually taken over. This solution usually involves a transmission pressure in the range 400–700 mbar in order to deliver at 100–200 mbar, utilizing Roots blowers and a carefully planned buried transmission line, laid with a minimum 3–5‰ slope (preferably in the flow direction) with a limited number of condensate wells. The bulk of condensation occurs in the first few wells within 100–500 m from the plant. Wells are equipped with valves for regular manual emptying, and the wells are usually designed to hold up to a weeks amount of condensate, meaning that the first wells must be quite large (typically 2 m3), while the last only need to be small (typically 100 liters).
Applications of the Anaerobic Digestion Process
21
At transmission distances up to a few hundred meters, high speed centrifugal blowers and a transmission pressure of approx. 100 mbar can be employed with advantage. 3.6.6 Fiber Separation
Manure and other types of biomass contain fibers (from straw) and other organic structures which are difficult for anaerobic organisms to access.A typical degradation “efficiency” of 50–60% is often seen in manure based systems, meaning that the digestate still contain 40–50% of the original organic dry matter content primarily as fibers. Some of the early plants included separators to take out a part of the fibers for production of commercial compost. Marketing this compost in relatively small scale proved quite difficult, and compost production has ceased in most plants. Some of the newest plants in Denmark again include separation equipment. The intention now is to produce a fiber fraction with a dry matter content in excess of 45%, which can be utilized as a supplementary fuel in wood chip boilers. This way the overall energy efficiency can be raised by approx. 15% through additional heat production. A side benefit, which in the near future might add to the feasibility of this scheme, is the removal of surplus phosphorous which is predominantly attached to the fibers. Screw type separators dominate, sometimes in combination with a pre-separation in order to increase the capacity (Fig. 13).
Fig. 13. Fiber collection wagon, with distribution screw
22
I. Angelidaki et al.
3.6.7 Odor Control Systems
One of the benefits from anaerobic digestion of manure is a considerable reduction in odor nuisance when spreading manure on fields. The biogas plant itself can, however, give rise to local odor nuisance if care is not taken to limit odor emissions. As a first preventive measure, it is of course desirable to situate the plant (if possible) without immediate neighbors. In addition all storage tanks should be covered. Due to service openings and as unloading/loading and internal pumping generates tank breathing, it is necessary to arrange a weak suction from the storage tanks and other contaminated sections of the plant. This air must be treated in order to avoid odor emission. Air treatment is usually chosen either as a biological cleaning in a compost filter or as a combustion. Compost filters are relatively cheap but operational experience. Careful attention and maintenance are required in order to ensure satisfactory cleaning, as filters sometimes fail due to acidification (oxidation of hydrogen-sulfide), might collapse and become too dense, or might run dry. Combustion in a boiler in conjunction with generating process heat for the biogas plant has been tried, but the need for combustion air is limited, enforcing a compromise on venting flow. Combustion in engines (together with the biogas) has so far not been attempted, due to the fear of affecting the engine negatively. In a few plants, a system involving combustion with regenerative heat exchanging has been employed, a technique commonly utilized for the combustion of gaseous solvents in industrial ventilation air.Although a rather expensive solution, results are positive. 3.7 Operational Experience and Results 3.7.1 Production Results
Figure 14 shows the average monthly biogas production during the first operational period of two thermophilic Biogas plants, both with a biomass treatment capacity of 100,000 ton biomass per year. The figure shows typical start-up periods of 4–6 months and illustrates the relative stability of the production, once normal operation is obtained. The sudden rise in biogas production of one of the biogas plants after month 13, was due to the introduction of residual fish oil as a co-substrate. Consumption of process energy of relatively new plants typically amount to approx. 4–5 kWh per m3 biomass treated for mixing, pumping, control etc., and a heat consumption of 15–25 kWh per m3 of biomass treated, including consumption for hot water, heating of buildings, etc. This should be compared to a biogas yield of approx. 30 m3 per m3 biomass, corresponding to a lower calorific value (LCV) yield of approx. 200 kWh, i.e., 10–15% of the energy produced is consumed in the process. These figures represent the state of the art level for large-scale manure based biogas plants in Denmark, achieved as a result of the technical development described shortly above.
Applications of the Anaerobic Digestion Process
23
Fig. 14. Average monthly biogas production from thermophilic manure based biogas plants:
from Thorsoe (1994) and Blaabjerg (1996)
4 Digestion of High-Solid Wastes A special problem arises when solid wastes are to be treated.Anaerobic treatment of waste water and other liquid types of wastes have been used for decades. However, treatment of high-solids wastes such as municipal solid waste (MSW) is a relatively new application of anaerobic treatment, and many technical aspects need special consideration when inhomogeneous urban waste is the source material. “High-solid wastes” might be defined as organic material with a content of solids between 10 to 40%, which is not fluid. The most important type of high solid wastes is Municipal Solid Waste (MSW) or Household Solid Waste (HSW). The most common way to dispose of MSW has until recently been land filling or incineration. However, in the recent years attention has focused on recycling. By recycling the valuable elements of waste are converted to useful products and returned to the supply chain, minimizing the pressure on land filling and incineration facilities, while also conserving primary resources. Recycling started as a positive but minor part of waste management, but has become a more and more dominant element in the waste policies and strategies in many countries. MSW is a heterogeneous waste, which can be divided into the following fractions: – an organic biodegradable fraction, consisting of food residues and sometimes garden wastes
24
I. Angelidaki et al.
– a combustible fraction, consisting of an organic recalcitrant fraction, such as wood, paper and plastic – an inert fraction, consisting of stones, glass, metal and other inorganic parts Waste composition varies enormously, both between countries and between locations in the same country. In urban built-up areas, MSW is less rich in organic biodegradable material compared to rural areas and the composition also varies with the social level of the area. The organic biodegradable fraction of the MSW can be treated by anaerobic digestion to produce energy and fertilizer. By anaerobic degradation, one ton of the organic fraction of MSW can give a net energy yield of 100 to 200 kWh electricity plus heat. Most often the technical aspects involved are: – – – – – –
source separation and collection separation (removal of plastic bags or other packing and foreign elements) transfer (transporting the material through the process/plant) heating the substrate reactor mixing sanitation of the biomass
Digestion of high solids wastes can take place as a continuous or batch process. Continuous processes are most desirable on a larger scale. There are many waste treatment systems specifically designed for treatment of municipal waste (MSW). The most treatment systems consist of the following steps: – Pretreatment – Biological treatment – Post-treatment 4.1 Pretreatment of Municipal Solid Waste
Municipal solid waste typically contains less than 1/3 organic material. In order to treat MSW biologically, a sorting of the waste is needed. The sorting can be achieved by either source-separation, mechanical separation, or by hand sorting techniques. Source-separated municipal waste (SSMSW), i.e., waste sorted into fractions where it is generated, is the preferred solution from a quality point of view, since it is difficult to derive a clean organic fraction once the waste has been mixed. Such systems are found only in limited areas. However, in several places in Europe, work is being done, in order to introduce source separation of the MSW. Source-separated collection was originally introduced to derive certain valuable components, due to the increased cost of raw materials and a wish to limit the use of primary resources and energy to process new raw materials. As a consequence, source-sorted collections were introduced mainly for paper, valuable metals (such as aluminum and copper), and glass etc. Lately sourceseparation has also included separation of an organic degradable fraction. This
Applications of the Anaerobic Digestion Process
25
is done by including separate waste bins and collection systems in the households. An alternative scheme, which has been implemented in some areas, is source separation into bags of different colors, but collected with only one truck and later sorted out into seperate fractions by an optical sorting machine. This way the extra cost of separate collection or multi compartment collection trucks can be avoided (by utilizing existing collection trucks), albeit at the expense of a sorting machine. Most of the MSW is, however, still not source-separated. Therefore, pretreatment is often needed to sort out the organic fraction of mixed MSW. The main goal in the different pretreatment systems is to separate the organic fraction from the inorganic material. Furthermore, effort is applied in recycling several useful components, such as ferrous metals. The fraction obtained by mechanical separation is usually more contaminated then the fraction obtained by source-separation, and a fraction of the organic material will be lost with the other fractions sorted out. It is especially the content of heavy metals and plastics that is higher in the mechanical separated MSW. Even in case of source separation, some form of pretreatment is often needed to guard the plant and end product against elements included in the organic fraction by mistake or carelessness. 4.2 Post Treatment
Following the biological treatment, the digested material (known as digestate or effluent) is usually dewatered to 50–55% TS with a screw press, filter press or other types of dewatering systems. The press cakes are refined with sieve and composted aerobically. At this point, the compost can be further cleaned by screening, to remove unwanted material such as small pieces of glass or plastic. The compost is often offered for sale as a soil conditioner or potting soil (nitrate approx. N = 0.25 kg/ton, total N approx. 7–8 kg/ton). Press liquid, which may contain high concentrations of volatile fatty acids, is centrifuged, recircled or sent to wastewater treatment. 4.3 Biological Treatment
There are various classification principles of the existing anaerobic systems for treatment of high solids wastes. There are batch and continuous systems according to the feeding procedure of the wastes to the reactors. Most systems are continuous systems. However, there also exist batch systems such as the Biocel process. The major difference between the various systems is whether a “wet” or “dry” methanization process is used. Digestion of the MSW can take place either in the mesophilic or the thermophilic temperature range, and the hydraulic retention time is 10 to 30 days, depending on the process temperature, the technology used, and the waste composition.
26
I. Angelidaki et al.
4.3.1 Wet Digestion Systems
With wet digestion systems the total solids content of the waste has to be reduced to concentrations below 20% in order to create a pumpable slurry. This can be done either by adding water, such as recycled process water, or by codigestion where MSW is mixed with more dilute streams such as sewage sludge or manure. Codigestion Process
Treating high solid waste requires technically complicated and expensive treatment systems. Therefore, it is desirable, if possible, to avoid treating the waste as a solid, but instead mix it with other more dilute waste. A new concept, which has been applied successfully in several biogas plants in Denmark (Vegger, Sinding Ørre, Studsgård and Århus) is mixing MSW with manure. Sweden's largest biogas plant in Kristianstad work with the same co-digestion concept: household waste is treated together with agricultural waste and industrial waste. The plant at Kristianstad has a capacity of 73,000 tons of biomass per year, of which 15% is OFMSW, 18% is industrial organic waste, and 67% is manure. The biogas yield is around 40 m3 biogas per ton feedstock and the production covers the heat requirements of 600 to 800 households. In this way some of the difficult technical problems treating high solid wastes are avoided. Also economy of scale can be achieved and the waste enters into an established nutrient/fertilizer recycling system. The plants generally operate at thermophilic temperatures. Another example of the excellent performance of a large-scale co-digestion process of the organic fraction of municipal solid waste is the sewage treatment plant in Grindsted, Denmark where MSW is mixed with sewage sludge. The Waasa Process
Another process using codigestion is the Waasa process. The process has been tested on a number of wastes. In the Waasa plant in the city of Vaasa, Finland the following types of wastes have been treated over the years since it started up in 1989; mechanically or source-separated MSW, sewage sludge, slaughterhouse waste, fish waste, and animal manure. The process operates with a TS content of approx. 10–15% and can operate both at mesophilic and thermophilic temperatures.At the Waasa plant both mesophilic and thermophilic treatment methods are in operation in two parallel reactors. Today the Waasa Process is also in operation in Kil, Sweden and outside Tokyo, Japan. Furthermore, a plant in Groningen, the Netherlands is under construction. One characteristic of the Waasa Process is its main reactor, which is divided into various zones in a simple way. The first zone is made up of a pre-chamber inside the main reactor. The mixing in the reactor is by pneumatic stirring, where biogas is pumped through the base of the reactor. A small part of the digestate is mixed into the newly fed bio waste to speed up the process by inoculation.
Applications of the Anaerobic Digestion Process
27
4.3.2 Dry Digestion Systems
In many cases, especially in large urban areas, it is difficult to find other wastes to codigest MSW with. Dry digestion systems can cope with solids as high as 35%. The VALORGA Process
The Valorga process was developed in France and is a semi-dry process (Fig. 15). The process is a mesophilic process and takes place in the following way: after pretreatment the waste is mixed with recycled process water. The process water is gained from separation of the reactor effluent by centrifugation, filtration or other types of separation.After mixing with process water, the influent is pumped into the reactor. The reactor is of the fully mixed reactor type. Mixing is taking place by pneumatic stirring, i.e., the produced biogas is compressed and sent through the contents of the reactor. Only a small amount of water is recirculated, and the total solids content of the waste in the reactor is still high. Other processes also use the same principle, i.e., recirculation of small or large amounts of process water. The Valorga process is a relatively widely used process. There are several full-scale plants worldwide, such as in Amiens, France (85,000 ton/year), Grenoble, France (16,000 ton/year), Tilburg, Netherlands (52,000 ton/year) (Fig. 16), Papeete, Tahiti (90,000 tons/year) and Tamara in French Polynesia (92,000 tons/year). The DRANCO Process
The Dranco (Dry Anaerobic Composting) process is a true dry-process for treatment of the organic fraction of MSW (Fig. 17). Indeed, this process requires a high total solids content in the reactor in order to have optimal performance.
Fig. 15. Principle diagram of the Valorga process
28
I. Angelidaki et al.
Fig. 16. View of a Valorga biogas plant (Tilburg, Holland). The plant has capacity of 52,000 tonnes per year of source separated VGF (Vegetable-Garden-Fruit) waste
Fig. 17. Principle diagram of the Dranco process
Therefore, it is often recommended to mix non-recyclable paper or garden waste into the MSW in order to achieve a sufficiently high TS content. The Dranco process is a thermophilic process. The process takes place in the following way: after the waste is pretreated and screened it is mixed with recirculating material from the reactor. Three quarters of the reactor content is recycled. Mixing of the waste with this large amount of digested material ensures inoculation of the incoming material. The biomass paste is pumped using piston pumps developed to pump concrete casting mixtures. The reactor is a downward plug-flow type reactor. In the reactor no significant mixing takes place. If the waste consists of high amounts of easily degradable
Applications of the Anaerobic Digestion Process
29
Fig. 18. View of a Dranco biogas plant
material, the degradation can cause liquefaction of some of the material in the reactor, preventing the plug flow principle with the risk that some particles will flow through the reactor without achieving the required retention time. Therefore, addition of recalcitrant ligno-cellulose material is usually recommended. The digested biomass is extracted from the bottom of the reactor by a special patented sliding frame, pulling material evenly from the reactor cross section into an extraction screw channel. Today, there are several DRANCO plants in operation, Fig. 18, such as the one in Brecht, Belgium (12,000 ton/year), Salzburg,Austria (20,000 ton/year), Bassum, Germany (13,500 ton/year), and Kaiserslautern, Germany (20,000 ton/year). The Kompogas Process
The Kompogas process is a dry process developed in Switzerland. The process operates in the thermophilic range with a hydraulic retention time of approx. 15 days. The reactor is a horizontal cylinder and the flow through the reactor is plug flow. In the reactor a stirrer provides some mixing of the waste. Recirculation of a part of the effluent to the incoming substrate ensures inoculation. 4.3.3 Multi-stage Anaerobic Digestion Systems
Most of the digestion systems for anaerobic treatment of MSW are single-stage reactor systems. However, multi-stage digestion systems are also used. These are usually two-stage systems, although 3-stage systems have also been proposed. The idea with the multi-stage systems is to separate the different phases of the anaerobic digestion process, in order to be able to apply optimal conditions in
30
I. Angelidaki et al.
each of them. Different operating conditions, such as pH and retention time can be kept in the different stages. The BTA Process
The BTA process was developed in Germany (Kubler and Schetler 1994). The process is a multi-stage process (Fig. 19). Pretreatment is centered on a hydro-pulper, which receives the source-sorted waste from a screw-mill, which opens bags and disintegrates larger agglomerated particles. In the hydro-pulper, the waste is mixed with recirculated process water and the organic material is dissolved through intense agitation. Floating light elements (plastic, wood etc.) are skimmed from the surface by a periodically operated rake and heavy items (metal, glass etc.) are removed by sedimentation. The pretreatment produces a pulp of approx. 10% total solids (TS). The pulp is pumped to a buffer tank where acidification occurs. The effluent from the acidification reactor is dewatered by centrifugation. The thin liquid fraction is fed into a high flow biofilm reactor; while the thick fraction of undissolved material is mixed with process water and fed to a CSTR reactor where further hydrolysis and acidification takes place. The effluent from the CSTR reactor is once again dewatered, and the liquid fraction fed into the biofilm reactor for methanization at mesophilic conditions. A part of the thick fraction is removed from the process to take out inert and undegradable material with a TS content of approx. 35%. This “compost” contains approx. 0.2–0.3 kg N per ton. Excess process water, which contains most of the nitrogen (about 4–8 kg N/ton original solid waste; see below) and other nutrients that were in the original waste, is disposed of to the sewer system or recycled to the agricultural sector as a fertilizer if possible. The volatile solids (VS) destruction is predicted to approx. 85%.
Fig. 19. Principle diagram of the BTA process
Applications of the Anaerobic Digestion Process
31
4.4 Summary of Processes Used for Anaerobic Treatment of Solid Wastes
There is a wide variety of configurations for anaerobic treatment of solid wastes. One way to classify these reactor systems is depicted in Fig. 20. The batch systems can be considered as accelerated landfill systems. These systems are simple and comparatively cheap. MSW is loaded batch wise in a closed vessel containing inoculum from a previous batch digestion. During the digestion period, leachate is recirculated for mixing purposes of substrate, microorganisms and moisture. When the digestion is complete, the digested material is unloaded and a new batch digestion is initiated. An alternative to batch digestion is the leach bed process, where the leachate from the base of the reactor is exchanged between established and new batches to facilitate start up, inoculation and removal of volatile acids in the active reactor.When the methanogenesis is established, the reactor is connected to another reactor containing new MSW. This concept has also been described as sequential batch anaerobic composting (SEBAC). The continuously operating systems can be divided into completely mixed and plug-flow systems. The completely mixed systems can again be classified as systems based on recirculation of process water for dilution of the incoming MSW and in systems based on the codigestion concept. Codigestion is especially well established in Denmark. Several systems are operating on the multi-stage digestion concept. However, one-stage systems are much simpler and cheaper and therefore, considerably more widespread.
Fig. 20. Classification of anaerobic solid waste digestion systems
32
I. Angelidaki et al.
5 References 1. McInerney MJ, Bryant MP, Stafford DA (1980) Metabolic stages and energetics of microbial anaerobic digestion. In: Stafford DA, Wheatley BI, Hudges DE (ed) Anaerobic digestion. Applied Science, London 2. Gujer W, Zehnder AJB (1983) Water Sci Technol 15:127 3. Allison MJ (1978) Appl Environ Microbiol 35:872 4. Switzenbaum MS, Giraldo-Gomez E, Hickey RF (1990) Enzyme Microb Technol 12:722 5. Ahring BK, Sandberg M, Angelidaki I (1995) Appl Microbiol Biotechnol 6. Dugba PN, Zhang H (1999) Bioresour Technol 68:225 7. Whitmore TN, Lazzari M, Lloyd D (1985) Biotechnol Lett 7:283 8. van Lier JB, Rebac S, Lettinga G (1996) In: Proceedings of the IASWQ-NVA Int. conf. on advanced wastewater treatment 9. Varel VH, Hashimoto AG, Chen YR (1980) Appl Environ Microbiol 40:217 10. Hashimoto AG (1982) Agriculural Wastes 4:345 11. Ahring BK (1995) Ant van Leeuw 67:91 12. Chen YR, Day DL (1986) Agriculural Wastes 16:313 13. Fang HHT, Wai-Chung Chung D (1999) Water Sci Technol 40:77 14. Archer DB (1983) Enzyme Microb Technol 5:162 15. Buhr HO, Andrews JF (1977) Wat Res 11:129 16. Varel VH, Isaacson HR, Bryant MP (1977) Appl Environ Microbiol 33:298 17. Hashimoto AG (1982) Biotechnol Bioeng 14:2039 18. Mackie RI, Bryant MP (1981) Appl Environ Microbiol 41:1363 19. Madamwar D, Patel A, Patel V (1990) J Ferment Bioeng 70:340 20. Casali GB, Senior E (1989) J Chem Tech Biotechnol 44:31 21. Hashimoto AG, Varel VH, Chen YR (1981) Agriculural Wastes 3:241 22. Hashimoto AG (1983) Biotechnol Bioeng 25:185 23. Nyns EJ, Schönborn W (1986) Biomethanation processes, Berlin: Wiley-VCH Weinheim, (8): Microbial degradations, p 207 24. Kato MT, Field JA, Kleerebezem R, Lettinga G (1994) J Ferment Bioeng 77:679 25. Rebac S, Ruskova J, Gerbens S, van Lier JB, Stams AJM, Lettinga G (1995) J Ferment Bioeng 80:499 26. Björnsson L (2000) Intensification of the biogas process by improved process monitoring and biomass retention, Dissertation, Lund University, Lund, Sweden 27. Zinder SH (1993) Physiology and ecology of methanogens. In: Ferry JG (ed) Methanogenesis. Ecology, physiology, biochemistry and genetics. Chapman and Hall, New York 28. Moosbrugger RE, Wentzel MC, Ekama GA, Marais GR (1993) Water SA 19:11 29. Mosey FE, Fernandes XA (1989) Water Science and Technology 21:187 30. Wilcox SJ, Hawkes DL, Guwy AJ (1995) Wat Res 29:1470 31. Angelidaki I, Ahring BK (1994) Water Res 28:727 32. Rozzi A (1991) Med Fac Landbouww Rijksuniv Gent 56:1499 33. Pretorius WA (1994) Water Science and Technology 30:1 34. Speece RE (1983) Environ Sci Technol 17:416 35. Angelidaki I, Ahring BK (1993) Appl Microbiol Biotechnol 38:560 36. Bhattacharya SK, Parkin GF (1989) J WPCF 61:55 37. Sprott GD, Shaw KM, Jarrell KF (1984) J Biol Chem 259:12602 38. Sprott GD, Patel GB (1986) System Appl Microbiol 7:358 39. Hansen KH, Angelidaki I, Ahring BK (1998) Wat Res 32:5 40. Koster IW (1986) J Chem Tech Biotechnol 36:445 41. Hansen KH, Angelidaki I, Ahring BK (1999) Wat Res 33:1805 42. Kayhanian M, Tchobanoglous G (1992) Biocycle 33:58 43. Hamzawi N, Kennedy KJ, McLean DD (1998) Environ Technol 19:993 44. Angelidaki I, Ahring BK (1998) Biodegradation 8:221
Applications of the Anaerobic Digestion Process
33
45. Imai T, Ukita M, Sekine M, Nakanishi H, Fukagawa M (1998) Water Science and Technology 38:377 46. Angelidaki I, Ahring BK (1992) Appl Microbiol Biotechnol 37:808 47. Rinzema A, Boone M, van Knippenberg K, Lettinga G (1994) Wat Environ Res 66:40 48. Hickey RF, Vanderwielen J, Switzenbaum MS (1987) Wat Res 21:1417 49. Mol N, Kut OM, Dunn IJ (1993) Water Science and Technology 28:55 50. Hickey RF, Vanderwielen J, Switzenbaum MS (1989) Wat Res 23:207 51. Hendriksen HV, Larsen S, Ahring BK (1992) Appl Environ Microbiol 58:365 52. Wu WM, Bhatnagar L, Zeikus JG (1993) Appl Environ Microbiol 59:389 53. Donlon BA, Razo-Flores E, Lettinga G, Field JA (1996) Biotechnol Bioeng 51:439 54. Bradley PM (2000) Hydrobiol J 8:104 55. Christansen N, Christensen SR, Arvin E, Ahring BK (1997) Appl Microbiol Biotechnol 47:91 56. Zhuang P, Pavlostathis SG (1994) Water Sci Technol 30:85 57. Hörber CH, Christansen N, Arvin E, Ahring BK (1998) Appl Environ Microbiol 64:1860 58. Donlon BA, Razo-Flores E, Luijten M, Swarts H, Lettinga G, Field JA (1997) Appl Microbiol Biotechnol 47:83 59. Bendixen HJ (1996) Copenhagen 1: 296 60. Ahring BK, Angelidaki I, Johansen K (1992) Water Sci Technol 25:311 61. Hedegaard M, Jaensch V (1999) Renewable Energy 16:1064 62. IEA Bioenergy (1994) Minister of Energy/Danish Energy Agency, Copenhagen, Denmark Received: June 2002
CHAPTER 1
Anaerobic Granular Sludge and Biofilm Reactors Ioannis V. Skiadas · Hariklia N. Gavala · Jens E. Schmidt · Birgitte K. Ahring The Environmental Microbiology and Biotechnology Group (EMB), Biocentrum-DTU, bldg 227, The Technical University of Denmark, 2800 Lyngby, Denmark E-mail:
[email protected] Present address. J. E. Schmidt, Environment & Resources DTU bldg 115, The Technical University of Denmark, 2800 Lyngby, Denmark
The long retention time of the active biomass in the high-rate anaerobic digesters is the key factor for the successful application of the high rate anaerobic wastewater treatment. The long solids retention time is achieved due to the specific reactor configuration and it is enhanced by the immobilization of the biomass, which forms static biofilms, particle-supported biofilms, or granules depending on the reactor’s operational conditions. The advantages of the high-rate anaerobic digestion over the conventional aerobic wastewater treatment methods has created a clear trend for the change of the role of the anaerobic digestion in the wastewater treatment plants from a pre-treatment method to the main biological treatment method. The application of staged high-rate anaerobic digesters has shown the larger potential among the recent developments in this direction. The most common high-rate anaerobic treatment systems based on anaerobic granular sludge and biofilm are described in this chapter. Emphasis is given to a) the Up-flow Anaerobic Sludge Blanket (UASB) systems, b) the main characteristics of the anaerobic granular sludge, and c) the factors that control the granulation process. Finally, the most innovative staged anaerobic digesters are also presented. Keywords. Granules, Biofilms, UASB, USSB, PABR
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
1
Introduction
2
Biofilm Reactors
3
Granular Sludge Reactors
3.1 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5
Characteristics of the Granular Sludge The Inorganic Composition . . . . . The Content of Extracellular Polymers Diffusional Resistance . . . . . . . . Metabolic Interspecies Transfer . . . Activity . . . . . . . . . . . . . . . . Factors Controlling Granulation . . Quality of the Inoculum . . . . . . . Nutrients . . . . . . . . . . . . . . . Temperature . . . . . . . . . . . . . Cations and Anions . . . . . . . . . Hydraulic Conditions . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . 39 . . . . . . . . . . . . . . . . . . . . . . 42 . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
44 45 46 47 49 50 50 50 52 52 53 55
Advances in Biochemical Engineering/ Biotechnology, Vol. 82 Series Editor: T. Scheper © Springer-Verlag Berlin Heidelberg 2003
36
I.V. Skiadas et al.
4
Innovative Granular Sludge and/or Biofilm Reactors
5
Conclusions
6
References
. . . . . . . 57
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Abbreviations ABR ANAMMOX COD CSTR ECP EGSB HRT OLR PABR SMA SRB TS UASB USSB VFA VSS vup
Anaerobic baffled reactor Anaerobic ammonium oxidation Chemical oxygen demand Continuously stirred-tank reactor Extracellular polymers Expanded granular sludge bed Hydraulic retention time Organic loading rate Periodic anaerobic-baffled reactor Specific methanogenic activity Sulfate-reducing bacteria Total solids Up-flow anaerobic sludge bed Up-flow Staged Sludge Bed Volatile fatty acids Volatile suspended solids Liquid up-flow velocity
1 Introduction Conventional anaerobic treatment systems have traditionally been used for the stabilization of sludge effluents such as sewage sludge and manure, during the last century. The most common configuration used is the continuously stirred-tank reactor (CSTR). The main problem of this reactor type is the fact that the active biomass grows in suspension and is continuously removed from the system thus leading to long retention times (10–30 days) and large volumes of reactors. However, intense research efforts concerning the anaerobic treatment technology and the microbiology of the anaerobic processes have resulted in the development of high-rate systems based on the retention and accumulation of the active biomass despite the short hydraulic retention time applied. High-rate anaerobic digestion has been successfully applied to the treatment of a wide range of industrial and municipal wastewater because it combines the advantages of the conventional anaerobic digestion (i.e., relatively low energy consumption, low production of excess stabilized sludge, energy recovery in the form of methane, low nutrients and chemicals requirements) with the additional advantages of: a) short hydraulic retention time (usually less than 1 day) and high organic loading rate
Anaerobic Granular Sludge and Biofilm Reactors
37
Fig. 1. Schematic representation of a static biofilm. According to Characklis and Marshall [1],
a static biofilm system includes the following five compartments: 1) substratum, 2) base film, 3) surface film, 4) bulk liquid and 5) gas. The base film and surface film constitute the biofilm
(usually more than 10 kg-COD L–1 d–1) resulting in the reduction of reactor size, space requirements, and capital cost, b) high efficiency, c) process stability and d) low or no requirement for mechanical mixing. The high retention time and concentration of the active anaerobic biomass in the high-rate digesters is achieved by the attached growth of the microbial cells, which form biofilms on the surface of solid materials present in the digesters, and/or the use of specially designed settling devices located at the effluent of the digesters.According to Characklis and Marshall [1], a biofilm consists of cells immobilized at a substratum and is a complex coherent structure of cells and cellular products like extracellular polymers. The substratum could be either a static solid surface (static biofilms, Fig. 1) or suspended carriers (particle-supported biofilms, Fig. 2). In the absence of a solid surface and under certain conditions the microbial cells can adhere to each other and form large, dense, self-supported biofilm particles usually called “granules” (Fig. 3). In this chapter, the terms “biofilm reactor” or “biofilm system” will only be used for anaerobic digesters where the active biomass is retained due to the formation of particle-supported and/or static biofilms. High-rate anaerobic wastewater treatment technology is mainly based on granular systems or static and particle-supported biofilm systems. Typical biofilm systems are the fixed bed, the fluidized bed and the expanded bed reactors and typical granular systems are the up-flow anaerobic sludge bed (UASB) and the expanded granular sludge bed (EGSB) reactors. Among them, the UASB reactor configuration is characterized by a) simplicity in construction and operation and b) reduced capital, operating, and maintenance
38
I.V. Skiadas et al.
Fig. 2. Particle supported biofilm [2]. Average diameter is 1.7 mm; bar = 1 mm
Fig. 3. Scanning electron micrograph showing the surface topography of an entire granule [3]
cost. To date, the UASB is the most popular and widely used high-rate anaerobic wastewater treatment system. In the sequel, a comparative description of the most common high-rate anaerobic treatment systems, which are based on anaerobic granular sludge and biofilm, will follow. Emphasis will be given on UASB systems, the granulation process, and the characteristics of the produced anaerobic granular sludge.
Anaerobic Granular Sludge and Biofilm Reactors
39
2 Biofilm Reactors Among the anaerobic reactors that are based on particle-supported biofilms one can distinguish the fixed bed (or fixed film), the fluidized bed and the expanded bed configuration. All three types contain an inert material on the surface of which microbial (mainly bacterial) growth takes place and they can be distinguished by the degree of medium (material and biomass) expansion in each one. The fixed film processes could be ‘up-flow’ or ‘down-flow’. The up-flow configuration is suitable for low suspended solid concentrations in the influent while the down-flow configuration could be applicable for treating wastewaters with higher concentrations of suspended material since it allows for easy and rapid wash-out. This may result in lower effluent quality but it prevents plugging of the reactor. In general, fixed film processes are suitable for treating wastewaters with an organic load in the range of 1–20 g-COD ◊ L–1. Especially for the higher strength wastewater it is recommended that an effluent recycle should be used in order to maintain an influent concentration between 8 and 12 g-COD ◊ L–1 [4]. The expanded and fluidized bed systems are characterized by an increase of the settled bed volume by 15% to 30% and 25% to 300%, respectively, depending on the upflow velocity applied. Small to medium particles with high surface-to-volume ratios should be used in order to improve the mass transfer characteristics of the reactor. In general and in comparison with the suspended growth systems, the biofilm reactors have the advantages of achieving high biomass concentrations and long suspended solids retention times and therefore are suitable for the treatment of low organic content wastewaters. Furthermore, there are no requirements for mechanical mixing while biogas production and effluent recycle could insure adequate mixing, and relatively uniform temperature, pH and substrate concentration. Smaller reactor volumes and therefore smaller land areas are required. On the other hand, these systems are not suitable for wastewater containing high concentrations of suspended solids. Furthermore, the packing material and/or the support system could be prohibitively expensive. In addition, there are high power requirements for bed expansion or fluidization and there is the risk of accidentally wash-out of the medium. In Tables 1 and 2 one can see the advantages and disadvantages of the biofilm reactors and expanded and fluidized bed reactors, respectively [5, 6]. Fluidized bed reactors were first used for aerobic oxidation and denitrification [7]. It was in the mid-1970s when it was realized that the fluidized bed configuration was very well suited for anaerobic processes as well. However, it took almost ten years, until 1984, before fluidized bed systems were applied in full-scale operation. Until then, many studies were performed on the laboratory/pilot scale [8]. More detailed information about full-scale applications of anaerobic fluidized bed reactors is given in a recent review study [6]. In general, biofilm and especially fixed bed (or fixed film) reactors have been used for the degradation of recalcitrant and xenobiotic compounds. In the study of Fatherpure and Vogel [9] a two-stage (anaerobic-aerobic) biofilm reactor was used for the complete dechlorination and degradation of polychlorinated hy-
40
I.V. Skiadas et al.
Table 1. Advantages and disadvantages of biofilm reactors compared to conventional, suspended growth systems
Advantages
Disadvantages
High biomass concentrations and long solid retention times achievable
Not suitable for high suspended solids wastewater
Smaller reactor volumes and consequently smaller land area required due to high organic loading rates
Provision may be required for periodic biomass removal
Suitable for wastewater with low organic content
Limited access to reactor interior for monitoring and inspection of biomass accumulation
No mechanical mixing required since biogas evolution and effluent recycle insure relatively uniform temperature, pH and substrate concentrations in reactor
Control of biofilm thickness is difficult
Relatively stable operation under variable feed conditions or toxic shocks
Cost of carrier medium and support systems are high
High biomass age and minimization of excess sludge production
Table 2. Advantages and disadvantages of expanded and fluidized bed reactors compared to
fixed film processes Advantages
Disadvantages
Excellent mass transfer characteristics Higher conversion capacities and better effluent quality due to the high biomass concentration and mass transfer area
Long start-up periods are required High power requirements for bed expansion or fluidization Risk of accidental wash-out of the medium
drocarbons such as hexachlorobenzene, tetrachloroethylene, and chloroform. Ascón and Lebeault [10] reported that a coupled aerobic-anaerobic recycle biofilm reactor proved to be very efficient for the degradation of the chloroaromatic compound 3,4-dichlorobenzoate.Also, in many cases, up-flow fixed film reactors as well as fluidized bed reactors have been used for the on-site biological remediation of contaminated ground water with petroleum hydrocarbons, monoaromatic hydrocarbons, chlorinated aliphatics and aromatics [11]. In the sequel, studies within the last decade and contributing to the knowledge about the anaerobic fluidized bed reactor and its potential are discussed. In Table 3 the operating conditions as well as the efficiency of anaerobic fluidized bed reactors in both per cent COD removal and methane production are presented based on recent studies. One major evolution during the last decade within the wastewater treatment technology was the recognition of the anaerobic ammonium oxidation process (ANAMMOX). This novel process had been observed for the first time in a flu-
35 °C
Slaughterhouse, clay particles of bentonite, lab-scale
8–0.5 h
20–1.1 d 1.05–0.87 d
2.91 d
35 °C
Wastewater from the production of protein isolates from extracted sunflower flour, saponite (magnesium silicate), lab-scale
Ice-cream wastewater, granular activated carbon, pilot-scale
Mesophilic (33 ± 2 °C)
Vinasse from ethanol distillery of sugar beet molasses, granular activated carbon, lab-scale
2.5–0.37 d
1.47 d
Thermophilic
Wine distillery, porous support medium
HRTs
Ice-cream wastewater, sand, pilot-scale
Temperature
Wastewater treated
methane production are presented % COD removal 96.5–81.5 93
98.3–80 80–48
98.9–75 55.7 63.5
OLR 5.88–32 kg-COD m–3 d–1 1.7 g-COD L–1 d–1
0.6–9.3 g-COD L–1 d–1 9.97–12.1 g-COD L–1 d–1
2.9–54 g-COD L–1 d–1 4.20 kg-COD m–3 d–1 2.20 kg-COD m–3 d–1
0.28 m3 kg-COD–1 added
0.19 m3 kg-COD–1 added
0.320 L g-COD–1 removed
0.209–2.32 L L–1 d–1 2.39–1.59 l L–1 d–1
360 ml g-COD–1 removed
1.08–9 m3 m–3 d–1
Methane production
[16]
[16]
[15]
[14]
[13]
[12]
Source
Table 3. Recent studies on anaerobic fluidized-bed reactors: wastewater treated, operating conditions and efficiency in both per cent COD removal and
Anaerobic Granular Sludge and Biofilm Reactors
41
42
I.V. Skiadas et al.
idized bed reactor treating the effluent from a methanogenic reactor [17]. Later studies showed that the microorganisms responsible of carrying out the ANAMMOX process grow extremely slowly (doubling time 11 days) [18, 19]. Hence it was the reactor configuration (high biomass concentration combined with long suspended solids retention times) as well as the environmental conditions (pH, temperature, nitrate concentration, redox potential) that allowed the growth and the establishment of these types of microbes. In the study of FDZ-Polanco et al. [13] a high, non-expectable, simultaneous removal of nitrogen and sulfur was observed during the treatment of diluted vinasse from an ethanol distillery of sugar beet molasses in an anaerobic fluidized bed reactor. The authors tried to explain this observation by considering a new degradation process that combines the ANAMMOX reaction [20] with sulfate reduction. Part of the possible new biochemical pathway was attributed to the presence of the granular, activated carbon used as biomass carrier. Perez et al. [12] investigated the thermophilic anaerobic digestion of wine distillery wastewater in a fluidized bed reactor. Besides the high efficiency of the system regarding COD removal and methane production, they discovered that extremely high active biomass concentrations could be maintained in the system compared with the concentration of the total volatile attached solids. This observation was attributed to the limitation of the thickness of the biofilm due to the high liquid flow rate applied.
3 Granular Sludge Reactors The most common reactor configuration based on the formation of granules is the Up-flow Anaerobic Sludge Blanket (UASB) reactor. The concept of the UASB was developed in the 1970s [21]. Typically, the interior of a UASB reactor is divided into four sections: (1) the granular sludge bed, (2) the fluidized zone, (3) the gas-liquid-solids separator, and (4) the settling area (Fig. 4a). The granular sludge bed is located in the bottom of the reactor where the biomass settles as granules. The wastewater is fed from the bottom of the reactor and flows upward through the sludge bed. Here the main part of the organic compounds is biologically degraded and biogas is produced. On top of the granular sludge bed, a fluidized zone is developed due to the biogas production within the granules. In this zone further biological degradation can take place. The produced biogas mixed with floated sludge is collected in the gas-solids separator located on the top section of the reactor, where it is separated and flows out of the reactor. Granules and sludge flocs with good settling abilities settle back to the granular sludge bed through the fluidized zone, while flocculent sludge and dispersed bacteria are washed-out of the reactor by the effluent stream. The formation of granular sludge is not a prerequisite for a UASB reactor, which can also be operated with well settling flocculent sludge of high methanogenic activity at low wastewater flow rates. However, the formation and stability of the granules are essential for successful operation of the UASB at high wastewater flow rates [22, 23]. Today the UASB has become the most popular high-rate digester and many plants for anaerobic biological treatment of wastewater based on the UASB con-
43
Anaerobic Granular Sludge and Biofilm Reactors
Biogas
Biogas
Effluent
Effluent
Gas-solid-liquid separation
Settling area
Fluidized zone
Recirculation
Granular sludge bed
Influent
a
Influent
b
Fig. 4. Schematic representation of the a UASB and b EGSB reactors
cept are in operation throughout the world. The UASB has been successfully applied to a wide range of industrial wastewaters, including wastewaters from sugar industries [21], potato processing [21, 24], breweries [25–27], ice cream factories [28], fruit and vegetable processing [29, 30], distilleries [25, 31], slaughterhouses [21, 32], paper mills [25, 33–37], ethanol production (vinasse) [38, 39], olive oil mills [40], soft drink manufacturers [41, 42], starch industries [25, 43] and dairy industries [21, 44, 45]. Also the UASB reactor has been proven effective for the treatment of certain xenobiotic compounds [35, 46–55]. High organic loading rates (20 kg-COD m–3 d–1) and low hydraulic retention times (1 day) can be applied for wastewater with no or low solid content, whereas moderate organic loading rates (< 5 kg-COD m–3 d–1) have to be used for wastewater with substantial solids content. Wastewater with a high proportion of insoluble COD (>15%) can result in poor granulation in the reactor; in these cases a two-stage system with separate solids hydrolysis can be the optimal choice [56]. The UASB process is also used for the treatment of domestic sewage and other low strength wastewater [57–59]. The characteristics of the anaerobic treatment of low strength wastewater in UASB reactors with emphasis on the treatment of sewage are summarized in the recent review studies of Lettinga [22], Seghezzo et al. (1998) [60], and Kalogo and Verstraete (1999) [61].
44
I.V. Skiadas et al.
Table 4. The main characteristics of the EGSB reactor in comparison with the UASB
– Application of higher liquid up-flow velocities (in the range of 4–10 m · h–1) and organic loading rates (up to 40 kg COD · m–3 · d–1) – Expansion or fluidization of the sludge bed – More suitable for soluble and dilute wastewaters – Granular sludge with good settling characteristics and high methanogenic activity – Good mixing conditions due to the high up-flow velocity and the increased biogas production
The main problems related to the use of the UASB for low strength wastewaters are the slow formation of granules, the reduced methanogenic activity, and the low biogas production, which lead to poor mixing conditions in the sludge bed (i.e., accumulation of flocculent sludge between the sludge granules and channeling of the fed wastewater through the sludge bed leading to dead volume) and low reactor performance. The same applies for the treatment of sewage. To overcome some of these problems a modified UASB reactor called the EGSB reactor (Fig. 4b) has been developed. This reactor is characterized by high liquid up-flow velocity (> 4 m h–1, achieved by recirculation and/or high feed flow rate), which in combination with the production of biogas causes expansion or even fluidization of the granular sludge bed. As a result, the content of an EGSB reactor is completely mixed and the hydraulic behavior of the EGSB is equivalent with that of a continuously stirred-tank reactor. This improves the contact between the biomass and the wastewater and significantly reduces the value of the apparent substrate saturation constant (KS) of the granular biomass [62, 63]. Consequently, the EGSB is suitable for the treatment of wastewaters with very low content of insoluble organic matter at high organic loading conditions. Kato et al. [63] demonstrated the feasibility of the EGSB for the treatment of dilute wastewaters with experiments at organic loading rates up to 12 g-COD L–1 d–1, influent COD concentrations as low as 100 g L–1 and removal efficiencies ranging from 80% to 97%. Furthermore, the EGSB systems enable the application of anaerobic treatment for low strength cold (i.e., < 10 °C) wastewater [64]. Finally, when the high upward liquid velocity is achieved by recirculation, the EGSB reactor looks very promising for the treatment of toxic compounds due to the influent dilution [60]. The results reported so far did not support the potential of this reactor for the treatment of raw (unsettled) sewage since the COD removal efficiency was low (30–40%). Influent suspended solids and flocculent sludge with poor settling characteristics are washed out of the EGSB system by the high velocity of the liquid and, consequently, the quality of the effluent from an EGSB is lower than that from a UASB reactor during the treatment of raw sewage [61]. The main characteristics of the EGSB reactor in comparison with the UASB are summarized in Table 4 [60]. 3.1 Characteristics of the Granular Sludge
Various types of conglomerates of microbes have been described, such as granules, pellets, flocs, and flocculent sludge. However, there is no clear distinction be-
45
Anaerobic Granular Sludge and Biofilm Reactors
tween the different conglomerates. Dolfing [65] used the following definitions of flocs, flocculent sludge, pellets, and granular sludge: pellets and granules are conglomerates with a dense structure. After settling, these conglomerates present a “well-defined appearance”. Flocs and flocculent sludge are conglomerates with a loose structure. After settling, they form one macroscopic layer. This gives a good descriptive definition, but gives no guidelines for determining the different types of conglomerates. The diameter of sludge granules varies from 0.14 to 5 mm, depending on the wastewater used, the operational conditions, and the analytical method. Granules cultivated on acidified substrates are generally smaller than granules grown on acidogenic substrates [66–72]. The granules vary widely in shape, depending on the conditions in the reactor [73–75] but they usually have a spherical form. Typical reported buoyant densities of granules range between 1.03 and 1.08 g ml–1, but densities up to 1.4 g ml–1 have been reported as well [72, 76–78]. The fact that the density of granules is in the same range with that of bacteria cells indicates that the observed settling abilities of the granules must be due to aggregation of the anaerobic bacteria together with the inorganic enclosements [23]. The linear liquid flow rate at which a granule with a given volume and buoyant density will be washed out of the reactor can be estimated by Stoke’s law: Vs =
D2 g (r p – r ) 18 h
,
(1)
Re p < 2
È (r p – r ) D1.6 g ˘ Vs = 0.153 Í ˙ 0.6 0.4 Î h r ˚
0.174
,
2 £ Re p £ 400
(2)
where Rep = Reynolds number (Vs rp D/h), Vs = settling velocity, g = gravimetric constant, D = diameter of granule, r = density of water, rp = density of granule, h = viscosity of liquid. Granules with different volumes and densities can be present in a reactor at a given linear flow rate; both small granules with high densities and larger granules with low and high densities will be present. Reported settling velocities for granular sludge are in the range of 18 to 100 m h–1 but typical values are between 18 and 50 m h–1 [74, 79–81]. 3.1.1 The Inorganic Composition
The ash content of the granules varies from 10% to 90% of the dry weight of the granules [69, 72, 77, 78, 82–86] and it strongly depends on the reactor operational parameters, i.e., temperature, complexity of the wastewater organic composition, pH, operational period, organic loading rate, and influent mineral concentration. As a result, the ash content of granule samples taken at various occasions from the same reactor can vary by up to 100% [79–87]. However, some generalizations can be made concerning the negative or positive effect of the operational parameters on the ash content of the granules and they are presented in Table 5.
46
I.V. Skiadas et al.
Table 5. The effect of the operational parameters on the ash content of the granules
Parameter
Effect
Source
Temperature Feed complexity pH Operational period Organic loading rate Feed mineral concentration
+ – + + + +
[83, 87] [67, 69, 70, 72, 75, 78, 81, 82, 84, 88, 89] [70, 90–93] [84] [85, 94] [95]
The structure of the anaerobic granular sludge is influenced by the location of inorganic precipitates. Precipitation can take place in the bulk liquid and/or in the granules. Bulk precipitation is governed by the composition of the bulk liquid which varies over the height of the reactor, especially in UASB reactors with a low up-flow velocity [62]. If wash-out of the forming precipitates occurs, no harmful effects on the sludge structure and quality are to be expected. If the formed precipitates are retained in to the reactor, new biofilms can develop on their surface. On the other hand, agglomeration of the formed precipitates can take place, which leads to channeling and clogging problems [97]. The accumulation of inorganic precipitates into the granules is thought to be essential for granule development in terms of maintaining high specific gravity and of providing support materials for bacterial cell attachment [72, 93]. However excess accumulation of inorganic deposits leads to a decrease in available reaction volume in the UASB reactor. If granules with a high ash content increase in size, too high density, low porosity and diffusion limitation will occur. Ash content below 60% is considered to be favorable for anaerobic sludge. For sludges with higher ash content, serious decreases of methanogenic activity have been reported [86, 97, 98]. The main components of the ash content of the granules are calcium, potassium, phosphorus, magnesium, sodium, iron, and trace elements such as Ni and Co [23]. Bhatti et al. [96] showed that not all the ash content of granular sludge fed with brewery wastewater was composed of metals but also contained materials similar to silica and/or silt. Weight calculations revealed that about 50% of the ash consisted of inorganic silica and silt while the other 50% consisted of metals. Silica and/or silt may also be an important ingredient of granular sludge and may help provide a three-dimensional matrix for bacterial aggregation. 3.1.2 The Content of Extracellular Polymers
Extracellular polymers (ECP) can usually be found in abundant quantities in natural systems. Understanding the physical and (bio)chemical characteristics of the extracellular polymeric matrix is important for the understanding of the structure and function of granules. Bacterial ECP are defined as polysaccharide-containing structures of bacterial origin lying outside the integral elements of the outer membrane of Gram-negative cells and the peptidoglycan of Gram-positive cells [99, 100]. The ECP in granules consist mainly of protein and polysaccha-
Anaerobic Granular Sludge and Biofilm Reactors
47
rides. The reported ECP content of granules is between 0.6–20% of the volatile suspended solids [30, 79, 85, 101–105] depending on the granular sludge examined and the extraction and analytical method used for ECP determination. Typically, the carbohydrate content of anaerobic sludge or granules ranges from 1–3.5% while that of proteins ranges from 3–16% of the volatile suspended solids [79, 93, 96]. The amount of lipids in ECP from granules grown under different conditions lies between 0.02 and 0.05% [88] and that of nucleic acids is around 0.4% [85] of the volatile suspended solids. Several researchers have shown, using microscopic observations, that the bacteria in granules are surrounded by ECP and it is generally accepted that the formation of granules is correlated with the production of ECP [79, 87, 90, 101, 103, 106–108]. ECP promote the formation of bacterial aggregates and mediate adhesion of bacteria in natural systems by forming a bridge between different surfaces, thereby producing a three-dimensional floc matrix [104]. The composition of ECP affects the surface properties of the bacterial flocs and the physical properties of the granular sludge [102, 109]. Dispersed bacteria are negatively charged and there is electrostatic repulsion between the cells. The production of ECP can change the surface charge of the bacteria, resulting in aggregation. The amount of ECP produced is affected by the conditions under which the granules are grown. The concentration of ECP is lower in thermophilically grown granules compared to mesophilically grown ones [84, 88]. The amount of ECP is also affected by the influent composition. Shen et al. [104] showed that the amount of carbohydrates extracted from granules increased with the addition of iron and yeast extract to the influent. The opposite effect was observed when iron was absent. It has also been shown [88, 105] that an increased C/N ratio stimulated the production of extracellular polysaccharide, resulting in improved bacterial attachment to solid surfaces. The ECP content of the anaerobic granules depends on the organic loading rate (OLR) as well. In experiments with anaerobic granular sludge treating industrial and synthetic wastewaters, Bhatti et al. [96] observed that granules operated at low OLR had the highest ECP content and generally, there was a tendency for an increase in ECP content with a decrease in OLR. Quarmby and Forster [110] found that the granules tended to become weaker as the OLR increased and the ECP from weak granules had lower polysaccharide content. Comparisons between the extracellular polysaccharides extracted from methanogenic granules and from Methanobacterium formicicum and Methanosarcina mazeii cultures showed that the extracellular polysaccharides from the methanogenic granules contained all the sugars found in the ECP content of Methanobacterium formicicum and Methanosarcina mazei indicating that these two methanogens contributed significantly to the production of the extracellular polymer of the anaerobic granules [105]. 3.1.3 Diffusional Resistance
One disadvantage of immobilized cells system such as the UASB reactor is that the overall conversion rate in the granules can be affected by external mass transfer resistance (i.e., mass transfer from the bulk liquid through the stagnant liq-
48
I.V. Skiadas et al.
uid film to the surface of a granule) and internal mass transfer resistance (i.e., mass transfer from the surface to the interior of a granule). The influence of the mass transfer resistance depends on the bulk substrate concentration, the degree of turbulence in the fluid, the morphology and the size of the granules, the cell density of the granular biofilm, the substrate saturation constant (KS), and the maximum specific activity of the biomass [76, 111, 112]. It has been shown that inter-granular chemical gradients are highly influenced by the effective diffusion rate of each substance within the biofilm and that the effective diffusion rate is a complex function of the diffusion rate in water and of the biofilm local VSS concentration through the local biofilm porosity [113]. Experimental results have revealed that the effective diffusion coefficient of LiCl in a dense matrix of deactivated methane-producing microorganisms is 22–33% of the diffusion coefficient in pure water [114]. Kitsos et al. in 1992 [115] estimated the effective diffusion coefficient of acetate in an anaerobic biofilm to be only 7% of that in pure water. Alphenaar and coworkers [116] showed that the substrate transport limitation increases with the diameter of the granules. However, insignificant difference in substrate affinity was observed between granules of different sizes. Diffusional resistance resulted in substrate depletion in the interior of the granules. Consequently, autolysis occurred in the core of large enough (2 mm) granules, producing hollow granules and thereby reducing the porosity of the interior layers. The average porosity of large granules was thereby lower than for small granules. The substrate diffusion velocity was overestimated because it was calculated over the total granule radius, whereas actually only the external layers were involved. For the granular samples tested, overall effective diffusion coefficients were found to be between 40 and 80% of the diffusion coefficient in pure water. Substrate transport limitation in large hollow granules can be reduced by convective flow.Van den Heuvel and coworkers [117] showed that gas bubbles entrapped in biocatalyst particles subjected to hydrostatic pressure oscillations, e.g., during recirculation in loop reactors, induce intraparticle liquid flows, and thereby enhance mass transfer in excess of diffusion. This ‘breathing particle’ mechanism was demonstrated in methanogenic granules, and led to a typical activity increase of 13% compared with static pressure conditions. Mass transfer resistance can be described by the effectiveness factor, h, which is the ratio of apparent substrate utilization rate of intact granules to the intrinsic substrate utilization rate of disintegrated granules. Substrate utilization rate increases with decreasing granule size.Wu et al. [118] investigated the diffusional resistance of brewery granules (1.8 to 3.0 mm) during acetate, propionate, and ethanol utilization.At low concentrations, the differences in rates between whole granules and flocs are small. At higher initial substrate concentration (6 mM acetate) the apparent utilization rate of the flocs is three times the rate of whole granules. Finally, the calculated h is 0.32, 0.41, and 0.75 for acetate, propionate, and ethanol, respectively. The effectiveness factor h for the acetate degradation in methanogenic granules is estimated between 0.57 and 0.62 using the pH profiles inside the aggregates [119]. Van Lier et al. [120] reported that crushing of the granules leads to a 2- to 3-fold increase in the maximum specific conversion rate with acetate or butyrate as substrate. On the other hand, Dolfing [112] showed
Anaerobic Granular Sludge and Biofilm Reactors
49
that there is no mass transfer resistance in granules from a mesophilic UASB reactor treating wastewater with acetate or propionate. Mass transfer resistance is only observed when hydrogen is the substrate. Similarly, Schmidt and Ahring [121] concluded that mass transfer resistance is not the limiting factor on the biodegradation of acetate in thermophilic granules (0.44 mm), whereas diffusional resistance is observed only when H2/CO2 is the substrate. The mass-transfer resistance is significant for granules/biofilm grown on acetate, ethanol and/or formate, when the diameter of the granules is high (> 2 mm) and the substrate concentration is low [76, 89, 112, 122, 123]. In conclusion, diffusional resistance increases with increasing granule size but whether the mass transfer is the limiting step of a substrate biodegradation depends on both the substrate bulk concentration and the size of the granules. External (liquid film) mass transfer resistance is related to weak liquid turbulence as a result of relatively low up-flow liquid velocity and low gas production. Consequently, liquid film resistance is not important for the EGSB reactors. On the contrary, it influences the rate of substrate utilization in UASB reactors with granular sludge. Kato et al. [124] reported an increase in the removal efficiency with the increase in up-flow velocity up to 14 m h–1. That could represent the typical effect of liquid film resistance in a mass transfer-limited region. According to Wu et al. [118] who investigated the liquid film resistance of brewery granules during acetate, propionate, and ethanol utilization, a Reynolds number of 159,000 is necessary to ensure good mixing and minimize liquid film resistance. Beyond that point, the effect of liquid film on mass transfer can be neglected. It is appeared, therefore, that liquid-film mass transfer resistance plays an important role in substrate utilization in the UASB reactor system. 3.1.4 Metabolic Interspecies Transfer
Under anaerobic conditions, the short-chained fatty acids other than acetate, such as propionate and butyrate, are oxidized to acetate and H2/CO2 by H2producing acetogenic bacteria. Because of unfavorable thermodynamics, oxidation of propionate and butyrate is only possible if the H2 partial pressure is kept very low by the H2-consuming methanogens. Propionate degradation is possible only below a partial pressure of 10–4 atm H2 [125–127]. In granules degrading propionate and/or butyrate, a slight increase in the partial pressure of hydrogen results immediately in a decrease in the degradation rate of the two volatile fatty acids. Low hydrogen partial pressure can only be achieved by interspecies transfer of molecular hydrogen from hydrogen-producing bacteria to hydrogen-oxidizing methanogens in microcolonies in the granules [128–131]. Thermodynamic and flux considerations have shown that the most effective degradation of propionate and butyrate will take place in these microcolonies where the distance between the syntrophic bacteria is small [132, 133]. Consequently, the disintegration of the granules leads to a decrease in the degradation rate of both propionate and butyrate, indicating the importance of microcolonies [128–131].
50
I.V. Skiadas et al.
3.1.5 Activity
High methanogenic activity is one of the characteristics of granular sludge and the specific methanogenic activity (SMA) of mesophilic granules is typically between 0.5 and 2 g-COD-CH4 g-VSS–1 day–1 (10 to 42 mmol-CH4 g-VSS–1 day–1), while the SMA of thermophilic granules has been observed up to 7.1 g-CODCH4 g-VSS–1 day–1 (148 mmol-CH4 g-VSS–1 day–1). The specific methanogenic activity of the biomass depends on the energy and carbon source that the granules were grown on. The maximum values are obtained when the test substrate is identical to the growth substrate or if the test substrate is an important intermediate of the anaerobic metabolism (e.g., H2 and formate). The activity of the granular sludge can be inhibited by high concentrations of fatty acids. Investigations with thermophilic granules degrading a mixture of acetate, propionate. and butyrate showed that propionate and acetate can be inhibitory for the degradation of other volatile fatty acids [134–138]. Furthermore, the granular sludge is able to maintain viability after extended periods without feeding. The methanogenic activity can easily be re-established after the UASB reactor has been shut down for many months [139–141]. 3.2 Factors Controlling Granulation
In spite of the intense research on the granulation of anaerobic sludge, the mechanisms by which granules are formed in UASB reactors are not yet well understood. Schmidt and Ahring [23] have reviewed the current knowledge concerning the microbiology of immobilized anaerobic bacteria and the suggested mechanisms of granule formation. Tay et al. [142, 143] propose a new theory for the molecular mechanism of sludge granulation together with a review of the different models that have been developed for the microbial sludge granulation. It is concluded that more research effort is necessary for gaining better knowledge of the mechanism of the granulation process in order to control it effectively. The presentation of these mechanisms and models are beyond the scope of this section and, therefore, a description of the main factors which control the granulation of the anaerobic sludge in UASB and EGSB reactors follows. 3.2.1 Quality of the Inoculum
The first problem faced when starting-up a UASB reactor is the choice of the seed material. When possible, the best inoculum is granular sludge from UASB reactors treating similar wastewater, but also granules from UASB reactors treating different wastewater can be used [144–146]. In any case, the use of pregranulated seed greatly reduces the required start-up time [147]. Stability problems may arise when a UASB reactor is started-up for treatment of wastewater of different composition and strength. Also, differences in operating conditions are important, e.g., the loading rate and temperature. The differences can lead to changes
Anaerobic Granular Sludge and Biofilm Reactors
51
in the microbial composition and topography of the granules and occasionally to total disintegration of the granules [148–151]. The main condition to be kept is the application of sludge loading rates well below 50% of the biomass maximum substrate utilization rate during the first weeks of reactor start-up at different operating conditions and/or wastewater characteristics [22, 118]. The sludge then can be easily adapted to the new situation. If granular sludge is not available as inoculum, other types of non-granular inocula can be used. Digested sewage sludge is most commonly used as the seed material in these cases. Other types of seeds include fresh or digested cow manure, raw sludge, mud, and aerobic activated sludge [67, 69, 82, 135, 141, 146, 148, 152–154]. The concentration of the non-granular sludge does not have a great influence on the probability of granulation but only on the rate of granulation [67, 76, 155–158]. Inoculation with seed of high total solids (TS) content (> 60 kg-TS m–3) and consequently of lower methanogenic activity (< 0.05 g-COD g-VSS–1 d–1) results in faster granulation than inoculation with seed of low TS content (< 40 kg-TS m–3) and of high methanogenic activity [25, 139, 158]. Inoculation with thicker types of sludge usually leads to a selective wash-out of particles and bacteria based on their settling abilities. On the contrary, the use of thin types of sludge leads to vigorous expansion of the sludge and results in considerable non-selective wash-out of the biomass and, therefore, a longer time is required for sludge granulation. The granulation process can usually be enhanced by the presence of a certain amount of soluble, rapidly acidifying chemical oxygen demand in the feed [22, 88, 159, 160]. The later combined with a low reactor liquid surface tension promotes the formation of layer-structured granules in which hydrophilic acidogens are predominant at the outer layer. This kind of granules appears to allow more stable reactor performance [159]. Grootard et al. [161] reported that the granular growth and the sludge bed stability of a lab-scale UASB reactor treating a completely acidified wastewater could be significantly increase by adding 20% of a sucrose/starch mixture at top of the original COD load and by adjusting the reactor liquid surface tension at low levels (by adding trace amounts of a surfactant). On the other hand, one must have in mind that the rate of the increase of the granule size in syntrophic/ methanogenic or acetoclastic flocs is much higher than in acidogenic flocs. Carbohydrate specific activity shows the best development on granules initiated by non-acidogenic flocs. Hence, solid inner granule layers of syntrophic and/or acetoclastic microorganisms are a prerequisite to obtain an efficient attachment of acidogens, which otherwise can preferentially be confined to flocculent (nongranular) or suspended growth [162]. The presence of inert particulate material to which microbes can be attached is usually beneficial for the initial granulation [25]. In this way, initial granulation can be enhanced by addition of crushed granulated sludge (up to 8%) or inert materials [76, 151, 157, 163], which then serve as granule precursors. The granulation process in UASB reactors inoculated with suspended anaerobic sludge can also be accelerated by the supplementation of the reactors with natural or synthetic polymers that act similarly to the extracellular polymeric substances in the granules [164]. When a non-granular inoculum is used, it is important that it is
52
I.V. Skiadas et al.
one that has been shown to have the ability to degrade the actual wastewater. Adaptation of the non-granular seed sludge can be beneficial [35, 153, 154]. However, one should bear in mind that the higher activity can lead to considerable wash-out if the start-up process is not followed carefully [153]. 3.2.2 Nutrients
The effects of nutrient limitation on microbial growth are well known, e.g., the action of some enzymes is influenced by co-factors containing certain metals. Nutrient requirements are related to conversion rates, which can vary according to the organic loading rates applied to the UASB reactor [165]. Studies have shown that satisfactory granulation and degradation of the carbon source can be achieved if the ratio COD:N:P in the wastewater is less than 250 : 5 : 1 [28, 92, 166]. Many metals have been found as components of essential enzymes that drive many anaerobic reactions. Fe, Ni, and Co are considered the most important metals for the anaerobic digestion process [104, 167]. Also, several researchers have reported that the addition of trace elements to wastewater has a positive effect on biofilm development [95, 104, 168–173]. On the other hand, Callander et al. [174] and Goodwin et al. [28] observed no effect on the granulation process after addition of trace elements. Usually, the required concentrations of trace elements are very low and external addition is not necessary as these metals are present in adequate concentrations in the wastewater to be treated. Combined treatment of several types of wastewater may compensate for the lack of nutrients in some of them. Singh et al. [175] have made an extended review of nutrient and trace metal requirements for the adequate treatment of different types of wastewater in UASB reactors.Also, they report about the dosing of nutrients and trace metals required for the acceleration of the granulation process. The production of ECP, which plays an important role in granule formation (see Sect. 3.1.2), is often activated by limitation of nutrients. ECP formation is generally enhanced by high C : N and/or C : P ratios [105, 170, 176]. Methanobacterium formicicum and Methanobrevibacter arboriphilus AZ, which have reported to be essential for the granulation process, produce more ECP under reduced nutrient (N and/or P) conditions [105, 177, 178]. This suggests that operation of a reactor under N and P deficiency accelerates the bacterial granulation process and is beneficial for the maintenance of the granular structure due to the enhanced ECP production [105, 176]. 3.2.3 Temperature
The temperature of UASB reactor can be an important factor affecting the granulation process. The effects of temperature fluctuations depend on the temperature, exposure time, and bacterial composition of the granular sludge. The investigation of the influence of temperature fluctuations on the formation of thermophilic (53–55 °C) granules grown on a mixture of acetate, propionate, and butyrate in lab-scale UASB reactors showed that especially propionate degrada-
Anaerobic Granular Sludge and Biofilm Reactors
53
tion is highly affected by temperature fluctuations [138]. Propionate degradation decreased concurrently with decreasing temperature (between 30 °C and 48 °C). However, the propionate removal efficiency was restored after stabilization of the temperature in the range of 53 to 55 °C. The recovery of the propionate degradation resulted in a 90% increase in the reactor’s biomass content and in a significant increase of the granules size indicating the strong effect that the right temperature conditions have on the granulation process. Lau and Fang [179] investigated the effects of temperature shock to the activity of thermophilic (55 °C) granules in UASB reactors fed with sucrose wastewater. Increase to 65 °C or decrease to 37 °C reduced the reactors’ COD removal efficiencies from 90% to 60% and 40% respectively. Severe biomass wash-out and volatile fatty acids accumulation (especially propionate) could take place but the reactors were able to fully recover in a short time. Results from specific methanogenic activity tests showed that acetotrophic methanogens are less sensitive to the shock than the other bacterial groups. Fang and Lau [151] noticed that during the start-up of thermophilic UASB reactors using mesophilic flocculent digested sludge or disintegrated granular sludge the impact of temperature increase was more severe on the reactors with flocculent sludge. Furthermore, the granulation process was completed much earlier in the reactor inoculated with the mesophilic disintegrated granular sludge, even that all the reactors encountered sludge wash-out and deterioration right after the temperature increase. Also, van Lier et al. [165] have investigated the influence of short-term temperature increases on the performance of mesophilic UASB reactors. Increases up to 45 °C had no detrimental effects on the process. Higher temperature (up to 61 °C) led to a sharp decrease of the methanogenic activity of the granules. Propionate degradation was the most sensitive process to temperature increases, whereas the acidogenesis was the least affected. The operation of the anaerobic reactors at a temperature below that for optimal bacterial growth stimulates the ECP production and subsequently the self-immobilization of bacteria in granules [105]. The granule formation provides an extra biomass buffer enhancing the stability of the reactors (especially thermophilic ones) in case of a sudden temperature increase or decrease. While the substrates utilization rates are strongly dependent on the temperature, the granular sludge is less sensitive than the flocculent sludge to temperature changes. This temperature compensation effect is lower at lower specific activity of the immobilized sludge. Regarding propionate degradation a strong temperature dependency was found irrespective the structure of the sludge [118, 180]. 3.2.4 Cations and Anions
The divalent cations such as Ca2+, Mg2+, Fe2+, Ba2+, Zn2+ are important for the production of bacterial flocs and enhance the granulation process [28, 94, 104, 156, 158, 181–184]. The fact that the divalent cations result in a relatively stronger effect of “van der Waals” attractive forces explains their positive effect on the flocculation of dispersed sludge. Cations also interact strongly with anionic
54
I.V. Skiadas et al.
groups present in the organic matrix of the granules [170, 181]. The anaerobic sludge is usually negatively charged due to the presence of ECP. These polymers interact with the divalent cations in the extracellular space producing a polymer matrix that binds the cells [185, 186]. Many of the microbes that participate in the granulation process undergo changes in morphology and growth rate under different concentrations of cations and therefore, changes of cations concentration could result in changes of the structure of the produced granules. Schmidt et al. [69, 90, 187] showed that the structure and microbial composition of thermophilic Methanosarcinae-dominated granules is greatly influenced by the magnesium concentration in wastewater. In the absence of magnesium, the conversion of acetate decreases and the granules become fluffy. The dominating bacterial flora changes from Methanosarcina spp. to rod-shaped bacteria immunologically related to Methanobacterium thermoautotrophicum GC1 and DH. At high Mg2+ concentrations (30 and 100 mM) the morphology of Methanosarcina spp. changes from clumps to single cells resulting in wash-out of the biomass from the UASB reactor. Other methanogenic bacteria, which can also change morphology from single rods to long filaments depending of the cation concentration, are Methanospirillum hungatei [188], Methanobacterium thermoautotrophicum DH, M. thermoautotrophicum BA, and Methanobacterium thermoformicicum sp. [189]. Ca2+ concentrations up to 3.75 mM appear to promote granulation [28, 157, 168, 183, 190] although no further improvement is observed at higher concentrations [157]. Moderate concentrations of Ca2+ (approximately 20 mM) have no significant effect on the granulation process [169, 170, 191]. However, high concentrations can lead to considerable precipitation of Ca2+ salts having a negative effect on the granulation process and resulting in severe problems such as scaling of the reactor walls, space occupation by inorganic precipitates and cementation of the sludge bed, sludge wash-out, decrease of specific methanogenic activity, and loss of buffer capacity [86, 97, 192]. The presence of Mg2+ has similar effect on the granulation of the anaerobic sludge [158, 168]. The optimal Mg2+ concentration for both granulation and degradation of acetate in thermophilic UASB reactors is proposed to be around 10 mM [69, 90, 187]. Monovalent cations such as NH4+ and Na+ may also affect the settlability of the anaerobic granular sludge. At neutral pH, Na+ concentrations of 5, 10, and 14 g L–1 caused 10, 50, and 100% inhibition, respectively, of the acetoclastic methanogenic activity of a granular sludge [193]. NH4+ at concentration of 1 g L–1 can be detrimental to the granulation process and at even higher concentrations it can be toxic to anaerobic bacteria [144, 181, 194]. Sulfate is also an ion that plays an important role in the granulation process. Sulfate and other sulfur compounds are reduced to sulfide by the sulfate-reducing bacteria (SRB) under anaerobic conditions. The SRB compete with the methanogens for the available hydrogen and acetate. In relatively high concentrations of sulfate the SRB predominate over the methanogens, which are the nucleation centers of the granulation process, resulting in reduced production of methane and slower granule formation. Additionally, the sulfide produced can be inhibitory for the methanogenic bacteria and can cause severe corrosion problems to
Anaerobic Granular Sludge and Biofilm Reactors
55
the digesters [144, 150, 195–199]. In order to avoid sulfate inhibition and corrosion problems a COD/sulfate ratio higher than 10 should be maintained [199]. 3.2.5 Hydraulic Conditions
The liquid up-flow velocity (vup) and the superficial biogas velocity have considerable effect on the sludge granulation process in the UASB reactors, thereby acting as a selection pressure in the cultivation of the biomass. The granule size development is the result of the dynamic equilibrium between the biomass detachment from the granule by fluid shear and granule colonization by bacterial growth. The high hydraulic loading rate and biogas production rate promote the granulation of biomass due to the washing out of a) bacterial flocs with poor settling characteristics, b) fine particles detached from the granules surface, and c) bacteria that are unable to form well settling granules [22, 75, 200, 201]. Recirculation or dilution of the wastewater can be a way to ensure the right hydraulic condition in UASB reactors offering many advantages such as lowering the influent COD concentration, adjustment of the alkalinity requirement and dilution of inhibitory or toxic components. Reported liquid up-flow velocities and superficial biogas velocities in UASB reactors under steady-state conditions are in the range of 0.04 to 6 m h–1 [57, 138, 141, 200, 202] and 0.3 to 2 m h–1 [67, 203] respectively. The selection of the optimal liquid up-flow velocity should be based on the characteristics of the wastewater treated in the reactor and on the operating conditions. Surprisingly, the aforementioned velocities are significantly lower than the reported settling velocities of granules that typically lay around 20 m h–1 [23]. This difference could be attributed to the reduced density of the active granules in operating reactors due to the entrapment of biogas in the granules. Consequently, the active granules have much lower settling velocity compared to the resting inactive granules usually used in settling velocity tests and lower liquid velocity will be enough to wash them out of a UASB reactor. The increased fluid shear at higher vup should enhance the biomass detachment resulting in a lower mean granule diameter or even granule disintegration. However, contradictory results exist concerning the influence of the vup on the granulation process, probably due to the different experimental conditions [68, 116, 138, 204–209]. Kosaric et al. [68] have showed that low vup (0.25 and 0.5 m h–1) is favorable for mesophilic granule development, while at high vup (1.0 and 1.5 m h–1) disintegration of the granules may be observed. On the other hand, Guiot et al. [204] showed that increasing the vup increases the rate of the granulation process under mesophilic conditions and only linear flow over 4.3 m h–1 results in a decrease in the size of the granules. Arcand et al. [208] observed a strong effect of the vup on the granule size during experiments in mesophilic upflow anaerobic sludge bed and filter reactors, which were operated at 0.9, 2.2, 4.4, and 6.6 m h–1. In addition, they reported that the final size of the granular biofilm decreased with increasing specific substrate removal rate and loading rate (g-COD g-VSS–1 d–1). This was associated with the strong shear stress caused by the rising biogas bubbles whose production increased proportionally with the
56
I.V. Skiadas et al.
substrate removal rate. The shear caused by the biogas was much higher than the one caused by the vup in the reactor. The proportional increase of the granule size with the vup was explained by the lowering of substrate mass transfer resistance at the granule surface, thus allowing better substrate penetration and higher bacterial growth rate at the interior of the granules. In other words, a deeper substrate penetration favored the attached growth over cell detachment leading to an increase in the mean granule diameter. On the other hand, Hwu et al. [210] reported that the granule size in a thermophilic UASB reactor (mean granule diameter 1 mm) operating at 7.9 m h–1 was half of the granule size (mean granule diameter 2 mm) in a similar reactor operating at 1.3 m h–1. In addition, the reactor with the smaller granule size had a higher substrate conversion rate. These observations are explained by the lower structural strength of thermophilic granules compared to the mesophilic ones [84] and the better mass transfer in granules with smaller size. Winther-Nielsen [138] observed that the size of thermophilic granules was larger in a UASB reactor with high liquid velocity compared to the granules in another UASB reactor with low liquid velocity. However, the specific methanogenic activities were the same for the two granular sludges and there was no significant difference in the COD removal in the reactors. The influent substrate concentration has been shown to influence the granule development. Step-wise increase of propionate concentrations from 2.4 to 6.9 gCOD L–1 in the influent of mesophilic UASB reactors increased the size of granules [71]. Similarly, increasing the concentration of acetate in the influent resulted in faster granulation [211]. Tay and Yan [212] investigated the influence of substrate concentration on microbial selection and granulation during the start-up of UASB reactors. The reactors were fed with a synthetic wastewater of glucose, peptone, and meat extract and inoculated with anaerobic digested sludge. Large granules (mean diameter 2.5–3.4 mm) dominated by Methanosaeta-like species were developed in the reactors operated at 1 to 5 g-COD L–1. The performance of the reactors was stable up to a loading rate of 30–40 g-COD L–1 d–1 and/or a hydraulic retention time of 1–2 h. On the contrary, much smaller granules (mean diameter 0.54 mm) dominated by Methanosarcina-like species were cultivated in the reactors fed with substrate at a concentration of 10 g COD L–1. The operation of these reactors was unstable at loading rates of 10–20 g-COD L–1 d–1 due to the accumulation of floating granules into the settling zone. On the other hand, low substrate concentrations can result in substrate insufficiency, especially in the center of large granules. This can result in hollow granules due to biomass autolysis. Hollow granules are more susceptible to shear stresses within the reactor [213, 214] and floatation due to gas accumulation. The hydrophobic bacteria have a higher capacity to adhere to negatively charged surfaces than hydrophilic cells [215, 216]. Bacteria isolated from granular sludge were all very hydrophobic [217] indicating that a selection of hydrophobic bacteria occur during granulation. Bacteria become more hydrophobic at high dilution rates in chemostats and tend to form flocs or stick to surfaces in the culture vessel [218]. This means that the hydraulic retention time applied in a UASB reactor should be low to induce high hydrophobicity of the bacteria but not so low that all particles are washed out of the reactor.
Anaerobic Granular Sludge and Biofilm Reactors
57
The viscosity of liquids generally decreases with increasing temperature. When comparing two identical UASB reactors with the same granular sludge, approximately 1.5-times higher liquid up-flow velocity can be applied at thermophilic temperatures than at mesophilic. This is consistent with the observations that granules under thermophilic conditions are smaller than under mesophilic conditions [82, 101, 130, 219]. In general, it is believed that the biomass with poor settling characteristics is usually washed out of the UASB reactors due to the selection pressure of the liquid up-flow velocity and the biogas superficial velocity. However, using Stoke’s law and under the assumption that the density of granules is 1040–1080 kg m–3, one can conclude that only conglomerates with a diameter under 1.8–2.5 mm will be washed out of the reactor under a typical linear flow. This means that a conglomerate has to consist of only a few bacteria or more, if it should be retained in the reactor. However, as has already been mentioned the diameter of the anaerobic sludge granules varies from 0.14 to 5 mm. The above observations indicate that the liquid up-flow velocity and the biogas superficial velocity are not the only factors responsible for the selection of bacteria that can be immobilized.
4 Innovative Granular Sludge and/or Biofilm Reactors The anaerobic digestion is a sustainable technology for the treatment of waste and wastewater and offers many substantial benefits compared to conventional aerobic wastewater treatment methods [22]. Conventional anaerobic digestion is traditionally used for the treatment of sludges and agro-industrial wastewaters. During the last two decades, the development of high-rate anaerobic treatment in granular sludge and/or biofilm reactors has given the possibility of employing the high-rate anaerobic digestion for the treatment of a wide range industrial and municipal wastewaters (see Sects. 2 and 3). To date, high-rate anaerobic treatment is usually applied to easily biodegradable, low to medium strength wastewaters with low solid content under mesophilic conditions. However, the clear advantages of the high-rate anaerobic digestion over the conventional aerobic wastewater treatment methods create a strong interest in expanding the application of the high-rate anaerobic digestion for the treatment of almost every type of wastewater under any temperature conditions (thermophilic, mesophilic, or psycrophilic). Furthermore, there is a clear trend for the change of the role of the anaerobic digestion in the wastewater treatment plants from a pre-treatment method to the main biological treatment method. Specific alterations in process layout, reactor configuration, or operational conditions are required to enable the use of high-rate anaerobic wastewater treatment under these conditions and for this purpose. The application of staged anaerobic digesters has shown the larger potential among the recent developments in this direction [220, 221]. In the sequel, a short description of the most innovative staged anaerobic digesters will follow. The Up-flow Staged Sludge Bed (USSB) reactor [222] is a vertically oriented multistage UASB reactor (Fig. 5), where gas is removed from each separate compartment. The retention of biomass in a USSB reactor is significantly improved
58
I.V. Skiadas et al.
Biogas
Effluent
Water seal Sludge
Feedstock Hot tap water Fig. 5. The Up-flow Staged Sludge Bed (USSB) reactor [221]
due to the lower biogas load on the final gas-liquid separator and the consequent optimal settling conditions at the top of the reactor. The advantage of using the USSB reactor was clearly demonstrated by van Lier et al. [222] in lab-scale experiments under extreme loading conditions. The organic loading rate of a thermophilic USSB reactor fed with a sucrose-VFA mixture could be increased up to 100 kg-COD m–3 d–1 with a COD removal efficiency exceeding 90% at an HRT of 2–2.5 h. A typical sequence in the degradation of the substrate was found. In the bottom compartment sucrose was converted followed by the conversion of butyrate and acetate in the upper compartments. The propionate oxidation, which thermodynamically is the most unfavorable, took place only in the last top compartment. So far, the USSB process has shown to be particularly attractive for the treatment of wastewaters under thermophilic or psycrophilic conditions at very high organic loading rates and short hydraulic retention times. The high efficiency of the USSB process under extreme operational conditions can be attributed to the effective biomass retention, the higher substrate concentration in the first stages, and the minimization of the substrate and product inhibition effects in the upper stages [220]. A staged anaerobic reactor suitable for the high-rate treatment of wastewaters with substantial solid content is the Anaerobic Baffled Reactor (ABR). The
59
Anaerobic Granular Sludge and Biofilm Reactors
Biogas Influent
Effluent
Fig. 6. The Anaerobic Baffled Reactor
ABR [223] consists of a cascade of baffled compartments where the wastewater flows upward through a bed of anaerobic sludge after being transported to the bottom of the compartment (Fig. 6). The suitability of the ABR for the treatment of wastewaters containing suspended solids is due to this compartmentalized structure, i.e., phase separation takes place with the hydrolysis and acidogenesis in the upstream compartments and the methanogenesis in the downstream compartments. The ABR does not require sludge to granulate in order to perform effectively, although granulation does occur over time [223]. In any case, the formation of granules improves the performance of the ABR due to the resulting higher sludge retention time and better contact of the wastewater with the biomass. To date, the ABR has not been tested in full scale. However, experiments with lab-scale reactors have shown that the ABR is very stable under hydraulic and organic shock loads [224, 225]. Furthermore, the ABR has many potential advantages, i.e., high biomass retention times without a gas-solids-separator, simple design and mechanical simplicity resulting in low capital and operating costs. However, and in contrast to the UASB and USSB, the ABR is not suitable for the treatment of high strength wastewaters under long-term feeding conditions due to the biomass wash-out caused by high gas production [223]. An anaerobic digester, which combines the advantages of the UASB and the ABR, is the Periodic Anaerobic Baffled Reactor (PABR). The PABR [226, 227] consists of two concentric cylinders. The area between the cylinders is compartmentalized so that the reactor resembles an ABR with the compartments arranged in a circular manner in the annular region (Fig. 7a). The wastewater enters the digester at the down-flow section of the feeding compartment, comes up at the up-flow section of the same compartment and passes on to the next compartment through outer tubing. The flow pattern is repeated at the next compartments and the wastewater eventually leaves the system after passing through the up-flow part of the effluent compartment. The role of the compartments is periodically changed and everyone serves as influent or effluent by proper switching (on or off) of the valves of the outer tubing (Fig. 7b). In the extreme of zero switching frequency (no switching) the reactor behaves as an ABR. In the other extreme (infinite frequency) the compartments become identical so that the reactor should behave like a UASB (given that at high loading rates the hy-
60
a
I.V. Skiadas et al.
b
Fig. 7. a Front view of a four compartment PABR. b Top view of a four compartment PABR.
Up-flow, Down-flow
draulic behavior of UASB reactor approaches that of a CSTR system [22]). By setting the switching frequency, a great flexibility is obtained, taking advantage of the optimal reactor configuration (UASB, ABR or “something in between”) depending on the loading conditions. Simulations have shown that, for high organic loading, the PABR is expected to perform better when operated at a high frequency of switching feed point whereas for smaller values of the organic loading the PABR should be operated at smaller frequencies approaching the behavior of an ABR. Also, it has been shown that, depending on the loading rate, which in principle may be varied, it is possible that the switching frequency could be manipulated accordingly, allowing for the PABR to be operated as an ABR, a UASB, or at an intermediate mode. Consequently, the PABR is best suited for handling time-varying loading rates since it allows for maximal conversion rates at all times [226].
5 Conclusions High-rate anaerobic wastewater treatment technology is mainly based on static or particle-supported biofilms and granular systems. Typical biofilm systems are the fixed bed, the fluidized bed, and the expanded bed reactors and typical granular systems are the up-flow anaerobic sludge bed (UASB) and the expanded granular sludge bed (EGSB) reactors. Among them, the UASB reactor configuration is characterized by a) simplicity in construction and operation and b) reduced capital, operating, and maintenance cost. To date, the UASB is the most popular and widely used anaerobic wastewater treatment system.
Anaerobic Granular Sludge and Biofilm Reactors
61
The formation of granular sludge is not a prerequisite for a UASB reactor, which can also be operated with well settling flocculent sludge of high methanogenic activity at low wastewater flow rates. However, the formation and stability of the granules are essential for successful operation of the UASB at high wastewater flow rates. Granules are biomass conglomerates with a dense structure. After settling, these conglomerates present a “well-defined appearance”. Anaerobic granular sludge is characterized by high settling velocity, relatively high content in mineral precipitates and high methanogenic activity. An important ingredient of granular sludge is the extracellular polymers, which mainly consist of carbohydrates, proteins, and lipids. These polymers are usually negatively charged and interact with the divalent cations in the extracellular space. The produced polymer matrix and the mineral precipitations provide the necessary substratum for bacterial aggregation and granule development. The granulation process is mostly affected by the quality of the inoculum, the nutrients content and the cations concentration of the wastewater, the temperature and the hydraulic conditions in the reactor. The clear advantages of the high-rate anaerobic digestion have created a strong interest in changing the role of the anaerobic digestion in the wastewater treatment plants from a pre-treatment method to the main biological treatment method. The application of staged anaerobic digesters has shown the larger potential among the recent developments in this direction.
6 References 1. Characklis WG, Marshall KC (1990) Biofilms: a basis for an interdisciplinary approach. In: Characklis WG, Marshall KC (eds), Biofilms. Wiley, New York 2. Nicolella C, van Loosdrecht MM, Heijnen SJ (2000) Trends In Biotechnology 18:312 3. MacLeod FA, Guiot SR, Costerton JW (1990) Appl Environ Microbiol 56:1598 4. Young JC, Yang BS (1989) Journal of Water Pollution Control Federation 61:1576 5. Hall ER (1992) Anaerobic treatment of wastewaters in suspended growth and fixed film processes. In: Malina JF, Pohland FG (eds),Water Quality Management Library, vol 7. Design of anaerobic processes for the treatment of industrial and municipal wastes. Technomic Publishing Company, Lancaster, Pennsylvania, USA 6. Nicolella C, van Loosdrecht MM, Heijnen JJ (2000) J Biotechnol 80:1 7. Cooper PF, Atkinson B (1981) Proc Congress, Manchester, April 1980, Ellis Horwood, Chichester 8. Heijnen JJ, Mulder A, Enger W, Hoeks F (1989) The Chem Eng J 41:B37 9. Fatherpure BZ, Vogel TM (1991) Appl Environ Microbiol 57:3418 10. Ascon MA, Lebeault JM (1999) Appl Microbiol Biotechnol 52:592 11. Langwaldt JH, Puhakka JA (2000) Environ Pollut 107:187 12. Perez M, Romero LI, Sales D (2001) Anaerobe 7:25 13. FDZ-Polanco F, FDZ-Polanco M, Fernandez N, Uruena MA, Garcia PA,Villaverde S (2001) Water Res 35:1111 14. Borja R, Gonzalez E, Raposo F, Millan F, Martin A (2001) Bioresource Technology 76:45 15. Borja R, Banks CJ, Wang Z (1995) Bioresource Technology 52:157 16. Hawkes FR, Donnely T, Anderson GK (1995) Water Res 29:525 17. Mulder A, van de Graaf AA, Robertson LA, Kuenen JG (1995) FEMS Microbiol Ecol 16:177 18. Strous M, Heijnen JJ, Kuenen JG, Jetten MSM (1998) Appl Microbiol Biotechnol 50:589
62
I.V. Skiadas et al.
19. Jetten MSM, Strous M, van de Pas-Schoonen KT, Schalk J, van Dongen UGJM, van de Graaf AA, Logemann S, Muyzer G, van Loosdrecht MCM, Kuenen JG (1999) FEMS Microbiol Rev 22:421 20. van Dongen U, Strous M, van de Pas-Schoonen K, van Loosdrecht M, Kuenen JG, Jetten M (2000) Proceedings of the Fourth International Symposium on Environmental Technology. Noordwijkerhout, the Netherlands 42 21. Lettinga G, van Velsen FM, Hobma SW, de Zeeuw WJ, Klapwijk A (1980) Biotechnol Bioeng 22:699 22. Lettinga G (1995) Anton Leeuwenhoek Int J Gen M 67:3 23. Schmidt JE, Ahring BK (1996) Biotechnol Bioeng 49:229 24. Wambeke Mv, Grusenmeyer S, Verstraete W, Longry R (1990) Process Biochem Inter 181 25. Hulshoff Pol LW, Lettinga G (1986) Water Sci Technol 11:41 26. Moosbrugger RE, Wentzel MC, Ekama GA, Marais GvR (1993) Water SA 19:4152 27. Wu WM, Hickey RF, Zeikus JG (1991) Appl Environ Microbiol 57:3438 28. Goodwin JAS, Wase DAJ, Forster CF (1990) Biol Wastes 32:125 29. Dold PL, Sam-Soon A, Palmer IH, Marais GvR (1987) Water Sci Technol 19:237 30. Ross WR (1984) Water SA 4:197 31. Moosbrugger RE, Wentzel MC, Ekama GA, Marais GvR (1993) Water SA 19:53 32. Sayed S, van Campen L, Lettinga G (1987) Biol Wastes 21:11 33. Dangcong P, Qiting J (1993) J Chem Tech Biotechnol 58:89 34. Kennedy KJ, Andras E, Elliott CK, Methven B (1991) Environ Technol 12:291 35. Kudo A, Kennedy KJ, Andras E (1991) Water Sci Technol 23:1919 36. Rintala J, Sanz Martin JL, Lettinga G (1991) Water Sci Technol 24:149 37. Russo SL, Dold PL (1989) Water Sci Technol 21:121 38. Costa FJCB, Rocha BBM, Viana CE, Toledo AC (1986) Water Sci Technol 18:135 39. Souza ME, Fuzaro G, Polegato AR (1991) Thermophilic anaerobic digestion of vinasse in pilot plant UASB. In: Sixth international symposium on anaerobic digestion (paper preprint). IAWPRC, São Paulo 40. Boari G, Brunetti A, Passino R, Rozzi A (1984) Agric Wastes 10:161 41. Cail RG, Barford JP (1985) Agric Wastes 14:291 42. Craveiro AM, Soares HM, Schmidell W (1986) Water Sci Technol 18:123 43. Paris JM,Vicent T, Balaguer MD, Cassú C, Cairo J, Canals J (1988) Stability and control of UASB reactors treating potato-starch wastewater; comparison of laboratory and full-scale results. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW (eds), Granular anaerobic sludge; microbiology and technology. Puduc Wageningen, Wageningen, The Netherlands 44. García PA, Rico JL, FDZ-Polanco F (1991) Environ Technol 12:355 45. Yan JQ, Lo KV, Pinder KL (1993) Biotechnol Bioeng 41:700 46. Hendriksen HV, Larsen S, Ahring BK (1991) Water Sci Technol 24:431 47. Hendriksen HV, Ahring BK (1992) Appl Microbiol Biotechnol 37:662 48. Hendriksen HV, Ahring BK (1993) Biodegrad 3:399 49. Kennedy KJ, Andras E, Elliot DC, Methven B (1991) Can J Civ Eng 18:53 50. Kennedy KJ, Lu J, Mohn WW (1992) Water Res 26:1085 51. Mohn WW, Kennedy KJ (1992) Appl Environ Microbiol 58:2131 52. Nowak J, Janke D (1990) Reductive dechlorination of 2,6-dichlorophenol in an anaerobic bioreactor. In: Hutzinger O, Fiedler H (eds), Organohalogen compounds, vol. 1: DIOXIN ’90 – EPRI – Seminar toxicology, environmental, food, exposure, risk. Ecoinforma, Bayreuth 53. Thiele JH, Zeikus JG (1988) Biotechnol Bioeng 31:521 54. Woods SL, Ferguson JF, Benjamin MM (1989) Environ Sci Technol 23:62 55. Wu W, Bhatnagar L, Zeikus JG (1993) Appl Environ Microbiol 59:389 56. Lettinga G, Hulshoff Pol LW (1986) Water Sci Technol 99 57. Lettinga G, Hulshoff Pol LW (1991) Water Sci Technol 24:87 58. van Last ARM, Lettinga G (1991) Anaerobic treatment of domestic sewage under moderate climatic (Dutch) conditions using up-flow reactors at increased superficial
Anaerobic Granular Sludge and Biofilm Reactors
59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80.
81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93.
63
velocities. In: Sixth international symposium on anaerobic digestion (poster preprint). IAWPRC, São Paulo Schellinkhout A (1993) Water Sci Technol 27(9):173 Seghezzo L, Zeeman G, van Lier JB, Hamelers HM, Lettinga G (1998) Bioresource Technol 65:175 Kalogo Y, Verstraete W (1999) World J Microbiol Biotechnol 15:523 Brito AG, Melo LF (1997) Environ Technol 18:35 Kato MT, Field JA, Lettinga G (1997) Water Sci Technol 36:375 van Lier JB, Rebac S, Lettinga G (1997) Water Sci Technol 35:199 Dolfing J (1987) Microbiological aspects of granular methanogenic sludge. Ph.D. thesis, Agricultural University of Wageningen, Wageningen, The Netherlands ten Brummeler E, Hulshoff Pol LW, Dolfing J, Lettinga G, Zehnder AJB (1985) Appl Environ Microbiol 49:1472 Wiegant WM, de Man AWA (1986) Biotechnol Bioeng 28:718 Kosaric N, Blaszczyk R, Orphan L, Valladares J (1990) Water Res 24:1473 Schmidt JE, Ahring BK (1993) Enzyme Microb Technol 15:304 Koornneef E, Macario AJL, Grotenhuis JTC, Conway de Macario E (1990) Fems Microbiol Ecol 73:225 Grotenhuis JTC, Kissel JC, Plugge CM, Stams AJM, Zehnder AJB (1991) Water Res 25:21 Fukuzaki S, Nishio N, Sakurai H, Nagai S (1991) J Ferment Bioeng 71:50 Malik RK, Tauro P, Dahiya DS (1989) Biotechnol Bioeng 33:924 Hulshoff Pol LW, van de Worp JJM, Lettinga G, Beverloo WA (1986) Physical characterization of anaerobic granular sludge. In: Anaerobic Treatment. A Grown-Up Technology. RAI Halls, Amsterdam de Zeeuw WJ (1988) Granular sludge in UASB reactors. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW (eds), Granular Anaerobic Sludge; Microbiology and Technology. Puduc Wageningen, Wageningen, The Netherlands Hulshoff Pol LW (1989) The phenomenon of granulation of anaerobic sludge. Ph.D. thesis, Agricultural University of Wageningen, Wageningen, The Netherlands Fukuzaki S, Nishio N, Nagai S (1991) J Ferment Bioeng 72:405 Fukuzaki S, Chang Y, Nishio N, Nagai S (1991) J Ferment Bioeng 72:465 Dolfing J, Griffioen A, van Neerven ARW, Zevenhuizen LPTM (1985) Can J Microbiol 31:744 Beeftink HH, van den Heuvel JC (1988) Physical properties of bacterial aggregates in a continuous-flow reactor with biomass retention. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW (eds), Granular anaerobic sludge; microbiology and technology. Puduc Wageningen, Wageningen Andras E, Kennedy KJ, Richardson DA (1989) Environ Technol Lett 10:463 Ahring BK, Schmidt JE, Winther-Nielsen M, Macario AJL, Conway de Macario E (1993) Appl Environ Microbiol 59:2538 Alphenaar PA, Perez MC, van Berkel WJH, Lettinga G (1992) Appl Microbiol Biotechnol 36:795 Quarmby J, Forster CF (1995) Enzyme Microb Technol 17:493 Fukuzaki S, Nishio N, Nagai S (1995) J Ferment Bioeng 79:354 Van Langerak EA, Gonzalez-Gil G, Van Aelst A, Van Lier JB, Hamelers HM, Lettinga G (1998) Water Res 32:1255 Dolfing J (1986) Water Sci Technol 18:15 Schmidt JE, Ahring BK (1994) Appl Microbiol Biotechnol 42:457 Alibhai KRK, Forster CF (1988) Enzyme Microb Technol 8:601 Schmidt JE, Macario AJL, Ahring BK, Conway de Macario E (1992) Appl Environ Microbiol 58:862 Chang Y, Nishio N, Maruta H, Nagai S (1993) J Ferment Bioeng 75:430 Bhatti ZI, Furukawa K, Fujita M (1993) J Ferment Bioeng 76:218 Fukuzaki S, Chang YJ, Nishio N, Nagai S (1991) J Ferment Bioeng 72:465
64
I.V. Skiadas et al.
94. Hassan MA,Wase DJ (1996) Asia-Pacific Journal of Molecular Biology and Biotechnology 4:13 95. Shen CF, Kosaric N, Blaszczyk R (1993) Appl Microbiol Biotechnol 39:132 96. Bhatti ZI, Furukawa K, Fujita M (1995) J Ferment Bioeng 79:273 97. ElMamouni R, Guiot SR, Mercier P, Safi B, Samson R (1995) Bioprocess Engineering 12:47 98. Keenan PJ, Iza J, Switzenbaum MS (1993) Wat Environ Research 65:181 99. Costerton JW, Irving RT, Cheng KJ (1981) Annu Rev Microbiol 35:299 100. Costerton JW, Marrie TJ, Cheng KJ (1900) Phenomena of bacterial adhesion. In: Savage DC, Fletcher M (eds), Bacterial Adhesion. Mechanisms and Physiological Significance. Plenum Press, New York 101. Grotenhuis JTC, Smit M, Lammeren AAMv, Stams AJM, Zehnder AJB (1991) Appl Microbiol Biotechnol 36:115 102. Morgan JW, Forster CF, Evison L (1990) Water Res 24:743 103. Guiot SR, Lavoie L, Rocheleau S, MacLeod FA, Costerton JW (1991) Extracellular polymeric substances in anaerobic granules. In: Sixth international symposium on anaerobic digestion (Poster paper). IAWPRC, São Paulo 104. Shen CF, Kosaric N, Blaszczyk R (1993) Water Res 27:25 105. Veiga MC, Jain MK,Wu WM, Hollingsworth RI, Zeikus JG (1997) Appl Environ Microbiol 63:403 106. Forster CF (1991) J Biotechnol 17:221 107. Morgan JW, Evison LM, Forster CF (1991) Environ Technol 12:421 108. Robinson RW,Akin DE, Nordstedt RA, Thomas MV,Aldrich HC (1984) Appl Environ Microbiol 48:127 109. Forster CF (1971) Water Res 5:861 110. Quarmby J, Forster CF (1995) Water Res 29:2449 111. Schmidt JE, Ahring BK (1991) Appl Microbiol Biotechnol 35:681 112. Dolfing J (1985) Appl Microbiol Biotechnol 22:77 113. Arcand Y, Chavarie C, Guiot SR (1994) Water Sci Technol 30:63 114. Nilsson BK, Karlsson HT (1989) J Chem Technol Biotechnol 44:255 115. Kitsos HM, Roberts RS, Jones WJ, Tornabene TG (1992) Biotechnol Bioeng 39:1141 116. Alphenaar PA, Perez MC, Lettinga G (1993) Appl Microbiol Biotechnol 39:276 117. van den Heuvel JC, Verschuren PG, Ottengraf SP (1997) Water Sci Technol 36:311 118. Wu MM, Criddle CS, Hickey RF (1995) Biotechnol Bioeng 46:465 119. deBeer D, Huisman JW, vandenHeuvel JC, Ottengraf SP (1992) Water Res 26:1329 120. vanLier JB, Martin JS, Lettinga G (1996) Water Res 30:199 121. Schmidt JE, Ahring BK (1991) Appl Microbiol Biotechnol 35:681 122. Henze M, Harremoës P (1983) Water Sci Technol 15:1 123. Tramper j, van Groenestijn JW, Luyben KChAM, Hulshoff Pol LW (1984) Some physical and kinetic properties in granular anaerobic sludge. In: Houwink EH, van der Meer RR (eds), Innovations in Biotechnology. Elsevier, Amsterdam 124. Kato MT, Field JA, Versteeg P, Lettinga G (1994) Biotechnol Bioeng 44:469 125. Bryant MP, Wolin EA, Wolin MJ, Wolfe RS (1967) Arch Microbiol 59:20 126. Dolfing J, Mulder J-W (1985) Appl Environ Microbiol 49:1142 127. Gujer W, Zehnder AJB (1982) Water Sci Technol 15:127 128. Schmidt JE,Ahring BK (1992) Effect of hydrogen on the degradation of volatile fatty acids (VFA) in granules from UASB reactors. In: Leinonen S (ed), Anaerobic Treatment of Municipal Solid Waste and Sludge for Energy Production and Recycling of Nutrients. University of Joensuu, Joensuu 129. Schmidt JE, Ahring BK (1992) Propionate and butyrate degradation in granules from a thermophilic UASB reactor degrading acetate, propionate and butyrate. In: Abstracts of the 92nd general meeting of the American society of Microbiology, 1992. American Society of Microbiology, Washington 130. Schmidt JE, Ahring BK (1993) Appl Environ Microbiol 59:2546 131. Schmidt JE, Ahring BK (1995) Antonie van Leeuwenhoek 68:339
Anaerobic Granular Sludge and Biofilm Reactors
65
132. Pauss A, Samson R, Guiot SR (1990) Appl Microbiol Biotechnol 33:88 133. Schink B, Thauer RK (1988) Energetics of syntrophic methane formation and the influence of aggregation. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW (eds), Granular Anaerobic Sludge; Microbiology and Technology. Puduc Wageningen, Wageningen 134. van Lier JB, Grolle KCF, Frijters CTMJ, Stams AJM, Lettinga G (1993) Appl Environ Microbiol 59:1003 135. Winther-Nielsen M,Ahring BK (1988) Experiences from start-up of a thermophilic UASBreactor on lactate as sole carbon and energy source. In: Tilche A, Rozzi A (eds), Fifth International Symposium on Anaerobic Digestion. Monduzzi editore, Bologna 136. Winther-Nielsen M,Ahring BK (1989) Thermophilic degradation of butyrate, propionate and acetate in granular sludge. In: Microbiology and Biochemistry of Strict Anaerobic Involved in Interspecies Hydrogen Transfer. CNRS-ORSTOM, Marseille, France 137. Winther-Nielsen M, Ahring BK (1990) Degradation of volatile fatty acids in a thermophilic UASB-reactor. In: Christiansen C, Munck L,Villadsen J (eds), 5th European Congress on Biotechnology. Munksgaard international publisher, Copenhagen 138. Winther-Nielsen M (1991) Mechanisms of importance for the formation of granules in UASB reactors. Ph.D. thesis, The Technical University of Denmark, Lyngby, Denmark 139. Lin K, Yang Z (1991) Intern J Environ Studies 39:203 140. Shin H, Bae B, Oh S (1993) Biotechnol Lett 15:537 141. Souza ME (1986) Water Sci Technol 18:55 142. Teo KC, Xu HL, Tay JH (2000) J Environ Eng-Asce 126:411 143. Tay JH, Xu HL, Teo KC (2000) J Environ Eng-Asce 126:403 144. Lettinga G, de Zeeuw WJ, Hulshoff Pol LW, Wiegant WM, Rinzema A (1985) Anaerobic wastewater treatment based on biomass retention with emphasis on the UASB process. In: China State Biogas Association (eds),Anaerobic digestion 1985. China State Biogas Association, Guangzhou 145. Pette KC,Versprille AI (1981) Application of the UASB-concept for wastewater treatment. In: Hughes DE (ed), Anaerobic Digestion 1981. Elsevier, Amsterdam 146. Wu W, Hu J, Gu X, Zhao Y, Zhang H (1987) Water Res 21:789 147. Goodwin JAS, Wase DAJ, Forster CF (1992) Bioresour Technol 41:71 148. Uemura S, Harada H (1993) Environ Technol 14:897 149. van Lier JB, Grolle KCF, Stams AJM, Conway de Macario E, Lettinga G (1992) Appl Microbiol Biotechnol 37:130 150. Visser A, Gao Y, Lettinga G (1991) Anaerobic treatment of synthetic sulfate containing wastewater under thermophilic conditions. In: Sixth international symposium on anaerobic digestion (paper preprint). IAWPRC, São Paulo 151. Fang HP, Lau IC (1996) Water Sci Technol 34:445 152. Laurino CN, Siñeriz F (1991) Biotechnol Lett 13:299 153. Sørensen AH, Winther-Nielsen M, Ahring BK (1991) Appl Microbiol Biotechnol 34:823 154. Wiegant WM, Lettinga G (1985) Biotechnol Bioeng 27:1603 155. de Zeeuw WJ (1981) Acclimation of digested sewage sludge during start-up of an upflow anaerobic sludge blanket (UASB) reactor. In: Proceeding of the 35th industrial waste conference, 1980. Purdue University, Lafayette, Indiana. Ann Arbor Science, 156. de Zeeuw WJ (1984) Acclimatization of anaerobic sludge for UASB-reactor start-up. Ph.D. thesis, Agricultural University of Wageningen, Wageningen, The Netherlands 157. Hulshoff Pol LW, Zeeuw Wd,Velzeboer CTM, Lettinga G (1983) Water Sci Technol 15:291 158. Lettinga G, van Velsen AFM, Hobma SW, Zeeuw Wd, Klapwijk A (1980) Biotechnol Bioeng 22:699 159. Thaveesri J, Daffonchio D, Liessens B,Verstraete W (1995) Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology 68:329 160. Thaveesri J, Liessens B, Verstraete W (1995) Appl Microbiol Biotechnol 43:1122 161. Grootaerd H, Liessens B, Verstraete W (1997) Appl Microbiol Biotechnol 48:304 162. ElMamouni R, Leduc R, Guiot SR (1997) Appl Microbiol Biotechnol 47:189 163. Yu HQ, Tay JH, Fang HP (1999) Environmental Technology 20:1095
66 164. 165. 166. 167. 168. 169.
170.
171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201.
I.V. Skiadas et al. El-Mamouni R, Leduc R, Guiot SR (1998) Water Sci Technol 38:341 van Lier JB, Rintala J, Sanz Martin JL, Lettinga G (1990) Water Sci Technol 22(9):183 Sam-Soon PALNS, Loewenthal RE, Wentzel MC, Marais GvR (1990) Water SA 16:165 Oleszkiewicz JA, Sharma VK (1990) Biol Wastes 31:45 Goodwin JAS, Wase DAJ, Forster CF (1990) Enzyme Microb Technol 12:877 Guiot SR, Gorur SS, Kennedy KJ (1988) Nutritional and environmental factors contributing to microbial aggregation during upflow anaerobic sludge bed-filter (UBF) reactor start-up. In: Hall ER, Hobson PN (eds),Anaerobic Digestion 1988. Pergamon Press, Oxford Guiot SR, Gorur SS, Bourque D, Samson R (1988) Metal effect on microbial aggregation during upflow anaerobic sludge bed-filter (UBF) reactor start-up. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW (eds), Granular Anaerobic Sludge; Microbiology and Technology. Puduc Wageningen, Wageningen Ilangovan K, Noyola A (1993) Environ Technol 14:795 Lauwers AM, Gorris LG, van der Drift C (1989) Biotechnol Lett 11:907 Oleszkiewicz JA (1989) Biol Wastes 27:217 Callander IJ, Clark TA, McFarlane PN (1987) Biotechnol Bioeng 30:896 Singh RP, Kumar S, Ojha CP (1999) Biochem Eng J 3:35 Punal A, Trevisan M, Rozzi A, Lema JM (2000) Water Res 34:2614 Sam-Soon Palns, Loewenthal RE, Marais GvR (1987) Water SA 13:69 Sam-Soon Palns, Loewenthal RE, Dold PL, Marais GvR (1988) Pelletization in upflow anaerobic sludge bed reactors. In: Hall ER, Hobson PN (eds), Anaerobic Digestion 1988. Pergamon Press, Oxford Lau IC, Fang HP (1997) Water Res 31:2626 van Lier JB, Martin JS, Lettinga G (1996) Water Res 30:199 Brunetti A, Boari G, Passino R, Rozzi A (1983) Physico-chemical factors affecting startup in UASB digestors. In: AWWT (eds), Anaerobic Waste Water Treatment. AWWT Symposium Secretariat, The Hague, The Netherlands Kosaric N, Mahoney EM, Varangu LK, Carins WL (1987) Wat Pollut Res J CAn 22:289 Mahoney EM, Varangu LK, Cairns WL, Kosaric N, Murray RGE (1987) Water Sci Technol 19(1/2):249 Patel GB, Roth LA, Sprott GD (1979) J Gen Microbiol 112:411 Fukuzaki S, Nishio N, Nagai S (1995) J Ferment Bioeng 79:354 Jia XS, Fang HP, Furumai H (1996) Water Sci Technol 34:309 Schmidt JE, Ahring BK (1991) Effect of magnesium on thermophilic acetate-degrading granules in UASB reactors. In: Anaerobic Microbiology and Anaerobic Biotechnolgical Processes. DTH, Lyngby Faguy DM, Koval SF, Jarrell KF (1993) Arch Microbiol 159:512 Hendriksen HV, Ahring BK (1991) FEMS Microbiol Ecol 85:241 Alibhai KRK, Forster CF (1986) Environ Technol Lett 7:193 Habets LHA, Knelissen JH (1985) Water Sci Technol 17:61 Keenan PJ, IZA J, Switzenbaum MS (1993) Water Environ Res 65:181 Rinzema A, van Lier J, Lettinga G (1988) Enzyme Microb Technol 10:24 Cail RG, Barford JP (1985) Biotechnol Lett 7:493 Koster IW, Rinzema A, de Vegt AL, Lettinga G (1986) Water Res 20:1561 Mendez R, ten Brummeler E, Hulshoff Pol LW (1989) Environ Technol Lett 10:83 Visser A, Nozhevnikova AN, Lettinga G (1993) J Chem Tech Biotechnol 57:9 Visser A, Alphenaar PA, Gao Y, van Rossum G, Lettinga G (1993) Appl Microbiol Biotechnol 40:575 Hulshoff Pol LW, Lens PNL, Stams AJM, Lettinga G (1998) Biodegradation 9:213 Samson R, Pauss A, Guiot SR (1990) Immobilized cell systems in anaerobic digestion processes. In: Tyagi RD,Vembu K (eds), Wastewater treatment by immobilized cells. CRC Press, Boca Raton Wiegant WM (1988) The “spaghetti theory” on anaerobic sludge formation, or the inevitability of granulation. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW
Anaerobic Granular Sludge and Biofilm Reactors
202. 203. 204.
205.
206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227.
67
(eds), Granular anaerobic sludge; microbiology and technology. Puduc Wageningen, Wageningen Wiegant WM (1986) Thermophilic anaerobic digestion for waste and wastewater treatment. Ph.D. thesis, Agricultural University of Wageningen, Wageningen Pereboom JF (1997) Water Sci Technol 36:141 Guiot SR, Pauss A, Bourque D, El Housseini M, Lavoie L, Beaulieu C, Samson R (1988) Effect of upflow liquid velocity on granule size distribution in an upflow anaerobic bedfilter (UBF) reactor. In: Tilche A, Rozzi A (eds), Fifth international symposium on anaerobic digestion. Monduzzi Editore, Bologna Hulshoff Pol LW, Heijnekamp K, Lettinga G (1988) The selection pressure as a driving force behind the granulation of anaerobic sludge. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW (eds), Granular anaerobic sludge; microbiology and technology. Puduc Wagenigen, Wageningen Shin HS, Paik BC (1990) Biotechnol Lett 12:469 Peyton BM, Characklis WG (1993) Biotechnol Bioeng 41:728 Arcand Y, Guiot SR, Desrochers M, Chavarie C (1994) Chem Eng J Biochem Eng J 56:B23 O’Flaherty V, Lens PL, deBeer D, Colleran E (1997) Appl Microbiol Biotechnol 47:102 Hwu CS, Molenaar G, Garthoff J, vanLier JB, Lettinga G (1997) Biotechnology Lett 19:447 Morvái L, Miháltz P, Czakó L, Péterfy M, Holló J (1990) Appl Microbiol Biotechnol 33:463 Tay JH, Yan YG (1996) Water Environ Res 68:1140 Beeftink HH, Staugaard P (1986) Appl Environ Microbiol 52:1139 Kosaric N, Blaszczyk R, Orphan L (1990) Water Sci Technol 22(9):275 van Loosdrecht MCM, Lyklema J, Norde W, Schraa G, Zehnder AJB (1987) Appl Environ Microbiol 53:1893 van Loosdrecht MCM, Lyklema J, Norde W, Zehnder AJB (1990) Microbiol Rev 54:75 Grotenhuis JTC, Plugge CM, Stams AJM, Zehnder AJB (1992) Appl Environ Microbiol 58:1054 van Loosdrecht MCM, Lyklema J, Norde W, Schraa G, Zehnder AJB (1987) Appl Environ Microbiol 53:1898 Schmidt JE, Ahring BK (1994) Unpublished work van Lier JB, van der Zee FP, Tan NCG, Rebac S, Kleerebezem R (2001) Water Sci Technol 44:15 Lettinga G, Field J, van Lier JB, Zeeman G, Pol LWH (1997) Water Sci Technol 35:5 van Lier JB, Boersma F, Debets MMWH, Lettinga G (1994) “High rate” thermophilic anaerobic wastewater treatment in compartmentalized upflow reactors. In: IAWQ (ed), Anaerobic Digestion, 1994. RSA Litho (Pty) Ltd, Goodwood Barber WP, Stuckey DC (1999) Water Res 33:1559 Nachaiyasit S, Stuckey DC (1997) Water Res 31:2737 Nachaiyasit S, Stuckey DC (1997) Water Res 31:2747 Skiadas IV, Gavala HN, Lyberatos G (2000) Water Res 34:3725 Skiadas IV, Lyberatos G (1998) Water Sci Technol 38:401
Received: April 2002
CHAPTER 1
Potential for Anaerobic Conversion of Xenobiotics A. S. Mogensen 1 · J. Dolfing 2 · F. Haagensen 1 · B. K. Ahring 1 1
2
BioCentrum-DTU, Building 227, The Technical University of Denmark, 2800 Lyngby, Denmark E-mail:
[email protected] E-mail:
[email protected] E-mail:
[email protected] Alterra, PO Box 47, 6700 AA Wageningen, The Netherlands
This review covers the latest research on the anaerobic biodegradation of aromatic xenobiotic compounds, with emphasis on surfactants, polycyclic aromatic hydrocarbons, phthalate esters, polychlorinated biphenyls, halogenated phenols, and pesticides. The versatility of anaerobic reactor systems regarding the treatment of xenobiotics is shown with the focus on the UASB reactor, but the applicability of other reactor designs for treatment of hazardous waste is also included. Bioaugmentation has proved to be a viable technique to enhance a specific activity in anaerobic reactors and recent research on reactor and in situ bioaugmentation is reported. Keywords. Anaerobic, Nitrate, Iron, Sulfate, Xenobiotics, Organic contaminants, Biodegrada-
tion, Mineralization, Surfactants, Polycyclic aromatic hydrocarbons, Phthalate esters, Polychlorinated biphenyls, Halogenated phenols, Pesticides, UASB, CSTR, Bioreactor, Bioaugmentation
. . . . . . . . . . . . . . . . . . . 72
1
Adverse Effects of Xenobiotics
1.1 1.2 1.3
Problems Related to Xenobiotics . . . . . . . . . . . . . . . . . . 72 Recalcitrance under Anaerobic Conditions . . . . . . . . . . . . . 73 Strategies for Removal of Xenobiotics . . . . . . . . . . . . . . . 74
2
Anaerobic Conversion of Aromatic Xenobiotics . . . . . . . . . . 75
2.1 2.1.1 2.1.1.1 2.1.1.2 2.1.1.3 2.1.2 2.1.2.1 2.1.2.2 2.2 2.2.1 2.2.1.1 2.2.1.2 2.2.1.3 2.2.2
Common Features of Aromatic Ring Biodegradation Factors Influencing Microbial Degradation . . . . . Number of Microorganisms . . . . . . . . . . . . . Microbial Growth Factors . . . . . . . . . . . . . . Bioavailability of Xenobiotics . . . . . . . . . . . . Aromatic Ring Cleavage . . . . . . . . . . . . . . . Peripheral Pathway . . . . . . . . . . . . . . . . . . The Benzoyl-CoA Pathway . . . . . . . . . . . . . . Anaerobic Conversion of Surfactants . . . . . . . . Surfactant Properties and Molecular Structure . . . Anionic Surfactants . . . . . . . . . . . . . . . . . Cationic Surfactants . . . . . . . . . . . . . . . . . Non-Ionic Surfactants . . . . . . . . . . . . . . . . Common Disposal Route . . . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
75 75 76 76 76 80 80 81 83 83 84 84 85 85
Advances in Biochemical Engineering/ Biotechnology, Vol. 82 Series Editor: T. Scheper © Springer-Verlag Berlin Heidelberg 2003
70 2.2.3 2.2.3.1 2.2.3.2 2.2.3.3 2.3 2.3.1 2.3.2 2.3.3 2.3.3.1 2.3.3.2 2.3.3.3 2.3.3.4 2.4 2.4.1 2.4.2 2.4.3 2.4.3.1 2.4.3.2 2.4.3.3 2.5 2.5.1 2.5.1.1 2.5.1.2 2.5.2 2.5.2.1 2.5.3 2.5.3.1 2.5.3.2 2.5.3.3 2.5.4 2.5.5 2.6 2.6.1 2.6.2 2.6.3 2.6.3.1 2.6.3.2 2.6.3.3 2.6.4 2.6.5 2.6.6 2.6.7 2.7 2.7.1 2.7.2 2.7.3 2.7.4
A.S. Mogensen et al.
Anaerobic Degradation of Surfactants . . . . . . . . . . . . Degradation of Anionic Surfactants . . . . . . . . . . . . . . Degradation of Non-Ionic Surfactants . . . . . . . . . . . . Degradation of Cationic Surfactants . . . . . . . . . . . . . Anaerobic Conversion of Polycyclic Aromatic Hydrocarbons Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . Environmental Contamination . . . . . . . . . . . . . . . . Anaerobic Biodegradation of PAHs . . . . . . . . . . . . . . Nitrate-Reducing Conditions . . . . . . . . . . . . . . . . . Fe(III)-Reducing Conditions . . . . . . . . . . . . . . . . . Sulfate-Reducing Conditions . . . . . . . . . . . . . . . . . Methanogenic Conditions . . . . . . . . . . . . . . . . . . . Anaerobic Conversion of Phthalate Esters . . . . . . . . . . Properties and Molecular Structure . . . . . . . . . . . . . . Common Disposal Route and Environmental Concentrations Anaerobic Degradation of Phthalate Esters . . . . . . . . . . Degradability . . . . . . . . . . . . . . . . . . . . . . . . . . Factors Affecting Degradability . . . . . . . . . . . . . . . . Degradation Pathway . . . . . . . . . . . . . . . . . . . . . Anaerobic Conversion of Polychlorinated Biphenyls . . . . . Physico-chemical Characteristics of PCBs . . . . . . . . . . PCB Classification . . . . . . . . . . . . . . . . . . . . . . . Prevalence of PCBs in the Environment . . . . . . . . . . . Anaerobic Degradation by Reductive Dechlorination . . . . Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . Priming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of FeSO4 . . . . . . . . . . . . . . . . . . . . . . . . . Intrinsic Priming . . . . . . . . . . . . . . . . . . . . . . . . Effect of Redox Conditions . . . . . . . . . . . . . . . . . . PCB-Dechlorinating Bacteria . . . . . . . . . . . . . . . . . Field Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . Halogenated Phenols . . . . . . . . . . . . . . . . . . . . . . Hydrophobicity of Chlorinated Phenols . . . . . . . . . . . Anaerobic Biodegradation of Halogenated Phenols . . . . . Redox Conditions . . . . . . . . . . . . . . . . . . . . . . . Nitrate-Reducing Conditions . . . . . . . . . . . . . . . . . Iron-Reducing Conditions . . . . . . . . . . . . . . . . . . . Sulfate-Reducing Conditions . . . . . . . . . . . . . . . . . Pure Culture Studies . . . . . . . . . . . . . . . . . . . . . . Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effect of Co-Substrates . . . . . . . . . . . . . . . . . . . . . Field Studies . . . . . . . . . . . . . . . . . . . . . . . . . . Anaerobic Degradation of Pesticides . . . . . . . . . . . . . Biodegradation Reactions . . . . . . . . . . . . . . . . . . . NSO Compounds . . . . . . . . . . . . . . . . . . . . . . . . Chlorinated Pesticides . . . . . . . . . . . . . . . . . . . . . Pesticides in Co-Digesters . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
87 87 90 91 92 93 93 95 95 96 97 98 99 99 100 100 100 101 101 103 103 104 104 105 105 106 107 107 108 108 109 109 110 110 112 112 113 113 113 114 115 115 116 117 117 119 120
71
Potential for Anaerobic Conversion of Xenobiotics
3
Treatment of Xenobiotics in Bioreactors . . . . . . . . . . . . . . 121
3.1 3.2 3.3 3.4
The UASB Reactor . . . . . . . . . . . . . . . . Continuous Stirred-Tank Reactor (CSTR) . . . Other Anaerobic Reactors . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . .
4
Bioaugmentation
4.1 4.2
In Situ Bioaugmentation . . . . . . . . . . . . . . . . . . . . . . . 126 Bioaugmentation in Reactors . . . . . . . . . . . . . . . . . . . . 127
5
Future Perspectives
6
References
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
122 123 124 125
. . . . . . . . . . . . . . . . . . . . . . . . . . 126
. . . . . . . . . . . . . . . . . . . . . . . . . 128
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Abbreviations AP APnEO ATP BBP C12-LAS COD CP CSTR DBP DCE DCP DEP DEHP DMP DOP GAC HRT LAS MEHP MSW NP NPnEO PA PAH PBB PE PCB PCP POE PVC
. . . .
Alkylphenol Alkylphenol polyethoxylate Adenosine triphosphate Benzyl butyl phthalate Linear dodecylbenzenesulfonate Chemical oxygen demand Chlorophenol Continuous stirred-tank reactor Dibutyl phthalate Dichloroethene Dichlorophenol Diethyl phthalate Diethylhexyl phthalate Dimethyl phthalate Di-n-octyl phthalate Granular activated carbon Hydraulic retention time Linear alkyl benzenesulfonates Monoethylhexyl phthalate Municipal solid waste Nonylphenol Nonylphenol polyethoxylates Phthalic acid Polycyclic aromatic hydrocarbon Polybrominated biphenyl Phthalate ester Polychlorinated biphenyls Pentachlorophenol Polyoxyethylene octylphenyl ether Polyvinyl chloride
72 QACs SAS TCE TCP UASB VFA WWTP
A.S. Mogensen et al.
Quaternary ammonium compounds Secondary alkanesulfonate Trichloroethene Trichlorophenol Upflow anaerobic sludge blanket Volatile fatty acid Wastewater treatment plant
1 Adverse Effects of Xenobiotics Chemicals with adverse effects have accompanied the life of human beings throughout history. Every combustion process is normally incomplete, leading to the formation of, for instance, polycyclic aromatic hydrocarbons, which are known to be both toxic and carcinogenic. Moreover, halogenated phenols can even be created by fungi, worms and ticks. In the last century, a vast number of chemicals has been developed for use in industry and households. The production of some has ceased due to ban or production of new and better formulations. However, at the beginning of the 21st century, still a large number of chemicals is produced and utilized worldwide despite the knowledge of their adverse effects in nature. These can be termed contaminants, or xenobiotics, defined as manmade chemicals which, when released to the environment, show adverse effects on the exposed organisms. The amount of xenobiotics is extensive and refers to chemicals with very diverse physical and chemical properties, and thus with different impact in the environment.A common property of many xenobiotics is that they do not show adverse effects on biological life until the point when they are disposed or when the object in which they have been used is disposed (though exceptions are easily found). At the point of release into the environment the lifetime of a xenobiotic compound varies greatly from a few seconds to several years or decades depending on the susceptibility to chemical and biological reactions. In addition, physical parameters such as, e.g., temperature and radiation have an important impact on the reactivity of the xenobiotic compound. Many factors can explain the persistence of xenobiotics in the environment; their recalcitrance is due to limited availability to the microbial cell by sorption or insolubility and to lack of suitable transport enzymes (permeases). Inadequate physical-chemical conditions will also impede bacterial activity, i.e., absence of proper electron acceptors, unfavorable temperature, light, pH, or redox potential. Growth limitation may also be related to inadequate transfer of nutrients, or formation of a toxic metabolite. 1.1 Problems Related to Xenobiotics
A release of a xenobiotic compound to the environment effects biological life, and during its lifetime, various organisms will be exposed to the compound. In order
Potential for Anaerobic Conversion of Xenobiotics
73
to perceive these one must consider transport, persistence, and bioavailability of the compound. With certain chemical and physical properties of a persistent chemical, even a point source release can result in a worldwide distribution, and hence a local environmental problem suddenly becomes global. On the other hand, a release of a xenobiotic to, e.g., soil might cause adverse effect on soil organisms only during a limited period of time due to a decreased bioavailability of the chemical by irreversible sorption. In such cases, the xenobiotic is present in the soil matrix but is not bioavailable, and thus no adverse effects will be noted. These aspects are carefully described by Alexander [1] and Wania and Mackay [2], and will not be further discussed here. Xenobiotic compounds are, as mentioned above, subjected to chemical, physical, and biological processes that alter the molecule with eventual elimination from the environmental compartments. Compounds that are resistant to chemical and physical transformation can be altered by microbial attack and a sequence of degradation steps will eventually lead to mineralization of the molecule. Common for any discharge of xenobiotics is the eventual contact between the chemical and microorganisms. Degradation of xenobiotics by aerobic microorganisms has been extensively studied, but anaerobic bacteria may be more important for the mineralization of especially those xenobiotics low in volatility. Such compounds often end up in environmental compartments that are low in oxygen, and the biological degradation is thus carried out by, e.g., bacteria that do not require oxygen in their metabolism of xenobiotics. Recently, research on anaerobic degradation of xenobiotics has shown a high tolerance of anaerobic bacteria to these compounds. Furthermore, anaerobic treatment processes such as the upflow anaerobic sludge blanket reactor and the continuous stirred-tank reactor have proved to be very effective in the treatment of industrial effluents high in the content of xenobiotics. 1.2 Recalcitrance under Anaerobic Conditions
Anaerobic degradation of organic contaminants is perceived as a poorly researched area and new findings tend to change the outlook. Recently, Alexander [1] listed a series of compounds (see Table 1) that had been reported to be resistant to anaerobic degradation [3]. He carefully pointed out that inclusion in the list did not necessarily imply that these compounds were de facto not biodegradable under anaerobic conditions. Indeed, it is presently possible to point at studies showing that these compounds are biodegradable under anaerobic conditions. However, this does not imply that these compounds are always degraded under anaerobic conditions, but it does imply that they can be degraded if the appropriate organisms are present in sufficient numbers and their requirements for growth are fulfilled (see Sect. 2.1.1). The main implication of this contrast opens the perspective of developing bioremediation techniques that lead to the destruction of otherwise long-lived pollutants. The second edition [1] indeed gives much more attention to the degradation of compounds that were until recently perceived as recalcitrant under anaerobic conditions. Comprehensive insight in this subject and its
74
A.S. Mogensen et al.
Table 1. Examples of compounds that were in 1998 in an authoritative oversight still perceived
as not biodegraded under anaerobic conditions, but have since then been shown to be transformed in the presence of at least one electron acceptor other than molecular oxygen Compound
Electron acceptor
Reference
Naphthalene
Sulfate Nitrate
[4, 5] [6]
2-Chlorobenzoate 3,3¢-Dichlorobenzidine Pyridine
Bicarbonate Bicarbonate Nitrate
[7] [8] [9]
Long-chain alkanes
Sulfate Bicarbonate
[10] [11]
Benzene
Sulfate Iron, bicarbonate Iron Sulfate
[12] [13] [14] [15]
environmental implications with special emphasis on, e.g., pesticides is however still lacking. 1.3 Strategies for Removal of Xenobiotics
Because of the vast use of organic contaminants in modern society, almost any waste from industrial processes or households contains these compounds, and disposal without proper treatment will therefore result in exposure to the environment. Once released, the removal is difficult, if not impossible, depending on the polluted area and the characteristics of the xenobiotics. Therefore, environmental contamination should be avoided by impeding the disposal of xenobiotics, an objective that can be achieved by declining the usage or by removing them from the waste, using, e.g., biological methods. Knowledge of waste composition regarding the presence of xenobiotics is therefore of primary importance. The chemical and physical characteristics of the xenobiotics combined with their ecotoxicological properties should be used to determine what kind of treatment process should be applied for an adequate waste treatment. In industrialized communities, solid and liquid waste effluents from households are collected and disposed with or without treatment. Treatment of wastewater from households prior to wastewater discharge may remove the xenobiotics to a certain extent by biological or physical processes. Such controlled processes facilitate the removal of xenobiotics from waste, and with increased focus of removal of xenobiotics, their discharge into the environment can be reduced markedly. The wastewater from any larger city contains a wide range of xenobiotics as, for example, PAHs, phthalate esters, and surfactants.According to the chemical and physical properties of the compounds within these three groups, different behavior during treatment can be observed. The main part of
Potential for Anaerobic Conversion of Xenobiotics
75
the hydrophobic organic contaminants (e.g., higher ring PAHs) is transferred to the sewage sludge in the primary settler and thus escapes the subsequent steps in the treatment facility. More hydrophilic compounds (e.g., lower ring PAHs) are to a larger extent found in the water phase and are thus subjected to the biological treatment. Hydrophobic xenobiotics discharged via household or industrial effluents that escape conventional treatment are eventually discharged into the environment by sludge application on farmland or landfilling. This disposal may be avoided by anaerobic sludge digestion that focuses on biotransformation of organic xenobiotics. Xenobiotics released to the environment by point sources may be removed by in situ bioremediation techniques, i.e., treatment of polluted soil and groundwater by amendment of specific microorganisms in order to enhance the biodegradability of the contaminants. Polluted soil and groundwater are otherwise removed from the site and subjected to remediation processes of either biological or physical/chemical nature, thus being more costly than in situ bioremediation.
2 Anaerobic Conversion of Aromatic Xenobiotics The anaerobic conversion of six different groups of aromatic xenobiotics, i.e., surfactants, phthalate esters, PAHs, PCBs, halogenated phenols, and pesticides is reported below. Aromatic xenobiotics share the aromatic ring as subunit in their molecular structure, and similarities regarding their chemical and physical properties and their potential for bioconversion, i.e., ring cleavage, exist. Obviously, molecular alterations of the aromatic ring, e.g., complex ring structures in polycyclic aromatic hydrocarbons or substitution with alkyl, carboxyl groups, and halogens, change the behavior of the xenobiotic compound in the environment. The microbiological attack on such compounds leads in most cases, however, to the formation of less substituted molecules with eventual cleavage of the aromatic ring. Factors that influence microbial degradation and the biochemistry of aromatic ring cleavage are the subjects of this chapter. 2.1 Common Features of Aromatic Ring Biodegradation 2.1.1 Factors Influencing Microbial Degradation
Biodegradation of xenobiotics is a function of characteristics of the particular system where the biological activity occurs; the laboratory study accomplishes the control of important parameters such as cell number, substrate and nutrient availability – which are almost impossible to control in ecosystems. The microbial degradation is regulated by the number of microorganisms capable of metabolizing the organic contaminant, the accomplishment of growth factors for these organisms (determined by temperature, pH, electron acceptors, nutrients
76
A.S. Mogensen et al.
and water content in the medium) and the bioavailability of the compound. Some parameters that are important for biodegradation of xenobiotics are discussed below. 2.1.1.1 Number of Microorganisms
When studying the microbial degradation of any organic molecule, it is obvious that the number of microorganisms capable of metabolizing the molecule should be of such a magnitude that the molecular transformations occur at rates relevant for the period of the experiment. Therefore, it is of major importance to assure the appropriate number of relevant microorganisms when assessing the biodegradability of xenobiotics. 2.1.1.2 Microbial Growth Factors
Several factors have been identified to be important for the extent and rate of biodegradation. The most important factors are listed below: – – – – –
Energy and carbon source, Temperature, pH, Micro- and macronutrients, Inhibition (substrate, product or specific).
Microbial degradation of xenobiotic compounds requires, as any microbial activity, the presence of an energy and carbon source to support growth. The degradation is carried out by direct or indirect involvement of microorganisms. The direct involvement is an enzymatic process with or without energy gain by the microorganism. The indirect involvement requires the presence of an electron acceptor in the oxidation of organic material (e.g., a xenobiotic). Temperature and pH are both major environmental factors controlling microbial growth, and microorganisms differ greatly in their specific temperature and pH-optimum. Another key factor influencing microbial growth and thus biodegradation of xenobiotics is the presence and availability of various microand macronutrients needed to support growth. Furthermore, microbial growth is influenced by inhibition – either partly or totally – of the substrate (the organic contaminant), a product formed due to biodegradation, or of another inhibitory compound. 2.1.1.3 Bioavailability of Xenobiotics
Biodegradation of a xenobiotic compound does not only require an appropriate number of microorganisms and convenient growth conditions. The compound also has to be available to the microorganisms, which is often a key problem in, e.g., bioremediation of hydrophobic chemicals. The bioavailability depends on:
Potential for Anaerobic Conversion of Xenobiotics
– – – –
77
Mass transfer rates, Dissolution rates, Chemical fate, Enzymatic processes.
It has often been neglected that the analytical depletion of aromatic compounds in soil was not only due to biodegradation but also caused by strong adsorption to the soil matrix, where it may become completely non-extractable, see Fig. 1. With time, most xenobiotics become more resistant to extraction as they are entrapped in macromolecular humus substances [16]. Also, aging and wetting/drying cycles of PAH-contaminated soil decrease the extent and the rate of mineralization [17], but differences among soils exists. Only few studies have been conducted with other xenobiotics than PAHs, but likely the same pattern is valid for chemicals with similar physico-chemical properties. In order to enhance biodegradation of the desorption-resistant fraction of xenobiotics, different approaches such as addition of surfactants, mixed bacterial cultures, and solvents have proven successful [18], see Fig. 2. Whether this entrapment is relevant in other solid media than soil, such as sludge, is still poorly examined. Biodegradation of sorbed contaminants can be a function of the mass transfer (from adsorbed to the aqueous phase) rates rather than the biodegradation rates [19], see Table 2.
Fig. 1. Fractionation of the 14C activity within a soil sample amended with radiolabeled an-
thracene and pyrene. Reproduced in part with kind permission of [16]. Copyright (1998) American Chemical Society
78
A.S. Mogensen et al.
% Mineralized
METHANOL
40
NO METHANOL
20
0 0
3
6
9
Days
NL-W
% Mineralized
27
18
9
RB-P
0 0
9
18
Days Fig. 2. Enhancement of biodegradation by solvent addition and effect of mixed culture. From [18]. Reprinted with kind permission from SETAC
In such cases, the concentration of the chemical in the water phase is often below the detection limit, and the fraction sorbed on organic matter in, e.g., soil, will theoretically desorp until only irreversibly sorbed molecules remain. It is important to note that the extractable fraction of xenobiotics in e.g., soil, can be much lower than the fraction available to microorganisms by desorption. Sorption of organic contaminants, which have higher solubility in water than PAHs, can also occur. In an anaerobic digester treating secondary stabilized sludge, the
79
Potential for Anaerobic Conversion of Xenobiotics
Table 2. Comparison of phenanthrene transfer rate constants obtained from experiments of biodegradation by strain S Phe Na 1 and abiotic transfer rates. From [19]. Reproduced with kind permission from Springer-Verlag GmbH & Co. Copyright (1995) Springer-Verlag GmbH & Co
Mode of substrate supply
Crystals Solution in silicone oil Solution in HMN
Phenanthrene transfer rate constants (h–1) in experiments on: Biodegradation
Abiotic transfer
0.018 0.084 0.015
0.030 0.076 0.014
amount of linear dodecylbenzenesulfonate (C12-LAS) in the water phase was more than 10-fold lower than the amount than could be extracted from a mixed sample with methanol (Fig. 3). Prior to any biodegradation experiment, whether in situ or ex situ, a study of the fate of the compound of interest will allow a more precise experimental set-up. The use of physico-chemical data such as water solubility, octanol-water partition coefficient, and Henry’s law constant helps evaluating in qualitative terms the environmental fate of a given xenobiotic compound. It has not been established if solubilization is required for attached organisms, i.e., using ad-
Fig. 3. The concentration of C12-LAS in an anaerobic digester continuously fed with secondary
stabilized sludge spiked with C12-LAS [20]
80
A.S. Mogensen et al.
sorbed contaminants, which do not solubilize prior to microbial degradation. 2.1.2 Aromatic Ring Cleavage
A large part of xenobiotics discharged into the environment contains on or more aromatic rings, e.g., polycyclic aromatic hydrocarbons, phthalate esters, surfactants and many pesticides. In addition, a variety of naturally occurring compounds contains the aromatic structure, e.g., amino acids, flavonoids and biopolymers. Biodegradation of the chemically stable aromatic ring under anaerobic conditions differs substantially from the oxidative steps known from aerobic catabolism. Different pathways have been found according to the substitutions on the ring, although benzoyl-coenzyme A has proved to be a central intermediate in the pathway of a variety of aromatic compounds. 2.1.2.1 Peripheral Pathway
Aromatic compounds serve as carbon and energy sources for anaerobic bacteria, and these growth substrates are eventually transformed into central inter-
Fig. 4. Transformation of some aromatic compounds by the peripheral pathway to benzoylCoA. From [21]. Reprinted with kind permission from Academic Press Ltd
Potential for Anaerobic Conversion of Xenobiotics
81
mediates recognized in the anaerobic degradation pathway of many different aromatic compounds. The peripheral pathway of some aromatic organic compounds to benzoyl-CoA is shown in Fig. 4. The aromatic molecule is generally more reduced than the products formed under anaerobic metabolism, and reducing equivalents must thus be disposed of. According to [21], only anaerobic respiratory bacteria, anaerobic photosynthetic bacteria, and fermenting bacteria that are capable of interspecies hydrogen transfer, can metabolize aromatic compounds under anaerobic conditions (only the latter will be discussed here). The highly hydroxylated aromatics with a redox state equivalent to the redox state of the fermentative products can be used by fermenting bacteria as source of carbon and energy [21]. Less oxidized aromatic molecules can be degraded by fermenting bacteria in a syntrophic co-culture. By disposing of the reducing equivalents as hydrogen, the pathway becomes thermodynamically feasible, and a wide range of aromatic substrates can therefore be used. The fermentative bacteria that biodegrade the aromatic molecule are coupled to, e.g., methanogenic bacteria, and such consortia can metabolize even highly chlorinated compounds. 2.1.2.2 The Benzoyl-CoA Pathway
According to Heider and Fuchs [21, 22], and Harwood and Gibson [23], benzoylCoA is the most common intermediate in the anaerobic degradation of aromatic compounds. Three variants of the anaerobic benzoyl-CoA pathway have been found in different anaerobic bacteria: the phototrophic bacterium Rhodopseudomanas palustris, the denitrifying Thauera aromatica, and the fermenting bacteria Syntrophus gentianae co-cultured with methanogenic bacteria. Benzoyl-CoA reductase is the key enzyme in the dearomatizing reaction where cyclohexa1,5-diene-1-carboxyl-CoA is produced in the reduction reaction. Eventually, benzoyl-CoA is metabolized in a series of hydration and dehydrogenation steps that are analogous to b-oxidation [22] resulting in the formation of three molecules of acetyl-CoA and one molecule of carbon dioxide (see Fig. 5). Harwood and coworkers [24] underline that the steps which today are believed to be the main reactions in the anaerobic degradation pathway of benzoate might only be side reactions; only a small part of the enzymatic activities demonstrated in benzoate-grown cells have in fact been shown to be essential in the benzoate degradation. Reduction of the aromatic ring requires a high energy input to overcome the energy barrier of 30 kcal/mol, and benzoyl-CoA reductase catalyze a two electron reduction of benzoyl-CoA [25]. This reduction is apparently energized by hydrolysis of two molecules of ATP. Recent studies indicate that the reduction of benzoyl-CoA not necessarily requires hydrolysis of ATP in fermenting bacteria, and that the overall energy gain for the benzoate fermentation might be as low as one third of an ATP [24].
82
A.S. Mogensen et al.
Fig. 5. The central benzoyl-CoA pathway of anaerobic aromatic degradation by R. palustris and T. aromatica. Reprinted from [24]. Copyright (1998), with permission from Elsevier Science
Potential for Anaerobic Conversion of Xenobiotics
83
2.2 Anaerobic Conversion of Surfactants
Molecules that contain a hydrophilic (polar) and a hydrophobic (non-polar) part have amphiphilic properties and are surface active. These compounds owe their chemical specificity to the hydrophilic and the hydrophobic groups of the molecule and are commonly termed surfactants. Surfactants are widely used for both domestic and industrial purposes as cleaning, emulsifying, and wetting agents and a very large number of different surfactants are commercially available. Surfactants can be divided into four categories [26] according to the property of the hydrophilic group being anionic, non-ionic and cationic, amphoteric (or weakly acid and weakly alkaline). The hydrophilic group of anionic surfactants carries a negative charge, as for example, the group of linear alkyl benzenesulfonates (LAS), which is one of the most widely used among the sulfonated hydrocarbons. The annual production is approximately 1¥106 t/year in the USA of a worldwide detergent production of 15¥106 t/year. Also, sulfates, sulfonated esters and amines, carboxylates, and phosphates belong to the group of anionic surfactants. Non-ionic surfactants carry no charge and are, e.g., non-nitrogenous, alkanoamides and ethoxylated dialkanoamides. Among non-ionic surfactants, the non-nitrogenous subgroup of alkylphenol polyethoxylates (APnEO) is one of the most widely used, with an annual production of about 0.5 ¥106 t nonylphenol polyethoxylates (NPnEO) as the most commonly used [27]. If the hydrophilic part of a surfactant has a positive charge it is termed cationic, e.g., quaternary ammonium salts or weak bases. Amphoteric surfactants have a hydrophilic part, which contains both positive and negative charges, e.g., carboxylates with weakly or strongly alkaline N and sulfonates with weakly or strongly alkaline N. Recently, surfactants have been applied in the bioremediation of sorbed organic contaminants due to increased bioavailability by solubilization of insoluble contaminants in presence of surfactants [28]. Surfactants have also been shown to increase the COD removal in anaerobic digestion of industrial waste thereby augmenting the biogas yield [29]. Many surfactants have a complex molecular structure including the hydrophilic and the hydrophobic part, and mineralization requires a consortium of bacteria that act on different levels of the molecular alteration process. For instance, the biodegradation of linear alkyl benzenesulfonate requires alteration of an alkyl chain, a benzene ring, and a sulfonate linkage.Anaerobic biodegradation of surfactants has received little attention, and the body of information regarding surfactant biodegradation is predominately dealing with aerobic bacteria [30, 31]. The current chapter collects the recent studies with the focus on anionic and non-ionic surfactants. 2.2.1 Surfactant Properties and Molecular Structure
Due to the amphiphilic nature of surfactants, they impair the function of biological membranes due to interactions with proteins and membrane lipids, re-
84
A.S. Mogensen et al.
sulting in decreased barrier functionality. Hence, surfactants have adverse effects on free living cells like bacteria and archaea in anaerobic digesters. Today it is still uncertain to which extent the activities of the different anaerobic bacteria are impaired by their presence, although methanogenic bacteria are seemingly more prone to surfactant inhibition than fermentative bacteria. The cell wall of archaea differs greatly from the bacterial, and peptidoglycan that is responsible for the strength of the wall in bacteria might be more resistant than the archaean cell walls that consist of polysaccharide, glycoprotein, or protein. 2.2.1.1 Anionic Surfactants
Anionic surfactants usually owe their negative charge to a sulfonate or a sulfate group attached to the lipophilic group, which commonly is an alkyl chain or an alkylaryl group. Linear alkyl benzenesulfonate (LAS) and secondary alkanesulfonates (SAS) are two high volume surfactants used in, e.g., domestic detergents (see Fig. 6 for molecular structure).With household wastewater, these surfactants enter the sewage treatment plants and are biodegraded mainly during aerobic treatment. A significant part of the surfactant load is not treated aerobically due to adsorption on solids in the primary settling tank. This part will be treated anaerobically by sludge digestion. SAS and LAS are removed physically from wastewater in the clarification by 16% (w/w) and 5–40%, respectively [32–35]. 2.2.1.2 Cationic Surfactants
Quaternary ammonium compounds (QACs) diversity rely on the various chemical groups that are attached to a nitrogen molecule, which can be, for example, three methyl groups and one alkyl group (see the molecular structure of two QACs in Fig. 7). Common properties for the positively charged QACs are their surface activity, adsorption on negatively charged solids, biocidal activity and
Fig. 6. Molecular structure of 4-(2-dodecyl)benzenesulfonate (C12-LAS) and [2-(hexadecyl)sul-
fonate] (C16-SAS)
Potential for Anaerobic Conversion of Xenobiotics
85
Fig. 7. The major commercial quaternary ammonium cations alkyltrimethylammonium and alkylbenzyldimethylammonium
ability to react with anionic surfactants [36]. QACs are used as fabric softeners, disinfectants, demulsifiers, emulsifiers and wetting agents. Hence, QACs are mostly found in wastewater. During wastewater treatment, the compounds absorb strongly to primary sludge and are therefore subjected to anaerobic treatment, if applied in the wastewater treatment plant (WWTP). 2.2.1.3 Non-Ionic Surfactants
The majority of non-ionic surfactants are hydrophilic polymerization products of 1,2-epoxyethane, which are introduced to a hydrophobe with a reactive hydrogen atom – being for example an amine or alkylphenol. The molecule does not contain any charged constituent, as seen in Fig. 8. APnEOs, as for example nonylphenol polyethoxylates, are used in industry in detergent, paint, pesticide, pulp and paper as well as in textile manufacturing [37, 38]. Biotransformation is reported under both aerobic and anaerobic conditions, and the hydrophobic metabolites AP1–2EO and alkylphenol are commonly found in sewage sludge due to increased sorption potentials compared to alkylphenol polyethoxylates.
Fig. 8. Molecular structure of alkylphenol polyethoxylate
2.2.2 Common Disposal Route
Surfactants are generally used in aqueous processes and their fate is therefore linked to the discharge route of the process water or household effluent. They are commonly water soluble compounds with very low vapor pressure and relatively high Kd values. In industrialized countries the major part of water with industrial
86
A.S. Mogensen et al.
and domestic use is discharged to a sewer system connected to a WWTP, though also discharge to surface water may be found. The loss of surfactant to the atmosphere is for selected surfactants negligible because of their low vapor pressure, and the surfactant partitions between the water and solid phase. Up to 50% of the surfactant can adsorb on solids or be removed by precipitation in the primary clarifier tank [35, 39–41] depending on water hardness. The remaining part is eventually subjected to aerobic treatment in, e.g., the activated sludge tank, and generally 50–100% is removed depending on the type of surfactant. The effect of sorption of surfactants during biological (aerobic) treatment has not been assessed. In general, primary degradation of surfactants in WWTP occurs readily, but ultimate degradation does not [42]. The discharge of nonylphenolic compounds – which are considered the primary intermediates of NPnEOs – from WWTP has been estimated to account for 60–65% of the amount that enters the plant [43]. A large part of the surfactant load is therefore associated with the sludge that formerly was dumped in the ocean, incinerated, or disposed by landfilling. These disposal routes are being abandoned favoring the use of sludge as a soil-improving agent on farmland. Due to the absorption of surfactants to sludge, the concentration in sludge can be as high as the g/kg range. Although very little is known about their effect of farmland, the concentration of, e.g., linear alkyl benzenesulfonates must be below 1300 mg/kg DM in order to fulfill the legislative requirements in many European countries. Since sludge is anaerobically stabilized at many wastewater treatment facilities, enhanced anaerobic bioremoval of surfactants as well as other contaminants is now becoming of major concern in order to minimize the waste generation in modern society. Anaerobic digestion of surfactants is therefore of great importance in order to avoid an increased load of lipophilic surfactants on farmland since surfactants that escape biodegradation during anaerobic treatment will consequently be found on farmland. The concentration of a variety of surfactants in sewage sludge is listed in Table 3.
Table 3. Concentration of surfactants in sewage sludge
Surfactants
Concentration in sewage sludge
Reference
Anionics
LAS SAS MBAS
0.05–18.8 mg/g dry solids 0.27–0.80 mg/g dry solids 1.9–7.9 mg/g dry solids
[41, 44–47] [32, 47] [48]
Non-ionics
Unspecified 4-OP 4-NP NPEO NP1EO NP2EO CTAS
1.1–4.7 mg/g dry solids 0–0.020 mg/g dry solids 0.0084–1.570 mg/g dry solids 2.2 mg/g dry solids 0.0039–0.680 mg/g dry solids 0.004–0.297 mg/g dry solids 0.5–1.8 mg/g dry solids
[41] [38, 47] [38] [43, 45] [38] [38] [48]
1.0–10.5 mg/g dry solids
[36, 49]
Cationics
Potential for Anaerobic Conversion of Xenobiotics
87
2.2.3 Anaerobic Degradation of Surfactants
The information accessible on anaerobic degradation of surfactants is limited and restricted to mainly the anionics and non-ionics. In Table 4, the available data are listed divided into surfactant type. Recent research has focused on the inhibition of methanogenic bacteria in a variety of environmental compartments, but most data concern sludge from anaerobic digesters. Data on cationic and amphoteric surfactants are still needed for unveiling their effects on anaerobic bacteria. Methanogenic bacteria are more sensitive to the presence of surfactants than fermentative bacteria, and LAS can impede their metabolism in concentrations of 10 ppm. In addition, the metabolites of NPnEO show adverse effects on acetate-utilizing methanogens, and methane formation in anaerobic digesters might be hampered. Surfactant sorption to organic matter is one of the primary mechanisms for removal in WWTP, and for certain surfactants, the concentration in sludge often increases during sludge processing. Therefore, high concentrations of surfactants have been measured in digested sewage sludge. This concentration is then used in the toxicity and biodegradation studies. The environmentally relevant, or the bioavailable concentration, is much lower than the typical experimental conditions (see Sect. 2.1.1.3), which might be the explanation of many unsuccessful attempts to demonstrate anaerobic surfactant biodegradation. Still insufficient data is available concerning anaerobic biodegradation of surfactants. Furthermore, pure cultures of microorganisms capable of degrading these compounds are needed in order to bioaugment, e.g., anaerobic reactors. 2.2.3.1 Degradation of Anionic Surfactants
Anionic surfactants with a sulfate group as the hydrophilic part of the molecule, e.g., alkyl sulfates, have shown inhibitory effects on methanogenic bacteria at 100 mg/L, and the alkyl sulfate molecule can be degraded under methanogenic conditions if the concentration is kept low. In many studies, the concentration is not related to the organic matter content in the test system, which is a fundamental parameter with influence on absorption (and bioavailability). Consequently, effect concentrations can hardly be extrapolated from one system to another. Glucose, acetate, and lactate catabolism was impeded by the presence of alkyl sulfate at 100 mg/L or higher using methane formation as the indicator of bacterial activity. However, linear alcohol sulfate was metabolized at 50 mg C/L using digester sludge, though bacterial adaptation was required. The sulfate-containing surfactants tetradecyl triethoxy sulfate and sodium tetradecyl sulfate are according to [50] readily degradable under anaerobic conditions.With mesophilic digester sludge as inoculum, between 80–88% of radioactivity was recovered as gas after 18 days of incubation of [14C]-(1,3-ethoxylate)-tetradecyl triethoxy sulfate and [1-14C]sodium tetradecyl sulfate. Acetoclastic bacteria was according to Shcherbakova and coworkers inhibited 50% by dodecyl sulfate at 573 mg/L when using granular sludge originating from a UASB reactor treating pulp and paper mill effluent [53].
96–99% primary degradation, though adaptation period required. No biodegradation after 60 days, and 50% methanogenic inhibition at 160 mg/L. 88% mineralization under mesophilic conditions with digester sludge as inoculum.
Linear alcohol sulfate (A24S, A45S)
Perfluorooctanesulfonic acid
Tetradecyl triethoxy sulfate (AES)
50% inhibition of acetate metabolism relative to control at 548 mg/L. Inhibition of glucose and lactate fed cultures measured as methane production. No biodegradation after 60 days.
n-(tert-octyl)phenol (Triton X-100)
Highly fluorinated oxethylate
Polyoxyethylene alkyl ether (E-LM75 and Increase in methane production with anoxic digester and polluted creek material as inoculum. [51] brij 35), Polyoxyethylene alkylphenol Adaptation required with digester sludge at 500 and 1000 mg/L (E-LM75, E-N90). ether (E-N90)
[55]
[53, 54]
Primary degradation of polyethoxylates observed. NP is very recalcitrant. Decrease production [37, 48, of methane during mesophilic glucose digestion at 100 mg/L NPE and inhibition of acetate 52] formation above 80 mg/L observed with mesophilic digester sludge.
[50]
[55]
[48]
Nonylphenol polyethoxylates (NPEO), octylphenol polyethoxylate (OPEO), nonylphenol (NP), octylphenol (OP)
Non-ionics:
Impeded glucose and acetate metabolism (measured as methane production), though increase [20, 41, in methane production with creek sludge at 100 mg LAS/L has been observed. 50% decrease in 51–53, methane formation observed at 100–500 mg/L. Biodegradability shown in UASB and CSTR 57] reactors (with formation of benzenesulfonic acid and benzaldehyde as LAS metabolites).
Linear alkylbenzene sulfonate (LAS)
[51]
Impairment of methane production at 10–100 mg/L and acetate accumulation at 100 mg/L with creek sludge as inoculum.
Sodium dodecylsulfonate
[50–54]
Reference
Up to 80% mineralization observed at 1 mg/L and increase in methane production. Inhibition on methanogenesis and impaired glucose and acetate metabolism at 40–1000 mg/g VSS.
Results
Sodium tetradecylsulfate and sodium dodecyl sulfate
Anionics:
Surfactants
Table 4. Overview of reports of anaerobic surfactant degradation
88 A.S. Mogensen et al.
Inhibitory towards methanogenesis at 200 mg/L. No inhibition of methane production in glucose and lactate fed cultures with nine different Tween surfactants. 60–83% of theoretical methane formation after 40–50 days. 96–97% primary degradation though adaptation period was required.
Polyoxyethylene alcohols (brij and witconol)
Polyoxyethylene sorbitan fatty acids esters (Tween)
Linear alcohol ethoxylate (LAE-8)
50% inhibition of acetate metabolism relative to control at 154 mg/L. 50% inhibition of acetate metabolism relative to control at 345 mg/L. Glucose digestion impeded by 90% after 30 days measured as methane production. Acetate formation was inhibited above 10 mg/L. 2–35% inhibition of biogas production after 200 days of incubation, primary degradation between 10 and 38%.
Alkamon DS, quaternary ammonium salt (90% primary substance)
Catamin AB, quaternary ammonium salt (50% primary substrate)
Tetradecyldimethylbenzyl ammonium chloride
Alkyl trimethyl ammonium halides and alkyl benzyl dimethyl ammonium halide
Cationics:
Results
Surfactants
Table 4 (continued)
[36]
[52]
[53]
[53]
[48]
[54]
[54]
Reference
Potential for Anaerobic Conversion of Xenobiotics
89
90
A.S. Mogensen et al.
The study of linear alkyl benzenesulfonate biodegradation under methanogenic conditions has been attempted throughout the world due to the high utilization of this surfactant. Nevertheless, biodegradation is only rarely reported, and only under certain conditions, e.g., under sulfate limited conditions [56], in UASB reactors [57], or in C12-LAS-enriched CSTR reactors [20, 58]. Sulfonated surfactants are extremely difficult to degrade under anaerobic conditions, and methanogenic activity is affected at concentrations as low as 10 mg/L of alkyl sulfonate or alkyl benzenesulfonate. Nevertheless, few studies report enhancement of methane production with addition of these surfactants [51, 59]. The stimulation of methane production was shown at low concentrations of linear alkyl benzenesulfonate using creek sludge as inoculum with 10 and 100 mg/L surfactant amendment or with sewage sludge as inoculum and 20 mg/L of C12-LAS. LAS was not quantified during these experiments, and since methane production was more abundant at the low concentration, it can be argued that the increment in gas production was caused by biodegradation of other organics. Enhancing the bioavailability of organics by addition of surfactants is a well-known phenomenon [28, 29, 60, 61]. Several studies contradict this finding and LAS is commonly found to impede methane production severely and does according to [41, 62] persist during anaerobic sludge treatment. However, one study has related the bioavailable fraction of LAS with LAS degradation during treatment of sewage sludge in a laboratory-scale CSTR reactor [20] (see Sect. 3.2). Moreover, LAS degradation in UASB reactors with formation of the metabolites benzenesulfonic acid and benzaldehyde has been shown under thermophilic conditions [57] (see Sect. 3.1). According to Shcherbakova and coworkers, these metabolites are virtually non-toxic to methanogenic bacteria [53]. The formation of benzenesulfonic acid during anaerobic degradation of LAS is presumably caused by beta-oxidation of the alkyl chain as known from the alkyl sulfates. Tanaka and Ichikawa showed that the activity of fermentative bacteria decreased by the presence of sulfonated surfactants at 100 mg/L, and that acetate formation was impeded at 40 mg/L [52]. Cook and coworkers could not show growth in three anoxic enrichments amended with octyl benzenesulfonate, but they cultured bacteria with other sulfonated compounds as carbon and energy source (taurine and cysteate) and nitrate as terminal electron acceptor [63]. Low inhibition of the acetoclastic bacteria by linear alkyl benzenesulfonate was found using pulp and paper mill granules [53]. LAS inhibited the methane formation by 50% at 270 mg/L, but from other studies it has been shown that 100 mg/L can inhibit methane production completely [59]. Secondary alkanesulfonate (SAS) has been reported to be stable under anaerobic conditions [47], and more hydrophobic SAS components are reported to be more susceptible to accumulation in sewage sludge. 2.2.3.2 Degradation of Non-Ionic Surfactants
Primary anaerobic biodegradation of alkylphenol polyethoxylates occurs in WWTP and in natural environments [38, 42], but mineralization is hampered by formation of recalcitrant intermediates, especially alkylphenol, that are more
Potential for Anaerobic Conversion of Xenobiotics
91
lipophilic, less water soluble and more toxic than the parent compounds [38, 64]. During wastewater treatment the number of ethoxy groups in APnEO is reduced, with formation of AP,AP1EO, and AP2EO. These are all more recalcitrant and hydrophobic than APnEO and the metabolites are therefore found in sludge from WWTP. It seems probable that NPEO or NP shows adverse effects on especially acetateutilizing methanogens, thereby hampering methane formation in bioreactors. Ejlertsson et al. [37] showed transformation of NP1–2EO using anaerobic digestor sludge as mesophilic inoculum with concomitant formation of NP1EO and NP, but no methane was formed. It was speculated that the phenol ring structure remained intact, though the amount of added alkylphenol ethoxylate might not have been sufficient for determination of methane increment relative to the controls. Salanitro and Diaz [48] found a high degree of primary degradation of NP9EO, but only 30 to 70% of the theoretical methane production was recovered and Prats et al. [41] found that non-ionic surfactants were removed by 65% during anaerobic treatment in a sludge digester. Yeh and coworkers [54] found a varying degree of inhibition of glucose and lactate fed cultures by 14 different surfactants from two classes, polyoxyethylene octylphenyl ether (POE) alcohols and POE sorbitan fatty acid esters (Tween surfactants). While the sorbitan fatty acid esters did not impair methane development, it was completely inhibited by presence of the alcohols. Lactate uptake was enhanced by addition of POE sorbitan fatty acid esters, possibly due to increased substrate transport across the cell membrane caused by increased membrane permeability. When no electron donor except the surfactant was present, POE sorbitan fatty acid esters were degraded by 14–46%, but POE alcohols were not degraded. Sludge from an anoxic digester and a polluted creek contained anaerobic bacteria capable of biodegrading non-ionic surfactants (polyoxyethylene alkyl ether and polyoxyethylene alkylphenol ether) as evidenced by the increment in methane development relative to unamended samples [51]. It was not assessed whether the increase in methane production was due to enhanced degradation of organic matter. A highly fluorinated oxethylate was totally inhibiting the methane production at a concentration of 100 mg C/L when added to an incubation vial with digester sludge [55]. 2.2.3.3 Degradation of Cationic Surfactants
In experiments carried out by Garcia and coworkers, no ultimate biodegradation was observed for alkyltrimethylammonium halide and alkylbenzyldimethylammonium halide with alkyl chain length of 12 to 16 carbons when tested in serum vials under mesophilic conditions [36]. The QACs were shown to inhibit methane formation. QACs with longer alkyl chains had lower inhibitory effects on biogas formation, but were also more recalcitrant. Primary degradation was found in the range of 19 to 38%. Also, it was found that anaerobic biomethanation of glucose was impeded by the presence of QACs, and acetate formation was inhibited at concentrations above 10 mg/L [36, 52]. In granular sludge, alkamon DS and cat-
92
A.S. Mogensen et al.
amin AB inhibited acetate metabolism by 50% at 154 mg/L and 345 mg/L, respectively [53]. 2.3 Anaerobic Conversion of Polycyclic Aromatic Hydrocarbons
Polycyclic aromatic hydrocarbons (PAHs) are widespread organic contaminants, which are found on the US EPA’s list of priority pollutants due to their carcinogenic and mutagenic properties. The contamination originates from sources such as oil seeps and water run-off from fire sites as well as anthropogenic sources, e.g., fossil fuel combustion. A variety of sources are listed in Table 5 [65]. PAHs are compounds that are found in significant amounts in a variety of soil environments as a result of contamination at gas works sites, fuel combustion and industrial processes, etc. PAHs are furthermore predominant among the organic contaminants in sewage sludge. Their wide environmental distribution, recalcitrance towards biodegradation, bioaccumulation abilities, and adverse effects on cellular macromolecules emphasize the importance of remediation with, e.g., biological means as a solution to an important environmental problem. It has been demonstrated that they are relatively easy biodegradable compounds under aerobic conditions [66–68]. Recent research demonstrates, however, that PAHs are degradable in other anoxic environments. PAHs in the environment are removed by both biotic (biodegradation) and non-biotic processes (e. g., chemical and photooxidation). Furthermore, a worldwide migration towards colder regions by sequential entrapment (by sorption to organic matter) and volatilization, with eventual bioaccumulation in lipidrich biota in the polar regions can be observed. Microbial degradation is the key decontamination process, despite the fact that biodegradation of PAHs in soil and
Table 5. Major sources of PAHs in the environment. Reprinted from [65]. Copyright (1992), with kind permission from Kluwer Academic Publishers
Natural oil seeps Refinery and oil storage waste Accidental spills from oil tankers and other ships Municipal and urban waste water discharge runoff River-borne pollution Atmospheric fallout of fly ash particulates Petrochemical industrial effluents Coal tar and other coal processing waste Automobile engine exhausts Combustion of fossil fuels Smoked, charcoal broiled, or pan fried foods Tobacco and cigarette smoke Forest and prairie fires Rural and waste incineration Coal gasification and liquefaction processes Creosote and other wood preservative waste Commercial and pleasure boating activities
Potential for Anaerobic Conversion of Xenobiotics
93
Fig. 9. Fate of polycyclic aromatic hydrocarbons. Reproduced from [65]. Copyright (1992), with kind permission from Kluwer Academic Publishers
aquifers is hampered by entrapment in micropores and partitioning on soil organic matter. These physical processes decrease the bioavailability of PAHs to microorganisms and the biodegradation in therefore impeded.With a higher number of rings in the molecule, the PAHs seem to be more recalcitrant, possibly due to tight sorption and low desorption rates compared to PAHs with fewer rings. In Fig. 9, the fate of PAHs is illustrated. The assessment of environmental quality objectives for soil and sediment for a vast part of PAHs is impossible due to lack of environmental data. The oxidative metabolism of PAHs can lead to formation of reactive metabolites responsible for the mutagenic and carcinogenic effects of PAHs [69, 70] by covalently binding to, e.g., DNA and other cellular macromolecules. Previous assessments regarding biodegradation of PAHs have excluded anaerobic microorganisms and the majority of data on PAH biodegradation is achieved with oxygen as the terminal electron acceptor. Nevertheless, some studies indicate that anaerobic biodegradation is possible, although the degradation rates are lower. 2.3.1 Properties
PAHs are lipophilic organic chemicals consisting of two or more aromatic rings. The recalcitrance towards microbial degradation increases with increasing number of rings, which is liked to the individual physicochemical properties (see Sect. 2.1.1.3). Table 6 shows the molecular structure and physico-chemical properties for selected PAHs. 2.3.2 Environmental Contamination
The use of wood preservatives agents such as creosote has resulted in contamination of soil, aquifers, and sediments. Typical creosote mixtures contain approximately 85% of PAHs from bi- to polycyclic aromatic hydrocarbons.
94
A.S. Mogensen et al.
Table 6. Properties and molecular structure of selected PAHs. Data from [71]
Name
Naphthalene
Anthracene
Phenanthrene
Fluoranthrene
Pyrene
Benzo[a]pyrene
128.17
178.23
178.23
202.26
202.26
252.31
Molecular structure
Molecular weight (g/mole) Water solubility (mg/L)
31
Log Kow
3.3
0.04
1.2
0.26
0.013
0.0038
4.45
4.46
5.16
4.88
6.13
Soil and sediment in lakes and rivers are accumulation sites for PAHs, and the rate of PAH-input to these environments exceeds often the rate of biodegradation. In these sites, the fast aerobic biodegradation is hampered by the slow transport of oxygen, and PAH removal depends often on the relatively slower anaerobic degradation. Commonly, a high concentration of PAH is observed in dredge sludge and sediment from harbors due to petrol contamination. Investigations have revealed PAH contamination in the g per kg range, but typically lower concentrations are detected [72–75]. The suspended matter of rivers is also contaminated with PAHs [76] as well as drinking water where water pipes are coal tar lined [77]. Table 7 lists the PAH concentration at different contaminated sites. Table 7. Concentration of PAHs in the environment
Site
PAHs
PAH concentration
Reference
Dredge river sediment
Flouranthrene, pyrene and phenanthrene
14, 12, and 10 mg/kg sediment
[73]
Harbor and sea sediment
Total PAH
145 mg/kg dry sediment
[74, 75]
European rivers
Total PAH
< 4 mg/L
[76]
Dredge sludge from a petrol harbor
Total PAH
2600 mg/kg
[72]
Contaminated Antarctic soil Total PAH
41–8105 mg/kg dry soil
[78]
Domestic sewage sludge
Total PAH
22 mg/kg dry matter
[79]
Rural Estonian soil
Total PAH
0.1 mg/kg DM
[80]
Estonian town
Total PAH
12 mg/kg DM
[80]
Soil at creosoting plant
Total PAH
501 mg/kg
[81]
Soil from gas work
Phenanthrene and anthracene
150 and 250 mg/kg soil
[82]
95
Potential for Anaerobic Conversion of Xenobiotics Table 8. Half-lives (in days) for PAHs in various systems. Data from [83]
Compound
Soil amended with PAH-containing sludge
Soil mixed with oily PAHs added to two PAHs added to waste, nutrients sandy loam soils lake sediment and microbes
Naphthalene Phenanthrene Pyrene
28 ± 16 124 ± 48 225 ± 92
– 24.4 19.4
2.1–2.2 16–35 199–260
16.8–30.8 28–126 238–630
2.3.3 Anaerobic Biodegradation of PAHs
It has earlier been suggested that PAHs were not biodegraded in the absence of molecular oxygen, which is needed for the monooxygenase, the required enzyme to initiate the alteration of the PAH molecule under aerobic conditions. In the last decade, however, we have learned that PAH biodegradation is possible also under anoxic conditions, although alterations occur at lower rates than when monooxygenase is involved. Below, the available data regarding biodegradation under nitrate-, iron-, sulfate-reducing, and methanogenic conditions are given. In Table 8 the half-lives in different environments for selected PAHs are given. Several studies on anaerobic PAH degradation have focused on facultative anaerobic microorganisms (nitrate-, and iron-reducing microorganisms), but capability of PAH degradation by strict anaerobes (sulfidogenic and methanogenic bacteria) has also been shown. 2.3.3.1 Nitrate-Reducing Conditions
Recalcitrance to biodegradation of PAHs under several anoxic conditions with a creosote contaminated sediment as inoculum has been observed [84]. Under nitrate-reducing conditions, only 2-methylanthracene was degraded. No biodegradation of 4- and 5-ring PAHs was seen under any redox conditions. With inoculum from a former gas plant, 13% degradation of naphthalene was observed in two out of 11 samples under nitrate-reducing conditions and a temperature of 22 °C [85]. This was confirmed by [14C]carbon dioxide development after inoculating under denitrifying conditions with a diesel fuel-contaminated aquifer and amending with [14C]-labeled naphthalene [86]. Biodegradation of 3- and 4-ring PAHs (anthracene, phenanthrene and pyrene) has been shown with three Pseudomonas putida strains with faster degradation rates under the anoxic than under aerobic conditions [87].Additionally, McNally found that single substrate biodegradation data are different from multisubstrate biodegradation data when using the Pseudomonas putida strains KBM-1 and SAG-R. Furthermore, phenanthrene (and pyrene) degradation was enhanced with the presence of naphthalene but hampered when pyrene (phenanthrene) was present [87].
96
A.S. Mogensen et al.
% of initial 14C-naphthalene or NO3–-remaining
120
Naphthalene Nitrate
100
80
60
40
20
0 Cell Free
Nap-4 (–NO3)
Nap-4 (+NO3)
Nap-3-1 (–NO3) Nap-3-1 (+NO3)
Culture Fig. 10. Nitrate-dependent anaerobic transformation of naphthalene. Reprinted from [6]. Copy-
right (2000), with kind permission from American Society for Microbiology
Naphthalene biodegradation has been observed in bioreactor systems inoculated with contaminated samples [88] but also by using nitrate-reducing bacterial strains [89]. Rockne and Strand found removal of naphthalene, phenanthrene and biphenyl in fluidized-bed reactors under nitrate- and sulfate-reducing conditions [90]. Specific PAH removal rates between 1.1 and 5.3 mg/g VSS/d and 0.1–0.5 mg/g VSS/d for nitrate- and sulfate-reducing conditions, respectively, were determined for the PAHs tested. A pure culture capable of metabolizing naphthalene coupled with reduction of nitrate to nitrite was obtained from this reactor [6] and is the first pure culture known to metabolize a PAH under nitratedependent, anaerobic conditions (see Fig. 10). 2.3.3.2 Fe(III)-Reducing Conditions
The biodegradation of PAHs under Fe(III)-reducing conditions is scarcely studied, but the available data show that degradation of naphthalene is possible in petroleum-contaminated aquifers [91]. Coates did not succeed in switching the terminal electron accepting process from sulfate reduction to Fe(III) reduction by addition of Fe(III) to a contaminated sediment with a sulfate-reducing microbial population [92]. The difference in the population density was possibly the cause since the sulfate-reducing microorganisms dominated the culture by a factor 103 over the iron-reducing microorganisms.
97
[14C]-CO2 (% of added [14C]-phenanthrene)
Potential for Anaerobic Conversion of Xenobiotics
20 mM MoO4 added at day 13
Time (day)
[14C]-CO2 (% of added [14C]-naphthalene)
10 mM SO4 added at day 0
10 mM SO4 added at day 9
Time (day)
production from naphthalene enrichment and inhibition of [14C]phenanthrene oxidation by molybdate. Reprinted from [4, 5]. Copyright (1996 and 1997), with kind permission from American Society for Microbiology
Fig. 11.
[14C]-CO2
2.3.3.3 Sulfate-Reducing Conditions
It has only recently been shown that PAH oxidation can occur under sulfate-reducing conditions. Coates found conversion of naphthalene, methylnaphthalene, phenanthrene, flourene and flouranthrene. The tests were performed with PAHcontaminated sediment radiolabeled with PAHs. Radiolabeled carbon dioxide and minor amounts of radiolabeled methane were detected. Addition of molybdate (which is an inhibitor of sulfate reduction) completely stopped PAH mineralization, and removal of sulfate resulted in termination of the oxidation process, see Fig. 11 [4, 5]. Thus, the sulfate reduction was at least necessary, if not responsible for the process.
98
A.S. Mogensen et al.
Naphthalene was biodegraded in a sulfidogenic enrichment derived from a laboratory microcosm with coal tar contaminated aquifer material as inoculum [84, 93]. [14C]CO2 was observed when amending with [14C]naphthalene and [35S]Na2SO4 was recovered as radiolabeled H2S during naphthalene transformation, but not in the control microcosms. It remains to be shown whether this process is common in sulfate-rich environments, such as many marine sediments. Data regarding microbiological community analysis are at present not available, rendering the study of biochemical reactions and genetics impossible. However, carboxylation seems to be an initial reaction in the metabolism of naphthalene and phenanthrene with the corresponding carboxylic acid as an intermediate [94]. 2.3.3.4 Methanogenic Conditions
Sludge contaminated with PAHs is treated under methanogenic conditions in wastewater treatment plants worldwide. Despite this common treatment process and the knowledge of the adverse effects of PAHs, little is known about their fate during this type of treatment. The physical transfer of PAHs from wastewater to sludge is typical for chemicals with similar chemical and physical properties, and their persistence towards microbial conversion under anaerobic conditions is obvious from the concentration in digested sludge. In a pilot study simulating anaerobic mesophilic digestion of municipal wastewater sludge spiked with 500 mg/L of six different two- and three-ring PAHs, solvent extraction revealed that 60 to 80% was removed. The secondary digestor, operated in batch mode, proved more efficient than the primary digester despite lack of temperature control and stirring. These presumably unfavorable environmental conditions (see also Sect. 2.1.1) question whether the removal was biologically or physically mediated. Genthner examined methanogenic degradation of two- to five-ring PAHs at substantially lower concentrations (3.6–36 mg/L). Despite the use of creosotecontaminated inoculum with methanogenic activity, PAH amendment severely inhibited methane production [84]. Still some degradation of bi-and tricyclic PAHs was observed, but the authors conclude that the potential for anaerobic bioremediation is limited. From one out of two inocula, Coates found development of 0.6% radiolabeled methane of the added 14C label after addition of five different radiolabeled PAHs, though no [14C]CH4 was produced when amending with [14C]acetate. The methane production could therefore be due to the mini-methane system in Desulfovibrio and Desulfotomaculum [4]. A continuous flow packed-bed column with a mixture of anaerobic soil, sediment and granular sludge in the bed was operated under methanogenic conditions, but no naphthalene degradation was observed despite addition of possible co-substrates [88].
99
Potential for Anaerobic Conversion of Xenobiotics
2.4 Anaerobic Conversion of Phthalate Esters
Phthalate esters (PEs) such as dibutyl phthalate (DBP), diethylhexyl phthalate (DEHP), and di-n-octyl phthalate (DOP) are plasticizers with widespread industrial use, e.g., additives in resins such as polyvinyl chloride (PVC). About 20 different PEs are used commercially, and their wide application is due to stability, fluidity and low volatility of the high molecular weight phthalates [95]. In 1994, 4.2 million tons of DEHP was produced [96] and was as such the most commonly used PE representing 40–50% of the annual production. DEHP is widely used in the packaging industry with concentrations of up to 7000 ppm in food packaging [97]. In wastewater, the dissolved PEs adsorb on suspended solids and are eventually removed from wastewater with the primary sludge settling [95]. The concentration of DEHP in sewage sludge has been observed within the range of 10–100 mg/kg dry matter [98]. Despite the fact that PEs are degraded under both aerobic and anaerobic conditions [99–101] they receive further attention due to continuous increase in production coupled with the suspicion of being endocrine disrupting in vertebrates [102]. 2.4.1 Properties and Molecular Structure
Most commercial PEs are liquid at ambient temperature, have low water solubility, and the Kow values increase with increased alkyl chain length resulting in greater lipophilicity. The common molecular structure is a phthalate acid molecule that is esterified by different alcohols (see Table 9).
Table 9. Molecular structure of dimethyl phthalate, butyl benzyl phthalate, and diethylhexyl phthalate. Data from [95]
Phthalate ester
Dimethyl Butyl benzyl phthalate (DMP) phthalate (BBP)
Diethylhexyl phthalate (DEHP)
Molecular weight (g/mole)
194.2
312.4
390.6
Water solubility (mg/L)
2179 – 4320
0.67–40.2
0.0006–1.2
log Kow
1.46–1.90
3.57–4.91
4.2–8.39
Molecular structure
100
A.S. Mogensen et al.
2.4.2 Common Disposal Route and Environmental Concentrations
The use of PEs in a variety of industrial and domestic products makes them a common pollutant in solid and liquid waste and the environmental fate of PEs is linked to organic waste and wastewater. Landfilling, anaerobic sludge digestion and certain methods for treatment of industrial wastewater are processes where PE-containing waste is exposed to anaerobic environments. If the solid fraction of municipal solid waste (MSW) is landfilled it will be subjected to anaerobic microbial activity after an initial aerobic phase in the landfill.Also, incomplete aeration of the organic fraction of MSW during composting reveals anaerobic zones where anaerobic PE degradation can occur. PEs in wastewater will be removed during the primary settling when entering the wastewater treatment plant due to their physico-chemical properties, and will eventually be subjected to anaerobic environments during sludge stabilization or digestion. DEHP has been found in sludge at concentration levels of 120 mg/kg dry matter [98], and up to 100 mg/L of diethyl phthalate (DEP), DBP, and benzyl butyl phthalate (BBP) have been found in landfill leachate. Hence, contamination of groundwater and farmland receiving sewage sludge as a soil-improving agent is possible. 2.4.3 Anaerobic Degradation of Phthalate Esters
The information available concerning methanogenic degradation of phthalate esters is limited, although increased attention has been seen in recent years from studies in landfills, anaerobic reactors and enrichments. The molecular backbone of the PEs, i.e., the phthalic acid, allows some general predictions concerning the potential anaerobic metabolism, though the side chain that gives the specific PE its chemical characteristics also seems to be a very important factor in PE biodegradability. 2.4.3.1 Degradability
Ejlertsson found that dimethyl phthalate (DMP) was transformed anaerobically within two days using MSW as inoculum [103]. Phthalic acid was formed and eventually converted to methane and carbon dioxide. Other phthalates (diethyl phthalate, DEP, and BBP) were not completely converted, but transformed to the corresponding mono-ester, and no phthalic acid formation was noted. DEHP was the most recalcitrant PE, and no intermediates were detected. Inocula from environments exposed to PEs contained the appropriate microbiological activity for degradation of DMP and phthalic acid [99]. Both were degraded under mesophilic methanogenic conditions using inoculum from different anaerobic reactors. Degradation of 50% of 350 mg/L phthalic acid and 420 mg/L of DMP could be achieved after 16 days using inoculum from UASB and CSTR reactors, respectively. DMP was eventually mineralized completely.
Potential for Anaerobic Conversion of Xenobiotics
101
Chauret and coworkers showed that in anaerobic subsurface environments, DBP was biodegradable under different redox conditions at 10 °C, although transformation rates were low [104]. Biodegradation of PEs in the presence of other xenobiotics or mixtures of PEs has been achieved in anaerobic reactors and serum vials. For example, DBP at 100 mg/L was biodegraded up to 78% in an anaerobic expanded-bed GAC reactor fed with different semivolatile compounds [105], and mixing with volatile and semivolatile compounds resulted in 95% DBP biodegradation at 1 mg/L [106]. Parker and coworkers found between 83 and 99% removal of DEP, DBP, BBP, and DEHP during treatment in a two-stage anaerobic digester system [107], while Ejlertsson and coworkers observed only 15% degradation of DEHP with adapted, landfilled MSW [108]. Generally, DEHP transformation seems to occur more slowly than transformation of other PEs, but different removal rates may be attributed not only to the use of different microbial consortia but also differences in physico-chemical properties. A typical intermediate from biodegradation of PEs (see below) – phthalic acid – shows recalcitrance to degradation under methanogenic conditions. Fajardo et al. [109] found that phthalic acid was not degraded after 12 days of incubation with anaerobic sludge adapted for degradation of the para-isomer – terephthalic acid. Kleerebezem et al. [100] showed that enrichment cultures grown on either one of the three phthalic acid isomers (ortho, iso and tere) were incapable of degrading the others. Benzoate was accumulated during exponential growth on the ortho-isomer (PA). According to [110], phthalic acid was degraded by a syntrophic culture and with substrate utilization impeded when hydrogen was added, i.e., inhibition of fermentative bacteria. 2.4.3.2 Factors Affecting Degradability
Anaerobic degradation of phthalate esters is according to Ejlertsson et al. [111] positively related to water solubility; DBP, BBP, BEHP (butyl 2-ethylhexyl phthalate), and DHP (dihexyl phthalate) with relatively high water solubilities were biodegraded by a BBP-degrading enrichment culture, while DEHP with low water solubility was left unaltered. Additionally, Ejlertsson and coworkers showed that long side chains of the phthalate esters seemed to hamper anaerobic transformation [103]. The water solubility of DEHP is at least a ten-fold lower that the other phthalate esters examined (see Table 9), and it could be questioned whether DEHP is left unaltered due to lower bioavailability or higher toxicity towards the anaerobic bacteria. The fact that PEs with long side chains are more susceptible to sorption (more lipophilic) could explain their recalcitrance towards biodegradation. 2.4.3.3 Degradation Pathway
The anaerobic degradation of DEP, BBP, DBP and DEHP was investigated with acidogenic and methanogenic landfilled MSW and mesophilic biogas reactor samples as inoculum [108]. The MSW activity resulted in 20–25% transforma-
102
A.S. Mogensen et al.
MEHP & PA (µM)
Accumulated CH4 ( µmoles)
tion of the added DEP. Phthalic acid (ortho-phthalic acid) and monoethyl phthalate were formed in samples from the methanogenic landfill stage only [108]. Phthalic acid transformation did not occur in the beginning of the methanogenic landfill phase, but complete phthalic acid transformation was observed in samples taken after six months. The lack of phthalic acid transformation is attributed to either bacteria adaptation or a too high hydrogen partial pressure impeding fermentative growth. Samples taken at the stable methanogenic phase contained bacteria capable of transforming DEP, DBP, BBP (70–87 mg/L) to their corresponding mono-esters after 110 to 278 days. Phthalic acid was transformed completely after 110 days. Benzoate, phthalic acid, monobenzyl phthalate, and monobutyl phthalate were formed during BBP biotransformation. Complete transformation of phthalic acid, DEP, DBP, and BBP was observed after 40–60 days with inoculum from a mesophilic anaerobic digester. Mono(2-ethylhexyl) phthalate, (MEHP), a presumed degradation product of DEHP, has been detected in landfill leachate at 10–30 mg/L [112]. MEHP has also been shown to be hydrolytically transformed to phthalic acid with subsequent methane formation (see Fig. 12). According to Ejlertsson et al., phthalate esters are hydrolyzed to the corresponding phthalic acid and alcohol [108]. This was confirmed by Kleerebezem et al., who proposed that the DMP mineralization occurs through the corresponding mono-methyl ester, phthalic acid, with eventual formation of methane and carbon dioxide, see Fig. 13 [99]. While the alcohol formed is further metabolized to methane and carbon dioxide following b-oxidation by methanogenic bacteria, it is suggested that phthalate acid degradation follows the coenzyme A, benzoyl-CoA and benzoate pathway forming hydrogen and acetate as described by Heider and Fuchs (see Sect. 2.1). Thus, the anaerobic biotransformation of phthalic acids requires syntrophic interaction between hydrogenotrophic methanogens and fer-
Time (days) Fig. 12. Degradation of MEHP () to methane () via the transient intermediate phthalic acid
() as compared with methane formed in inoculated bottles without MEHP. From [112]. Copyright (1996), with kind permission from Kluwer Academic Publishers
Potential for Anaerobic Conversion of Xenobiotics
103
Fig. 13. Proposed pathway for anaerobic biodegradation DMP and BBP. The pathway is not balanced. Abbreviations: DMP – dimethyl phthalate, BBP – butyl benzyl phthalate, MMeP, monomethyl phthalate, MBeP – monobenzyl phthalate, MBuP – monobutyl phthalate, PA – phthalic acid ester. Reproduced from [99, 108]
mentative bacteria in order to keep the partial pressure of hydrogen low enough to make the benzoate transformation thermodynamically favorable. 2.5 Anaerobic Conversion of Polychlorinated Biphenyls
It is estimated that approximately 600 million kilograms of polychlorinated biphenyls (PCBs) have been produced worldwide and that several million kilograms have been released into the environment [113]. Commercial PCBs were used as dielectric fluids in capacitors and transformers and as flame retardants. They were manufactured as complex mixtures of chlorine-substituted biphenyl molecules, typically consisting of 60–90 of possibly 209 PCB congeners. These mixtures are distributed throughout the global ecosystem at relatively low concentrations, but can be found at much higher concentrations at specific locations, often sediments. 2.5.1 Physico-chemical Characteristics of PCBs
Polychlorinated biphenyls have a series of characteristics that make them a prime example of a class of compounds that are tremendously undesired in the envi-
104
A.S. Mogensen et al.
ronment. Due to their hydrophobicity, they adsorb to organic soils and sediments and tend to accumulate in biota. They have however also a series of very interesting characteristics such as a low chemical reactivity, heat stability, non-flammability, and high electrical resistance, hence their industrial applications.A significant fraction of the worldwide production of PCBs has ended up in the environment, where they turned out to be persistent and toxic. Therefore, their use is now stringently regulated, but we still have the heritage of past spills. 2.5.1.1 PCB Classification
PCBs can be classified into ortho or non-ortho congeners based on the presence or absence of ortho substituents. Because the electrons on the ortho chlorine repel the electron density on the aromatic phenyl ring, the conformation of the PCB molecule is determined by the presence of ortho chlorines. The meta and para chloro-substituents are, on the contrary, too far away from the aromatic nucleus to cause this effect. Thus, the angle between the two phenyl rings is determined by the number of ortho chlorines. In general, the PCBs with ortho chlorines are loosely referred to as non-planar and those without ortho chlorines are coplanar. The coplanar PCBs are of the most toxicological interest because they resemble, both in their structure and in their activities, the infamous 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD). The non-planarity of the ortho-substituted phenyl rings restricts their free spinning along the biphenyl-bridge. This conformation forces the electron cloud of the ortho chlorine on one phenyl ring to be hanging on top of the other phenyl ring, which is likely to hinder the reduction of the molecule [114]. 2.5.1.2 Prevalence of PCBs in the Environment
Typical concentrations of PCBs in surface sediments of rivers in Western Europe are at levels of approximately 500 mg/kg dry weight, while on the other hand the corresponding concentrations in the pore water are extremely low. For the heavy polluted sediments of the Hudson and the Sheboygan River, concentrations of 10,000 and 150,000 mg/kg have been reported [115]. These high concentrations are however significantly lower than the concentrations occasionally found biota, e.g., birds and marine animals [116]. Presently there is a slow but steady decrease of PCBs in most environments [117–119]. A long-term data set on PCB trends in Lake Superior indicates that PCBs are decreasing in the water column at a rate of 0.20 per year since 1980 with a water concentration of 10 ng S25 PCB/L in 1992. The decrease can be attributed primarily to volatilization [117, 120]. The decrease reflects the impact of the ban on PCB production implemented in 1979. Due to its remote location Lake Superior received most of its chemical inputs including PCBs via the atmosphere. Nowadays this relatively pristine environment acts however as a source rather than a sink for PCBs, illustrating that PCBs are volatile. The net effect of this phenomenon is that the higher latitudes act as the ultimate sink for PCBs due to the
Potential for Anaerobic Conversion of Xenobiotics
105
effect of lower temperatures on the air-water and air-terrestrial processes [121]. The long-range transport results in the presence of PCBs and other persistent semi-volatile compounds in remote areas on earth; global atmospheric transport processes result in “distillation” of semivolatile organic compounds from warmer source areas to the colder polar regions [121, 122]. As a result of their chemical characteristics PCBs are widely distributed in the environment. They are dispersed through the atmosphere and sorbed to particles in the water phase. In this way they frequently end up in sediments and they therefore are frequently measured in anaerobic sediments. Other sources, e.g., due to spills, are better defined, resulting in locally high concentrations. Interestingly, reductive dechlorination results in the formation of congeners with a higher aqueous solubility and a lower octanol/water partitioning coefficient [115, 123] which is an indication of the sediment/water partitioning coefficient and the related mobility of the compound in the aquatic and terrestrial environments [124]. Thus reductive dechlorination of highly chlorinated PCBs to lesser chlorinated congeners can be a blessing in disguise as it yields compounds that are more easily bioavailable and more mobile [115]. 2.5.2 Anaerobic Degradation by Reductive Dechlorination
Anaerobic biodegradation of PCBs is possible, but transformation rates in the environment are generally slow. Half-life times for PCBs in anaerobic sediments are in the order of years. The predominant conversion pathway of PCBs in such environments is via a series of reductive dechlorination reactions in which chlorine atoms are sequentially removed from the aromatic ring and replaced by hydrogen. Under environmentally realistic conditions, this reaction type is highly exergonic [123]. From a thermodynamic point of view, it is therefore to be expected that microorganisms can obtain energy for growth from catalyzing this reaction. To obtain enrichments capable of dechlorinating PCBs is indeed possible, but somehow isolation in pure form has been elusive so far. The cause is possibly that PCBs are not very bioavailable, which will affect diffusion and uptake rates. 2.5.2.1 Pathways
In the early days of research on anaerobic degradation of PCBs, much attention was paid to their dechlorination pathways. The knowledge gained was generally based on product patterns. It was concluded that the dechlorination pathways, designated “processes”, varied for different locations, and that they could not be adequately described by the simple terminology ortho, meta, and para dechlorination. Further refinements were deemed necessary to address the effect of the presence of neighboring chlorine substituents. Thus while, for example, meta dechlorination is known as process M, there is also a more restrictive process N in which only the flanked meta positions are dechlorinated. A detailed description of all processes and their nomenclature is given in a recent review by Bedard and Quensen [125].
106
A.S. Mogensen et al.
The most frequently observed dechlorination processes are, in addition to processes M and N, the processes Q (flanked and unflanked para, inclusive meta dechlorination in 2,3), P (flanked para) and C (flanked and unflanked meta and para, and hence a combination of M and Q), while ortho dechlorination is rare. The observed differences in pathways are ascribed to the presence of different organisms with different substrate specificities. In many cases, the habit to think in terms of processes does not allow conclusions on the pathway. In process M, for example, it is not clear which step is taken first, removal of the flanked or of the unflanked meta chlorines. One approach to predict such a preference would be by comparing the energetics of the various dechlorination steps. The hypothesis is then that a dechlorination step with a higher energy gain has preference over a dechlorination step with a lower energy gain [126]. Dechlorination studies with vitamin B12 have indicated the value of this approach: product ratios of PCBs are predicted well by the redox potential of the various redox couples [127]. The chemical system also showed that the formation of energetically less favorable reactions still occurred. 2.5.3 Priming
In the early 1990s there was a report that chlorobenzene-adapted cultures also showed substantial PCB degradation, but detailed information on cross adaptation in PCB dehalogenating cultures has been obtained only recently. In a series of studies with brominated biphenyls (PBBs) Bedard and coworkers have shown that it is possible to prime sediments for dechlorination of PCBs [128]. In all cases, the authors used sediments with a history of PCB contamination. To these sediments, PBBs were added in concentrations of up to 0.45 mM. These amounts were fairly rapidly (60 days) debrominated, and subsequently dechlorination of PCBs was enhanced. The best inducer of PCB dehalogenation was 2,6-dibromobiphenyl [128, 129]. The rational for this research was the development of strategies for the clean up of PCB-contaminated sediments. Priming with PBBs for the in situ treatment of PCB-contaminated sediments is inconvenient, but the approach may be useful for the enrichment of dechlorinating populations for subsequent use in on site bioreactors. Addition of PCB congeners and brominated (but not chlorinated or iodinated) chlorobenzoates stimulated specific dechlorination routes of PCBs [130, 131]. Compounds that prime dehalogenation of PCBs do not have to be their structural analogues. Brominated and iodinated benzoates, for example, primed dechlorination of PCBs in Housatonic River sediments while the fluorinated and chlorinated analogues were ineffective, even though they too were dehalogenated.Among the halogenated benzoates tested, 4-iodo- and 4-bromobenzoate, were the most effective, priming process N, which encompasses meta dechlorination. None was as effective though as 2,6-dibromobiphenyl. DeWeerd and Bedard rightly caution that the priming of PCB dechlorinating organisms with halobenzoates is less certain than priming with halobiphenyls. The possibility exists that multiple halobenzoate dehalogenating populations are present in sediment and that only a subset of these can dechlorinate PCBs. The
Potential for Anaerobic Conversion of Xenobiotics
107
carbon-iodide bond is very labile, so 4-iodobenzoate is potentially more susceptible to reductive dehalogenation by a greater number of sediment microorganisms. This decreases the likelihood that a microbial population capable of dehalogenating both 4-iodobenzoate and PCBs will be enriched by this compound. The use of halogenated benzoates instead of halogenated biphenyls offers several advantages over 2,6-dibromobiphenyl. The mineralization of benzoate is well documented but the removal of biphenyl has not been reported. Also, halogenated benzoates are less expensive than brominated biphenyls and are commercially available in bulk quantities. Finally, the use of brominated or iodinated benzoates may enrich dehalogenating populations as well as provide necessary carbon and electron flow in energy-starved environments. 2.5.3.1 Effect of FeSO4
Zwiernik and coworkers described a stimulating effect of FeSO4 addition on meta and para dechlorination of Aroclor 1242 in contaminated sediments [132]. The authors proposed that sulfate stimulated the growth of sulfate-reducing organisms responsible for PCB dechlorination, while Fe reduces sulfide toxicity by forming the insoluble precipitate FeS. The latter contention was supported by the fact that similar results were obtained with PbCl2/Na2SO4 . The authors hypothesize that once sulfate is consumed an increased number of sulfate reducers utilizes PCBs as an alternative electron acceptor. In the past, numerous reports have shown that the presence of sulfate inhibits PCB dechlorination and this phenomenon was also observed by Zwiernik and coworkers.Apparently, sulfate was preferred over PCBs as electron acceptor by potentially PCB dechlorinating, sulfate-reducing bacteria. Growth of these bacteria was however stimulated by the availability of sulfate, and once sulfate was depleted, the organisms switched back to using PCBs as electron acceptor, which proceeded at higher rates than in unamended controls. FeSO4 appears to be an effective, inexpensive, and innocuous amendment for stimulating PCB dechlorination. 2.5.3.2 Intrinsic Priming
Enhancement of PCB bioavailability can be labeled as “intrinsic priming”. The idea is to add an agent (a surfactant) to the soil or sediment in order to increase the concentration of PCBs in the liquid phase (see Sect. 2.1.1.3). An important consideration in such an approach is to work well above the critical micelle concentration. A potential problem may then be that the surfactants are frequently toxic (see Table 4). Thus, the outcome of such an approach may very well be a technology in which the PCB-containing micelles are washed out of the system and degraded in a bioreactor. Recently such a strategy was described for the release and subsequent (aerobic) degradation of polycyclic aromatic hydrocarbons [133]. The importance of (microbial) bioavailability was also demonstrated in a recent study [134] with sediments of Silver Lake. It was shown that residual petroleum in sediments reduces the bioavailability and hence the rate of reductive dechlorination of PCBs.
108
A.S. Mogensen et al.
Dissolved organic matter in pore water on the other hand enhances the aqueous solubility of PCBs [135]. To what extent the availability of PCBs for microbial degradation is increased by this process is not known. 2.5.3.3 Effect of Redox Conditions
Research by Alder and coworkers showed that addition of sulfate to a culture that dechlorinated PCBs under methanogenic conditions blocked dechlorination [136, 137]. It has been suggested that dechlorination did occur [138], but Bedard and Quensen stated that dechlorination occurred only after depletion of sulfate [125]. Similarly, they have stated that evidence suggests that addition of sulfate inhibited dechlorination in the experiments of [139]. The observed dechlorination may be due to dechlorination that occurred after sulfate had been depleted. The effect of redox conditions on the rate and pathway of Aroclor 1242 dechlorination by Hudson River microorganisms has showed that sulfidogenic conditions decreased dechlorination; nitrate-reducing conditions completely blocked dechlorination [140]. Future experiments should take these considerations into account and should preferably be carried out in homogeneous, sediment-free media to avoid the occurrence of microniches. 2.5.4 PCB-Dechlorinating Bacteria
Pure cultures of PCB-dechlorinating bacteria are not yet available, but in the last few years considerable progress has been made towards their identification. As indicated above, the occurrence of various dechlorination patterns and the possibility to prime dechlorination rather specifically all point towards the existence of specific groups. Other lines of evidence include the finding that pasteurization or enrichments started from the same inoculum, but set up at different temperatures led to rather specific dechlorinating populations [141–143].Wu et al found that the incubation temperature affects the dechlorination pattern of 2,3,4,6tetrachlorobiphenyl, which was attributed to the differences between sediments and to differences in microbial communities [129]. 2,3,4,6-Tetrachlorobiphenyl can prime dechlorination of Aroclor 1260 [144] and temperature determines the subsequent dechlorination pattern of this mixture at meta and para positions; ortho was not primed, even though 2,3,4,6-tetrachlorobiphenyl itself was eventually also ortho dechlorinated. Wu and coworkers used sediment for dechlorination of PCBs although this sediment had no known history of PCB contamination [130]. In some cases, it was possible to achieve dechlorination in sediment-free cultures [137, 145].Also, it is possible to obtain colonies on agar that subsequently dechlorinate PCBs in liquid culture [138]. The possibility to enrich is not only based on rates but also on increases in bacterial number (measured as MPN). Using a combination of inhibitors for eubacteria and substrates for methanogens Ye and coworkers have presented evidence that methanogenic bacteria eluted from a PCB-dechlorinating sediment are among the physiological groups capable to para dechlorinate PCBs [142]. There are however, no subse-
Potential for Anaerobic Conversion of Xenobiotics
109
quent reports that pure cultures of methanogens isolated from these sediments were capable of dechlorination of PCBs. To what extent methanogens are involved in dechlorination in real life is not known either. In principle, it is conceivable that these results of Ye are a laboratory artifact, i.e., that dechlorination of PCBs by methanogens is a co-metabolic side reaction that occurs only at low rates, and that para dechlorination of PCBs in this and other sediments is catalyzed by other physiological groups. A supportive evidence for a role of methanogens in PCB dechlorination is the observation of May and coworkers, who found that enrichments repeatedly plated on solid medium simultaneously lost the ability to produce methane and to para dechlorinate selected PCB congeners, while the ability to meta dechlorinate was not lost [138]. Pasteurization and ethanol treatment of Hudson River sediment has brought a non-methanogenic, spore-forming culture to the fore, which dechlorinates PCBs at the meta position [141], and it was speculated that sulfate reducers were responsible for dechlorination in this culture.Attempts to further enrich this culture by using sulfate have failed. Presently, all known spore-forming sulfate reducers belong to the genus Desulfotomaculum. The use of FeSO4 or specific PCB congeners and halobenzoates in conjunction with a source of reducing equivalents would enhance PCB degrading activity. Modern molecular techniques such as denaturing gradient gel electrophoresis (DGGE) [132, 146] and other cultureindependent surveys [147, 148] may help to identify the organisms responsible for the dechlorination of PCBs. 2.5.5 Field Tests
Evidence for in situ dehalogenation is hard to obtain since mass balances are hard to make and isolation of organisms is difficult. It is however possible to see shifts in the dechlorination pattern which in combination with activity in sediment cultures is good evidence. Tiedje and coworkers stated in 1993 that it, already then, was worthwhile to field test methods for reductive dechlorination of PCBs in cases where (i) the PCB concentration was in such a range that regulatory standards could be directly achieved by dechlorination, (ii) where a subsequent aerobic treatment is feasible, (iii) where any co-contaminants do not pose an inhibitory problem, and (iv) where anaerobic conditions can be established [149]. Since that time, we have learned how to prime populations, how to quantify dechlorinating organisms, how to selectively stimulate dechlorinators, and how to achieve complete dechlorination. Nevertheless, there are no convincing reports of successful anaerobic treatment of PCBs in the field. The only related work is that of Natarajan et al. [150] that is discussed in Sect. 4.2. 2.6 Halogenated Phenols
Chlorophenols are industrially produced by chlorination of phenol at high temperatures. They are also formed as by-products of chlorine bleaching in paper
110
A.S. Mogensen et al.
mills and during chlorination of municipal and industrial waste. In industry, chlorophenols are used to produce polymers, dyes, antiseptics and disinfectants. In Nature, halogenated phenols can be produced by a variety of organisms [151], including fungi [152], worms [153] and ticks. Thus halogenated phenols are not necessarily xenobiotic and it is not unexpected that microorganisms have developed catabolic activity to degrade these compounds. In the past, chlorinated phenols have been emitted into the environment on a rather large scale as waste from saw, pulp and paper mills. With our present day knowledge it is now possible to develop bioremediation strategies to clean up these polluted sites.Another obvious application where this knowledge is used effectively, is the design of reactor systems to clean up chlorophenol-containing industrial waste streams. Chlorophenols are on the lists of priority pollutants in most industrialized countries, although they are not known to be mutagenic of carcinogenic. Häggblom and Valo have recently reviewed the bioremediation of chlorophenol wastes with the emphasis on soil treatment. Intricacies of water treatment and intrinsic bioremediation were the focus there [154]. 2.6.1 Hydrophobicity of Chlorinated Phenols
Chlorophenols are hydrophobic and not very volatile [124, 155] with their log Kow values varying between 2 and 5. Due to their hydrophobicity they tend to sorb to organic matter such as, for example, biomass. This can be an issue when evaluating the efficiency of biodegradation in bioreactor systems, since the chlorophenols will be partly removed by physical processes (see Sect. 2.1.1.3). The acidity of the OH group depends on the degree of chlorination of the compound, and this factor also plays an important role in the sorption behavior of these compounds [124]. 2.6.2 Anaerobic Biodegradation of Halogenated Phenols
All 19 possible chlorophenol congeners are biodegradable under anaerobic conditions. The first step in their degradation pathway is generally a (series of) reductive dechlorination(s), resulting in the formation of phenol. These steps are exergonic and it has been shown that microorganisms are indeed able to obtain energy for growth from catalyzing them. Nevertheless, anaerobic incubation studies with soils and sediments frequently show that (some) chlorophenols are not readily degradable under the conditions imposed. More detailed studies are needed to explore the reason for this apparent contradiction; it is not clear whether the organisms that catalyze these reactions are not ubiquitous or whether certain environmental conditions impede the degradation of these compounds (see Sect. 2.1.1). The most likely explanation is that the organisms are not always present. In dechlorination of PCP is described by three different pathways, in which a chlorine atom is cleaved from the ortho-, meta-, or para position of the mother compound (PCP). The three different pathways are illustrated in Fig. 14.
Potential for Anaerobic Conversion of Xenobiotics
111
Fig. 14. Anaerobic PCP degradation pathways
In the first systematic study with sewage sludge, Shelton and coworkers found that 4-chlorophenol was recalcitrant, while 2- and 3-chlorophenols were degradable. In a more recent study, it was reported that chlorophenols were readily degradable in laboratory sludge that had not been exposed to synthetic chemicals [157]. With this sludge, the degradation of 4-chlorophenol was about ten times slower than the degradation of 2- and 3-chlorophenols.
112
A.S. Mogensen et al.
The presence of dechlorinating activity in the seed material is of course an important consideration in biodegradation studies as well as when starting up a bioreactor. Much attention has been paid to this issue and to the possibility to mix seed material with different (ortho- meta- and para-) dechlorinating activities to obtain complete dechlorination of multi-chlorinated phenols. Originally this issue was studied with digested sewage sludge as seed material, but recently granular sludge has been described that dechlorinate chlorinated phenols in UASB reactors. Dechlorinating activity does not necessarily have to be present in the seed material since incorporation of this activity into granular sludge is a viable strategy when dechlorinating granules are not readily available. Christiansen and Ahring, for example, have introduced pentachlorophenol-transforming activity into anaerobic granular sludge using the dechlorinating bacterium DCB-2 [158]. 2.6.3 Redox Conditions 2.6.3.1 Nitrate-Reducing Conditions
Dechlorination under nitrate-reducing conditions was present in soil tested by Sanford and Tiedje [159]. However, dechlorination activity could only be serially transferred in enrichments with an added electron donor such as acetate. Simultaneous dechlorination and denitrification was observed and could be main-
Table 10. Proven dechlorination of chlorinated phenols under various reducing conditions.
(+) indicates dechlorination. (–) indicates no dechlorination observed Compound
Fe3+
NO3–
SO42–
HCO3–
2-chlorophenol 3-chlorophenol 4-chlorophenol 2,3-dichlorophenol 2,4-dichlorophenol 2,5-dichlorophenol 2,6-dichlorophenol 3,4-dichlorophenol 3,5-dichlorophenol 2,3,4-trichlorophenol 2,3,5-trichlorophenol 2,3,6-trichlorophenol 2,4,5-trichlorophenol 2,4,6-trichlorophenol 3,4,5-trichlorophenol 2,3,4,5-tetrachlorophenol 2,3,4,6-tetrachlorophenol 2,3,5,6-tetrachlorophenol pentachlorophenol
+ + + – – – – – – – – – – – – – – – –
+ + + + + + + + – – – – – – – – – – –
+ – – – – – – – – – – – – – – – – – –
+ + + + + + + + + + + + + + + + + + +
Potential for Anaerobic Conversion of Xenobiotics
113
tained. Dechlorination and denitrification were mediated by two separate microbial communities; one that dechlorinated without use of nitrate and one that denitrified while oxidizing the dechlorinated aromatic ring. This is interesting, as in the environment the dechlorinator has to compete for reducing equivalents. A comparison with the nitrate/nitrite redox couple shows that the amount of energy is virtually identical for dechlorination and nitrate reduction. In Table 10 the chlorinated phenols, which until today have been shown to be dechlorinated is listed. 2.6.3.2 Iron-Reducing Conditions
In 1995 the first report in the degradation of monochlorophenols under iron-reducing conditions was published [160], but from this work it was not clear whether the first step in their degradation pathway was actually a reductive dechlorination step [161]. A subsequent study with brominated phenols showed indeed production of phenol as a transient intermediate, thus demonstrating that reductive dehalogenation is the initial step in the biodegradation of halophenols when their biodegradation is coupled to iron reduction [162]. 2.6.3.3 Sulfate-Reducing Conditions
In 1991 Madsen and Aamand reported that sulfate inhibited the dechlorination of PCP and its potential metabolites in a methanogenic PCP degrading enrichment culture [163]. The authors put forward the hypothesis that the inhibitory effect of sulfate was due to competition for reducing equivalents between sulfate reduction and dechlorination, the idea being that sulfate reduction would be energetically more favorable than reductive dechlorination. This is however not the case [155] and subsequent studies by Häggblom and Young [164] have shown that degradation of halogenated phenols by sulfate-reducing consortia is indeed possible, especially in estuarine and marine environments. Under sulfatereducing conditions reductive dechlorination as the initial step in chlorophenol degradation has been confirmed with the use of chlorofluorophenols [165]. Interestingly sulfate reduction was a sine qua non for dechlorination to occur in this culture. All three monochlorophenol isomers can be dechlorinated under sulfate-reducing conditions. Enrichment cultures have been maintained on these compounds for at least five years [164]. Surprisingly, there are no reports on dechlorination of multi-halogenated phenols under sulfate reducing conditions. 2.6.4 Pure Culture Studies
A series of organisms has been isolated that are able to grow with chlorinated phenols as electron acceptors (Table 11). This collection does however not cover all of the 19 individual chlorophenol congeners.
114
A.S. Mogensen et al.
Table 11. Pure cultures of microorganisms able to grow with chlorinated phenols as electron
acceptor Organism
Electron acceptor
Reference
Desulfitobacterium dehalogenans Desulfitobacterium strain PCE1 Desulfitobacterium hafniense Desulfitobacterium chlororespirans Desulfomonile tiedjei strain DCB-1 Desulfitobacterium frappieri Strain 2 CP-1
2,4-DCP 2-CP, 2,4,6-TCP 2,4-DCP, 3,5-DCP 2,3-DCP
[166] [167] [168] [169] [170] [171] [172]
2-CP
The first dechlorinating bacterium isolated, Desulfomonile tiedjei DCB-1, can dechlorinate the meta position of polychlorinated phenols, but this organism has not been shown to obtain energy from chlorophenol dechlorination, even though it can obtain energy for growth from the dechlorination of 3-chlorobenzoate [173–175]. In contrast to DCB-1, the second anaerobic dechlorinator to be isolated, DCB-2 [176], which was later characterized under the name Desulfitobacterium hafniense [168] is a spore-forming bacterium. Other members of this genus are Desulfitobacterium dehalogenans [166], Desulfitobacterium frappieri [171] and Desulfitobacterium sp. strain PCE1 [167], an anaerobic bacterium that was isolated for its ability to obtain energy for growth from the reductive dechlorination of tetrachloroethene, and was subsequently shown to also dechlorinate ortho-chlorinated phenols. In D. frappieri the rates of chlorophenol degradation are up to 1 mmol/min/g protein, i.e., up to 100-fold higher than in D. tiedjei [177]. D. frappieri appears to contain two enzyme systems for PCP dechlorination, one for ortho-dechlorination, and one for meta- and para-dechlorinations. The Desulfitbacteria are generally able to obtain energy for growth from the dechlorination of chlorophenols, but a co-metabolic dechlorination of pentachlorophenol by D. frappieri is not excluded [178]. In our opinion one of the most intriguing observations on the suite of dechlorinating organisms isolated thus far is that these organisms never mineralize their substrates. For the supply of reducing equivalents needed for the reductive dechlorination reaction these organisms all require an additional source of electrons. 2.6.5 Pathways
Kennes and coworkers worked with methanogenic pentachlorophenol-degrading granules from a laboratory-scale UASB reactor, and reported a pathway for PCP mineralization in which the parent compound was first meta dechlorinated to 2,4,6 TCP followed by ortho dechlorination to 2,4-DCP [179]. This compound was either ortho or para dechlorinated to 4- or 2-CP, respectively. These compounds were then dechlorinated to phenol, which was subsequently mineralized to methane and carbon dioxide. More interesting than this pathway per se, is that
Potential for Anaerobic Conversion of Xenobiotics
115
the presence of another pathway giving rise to the formation of 3,4,5-TCP as an intermediate could not be ruled out [180].A survey of the literature indeed shows that various pathways have been shown to occur, sometimes, as indicated above, even in parallel in the same macroscopic system [157, 181, 182]. The route is thus not necessarily explained by thermodynamics but the chemical characteristics of the substituent also play a role [115]. It is nowadays generally tacitly assumed that reductive dechlorination is the first step in the anaerobic degradation pathway of chlorinated phenols [161]. Recent work by Becker and coworkers however indicates that this is not necessarily correct. These authors showed that biotransformation of 2-chlorophenol by a methanogenic sediment community resulted in the transient accumulation of 3-chlorobenzoate, which was probably formed via a para carboxylation of 2-chlorophenol to 3-chloro-4-hydroxybenzoate followed by a dehydroxylation to 3-chlorobenzoate [183]. 2.6.6 Effect of Co-Substrates
Reductive dechlorination of course requires reducing equivalents. In the literature, much attention is paid to identify whether the addition of an external source of reducing equivalents stimulates dechlorination, but the results are contradictory. Sources that stimulate in one study can be neutral or even inhibitory in another study. The lesson to be learned is thus that the choice of external sources of reducing equivalents is case specific. Furthermore, it cannot be stressed enough that the dechlorination product phenol can be mineralized to carbon dioxide. The reducing equivalents that become available during this mineralization process can be used for the subsequent dechlorination of additional chlorophenol molecules. Thus far, no organisms have been described that can do more with the chlorophenol molecule than dechlorinate it. The transformation products are always (lesser chlorinated) phenol(s). Hence, the dechlorinating organisms will always have to compete with other organisms for (an external source of) reducing equivalents. In this context, it is therefore intriguing that a (mixed) culture was described in which 2-chlorophenol degradation resulted in the formation of acetate [184]. Thermodynamic calculations [155] indicate that this is indeed an exergonic reaction, which yields enough energy (182 kJ/mole 2-chlorophenol) to sustain growth of 2-chlorophenol fermenting organisms, either alone or in mixed culture. Even more energy would be obtained when the organisms would grow on multi-halogenated phenols. 2.6.7 Field Studies
Since the removal of a chlorine atom from the aromatic ring is a reductive process, the input of electrons is required. In the environment these electrons will have to come from other organic substrates. Hence, it seems logical that the extent and the rate of dechlorination are affected by the concentration and the degradability of organic substrates within a given environment. Gibson and Su-
116
A.S. Mogensen et al.
flita have indeed noted shorter acclimation periods for dechlorination when groundwater, collected from a methanogenic site down-gradient from a municipal landfill, was brown in color relative to water that was clear following rain [185]. The brown color was presumed to result from a higher organic content that consisted of leachate components and humic materials originating from the site. This observation suggests that in, e.g., marine and subsurface environments where marginally low levels of organic carbon are available, chlorophenol degradation may be severely limited by the level of carbon substrate.Addition of yeast extract to slurries made from subsurface material has indeed been shown to stimulate dechlorination of chlorophenols, but it cannot be excluded that this was due to the presence of some other nutrient in the yeast extract. In experiments with marine sediment slurries, with a dissolved organic carbon content of 1–2 mg/L, stimulation of chlorophenol degradation by yeast extract was not observed [186]. Thus dechlorination cannot always be predicted by the total organic carbon concentration at a site. An important ramification of this work is that it is reasonable to assume that the addition of an organic substrate in the form of yeast extract will have a large benefit in enhancing the dechlorination rate. The amount of substrate to be added may only have to be the amount needed to initiate dechlorination of the carbon backbone of the contaminant, which subsequently also can be degraded to produce electrons to support continued dechlorination. Thus ideally, i.e., when no competition for reducing equivalents would occur or when the dechlorinators would obtain all the reducing equivalents from the additional substrate, the amount of additional organic substrates required to support complete degradation of the chlorinated substrate would be determined from the difference between the number of electrons required for dechlorination and the number of electrons recovered from the carbon backbone. For phenols the latter number is 28, which is far more than the two electrons needed to remove one chlorine atom. Chlorophenols are generally readily degradable by slurries from anoxic estuarine sediments [182, 187]. The intermediates in these studies, which were conducted under sulfate-reducing conditions, as well as in a study with unadapted methanogenic sludge indicate that the chlorine atoms at the ortho position are the most easily dechlorinated, whereas the dechlorination rates at the para position are the lowest. This implies that the hypothesis that in microbial communities the thermodynamically most favorable pathways are selected is not applicable to chlorophenol transformation. 2.7 Anaerobic Degradation of Pesticides
Pesticides, also known euphemistically as crop protection agents, are used against a wide range of pests. Their targets and hence their structures are very diverse and it is therefore virtually impossible to generalize about their biodegradability. Nevertheless, it is possible to make some statements. Classical pesticides are frequently characterized by the presence of a chlorine group. Hence, these compounds, e.g., DDT and chlordane, are transformed under anaerobic conditions by means of reductive dechlorination reactions.
Potential for Anaerobic Conversion of Xenobiotics
117
2.7.1 Biodegradation Reactions
Three general types of reactions can be envisaged for pesticides: reductions, oxidations, and neutral reactions in which the oxidation state of the molecule remains unchanged. It has been stated that reduction of a pesticide usually gives rise to a product that is less polar, and hence less water-soluble than the parent compound [188], though halogenated pesticides tend to be exceptions to this rule. The metabolites formed during the removal of chlorine atoms by reductive dechlorination are more polar compounds [115] and therefore more watersoluble. Other examples of reductive reactions of importance for pesticides in the environment are sulfoxide reduction and nitro group reduction [189]. For these compounds, the metabolites tend to be of at least as much concern as the parent compound. Pentachloronitrobenzene, for example, is less toxic than its transformation product pentachloroaniline [190]. The reduction of the nitro group to an amine is a very common transformation reaction in anaerobic environments [191]. Neutral reactions of potential importance in anaerobic environments are nucleophilic substitutions by H2O. Examples are the transformation of aldicarb in aldicarb oxime, or the hydrolysis of 1,3-dichloropropylene to 3-chloroallyl alcohol [189]. It is not always clear to what extent transformation reactions are biologically catalyzed in the environment. In some cases the reactions are dependent on microbial activity, but only indirectly. An example of such a transformation is the conversion of ethylene dibromide into 1,2-dithioethane catalyzed by microbially produced HS– [192]. Another example, where microbial activity was required to obtain redox conditions conducive to anaerobic biodegradation of a herbicide, is the anaerobic degradation of the nitrogen heterocyclic herbicide picloram. This compound was transformed under methanogenic conditions in freshwater sediments, but transformation was inhibited by the presence of sulfate or nitrate [193]. 2.7.2 NSO Compounds
Trifluralin [2,6-dinitro-N,N-dipropyl-4-(trifluoromethyl)-benzenamine] ranks among the most commonly used pesticides in the United States. The compound is rapidly degraded in the presence of Fe(II), where its nitro groups are rapidly reduced to amines [194], see Fig. 15. This is a typical example of an environmental transformation reaction that can be catalyzed not only purely chemically but also by microorganisms. Be it directly or indirectly, microbial activity is always the driving force behind the occurrence of this reaction, since it is microorganisms that make the environmental conditions conducive to its incidence. From an environmental point of view, it is relevant that trifluralin degradation occurred with 3-fold lower rates under nitrate-reducing conditions than under iron-reducing conditions, since nitrate is a frequent contaminant in (anaerobic) groundwater.
118
A.S. Mogensen et al.
Fig. 15. Time course for electron acceptors [NO3–/Fe(III)] and trifluralin degradation (% of total) in aerobic and flooded soil. a Loss of nitrate () and accumulation of Fe(II) (). b Trifluralin loss in flooded () soils versus aerobic soils (). Reprinted with kind permission of [194]. Copyright (2000) American Chemical Society
Potential for Anaerobic Conversion of Xenobiotics
119
Adrian and Suflita have studied anaerobic biodegradation of a series of N-, S-, and O-heterocyclic compounds in aquifer slurries [195].Well-known examples of this class of compounds are atrazine and triazine. The authors found that nonsubstituted and oxygenated heterocyclic compounds tended to resist anaerobic biodegradation whereas carboxylated and oxygenated compounds were susceptible. The brominated heterocyclic compounds were similar to their homocyclic counterparts, in that the latter were more readily dehalogenated than the corresponding chlorinated analogues. Cross acclimation to 4-chloro- or 5-chlorouracil from microorganisms acclimated to the biodegradation of 5-bromouracil was not observed. Others have on the other hand found that microbial populations acclimated to the biodegradation of a particular substrate can metabolize other structurally related compounds [196, 197]. In spite of some well-documented reports on the anaerobic transformation of atrazine there is, however, extensive field evidence that this compound is persistent in the environment. Thus, it is clearly necessary to determine the extent to which the concentrations of triazines and other herbicides are influenced by redox conditions in groundwater. 2.7.3 Chlorinated Pesticides
Halogenated compounds, especially the highly chlorinated ones, are however rather insoluble in water, while having a high affinity for the organic carbon that is an inherent part of the soil matrix. Thus, the actual concentrations in solution are relatively low, which reduces their bioavailability, and impedes induction of the enzymes required to catalyze degradation. Generally reductive dechlorination is the first step in the degradation pathway of chlorinated compounds under strict anaerobic conditions (sulfate-reducing or methanogenic conditions). A known exception to this rule is dicamba (3,6-dichloro-2-methoxybenzoic acid). Dicamba is first converted into 3,6-dichlorosalicylic acid before being dechlorinated [198]. The presence and nature of other aryl substituents impacts the susceptibility of substrates to dehalogenation. Aryl dehalogenation reactions seem to proceed easier when the ring is substituted with electron destabilizing groups such as carboxy, hydroxy, or cyano groups [190]. An important issue is that the redox conditions can affect the rate and extent of pesticide degradation in anaerobic environments. Milligan and Häggblom, for example, found that dicamba was completely mineralized under methanogenic conditions, see Fig. 16 [199]. Nitrate, on the other hand, inhibited dicamba degradation in their cultures. This finding has environmental implications, especially for agricultural areas where dicamba is used extensively and where nitrogen from nitrate in groundwater is often high. The possibility for transport of dicamba in soils, resulting in subsequent groundwater pollution is potentially high. Both dicamba and its initial transformation product 3,6-dichlorosalicylate have pKa values of 1.95. The high solubility of these weak acids at neutral to high pH makes them mobile in most soils. Biodegradation of dicamba under aerobic and methanogenic condi-
120
A.S. Mogensen et al.
Fig. 16. O-Demethylation and dechlorination of dicamba under methanogenic conditions. In methanogenic golf course sediment microcosms (20% inoculum), dicamba () was Odemethylated to 3,6-dichlorosalicylate (), 6-chlorosalicylate (), and salicylate (), which was depleted by day 135. There was no loss in the dicamba sterile control ().Values represent the means of triplicate cultures (standard deviation ±10 M). Reprinted with kind permission of [199]. Copyright (1999) American Chemical Society
tions is well documented, and the problematic issue is thus the recalcitrance of the intermediate under certain redox conditions. The transformation pathways observed by Milligan and Häggblom are shown in Fig. 17. Another point from this study is that 6-chlorosalicylate tends to accumulate when dicamba is degraded. It took a long period before complete degradation occurred. Another compound that is of wide interest is lindane (gammahexachlorocyclohexane). This compound was recently shown to be dehalogenated by anaerobic bacteria from marine sediments and by sulfate-reducing bacteria [200]. 2.7.4 Pesticides in Co-Digesters
A rather neglected issue is the anaerobic degradation of plant-protecting agents in anaerobic digesters. Post-harvest fungicides used in fruit production such as ortho-phenylphenol and thiabendazole have been found in high concentrations in southern fruit peels. When these fruit peels are used for biogas production in combination with manure in co-digestion systems the plant protecting agents should be degraded otherwise the farmers will be reluctant to use the
Potential for Anaerobic Conversion of Xenobiotics
121
Fig. 17. Transformation and degradation pathways of dicamba under different reducing conditions. Reprinted with kind permission of [199]. Copyright (1999) American Chemical Society
digested product as fertilizer. Preliminary studies however indicate that these compounds as well as other frequently used pesticides such as benomyl and endosulfan were not degraded [201]. Thus there is a need for cultures that can degrade these compounds in order to inoculate co-digesters exposed to such compounds.
3 Treatment of Xenobiotics in Bioreactors The performance of anaerobic reactor systems has improved steadily in the recent decades and the range of treated recalcitrant waste and wastewater has broadened significantly. Hence, the impact of anaerobic waste treatment systems on waste treatment is becoming more and more important as a result of improved reactor design, operating conditions and the use of specialized microbial consortia [202]. Many of the bacteria in anaerobic reactor systems can catalyze different transformation reactions leading to the formation of biogas; e.g., hydrolysis, decarboxylation, dechlorination, demethoxylation, and deamination. Regarding the dechlorination of PCP, continuous-tank reactors [203], biofilm reactors [204, 205] and UASB reactors [206] have all shown dechlorination activity. The operating conditions of these systems are profoundly different, and this will obviously affect the population build up and the stability of the treatment process. Regarding the relationship between the composition of the microflora and the stability of the treatment process, it is interesting to note that the microflora in apparently stable methanogenic bioreactors showed dramatic changes in population without any effect on the treatment efficiency of the system.
122
A.S. Mogensen et al.
3.1 The UASB Reactor
A number of studies on the anaerobic degradation of chlorinated aromatics in UASB reactor systems has been summarized by Christiansen [206], see Table 12. In general, an almost complete removal of the chlorinated aromatics has been obtained using different synthetic waste types and occasionally with addition of an easily degradable carbon source. Hendriksen and coworkers worked with a UASB and a fixed film reactor at an HRT of 2 days and PCP concentrations of up to 4.5 mg/L. They found that the source of the inoculum greatly influenced the general performance and the stability of the different reactor systems [209]. In addition, the UASB reactor proved to be more stable than the fixed film reactor.A degradation of 52% of 2,4-DCP was reported for treatment of a synthetic wastewater amended with 2,4-DCP and glucose in a UASB reactor with organic loading rate of 0.18 kg/m3/day and a retention time of 26 hours [210]. A series of trichlorinated phenols was reported to be removed in a UASB reactor [211] at an HRT of 1 day with efficiencies of 50 to 99% at specific loading rates of up to two mg chlorinated phenol per gram VSS (biomass) per day. Sorption of chlorophenols to active dechlorinating granules was insignificant. There are indications that the UASB reactor is superior to the other reactor types regarding dechlorination efficiency [209]. The results indicated above show that complete degradation is not always achieved, the reactors are not always stable, and that the loading rates vary. The important conclusion from this research is that efficient anaerobic treatment of chlorinated phenols in bioreactors is possible but far from trivial. Thus, it is important to direct future work towards optimization strategies, by formation of new granules, or by incorporation of dechlorinating organisms into existing granules. Anaerobic degradation of linear alkylbenzene has been investigated in many anaerobic digesters, though assessment of its bioavailability was generally absent. The bioavailability is maximized using the UASB reactor, and Mogensen and coworkers showed that anaerobic LAS degradation is indeed possible. Improved degradation was observed under thermophilic conditions compared to mesophilic conditions [57, 58, 212]. Under thermophilic conditions the C12 homologue of LAS was removed by 50–85% with a maximum removal rate of 5.7 µg/hour/ mL biomass in a lab-scale reactor. Benzaldehyde (see Fig. 18) and benzene sulphonic acid were found as intermediates [212]. Table 12. Summary of studies in which the UASB reactors was used for degradation of chlori-
nated aromatics. Data from [156] Compound
Influent conc. [mg/L]
Removal
Auxiliary carbon source
Reference
PCP 3-CP PCP PCP
3.2–8.1 10, 20 0.1 4.5
35% 100%, > 90% 95% 99%
none none acetate, methanol, glucose glucose
[207] [208] [205]
123
Potential for Anaerobic Conversion of Xenobiotics Abundance
Average of 5.218 to 5.224 min.: 0202004.D 77 106
8000 6000 4000 50 2000 37 44 m/z Æ
63
88 94
0 40
60 80 #63686: Benzaldehyde 77
Abundance
100 106
8000 6000 51 4000 2000
29
39 45
63
85 91
0 m/z Æ
40
60
80
100
Fig. 18. Mass spectrum of a sample from a UASB reactor compared with a benzaldehyde reference spectrum. From [212]
Anaerobic transformation of p-cresol in UASB reactors has been studied by Kennes and coworkers [213]. In one study, p-cresol was added to a synthetic media (200–650 ppm) and fed to two UASB reactors with a hydraulic retention time of 0.67–0.8 days. The first reactor was amended solely with p-cresol whereas the second one was amended with volatile fatty acids in addition to p-cresol (1:1). The study showed an overall degradation of p-cresol from 80–100%, with highest degradation when VFA was included in the reactor feeding. Treatment of biomass gasification wastewater with 13 g/L phenols was attempted using a mesophilic UASB reactor inoculated with pulp and paper mill granular biomass [214]. Methyl- and methoxy-substituted phenols were degraded up to a degree of 100%. The EC50 value of the gasification wastewater on nitrifying sludge changed from 0.6 mL/L prior to treatment to 51.7 mL/L after UASB reactor treatment. Hence, an overall detoxification of the wastewater was achieved. 3.2 Continuous Stirred-Tank Reactor (CSTR)
Studies on the transformation of xenobiotic compounds during anaerobic digestion of sewage sludge have received increased attention since sewage sludge was applied as a soil-improving agent on a larger scale. Sewage sludge contains a wide range of xenobiotics, e.g., phthalates, PAHs, and surfactants, and today the
124
A.S. Mogensen et al.
Fig. 19. Influent and effluent concentrations of C12-LAS in mg/L from a laboratory-scale CSTR reactor fed with digested sewage sludge. From [215]
adverse effects of certain xenobiotics on sludge amended farmland remain unknown. The biodegradation of xenobiotics present in solid waste, e.g., sewage sludge, treated in continuous stirred-tank reactors is therefore crucial in order to recycle organic waste on farmland. In a laboratory scale reactor, C12-LAS (100 mg/L) was added to a sludge slurry and fed to a mesophilic CSTR reactor having a hydraulic retention time of 15 days [215]. Haagensen and coworkers observed 25% transformation in the study, being the first to show anaerobic transformation of LAS in a continuous reactor system. LAS degradation was related to its bioavailability to the microorganisms. In Fig. 19, the influent and effluent concentrations from the study are shown during a test period of approximately 90 days. 3.3 Other Anaerobic Reactors
A fluidized bed granular activated carbon (GAC) bioreactor showed 99.9% PCP removal efficiency at PCP loadings of 4 g/kg GAC per day and influent concentrations of 1.3 g/L [216]. The main transformation product of PCP, 4-chlorophenol, inhibited methanogenesis at concentrations as low as 116 mg/L. Toxicity was minimized by controlling the influent PCP concentration, since this was the primary source of the inhibitory products. Toxicity of PCP and its degradation products was also the subject of a study [217]. These authors found that 35 mg/L PCP
Potential for Anaerobic Conversion of Xenobiotics
125
allowed greater than 99.5% glucose removal in chemostats. Subsequent studies with batch cultures however showed that, in serum bottle experiments, this PCP concentration partially inhibited glucose degradation. Sorption to biomass was the dominant mechanism for PCP removal in these acidogenic cultures. It cannot be concluded that acidogenic cultures are not conducive to dechlorination of phenols, but the results indicate that further research on factors stimulating PCP degradation in anaerobic non-methanogenic environments is still necessary. Complete (> 99%) degradation of pentachlorophenol (0.015 kg/m3/day) in a fixed film reactor was reported [218] at a hydraulic retention time of 17 hours. An interesting observation in this study was also that in a semi-continuous reactor PCP biodegradation was unstable and necessitated periodic additions of unacclimated anaerobic waste sludge to restore the activity. In continuous flow reactors PCP degradation activity was more stable when a mixture of sodium formate was used as secondary source of carbon and energy instead of glucose alone. A classical anaerobic fixed-bed bioreactor was constructed and fed with a mixture of hydroxylated aromatic compounds [219]. The methanogenic consortia contained in the reactor was able to transformed a toxic mixture of 2,4,6-TCP, 1-naphthol and 2-nitrophenol at 25 mg/L, 18 mg/L, and 22 mg/L per day, respectively. The bioreactor was run at 35 °C, 22 days retention time and citrate was added as an auxiliary carbon and energy source.Another interesting observation in this context was made by Themel and coworkers who report the anaerobic dehalogenation of 2-chlorophenol by mixed bacterial cultures in the absence of methanogenesis [184]. Acetate was found to be the end product of chlorophenol degradation in their cultures, which did not require sulfate or nitrate as terminal electron acceptor. A removal of over 80% of 2,4,6-TCP in a fluidized-bed reactor was reported [220] at influent concentrations of 25 mg/L; in this study no accumulation of lower chlorinated phenols was observed. Mohn used sewage sludge to start up a fluidized-bed reactor for the degradation of PCP [170]. It was observed that it was necessary to reinoculate the reactor with monochlorophenol-fed enrichments. After 8 months of operation more than 96% dechlorination and 68% mineralization of PCP was observed at an influent concentration of 2 mg/L. In another recent study dechlorination rates of 2,4,6-TCP were shown to decrease significantly at higher agitation rates [221]. This is in agreement with the concept that dechlorinating organisms require reducing equivalents and “nutrients” from the other members of the microbial community, and that fixed spatial positions are conducive to efficient exchange of these and other products. In addition to HRT and influent concentration the specific chlorophenol loading rate is also a relevant parameter. 3.4 Conclusions
As illustrated above, there is a widespread applicability of anaerobic reactor systems, which can treat and biodegrade a vast number of xenobiotics in different types of waste; e.g., sludge, soil, and water. Important aspect, when addressing anaerobic microbial processes for treatment of xenobiotics, are factors
126
A.S. Mogensen et al.
such as type of xenobiotics, the type of media, bioavailability of the xenobiotics, and presence or availability of microorganisms with the required metabolic capabilities.
4 Bioaugmentation Bioaugmentation as a bioremediation technology is applied to enhance the microbial activity in bioreactors and at contaminated sites. It counteracts the lack of activity by the indigenous microorganisms caused by unfavorable environmental parameters impeding their growth, low number of microorganisms, or absence of the required microbial activity in the soil or bioreactor. In the latter case, bioaugmentation might be a valuable tool for remediation, since microorganisms with specific properties are added to the contaminated site or reactor in order to enhance the required and often very specific activity. A contrast to bioaugmentation is the use of the indigenous microorganisms. Their acclimation to a change in the environment due to, e.g., a contamination is remarkable, since the environmental stress imposed on the microflora is believed to enhance mutation rates. Therefore, certain bioremediation procedures proceed without bioaugmentation. The physicochemical properties of the xenobiotics subjected to bioremediation are of primary importance, and their bioavailability should be considered prior to any bioremediation, including bioaugmentation. Obviously, strongly sorbed chemicals such as PCBs and PAHs do not become more available for biodegradation by addition of bacteria, but other means should be applied. The factors influencing microbial activity are discussed in Sect. 2.1.1. Furthermore, the microbial toxicity of the xenobiotics, soil characteristics, microbial ecology (e.g., presence of predators, interspecies competition), and bioaugmentation methodology are factors different at each contamination occurrence, and specific bioaugmentation strategies are therefore required. 4.1 In Situ Bioaugmentation
Reports on successful anaerobic bioaugmentation projects are scarce, but recently a successful anaerobic bioaugmentation was carried out on a trichloroethene (TCE)-contaminated aquifer at Dover Air Force Base in the USA [222], where the indigenous bacteria were unable to dechlorinate TCE beyond cis1,2-dichloroethene (c-DCE). The bioaugmentation consisted of the injection nutrients, substrates, and an enrichment culture capable of reductively dechlorinating TCE to ethene [223]. In the project, carried out by the Remediation Technologies Development Forum, a consortium composed of industrial corporations and government agencies, it was shown that the injected culture survived in the new environment and that it was physically transported throughout the study area. After a lag period of about 90 days, vinyl chloride and ethene began to appear in the wells. By day 509, TCE and cis-DCE were fully converted to ethene. This was the first report of a successful bioaugmentation study of this group. As recently as 1998 the same group discussed the perspectives on microbial de-
Potential for Anaerobic Conversion of Xenobiotics
127
halogenation of chlorinated solvents, and in that treatise bioaugmentation was not mentioned [224]. Another example of a successful in situ bioaugmentation pilot was the introduction of Pseudomonas stutzeri KC, a denitrifying bacterium that degrades carbon tetrachloride, in a nitrate- and carbon tetrachloride-contaminated aquifer in Michigan. The demonstration showed subsurface transport of strain KC, assimilation of the organism into the aquifer community, and removal of carbon tetrachloride [225]. 4.2 Bioaugmentation in Reactors
Natarajan and coworkers have developed PCB-dechlorinating granular sludge in a UASB reactor, which when added to sediment catalyzes extensive reductive dechlorination of PCBs, including dechlorination at the ortho position [150, 226]. The granules were developed by incorporating bacteria from PCB-dechlorinating enrichments into pentachlorophenol-dechlorinating granular sludge. Thus these granules dechlorinate both PCBs and PCP. The granules appear to be very stable. They have been maintained for more than 7 years now without any loss of activity. The Michigan Biotechnology Institute where the technology is being developed reports that this bioaugmentation method has been shown to also work in other PCB-contaminated environments [227]. The process is expected to be cheaper than landfilling, thermal-desorption and incineration methods, but recently doubts were raised about the cost effectiveness of this approach [228]. Definitely exciting is the report of Natarajan and coworkers that these granules (which are available from the American Type Culture Collection under catalogue number ATCC 55616) can mineralize biphenyl to methane and carbon dioxide [229]. Another approach to treat PCB-contaminated sediment would be to use an anaerobic bioreactor system. This was done by using sanitary landfill leachate as a novel nutrient, carbon, and/or microbial source [230]. This was the first reported occurrence of anaerobic dechlorination of PCB-contaminated sediment in a low-cost bioreactor system, but the overall success of the system was rather limited. Another example, also on a laboratory scale, is the novel soil bioreactor developed by Agathos and coworkers in Belgium. These authors showed that addition of Desulfomonile tiedjei to a 3-chlorobenzoate-containing soil in a pilotscale-sized bioreactor with a volume of 500 liters enhanced dechlorination of 3-chlorobenzoate. D. tiedjei was shown to be able to maintain itself in the soil, but only at the top surface [231] of the soil. The results of bioaugmentation studies are often unequivocal. Beaudet and coworkers, for example, observed a distinct increase in the rate of PCP removal when a PCP-contaminated soil was inoculated with D. frappieri [232]. After 21 days however the difference was negligible. The experiments were performed in soil slurries under an anaerobic headspace. Molecular techniques were used to show that D. frappieri was present at the same level throughout the three weeks of biotreatment.
128
A.S. Mogensen et al.
In another bioaugmentation experiment, D. frappieri was added to granular sludge in a UASB reactor. Its population increased from 106 to 1010 cells/g of volatile suspended solids within a period of 70 days. This proliferation was paralleled by a substantial increase in the volumetric PCP load from 5 to 80 mg/L of reaction volume/day. A PCP removal efficiency of 99% and a dechlorination efficiency of at least 90.5% were observed throughout the experiment, with 3-chlorophenol and phenol as dechlorination intermediates [178].
5 Future Perspectives The thesis that microbes are capable of degrading every compound that Nature (including man) can synthesize is an overstatement, but the last decade has shown us that it is true for a remarkable subset of even xenobiotic compounds. Whether a compound is xenobiotic or not is indeed no longer a true a false discussion but also a matter of concentration: some formerly xenobiotic compounds are today known to be produced naturally [233], but generally in extremely low concentrations. Thus, the discussion on the xenobiotic nature of a compound is becoming academic. What counts in terms of practical (environmental and biotechnological) purposes is whether microorganisms can degrade that compound or not. The experience that most compounds are indeed biodegradable does not of course imply that the compound is actually biodegraded in the environment. For biotechnological clean-up processes biodegradability is encouraging but still only a first step. The challenges are (i) to find the requisite “best” microorganisms and (ii) to develop systems in which these microorganisms function as we want them to. An interesting example of the first challenge is the degradation of perchloroethylene, a frequent pollutant of soils and groundwater. Perchloroethylene is degraded via a series of reductive dechlorination steps to ethene and ethane, with vinyl chloride as a prominent and frequently rate-limiting intermediate. Thermodynamic calculations however indicate that reductive dechlorination is not the only route by which the organisms can obtain energy for growth on chlorinated ethylenes, and there is indeed evidence that certain microorganisms use these alternative pathways [234]. The organisms themselves have been elusive thus far. Our knowledge of reactor systems in general, and the UASB reactor in particular, has been developed sufficiently to encourage us in meeting the second challenge, especially now that we are able to introduce new degradative capacities into the microbial consortia on which these reactor systems depend [158]. Thus far, introduction of new degradative capacities has focused on, but not been limited to, the introduction of one organism or activity. Mineralization processes in anaerobic ecosystems are however integrated processes involving obligatory interactions between different groups of organisms. Thus, introducing the capacity to mineralize, e.g., halogenated compounds by appropriate consortia is more elegant than merely introducing the capacity to dehalogenate these compounds. Another aspect that deserves more attention is the degradation of mixtures of “xenobiotic” compounds [235]. Finally, an inventory of microorganisms that degrade (all) the compounds that need to be degraded would provide
Potential for Anaerobic Conversion of Xenobiotics
129
us with an excellent tool when facing a clean up of an environmental contamination. With the use of molecular techniques, we can improve our understanding of population dynamics in, e.g., reactor systems. Just recently, we have learned that even apparently stable systems can experience large changes in the microbial population [236]. We are also beginning to learn that microorganisms (in biofilms) communicate in subtle ways, and there is every reason to believe that better insights into their language will eventually allow us to develop more stable systems [237, 238].
6 References 1. Alexander M (1998) Biodegradation and bioremediation, 2nd edn. Academic Press, San Diego 2. Wania F, Mackay D (1996) Environ Sci Technol 30:390 3. Alexander M (1994) Biodegradation and bioremediation. Academic Press, San Diego 4. Coates JD,Woodward J,Allen J, Philp P, Lovley DR (1997) Appl Environ Microbiol 63:3589 5. Coates JD, Anderson RT, Lovley DR (1996) Appl Environ Microbiol 62:1099 6. Rockne KJ, Chee-Sanford JC, Sanford RA, Hedlund BP, Staley JT, Strand SE (2000) Appl Environ Microbiol 66:1595 7. Genthner BS (1999) Biodegradation 10:27 8. Nyman MC, Nyman AK, Lee LS, Nies LF, Blatchley ER (1997) Environ Sci Technol 31:1068 9. Ronen Z, Bollag JM (1992) Appl Microbiol Biotechnol 37:264 10. Aeckersberg F, Rainey FA, Widdel F (1998) Arch Microbiol 170:361 11. Anderson RT, Lovley DR (2000) Nature 404:722 12. Phelps CD, Young LY (1999) Biodegradation 10:15 13. Kazumi J, Caldwell ME, Suflita JM, Lovley DR, Young LY (1997) Environ Sci Technol 31:813 14. Lovley DR, Woodward JC, Chapelle FH (1994) Nature 370:128 15. Lovley DR, Coates JD, Woodward JC, Phillips EP (1995) Appl Environ Microbiol 61:953 16. Eschenbach A, Wienberg R, Mahro B (1998) Environ Sci Technol 32:2585 17. White JC, Kelsey JW, Hatzinger PB, Alexander M (1997) Environ Toxicol Chem 16:2040 18. White JC, Alexander M (1996) Environ Toxicol Chem 15:1973 19. Bouchez M, Blanchet D, Vandecasteele JP (1995) Appl Microbiol Biotechnol 43:952 20. Haagensen F (2000) Master Thesis. Technical University of Denmark 21. Heider J, Fuchs G (1997) Anaerobe 3:1 22. Heider J, Fuchs G (1997) Eur J Biochem 243:577 23. Harwood CS, Gibson J (1997) J Bacteriol 179:301 24. Harwood CS, Burchhardt G, Herrmann H, Fuchs G (1998) FEMS Microbiol Rev 22:439 25. Boll M, Fuchs G (1995) Eur J Biochem 234:921 26. Cullum DC (1994) Introduction. In: Cullum DC (ed), Introduction to Surfactant Analysis. Blakie Academic & Professional, London 27. Di Corcia A, Costantino A, Crescenti C, Marinoni E, Samperi R (1998) Environ Sci Technol 32:2401 28. Haigh SD (1996) Sci Total Envir 185:161 29. Patel P, Madamwar D (1998) Process Biochem 33:199 30. White GF, Russel NJ (1994) Biodegradation of anionic surfactants and related molecules. In: Ratledge C (ed), Biochemistry of Microbial Degradation. Kluwer Academic Publishers, Dordrecht 31. Swisher RD (1987) Metabolic pathways and ultimate biodegradation. In: Surfactant biodegradation. Marcel Dekker, New York 32. Field JA, Field TM, Pioger T, Siegrist H, Giger W (1995) Water Res 29:1301 33. Painter HA, Zabel T (1989) Tenside Surfactants Deterg 26:108
130
A.S. Mogensen et al.
34. Berna JL, Ferrer J, Moreno A, Prats D, Bevia FR (1989) Tenside Surfactants Deterg 26:101 35. Mackay D, di Guardo A, Paterson S, Kicsi G, Cowan CE, Kane DM (1996) Environ Toxicol Chem 15:1638 36. Garcia MT, Campos E, Sanchez-Leal J, Ribosa I (1999) Chemosphere 38:3473 37. Ejlertsson J, Nilsson ML, Kylin H, Bergman A, Karlson L, Oquist M, Svensson BH (1999) Environ Sci Technol 33:301 38. Bennie DT (1999) Water Qual Res J Can 34:79 39. Greiner P, Six E (1997) Tenside Surfactants Deterg 34:250 40. McAvoy DC, Dyer SD, Fendinger NJ, Eckhoff WS, Lawrence DL, Begley WM (1998) Environ Toxicol Chem 17:1705 41. Prats D, Ruiz F, Vazquez B, RodriguezPastor M (1997) Water Res 31:1925 42. Maguire RJ (1999) Water Qual Res J Can 34:37 43. Ahel M, Giger W, Koch M (1994) Water Res 28:1131 44. Holt MS, Bernstein SL (1992) Water Res 26:613 45. Brunner PH, Capri S, Marcomini A, Giger W (1988) Water Res 22:1465 46. Jensen J (1999) Sci Total Envir 226:93 47. Field JA, Barber LB, Thurman EM, Moore BL, Lawrence DL, Peake DA (1992) Environ Sci Technol 26:1140 48. Salanitro JP, Diaz LA (1995) Chemosphere 30:813 49. Gerike P, Klotz H, Kooijman JA, Matthijs E, Waters J (1994) Water Res 28:147 50. Nuck BA, Federle TW (1996) Environ Sci Technol 30:3597 51. Wagener S, Schink B (1987) Water Res 21:615 52. Tanaka S, Ichikawa T (1993) Water Sci Tech 28:103 53. Shcherbakova VA, Laurinavichius KS, Akimenko VK (1999) Chemosphere 39:1861 54. Yeh DH, Pennell KD, Pavlostathis SG (1998) Water Sci Tech 38:55 55. Remde A, Debus R (1996) Chemosphere 32:1563 56. Denger K, Cook AM (1999) J Appl Microbiol 86:165 57. Mogensen AS, Ahring BK (2002) Biotechnol Bioeng 77:483 58. Berna JL, Battersby N, Cavalli L, Fletcher R, Guldner A, Schowanek D, Steber J (2001) Tenside Surfactants Deterg 38:86 59. Verge C, Moreno A, Bravo J, Berna JL (2001) Chemosphere 44:1749 60. Macur RE, Inskeep WP (1999) Environ Toxicol Chem 18:1927 61. Volkering F, Breure AM, Rulkens WH (1998) Biodegradation 8:401 62. Federle TW, Schwab BS (1992) Water Res 26:123 63. Mikosch CARM, Denger K, Schafer EM, Cook AM (1999) Microbiology-Uk 145:1153 64. Servos MR (1999) Water Qual Res J Can 34:123 65. Cerniglia CE (1992) Biodegradation 3:351 66. Volkering F, Breure AM, van Andel JG (1993) Appl Microbiol Biotechnol 40:535 67. Stringfellow WT, Aitken MD (1995) Appl Environ Microbiol 61:357 68. Mahajan MC, Phale PS, Vaidyanathan CS (1994) Arch Microbiol 161:425 69. Kalf DF, Crommentuijn T, van de Plassche EJ (1997) Ecotoxicol Environ Safety 36:89 70. Dean-Ross D, Cerniglia CE (1996) Appl Microbiol Biotechnol 46:307 71. Connell DW, Hawker DW, Warne MStJ, Vowles PP (1997) Polycyclic aromatic hydrocarbons. In: Connell DW (ed), Basic Concepts of Environmental Chemistry. Lewis, New York 72. Cuypers MP, Grotenhuis JT, Rulkens WH (1998) Water Sci Tech 37:157 73. Johnson K, Ghosh S (1998) Water Sci Tech 38:41 74. Wong CC, Yeung HY, Cheung RH, Yung KL, Wong MH (2000) Arch Environ Contam Toxicol 38:486 75. Yang GP (2000) Environ Pollut 108:163 76. Fernandes MB, Sicre MA, Broyelle I, Lorre A, Pont D (1999) Hydrobiologia 410:343 77. Maier M, Maier D, Lloyd BJ (2000) Water Res 34:773 78. Aislabie J, Balks M, Astori N, Stevenson G, Symons R (1999) Chemosphere 39:2201 79. Berset JD, Holzer R (1999) J Chromatogr A 852:545 80. Trapido M (1999) Environ Pollut 105:67 81. Guerin TF (1999) J Hazard Mater 65:305
Potential for Anaerobic Conversion of Xenobiotics
131
82. Paterson IF, Chowdhry BZ, Leharne SA (1999) Chemosphere 38:3095 83. Shuttleworth KL, Cerniglia CE (1995) Appl Biochem Biotech 54:291 84. Genthner BS, Townsend GT, Lantz SE, Mueller JG (1997) Arch Environ Contam Toxicol 32:99 85. Durant ND, Wilson LP, Bouwer EJ (1995) J Contam Hydrol 17:213 86. Bregnard TP, Höhener P, Häner A, Zeyer J (1996) Environ Toxicol Chem 15:299 87. McNally DL, Mihelcic JR, Lueking DR (1999) Chemosphere 38:1313 88. Langenhoff AA, Zehnder AJ, Schraa G (1996) Biodegradation 7:267 89. Grishchenkov VG, Townsend RT, McDonald TJ, Autenrieth RL, Bonner JS, Boronin AM (2000) Process Biochem 35:889 90. Rockne KJ, Strand SE (1998) Environ Sci Technol 32:3962 91. Lovley DR, Anderson RT (2000) Hydrogeol J 8:77 92. Coates JD, Anderson RT, Woodward J, Phillips EJ, Lovley DR (1996) Environ Sci Technol 30:2784 93. Bedessem ME, SwobodaColberg NG, Colberg PS (1997) FEMS Microbiol Lett 152:213 94. Zhang XM, Young LY (1997) Appl Environ Microbiol 63:4759 95. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:667 96. Ejlertsson J, Houwen FP, Svensson BH (1996) Swed J Agr Res 26:53 97. Balafas D, Shaw KJ, Whitfield FB (1999) Food Chem 65:279 98. Schnaak W, Kuchler T, Kujawa M, Henschel KP, Sussenbach D, Donau R (1997) Chemosphere 35:5 99. Kleerebezem R, Pol LH, Lettinga G (1999) Biodegradation 10:63 100. Kleerebezem R, Pol LH, Lettinga G (1999) Appl Environ Microbiol 65:1152 101. Roslev P, Madsen PL, Thyme JB, Henriksen K (1998) Appl Environ Microbiol 64:4711 102. Colborn T, Saal FV, Soto AM (1993) Environ Health Perspect 101:378 103. Ejlertsson J, Johansson E, Karlsson A, Meyerson U, Svensson BH (1996) Anton Leeuwenhoek Int J Gen M 69:67 104. Chauret C, Inniss WE, Mayfield CI (1996) Ground Water 34:791 105. Narayanan B, Suidan MT, Gelderloos AB, Brenner RC (1993) Water Res 27:171 106. Narayanan B, Suidan MT, Gelderloos AB, Brenner RC (1995) Water Environ Res 67:46 107. Parker WJ, Monteith HD, Melcer H (1994) Water Res 28:1779 108. Ejlertsson J, Meyerson U, Svensson BH (1996) Biodegradation 7:345 109. Fajardo C, Goyot JP, Macarie H, Monroy O (1997) Water Sci Tech 36:83 110. Kleerebezem R, Pol LH, Lettinga G (1999) FEMS Microbiol Ecol 29:273 111. Ejlertsson J, Alnervik M, Jonsson S, Svensson BH (1997) Environ Sci Technol 31:2761 112. Ejlertsson J, Svensson BH (1996) Biodegradation 7:501 113. Hutzinger O, Safe S, Zitko V (1983) The chemistry of PCBs. Robert Krieger Publishing Company, Malabar, FL 114. Yak HK, Lang QY, Wai CM (2000) Environ Sci Technol 34:2792 115. Dolfing J, Beurskens JM (1995) Advan Microbial Ecol 14:143 116. Pearson CR (1982) In: Hutzinger O (ed), The Handbook of Environmental Chemistry. Springer, Berlin 117. Jeremaison JD, Hornbuckle KC, Eisenreich SJ (1994) Environ Sci Technol 28:903 118. Jones KC, Duarte-Davidson R, Cawse PA (1995) Environ Sci Technol 29:272 119. Harner T, Mackay D, Jones KC (1995) Environ Sci Technol 29:1200 120. Jeremaison JD, Eisenreich SJ, Baker JE, Eadie BJ (1998) Environ Sci Technol 32:3249 121. Wania F, Mackay D (1993) Ambio 22:10 122. Muir DG, Omelchenko A, Grift NP, Savoie DA, Lockhart WL, Wilkinson P, Brunskill GJ (1996) Environ Sci Technol 30:3609 123. Holmes DA, Harrison BK, Dolfing J (1993) Environ Sci Technol 27:725 124. Schwarzenbach RP, Gschwend PM, Imboden DM (1993) Environmental Organic Chemistry. Wiley, New York 125. Bedard DL, Quensen JF (1995) In: Young LY, Cerniglia CE (ed), Microbial transformation and degradation of toxic organic chemicals. Wiley-Liss, New York 126. Dolfing J, Harrison BK (1993) FEMS Microbiol Ecol 13:23
132
A.S. Mogensen et al.
127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139.
Woods SL, Trobaugh DJ, Carter KJ (1999) Environ Sci Technol 33:857 Bedard DL, Van Dort H, Deweerd KA (1998) Appl Environ Microbiol 64:1786 Wu QZ, Bedard DL, Wiegel J (1999) Environ Sci Technol 33:595 Wu QZ, Bedard DL, Wiegel J (1997) Appl Environ Microbiol 63:2836 Deweerd KA, Bedard DL (1999) Environ Sci Technol 33:2057 Zwiernik MJ, Quensen JF, Boyd SA (1998) Environ Sci Technol 32:3360 Pinto LJ, Moore MM (2000) Environ Toxicol Chem 19:1741 Zwiernik MJ, Quensen JF, Boyd SA (1999) Environ Sci Technol 33:3572 Hunchak Kariouk K, Schweitzer L, Suffet IH (1997) Environ Sci Technol 31:639 Alder AC, Haggblom MM, Oppenhelmer SR,Young LY (1993) Environ Sci Technol 27:530 Wu QZ, Sowers KR, May HD (2000) Appl Environ Microbiol 66:49 May HD, Boyle AW, Price WA, Blake CK (1992) Appl Environ Microbiol 58:4051 Morris PJ, Mohn WW, Quensen JF, Tiedje JM, Boyd SA (1992) Appl Environ Microbiol 58:3088 Rhee GY, Bush B, Bethoney CM, Denucci A, Oh HM, Sokol RC (1993) Environ Toxicol Chem 12:1033 Ye DY, Quensen JF, Tiedje JM, Boyd SA (1992) Appl Environ Microbiol 58:1110 Ye DY, Quensen JF, Tiedje JM, Boyd SA (1995) Appl Environ Microbiol 61:2166 Ye DY, Quensen JF, Tiedje JM, Boyd SA (1999) Appl Environ Microbiol 65:327 Wu QZ, Bedard DL, Wiegel J (1997) Appl Environ Microbiol 63:4818 Cutter L, Sowers KR, May HD (1998) Appl Environ Microbiol 64:2966 Muyzer G, Dewaal EC, Uitterlinden AG (1993) Appl Environ Microbiol 59:695 Holoman TP, Elberson MA, Cutter LA, May HD, Sowers KR (1998) Appl Environ Microbiol 64:3359 von Wintzingerode F, Selent B, Hegemann W, Gobel UB (1999) Appl Environ Microbiol 65:283 Tiedje J, Quensen JF, Chee-Sanford JC, Schimel JP, Boyd SA (1993) Biodegradation 4:231 Natarajan MR, Nye J, Wu WM, Wang H, Jain MK (1997) Biotechnol Bioeng 55:182 Hoekstra EJ, De Weerd H, De Leer EB, Brinkman UT (1999) Environ Sci Technol 33:2543 Ando K, Kato K, Susuki S (1970) Biochem Biophys Res Commun 39:1104 Steward CC, Dixon TC, Chen YP, Lovell CR (1995) Can J Microbiol 41:637 Haggblom MM,Valo RJ (1995) In: Young LY, Cerniglia CE (ed), Microbial transformation and biodegradation of toxic organic chemicals. Wiley-Liss, New York Dolfing J, Harrison BK (1992) Environ Sci Technol 26:2213 Christiansen N, Mathrani I, Ahring BK (1997) Granular sludge consortia for bioremediation. In: Sheehan D (ed), Methods in Biotechnology, vol 2: Bioremediation Protocols. Humana Press Inc, Totowa, NJ Takeuchi R, Suwa Y, Yamagishi T, Yonezawa Y (2000) Chemosphere 41:1457 Christiansen N, Ahring BK (1996) Anton Leeuwenhoek Int J Gen M 69:61 Sanford RA, Tiedje JM (1996) Biodegradation 7:425 Kazumi J, Haggblom MM, Young LY (1995) Appl Environ Microbiol 61:4069 Dolfing J (1996) Appl Environ Microbiol 62:3554 Monserrate E, Haggblom MM (1997) Appl Environ Microbiol 63:3911 Madsen T, Aamand J (1992) Appl Environ Microbiol 58:557 Haggblom MM, Young LY (1995) Appl Environ Microbiol 61:1546 Haggblom MM (1998) FEMS Microbiol Ecol 26:35 Utkin I, Woese C, Wiegel J (1994) Int J Syst Bact 44:612 Gerritse J, Renard V, Gomes TP, Lawson PA, Collins MD, Gottschal JC (1996) Arch Microbiol 165:132 Christiansen N, Ahring BK (1996) Int J Syst Bact 46:442 Sanford RA, Cole JR, Loffler FE, Tiedje JN (1996) Appl Environ Microbiol 62:3800 Mohn WW, Kennedy KJ (1992) Appl Environ Microbiol 58:1367 Bouchard B, Beaudet R, Villemur R, McSween G, Lepine F, Bisaillon JG (1996) Int J Syst Bact 46:1010 Cole JR, Cascarelli AL, Mohn WW, Tiedje JM (1994) Appl Environ Microbiol 60:3536
140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172.
Potential for Anaerobic Conversion of Xenobiotics
173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219.
133
Dolfing J, Tiedje J (1987) Arch Microbiol 149:102 Dolfing J (1990) Arch Microbiol 153:26 Mohn WW, Tiedje J (1990) Arch Microbiol 153:267 Madsen T, Licht D (1992) Appl Environ Microbiol 58:2874 Dennie D, Gladu I, Lepine F, Villemur R, Bisaillon JG, Beaudet R (1998) Appl Environ Microbiol 64:4603 Tartakovsky B, Levesque MJ, Dumortier R, Beaudet R, Guiot SR (1999) Appl Environ Microbiol 65:4357 Kennes C, Wu WM, Bhatnagar L, Zeikus JG (1996) Appl Microbiol Biotechnol 44:801 Wu WM, Bhatnagar L, Zeikus JG (1993) Appl Environ Microbiol 59:389 Law RJ, Kelly CA, Nicholson MD (1999) Polycycl Aromat Compound 17:229 Masunaga S, Susarla S, Gundersen JL, Yonezawa Y (1996) Environ Sci Technol 30:1253 Becker JG, Stahl DA, Rittmann BE (1999) Appl Environ Microbiol 65:5169 Themel K, Sparling R, Oleszkiewicz J (1996) Environ Technol 17:869 Gibson SA, Suflita JM (1990) Appl Environ Microbiol 56:1825 Hale RC, Aneiro K (1997) J Chromatogr A 774:79 Liu SM, Kuo CE, Hsu TB (1996) Chemosphere 32:1287 Somasundaram L, Coats JR (1991) Pesticide Transformation Products: Fate and significance in the environment. In: ACS Symposium Series 459. American Chemical Society, Washington DC Barbash D, Resek EA (1996) Pesticides in Ground Water. Ann Arbor Press, Chelsea, MI Kuhn EP, Suflita JM (1989) Reactions and movements of organic chemicals in soils. Soil Science Society of America and American Society of Agronomy, Madison, WI Suflita JM, Townsend GT (1995) In: Young LY, Cerniglia CE (ed), Microbial transformation and degradation of toxic organic chemicals. Wiley-Liss, New York Barbash JE, Reinhard M (1989) Environ Sci Technol 23:1349 Ramanand K, Nagarajan A, Suflita JM (1993) Appl Environ Microbiol 59:2251 Tor JM, Xu CF, Stucki JM, Wander MM, Sims GK (2000) Environ Sci Technol 34:3148 Adrian NR, Suflita JM (1994) Environ Toxicol Chem 13:1551 Peijnenburg WM, Thart MJ, Denhollander HA, Vandemeent D, Verboom HH, Wolfe NL (1992) Environ Toxicol Chem 11:289 Struijs J, Rogers JE (1989) Appl Environ Microbiol 55:2527 Taraban RH, Berry DF, Berry DA, Walker HL (1993) Appl Environ Microbiol 59:2332 Milligan PW, Haggblom MM (1999) Environ Sci Technol 33:1224 Boyle AW, Haggblom MM, Young LY (1999) FEMS Microbiol Ecol 29:379 Forster M, Vorkamp K, Taube J, Herrmann R (1999) Water Sci Tech 40:371 Verstraete W, Vandevivere P (1999) Crit Rev Environ Sci Technol 29:151 Guthrie EA, Pfaender FK (1998) Environ Sci Technol 32:501 Hakulinen R, Salkinoja-Salonen MM (1982) Process Biochem 18: Hendriksen HV, Larsen S, Ahring BK (1992) Appl Environ Microbiol 58:365 Christiansen N, Hendriksen HV, Järvinen KT, Ahring BK (1995) Water Sci Tech 312:249 Krumme ML, Boyd SA (1988) Water Res 22:171 Woods SL, Ferguson JF, Benjamin MM (1989) Environ Sci Technol 23:62 Hendriksen HV, Ahring BK (1993) Biodegradation 3:399 Atuanya EI, Purohit HJ, Chakrabarti T (2000) World J Microbiol Biotechnol 16:95 Droste RL, Kennedy KJ, Lu JG, Lentz M (1998) Water Sci Tech 38:359 Mogensen AS, Haagensen F,Ahring BK.Accepted for publication in Environ Toxicol Chem Kennes C, Mendes R, Lema JM (1998) Water Res 31:1549 Mogensen AS (2001) PhD Thesis. Technical University of Denmark Haagensen F, Mogensen AS,Angelidaki I,Ahring BK (2002) Water Sci Tech 46(10):159–165 Khodadoust AP, Wagner JA, Suidan MT, Brenner RC (1997) Water Res 31:1776 Piringer G, Bhattacharya SK (1999) Water Res 33:2674 Juteau P, Beaudet R, McSween G, Lepine F, Milot S, Bisaillon JG (1995) Appl Microbiol Biotechnol 44:218 El Fantroussi S, Giot R, Naveau H, Agathos SN (1998) Chemosphere 36:1575
134
A.S. Mogensen et al.: Potential for Anaerobic Conversion of Xenobiotics
220. Vallecillo A, Garcia-Encina PA, Pena M (1999) Water Sci Tech 40:161 221. Chang BV, Chiang CW, Yuan SY (1999) J Environ Sci Health B-Pestic 34:491 222. Ellis DE, Lutz EJ, Odom JM, Buchanan RJ, Bartlett CL, Lee MD, Harkness MR, Deweerd KA (2000) Environ Sci Technol 34:2254 223. Harkness MR, Bracco AA, Brennan MJ, Deweerd KA, Spivack JL (1999) Environ Sci Technol 33:1100 224. Lee MD, Odom JM, Buchanan RJ (1998) Annu Rev Microbiol 52:423 225. Dybas MJ, Barcelona M, Bezborodnikov S, Davies S, Forney L, Heuer H, Kawka O, Mayotte T, Sepulveda-Torres L, Smalla K, Sneathen M, Tiedje J, Voice T, Wiggert DC, Witt ME, Criddle CS (1998) Environ Sci Technol 32:3598 226. Natarajan MR, Wu WM, Nye J, Wang H, Bhatnagar L, Jain MK (1996) Appl Environ Microbiol 46:673 227. Natarajan MR, Wu WM, Wang H, Bhatnagar L, Jain MK (1998) Water Res 32:3013 228. Wiegel J, Wu QZ (2000) FEMS Microbiol Ecol 32:1 229. Natarajan MR, Wu WM, Sanford R, Jain MK (1999) Biotechnol Lett 21:741 230. Pagano JJ, Scrudato RJ, Roberts RN, Bemis JC (1995) Environ Sci Technol 29:2584 231. El Fantroussi S, Belkacemi M, Top EM, Mahillon J, Naveau H, Agathos SN (1999) Environ Sci Technol 33:2992 232. Beaudet R, Levesque MJ,Villemur R, Lanthier M, Chenier M, Lepine F, Bisaillon JG (1998) Appl Microbiol Biotechnol 50:135 233. Keppler F, Eiden R, Niedan V, Pracht J, Scholer HF (2000) Nature 403:298 234. Dolfing J (2000) Microb Ecol 40:2 235. Fathepure BZ, Vogel TM (1991) Appl Environ Microbiol 57:3418 236. Fernandez A, Huang SY, Seston S, Xing J, Hickey R, Criddle C, Tiedje J (1999) Appl Environ Microbiol 65:3697 237. Davies DG, Parsek MR, Pearson JP, Iglewski BH, Costerton JW, Greenberg EP (1998) Science 280:295 238. Watnick P, Kolter R (2000) J Bacteriol 182:2675
Received: April 2002
CHAPTER 1
Monitoring and Control of Anaerobic Reactors Peter F. Pind 1 · Irini Angelidaki 1 · Birgitte K. Ahring 2 · Katerina Stamatelatou 3 · Gerasimos Lyberatos 3 1 2 3
Environment and Resources DTU, Technical University of Denmark, Building 115, 2800 Lyngby, Denmark. E-mail:
[email protected] BioCentrum-DTU, Technical University of Denmark, Building 227, 2800 Lyngby, Denmark. E-mail:
[email protected] Department of Chemical Engineering, University of Patras, 26500, Patras, Greece. E-mail:
[email protected]
The current status in monitoring and control of anaerobic reactors is reviewed. The influence of reactor design and waste composition on the possible monitoring and control schemes is examined. After defining the overall control structure, and possible control objectives, the possible process measurements are reviewed in detail. In the sequel, possible manipulated variables, such as the hydraulic retention time, the organic loading rate, the sludge retention time, temperature, pH and alkalinity are evaluated with respect to the two main reactor types: highrate and low-rate. Finally, the different control approaches that have been used are comprehensively described. These include simple and adaptive controllers, as well as more recent developments such as fuzzy controllers, knowledge-based controllers and controllers based on neural networks. Keywords. Anaerobic Reactors, Control methods, Control objectives, Control variables, Process
outputs
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
1
Introduction
1.1
Reactor and Waste Types
2
Basic Structure of a Control System
3
Process Measurements . . . . . . . . . . . . . . . . . . . . . . . . 144
3.1 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.1.6 3.1.7 3.1.8 3.2 3.3 3.3.1 3.3.2 3.3.3 3.3.4
Gas Production and Gas Composition . . . . . . . Gas Flow . . . . . . . . . . . . . . . . . . . . . . . Methane and Carbon Dioxide . . . . . . . . . . . . Carbon Dioxide and Bicarbonate . . . . . . . . . . Hydrogen Gas . . . . . . . . . . . . . . . . . . . . Hydrogen in the Liquid Phase . . . . . . . . . . . . Hydrogen Sulfide . . . . . . . . . . . . . . . . . . . Carbon Monoxide . . . . . . . . . . . . . . . . . . Electronic Noses for Gas Measurements . . . . . . Intermediate Species . . . . . . . . . . . . . . . . . Inorganic Chemical Components and Their Activity Redox Potential . . . . . . . . . . . . . . . . . . . . Ammonia . . . . . . . . . . . . . . . . . . . . . . . pH . . . . . . . . . . . . . . . . . . . . . . . . . . . Alkalinity . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . 137 . . . . . . . . . . . . . . . . 142 . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
145 145 146 146 147 148 149 149 149 150 152 152 153 154 154
Advances in Biochemical Engineering/ Biotechnology, Vol. 82 Series Editor: T. Scheper © Springer-Verlag Berlin Heidelberg 2003
136
P.F. Pind et al.
3.4 3.5 3.5.1 3.5.2 3.5.3
Indirect Measurements of Organic Matter Metabolic Activity Measurements . . . . Microbial Techniques . . . . . . . . . . . Molecular Techniques . . . . . . . . . . Chemical Indicators . . . . . . . . . . .
4
Manipulated Variables . . . . . . . . . . . . . . . . . . . . . . . . 160
4.1 4.2 4.3 4.4 4.5
Hydraulic Retention Time, Sludge Retention Time and Organic Loading Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bicarbonate Alkalinity . . . . . . . . . . . . . . . . . . . . . . Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . Waste-Management or Co-Digestion . . . . . . . . . . . . . .
5
Control Approaches
5.1 5.2 5.3 5.3.1 5.3.2 5.3.3
Simple Controllers . . . . . . . . Adaptive Controllers . . . . . . . Other Control Schemes . . . . . Fuzzy Logic Controllers . . . . . Knowledge-Based Expert Systems Neural Networks . . . . . . . . .
6
Concluding Remarks
7
References
. . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
155 157 157 158 158
160 161 162 163 163
. . . . . . . . . . . . . . . . . . . . . . . . . 164 . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
164 168 173 173 175 176
. . . . . . . . . . . . . . . . . . . . . . . . 178
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Abbreviations BA BOD COD CSTR GC GCMS HPLC HRT OFMSW OLR PID SRT TA TOC TS TSS UASB VFA VS VSS
. . . . .
Bicarbonate Alkalinity Biological Oxygen Demand Chemical Oxygen Demand Continuously Stirred-Tank Reactor Gas Chromatograph(y) Gas Chromatography-Mass Spectrometry High-Pressure Liquid Chromatograph(y) Hydraulic Retention Time Organic Fraction of Municipal Solid Waste Organic Loading Rate Proportional Integral Differential Sludge Retention Time Total Alkalinity Total Organic Carbon Total Solids Total Suspended Solids Up-flow Anaerobic Sludge Blanket Volatile Fatty Acids Volatile Solids Volatile Suspended Solids
Monitoring and Control of Anaerobic Reactors
137
1 Introduction This review will cover recent developments in monitoring high-rate anaerobic reactors such as the up-flow anaerobic sludge blanket (UASB), the fluidized beds and anaerobic filters as well as low-rate systems, such as the continuously stirredtank reactor (CSTR). Special reactor types such as membrane reactors and hybrids are not included in this review, although many aspects discussed are readily applicable for such systems as well. Application of process control theory will be described in general terms, and examples of use will be referred to. The reader is referred to the literature for a more detailed description of control theories. It is assumed that the reader has a common knowledge on the anaerobic process. The main emphasis will be placed on the process outputs used to gain information on the anaerobic process and the control variables used to control the evolution of the process. The anaerobic degradation process is a highly complex and dynamic process, which makes it quite difficult to control in a simple manner. Many applications of the anaerobic process have been made through time, giving many reactor concepts and treating many different kinds of wastes and organic matter. Reactor design and waste type play an important role in the ability to monitor and control the anaerobic process [1]. It is, therefore, essential to understand the limitations and differences in reactor designs and waste types, in order to be able to make appropriate comparisons between results obtained in different systems. 1.1 Reactor and Waste Types
Anaerobic reactor concepts can generally be divided into two types of systems; 1. “high-rate” with a hydraulic retention time (HRT) less than 5 days; 2. “low-rate” with HRT > 5 days and usually more than 10 days. High-rate systems are characterized by having a special process design allowing retention of the viable biomass despite low HRT (often below 24 hours). The most common high-rate systems are the UASB, the expanded/fluidized-bed and the anaerobic filter reactors (Fig. 1). The biomass inside the UASB reactor is retained
Fig. 1. Principle diagram for: UASB, fluidized-bed reactors, anaerobic filter and CSTR
138
P.F. Pind et al.
by forming granular sludge particles of high density. This unique biomass formation makes it possible to withstand high concentrations of otherwise inhibitory compounds. At the same time, some waste compositions can cause deterioration of granules or prevent renewal of the granules. The application of UASB reactors is, therefore, highly dependent on the waste characteristics, generally requiring low concentrations of suspended solids and good capability for granulation. Low retention times can cause sudden shifts in concentration profiles inside the reactor (both build-up and wash-out). The need for monitoring and controlling the feed is, therefore, essential when dealing with UASB reactors. Sampling and analysis of a UASB reactor effluent is often reasonably feasible due to the naturally low concentration of suspended material. However, the effluent composition does not necessarily reflect the effect of chemical composition on different regions of the biomass and the reactor. Due to the high density of the granules and the packing of the reactor with granules, significant gradients may be observed throughout the reactor. Control of the process behavior is most often limited to controlling the organic loading rate (OLR) by substituting waste with recycled effluent, and by changing the characteristics of the waste by addition of other components. Control experiments with UASB reactors can be found in the literature [2]. Modifications of the basic UASB concept, such as increased retention of suspended solids and alternative gas separation designs, have been made in an effort to improve the reactor design, giving thus rise to new hybrid reactors [3, 4]. The fluidized-bed reactor concept uses a special distribution network for the inlet flow, allowing a bed of fine-grained medium to expand or fluidize inside the reactor. The need for the biomass to attach to the carrier medium calls for careful selection of carrier media depending on the specific wastes [3]. Due to the high density of the carrier material, introducing suspended solids, as opposed to the UASB reactor, does not necessarily wash out the fluidized-bed biomass. The need for monitoring and control is otherwise much the same as for UASB reactors, except that the HRT is often smaller and the mixing is closer to perfect. Sampling from the recycle stream is very often used, generally giving a good picture of the chemical concentration affecting the biomass inside the reactor. Process control of the same type as in UASB can be used but has to be more rapid in response due to the low HRT. Therefore, a wide range of control design experiments has been carried out [5–15]. Anaerobic filters differ from UASB reactors and fluidized beds, as they use almost complete entrapment of the biomass, employing a densely packed support material or the capability to form biofilms on support material in low turbulence regions. This special characteristic gives rise to a pronounced plug-flow, which again leads to a concentration gradient inside the reactor. In order to minimize the gradient effect, recycling can be used, but the volumetric loading has to be kept low in order to prevent wash-out of the biomass. Separating the process into a series of smaller filter reactors can also be a solution. Clogging inevitably becomes a major problem, due to low flow rates and densely packed material. Monitoring of the anaerobic filters is, therefore, often more difficult and may require sampling from various different locations in the reactor. Control of the process is in the same manner restricted to manipulating the organic loading rate (OLR),
Monitoring and Control of Anaerobic Reactors
139
the HRT and the waste composition. Only few monitoring and control experiments have been carried out on anaerobic filters [16, 17], mainly focusing on alkalinity. High-rate reactor systems are mostly used to treat industrial wastewater containing easily degradable soluble organics, usually with low alkalinity. To maintain constant loading of the reactor, an equalizing tank is often placed before the reactor. The retention time of the equalization tank can vary from a few hours to 72 hours. Despite the presence of an equalization tank, the treatment of the waste can only be delayed for relatively short periods of time. Changing (lowering) the dilution rate can, therefore, only be accomplished through by-passing part of the waste. This is not desirable, and controlling the process stability through other control variables is therefore needed. The equalization tank is sometimes used as a hydrolysis step resulting in conversion of some of the organic matter to volatile fatty acids (VFA).When fermentation is allowed to take place in the equalization tank, the anaerobic treatment facility is commonly called a two-step process. Production of VFA without further conversion together with low alkalinity can cause a drop in the pH of the waste. Low pH from the equalization tank weakens the ability of the main reactor to withstand organic overloading without pH control. This variance in pH is typically overcome by using a separate pH controller on the equalization tank. The use of pH control and more general control schemes of high-rate reactor systems will be covered in more detail later. An extensive review on the start-up and operation of high-rate anaerobic reactor systems may be found in [1, 3] and a more recent review of future aspects on hybrids of high-rate systems can be found in [4]. Low-rate anaerobic reactor systems are most typically CSTRs, which have the benefit of being able to handle high concentrations of suspended matter and even fibers, sand, etc., often in the range of 10 to 80 g · L–1. The biomass is therefore completely suspended in the waste and will be washed out together, which makes it necessary to maintain a high HRT, usually more than 10–15 days, in order to prevent wash-out and reasonable biomass concentrations. In the effort to reduce pumping costs, CSTR systems are often sequence batch loaded with smaller portions 10–200 times during the HRT, rather than continuously loaded. The reader should be aware that, despite the use of sequence batch loading, the systems are commonly referred to (erroneously) as CSTR systems because of the lack of biomass immobilization within the reactor, a phenomenon characteristic of a CSTR. Depending on the duration between loading periods, process variables can display transient behavior of a more or less pronounced nature. Inhomogeneous and viscous nature of the waste and reactor content make it difficult to obtain representative samples on-line, compared to sampling from high-rate systems. Analyzing both on-line and off-line often causes fouling of sensors and sample systems. Controlling the process is mainly done using the HRT, which equals the sludge retention time (SRT) of the biomass. Plants treating municipal wastewater sludge often separate the sludge using a settler after the reactor, allowing thus recycling of the sludge, and thereby maintaining a higher SRT than HRT. Plants treating animal waste, heavy industrial waste, and the organic fraction of municipal solid waste (OFMSW) have storage facilities allowing for larger variations in the HRT than most high-rate systems. The organic composition of the waste
140
P.F. Pind et al.
treated in CSTR systems, makes it often necessary to optimize reactor conditions towards hydrolysis of the suspended matter, since the concentration of easily degradable matter is normally relatively low (less than 5 g · L–1). Slow degradation (giving a low VFA/CO2-production ratio), together with usually high alkalinity of the treated waste, often gives high alkalinity in the reactor and makes it possible to withstand high concentrations of VFAs without significant drops in the pH. Control of pH is therefore only used in systems treating special wastes (low alkalinity, easy degradable organics), something rather nontypical for full-scale CSTRs. Slow degradation rates and high HRTs give slow responses to changes in both OLR and HRT within certain limits. This, together with the difficulties in sampling on-line, has made it common to use manual sampling and human intuition rather than advanced control systems when operating these plants [18]. Most testing and development in control systems has therefore been focused on laboratory experiments with more or less synthetic wastes [19–23] and fewer experiments with sludge [24] and OFMSW [25]. Furthermore, using sequence batch loading in full-scale applications complicates the interpretation of process variables due to transient behavior of the different process variables [26]. Waste composition often decides the choice of reactor system and the possible on-line monitoring and thereby feasible control systems. As shown in Fig. 2,
Fig. 2. Degradation pathways in anaerobic digestion of organic matter. The parallel time axis corresponds to the conversion time of the organic matter through each degradation step
141
Monitoring and Control of Anaerobic Reactors
the time horizon can extend from a few minutes to several days depending on the content of organic polymers and their degradability. Furthermore, the presence or lack of specific microorganisms that degrade specific compounds can give rise to large differences in the time between adding the organic compounds to the reactor and observing the actual changes in the microbial response. This time delay has to be taken into account, when one decides which reactor design and control strategy to use. A simple way of illustrating the importance of the waste composition is to consider the following situation: Consider a reactor treating wastewater, which contains both easily degradable matter and slowly degradable organic polymers from a food factory. The plant operator and control systems have access to on-line monitoring of the total easily degradable compounds, such as the VFAs. An increase in the content of organic matter in the waste causes the VFAs to increase, and the optimal control response is to decrease the loading. If the increase in the VFAs is due to an increase in organic polymers, the VFA level is maintained high for a long time or even increased for a while (organic polymers hydrolyzed into VFA). This delayed response (of sometimes several days) of the biological process to an increased loading, could be misinterpreted by a control system as an ongoing increase in the VFA in the waste. This misinterpretation by the control will erroneously decrease the loading rate, perhaps until actual starvation could occur. In this situation it would be optimal to react in a more supple way. On the other hand, if the increase were due to increased VFAs in the waste being fed, the response would have been correct in order to prevent complete substrate inhibition of methanogenesis. It is these time delays together with the complex interaction between the many degradation pathways that has made it a great challenge to develop appropriate control systems for anaerobic reactors. When looking at the latest developments in control systems (Table 1) for anaerobic systems, it is evident that high-rate systems, treating industrial waste, have received the highest attention. This is mainly due to the application of fluidized-bed systems to the treatment of industrial wastewater with relatively high variations in organic concentrations. The low-rate CSTR systems have mostly been used to test new control theories in anaerobic environments with complete mixing, allowing for simplifying assumptions on growth conditions, etc. Table 1. An overview of some of applications of control systems to different anaerobic systems and waste types reported in the literature
Type of waste
UASB
Wine distillery waste Molasse waste Baker’s yeast waste Whey powder Ice cream wastewater Synthetic waste [2, 30] Municipal wastewater sludge OFMSW Citric acid fermentation waste
Fluidized-bed
Anaerobic filter
CSTR
[5–7, 11, 12, 14, 27] [8, 28, 29] [13, 15] [9, 10] [16, 17] [19–21] [24] [25] [22, 23]
142
P.F. Pind et al.
2 Basic Structure of a Control System Automatic handling of process information with the purpose of control is generally referred to as process control. The aims of automatic control are to ensure process stability, suppress the effect of the external disturbances, as well as to enhance the process performance. Anaerobic digestion is a sensitive process to environmental factors and prone to instability, and as a result there is an absolute need to develop robust control strategies. The theoretical aspects of process control are beyond the scopes of this article. However, it would be appropriate for the sake of clarity and comprehension, to describe the basic structure of a control system and refer to its principles and operation. In order to design a control system, it is necessary to define which of the process variables are input or output (Fig. 3). The input variables express the effect of the environment on the process and they are distinguished into the manipulated inputs (the one that can be adjusted by the operator, e.g., dilution rate, or another control system, e.g., pH) and the disturbances (which cannot be adjusted, e.g., substrate composition or concentration, inhibitors, etc.). The output variables express the effect of the process on the environment and they are split into the measured and unmeasured output variables. The measured output variables consist of the controlled variables (the ones that the controller is designed to regulate by the proper adjustment of the manipulated inputs) and other measured variables that may help the controller in its work, with additional information about the process state. Other important steps for the control system design are: – The definition of the control objective(s). The aim of a controller may be the accomplishment of a simple requirement or a combination of requirements. In the latter case, a supervisory controller is needed to select which of the par-
Fig. 3. Input and output variable of a process
Monitoring and Control of Anaerobic Reactors
143
tial objectives is important or feasible and activate or inactivate the appropriate controllers for this task. As illustrated in Fig. 4, the supervisory controller can decide the set point of lower level controllers, defining thus a control hierarchy. – The selection of the measurements. The measurements must be selected on their capability to reflect the process state, their sensitivity to the possible changes of the process state due to the disturbances, as well as the reliability, the time delay and the simplicity of the measurement method. Typical parameters used as measured outputs are the pH, the alkalinity, dissolved hydrogen concentration, biogas production rate and composition, etc. Other measurements, if available, may concern the disturbances, for example the substrate concentration in the feed. – The selection of the manipulated inputs. The choice of the manipulated variables is very important since they must have a great effect on the process state. In most cases in anaerobic digestion control, the manipulation input is chosen to be the dilution rate (reciprocal of the hydraulic retention time). Other manipulated inputs may be the rate of alkali or acid addition, the heating element, etc. – The selection of the control configuration. The control configuration (or control law) is simply the structure of the information flow from the measure-
Fig. 4. Scheme of the supervisory control acting on the lower level direct controllers
144 a
P.F. Pind et al.
b
Fig. 5. a General feedback control configuration, b General inferential control configuration
Fig. 6. General feedforward control configuration
ments to the manipulated inputs. Generally, there are three types of control configuration: the feedback, the inferential, and the feedforward. – Feedback control configuration. The measurements of the controlled outputs are used directly to estimate the proper value of the manipulated inputs (Fig. 5a). – Inferential control configuration. The measurements of other outputs (other than the controlled ones) are used by an estimator that computes (usually via mass or energy balances of the process) the state of the process and the controlled variables (if not measured). The controller based on these estimations, determines the value of the manipulated inputs (Fig. 5b). – Feedforward control configuration. The measurements of the disturbances are directly used by the controller to calculate the values of the manipulated inputs (Fig. 6). The ultimate control configuration may be simply one of those basic types or a combination of them (Fig. 4). – The algorithm of the controller (or control law).
3 Process Measurements For process optimization and control, it is necessary to evaluate the process status by observing important process inputs and outputs. Optimally, the process
Monitoring and Control of Anaerobic Reactors
145
outputs to be used for process control should be easily measurable and have a pronounced response to changes in the process loading and stability. Ideally, a process output would change value proportionally to the activity of the different microorganisms involved. However, due to the non-linear behavior of biological degradation, this is often not the case. Some process outputs respond in an almost on/off or stepwise manner when certain process instabilities occur. These process outputs are sometimes called “process indicators”. An overview of the different process measurements that are being used for process monitoring together with the possible new ways of obtaining process information in anaerobic reactors will be presented in the following sections. 3.1 Gas Production and Gas Composition 3.1.1 Gas Flow
Gas measurements can be conducted quite easily compared to analytical measurements in water. When running a fairly stable process, the ratio between CO2 and CH4 will be close to constant and the total biogas production will therefore reflect the process activity and yield. Inhibition or overloading of methanogenesis can cause build-up of intermediate compounds such as VFAs, giving a lower gas production. Inhibition would simply cause decreased production and overloading would give rise to increased biogas production in the beginning, followed by a decrease when VFA has accumulated. Gas production is, therefore, the earliest and most commonly used parameter for monitoring and control of the anaerobic process, and the use of gas flow as an on-line measured parameter in control systems is widespread [6, 11, 12, 14, 22, 23, 25], including the use of gas production for model prediction [13]. Reactor systems with high HRTs do, however, suffer from the fact that the biogas production is delayed, covering a complete HRT period or even more. Industrial gas measuring instruments of many types are developed today, all requiring high and constant flow of more than approximately 5 L · min–1. Recently, an on-line flow metering system was developed that could measure down to 0.1 mL · min–1 [31]. For laboratory use, volumetric displacement instruments are more common [32, 33]. These instruments can measure gas flow down to a few mL · h–1, depending on calibration and setup. Fluctuation noise in the gas flow measurements is common, and data filtering is often necessary if smooth gas measurements are needed for control purposes. Since foaming can cause fluctuation in the gas flow, supervision of the gas flow by sophisticated data analysis systems has been used for fault and foaming detection [5, 27]. One of the most recent applications of gas production for control purposes is the use of disturbance monitoring using gas production as the only output [7] (Fig. 7). By increasing the influent flow rate for a short period of time, the increased biogas yield can be compared with the expected. Overloading or inhibition will be reflected by an unsatisfactory gas yield.
146
P.F. Pind et al.
Fig. 7. Monitoring gas flow rate, when inducing a disturbance on the influent flow rate. In-
creased gas production is compared with the expected gas potential of the influent. The evaluation is termed disturbance monitoring [7]
3.1.2 Methane and Carbon Dioxide
Methane (CH4) can be used for energy production, and the yield in volume and percentage of CH4 is therefore interesting. Normally, the total redox potential of the degraded waste controls the ratio between CH4 and CO2 being produced. However, the pH and alkalinity inside the reactor control the evaporation of CO2 , so controlling pH and alkalinity can to some extent control the gas percentage of CO2 . Changing pH and alkalinity will also affect the general process activity, and control of CH4/CO2 ratio is therefore unsuitable for anaerobic digestion [10]. However, fluctuations in alkalinity due to pH control in low-buffered waste can cause noisy gas flow due to deflected fluctuations in CO2 evaporation. Therefore, it is desirable to use the production rate of CH4 alone, since it is not disturbed by this phenomenon. By passing the biogas through a column/bed of soda lime, the CO2 can be scrubbed off, and a simple volumetric measurement of the methane can be conducted.Alternatively, the CH4 and CO2 percentage can be measured using gas chromatographic methods or infrared measurements. The use of methane production as a process output for control is commonly employed [9–12, 19–25, 34]. 3.1.3 Carbon Dioxide and Bicarbonate
Bicarbonate is usually the predominant buffer affecting the alkalinity of the reactor due to the production of CO2 [35]. Therefore, it is interesting to follow the evolution of CO2 production if alkalinity or pH control is desired. This is
Monitoring and Control of Anaerobic Reactors
147
why some have decided to use this process measurement as input to a control system [11, 12] or included it for model prediction [13]. Because of the evident interaction between CO2 in the gas phase and the bicarbonate concentration, it would be desirable to measure bicarbonate indirectly as a function of the CO2 concentration. However, due to both chemical and physical transfer limitations, accurate balance calculations and determination of transfer limitation are not easily accomplished [36, 37]. Measurements of bicarbonate in the water phase can be accomplished by controlled stripping of CO2 from the liquid followed by measurement of the amount of CO2 stripped. Automatic methods for measuring bicarbonate have been developed [35] and used later for control [16]. 3.1.4 Hydrogen Gas
Hydrogen, an intermediate in the anaerobic process, is another important gas component [38]. Anaerobic degradation of organic matter generates hydrogen, which together with carbon dioxide is further converted into methane. Degradation of propionate into acetate is one of the reactions that produce hydrogen. This conversion is energetically favorable only if the concentration of hydrogen is less than 40 nM, corresponding to less than 6 Pa at 35 °C [39]. Assuming that the microorganisms producing and utilizing hydrogen are well balanced in a stable reactor, overloading would give rise to increased hydrogen levels and perhaps reach inhibiting levels, causing build-up of other intermediates. Likewise, underloading would decrease hydrogen concentration. This knowledge, together with the fact that hydrogen has a low solubility in water, has made it natural to study the use of hydrogen concentration in the gas as a process variable. However, the results of earlier studies were inconsistent and have revealed a hydrogen response of a much more complex nature. Summarizing the results, [40] showed that hydrogen concentration could indicate changes in reactor balances, at an early point after the imbalance has occurred in the system. However, the level and time drifts in hydrogen concentration do not necessarily reflect the degree of loading or inhibition. These conclusions were supported by experiments on a laboratory CSTR system loaded once per day [41]. Daily variations between 20–120 ppm (approximately 2–12 Pa) in the headspace concentration of hydrogen were common. When changing the HRT from 20 to 8 days, significant changes in the hydrogen level were observed more than 500 hours later, a time by which the reactor had completely failed. Similar results have recently been reported in experiments with fluidized beds [42]. In Fig. 8, large drifts in hydrogen concentrations are evident, when a new feed portion was used. Hydrogen concentration in the gas phase can be measured by using a mercury-mercuric oxide detector cell, having a detection limit down to 0.01 ppm or 1 mPa [39]. Cheaper measurements down to 0.1 Pa can be made using an Exhale Hydrogen Monitor [43] or palladium metal oxide semiconductors (Pd-MOS) [39, 44–46]. Using thermistor thermal conductivity, a detection limit of about 1 Pa is possible [47]. Due to sensitivity to thiols and sulfides, the gas is often rinsed of these components prior to analysis, in order to avoid interference.
148
P.F. Pind et al.
A
B
C
D
Biogas (ml/min)
Hydrogen Concentration (ppm)
Hydrogen
Biogas
Fig. 8. Hydrogen and biogas production from a fluidized-bed reactor exposed to step increases in the organic loading by increasing the feeding rate. At point A loading increases from 0 to 12.3 kg COD m–3 day–1, at point B to 24.8 kg COD m–3 day–1, at point C to 30.8 kg COD m–3 day–1, at point D to 38.5 kg COD m–3 day–1 [42]
3.1.5 Hydrogen in the Liquid Phase
The inconsistency in hydrogen gas measurements is partly due to highly dynamic and non-linear air-liquid transfer of hydrogen, as documented in intensive studies of hydrogen concentrations in both the gaseous and liquid phases of sludge bed reactors [44, 48–50]. Other investigations reported that gaseous hydrogen concentrations could not reflect the long-term inhibition [11, 12], and a stable process could not be obtained using this parameter alone. Dissolved hydrogen, as a process variable, does not suffer from the same limitations as gaseous hydrogen. Studies of dissolved hydrogen have shown better correlation to the OLR [48]. A non-linear adaptive controller using dissolved hydrogen as the process variable did prove successful in simulation experiments [51]. Simple control using a set point level for dissolved hydrogen in a laboratory CSTR reactor continuously loaded with glucose media was also successful [52], but when shifting to the more common batch sequence loading of the system, the hydrogen response showed a transient behavior requiring a much more complex control scheme. It has been reported that the hydrogen level sometimes returns to so-called stable levels before other parameters have reached stability [52, 53]. These unexpected changes of dissolved hydrogen concentration can be attributed to the presence of significant amounts of hydrogen scavengers, which change their uptake of hydrogen quickly whenever the concentration increases dramatically [41].
Monitoring and Control of Anaerobic Reactors
149
Dissolved hydrogen can be measured down to 50 nM using an amperometric probe [54] or hydrogen/air fuel cell [39] down to 80 nM. Mass spectrometry can also be used, provided that the medium is relatively clean [55]. A relatively new technique, using a coiled silicone tubing inside the reactor to create a fast interface action between gas and liquid and then measure the hydrogen concentration in the gas collected, has been used to detect concentrations as low as 10 nM [52]. Unfortunately, silicone membranes and tubings are susceptible to fouling, i.e., attachment of biofilm when exposed to the biomass. Teflon membranes hold the advantage of being less permeable for H2S and can therefore be advantageous compared to silicone [44]. However, when using membranes in reactor systems with a high content of suspended solids and biomass, fouling seems inevitable [44]. 3.1.6 Hydrogen Sulfide
Hydrogen sulfide also ends up in the produced gas in small concentrations of 1000–3000 ppm, due to the degradation of proteins or other sulfide-containing components. The concentration of hydrogen sulfide in the gas can therefore reflect the current presence and degradation of sulfide containing compounds, if the pH is known.When dealing with concentrations well below inhibiting levels, the measurement of hydrogen sulfide is, however, not interesting from a control point of view. Since hydrogen sulfide is easily converted into sulfuric acid, it can cause corrosion, and it may be necessary to monitor the levels of hydrogen sulfide in the gas, using electronic sensors or gas measurements. 3.1.7 Carbon Monoxide
The presence of carbon monoxide in concentrations of 10–100 ppm has also been reported to be of interest, due to the low solubility and easy measurement [40]. However, use of carbon monoxide concentration for process control has not yet been reported. 3.1.8 Electronic Noses for Gas Measurements
Recent developments in electronic noses or gas sensors have made it possible to use such gas sensors to measure metabolic activity, indirectly. The sensors consist of several different electronic semiconductors that give different responses when subjected to different gas compositions. These sensors are highly sensitive towards even small changes in the gas compositions. However, the data obtained are not easily interpreted, but have to be correlated through data analysis with known observations. Experiments have shown good correlation between data obtained from gas sensors during the growth phases of Methanobacterium formicicum and the evolution in acetate and propionate [56, 57]. These new measuring techniques hold promising potential for new robust sensors to anaerobic
150
P.F. Pind et al.
systems since they do not require direct contact with the waste. However, the understanding of liquid-gas phase equilibrium is still limited in anaerobic systems, and much research is needed if the electronic noses are to be used for processes involving more complex wastes. 3.2 Intermediate Species
Degrading soluble and insoluble carbohydrates, proteins and lipids produces a wide range of intermediate species, which are further degraded into other intermediates or the end products methane and carbon dioxide. Monitoring these intermediate species can give vital information on microbial activity. Insoluble carbohydrates and highly branched polymers such as cellulose can be measured using HPLC or GCMS, but require quite complicated pretreatment methods, while on-line measurements are not feasible. Measuring soluble carbohydrates such as glucose, xylose, mannose, etc. is in principle quite easy with HPLC and could be done on-line, if the amount of suspended materials in the samples is low or can be removed by filtering. But the costs are so high that these types of measurements are only used for laboratory studies. Analyzing for proteins, amino acids, lipids and long chain fatty acids (LCFA) holds the same problems, requiring purification or extraction before analysis. Some of the most interesting intermediates are the volatile fatty acids (VFAs) that through the acetogenic and acetoclastic steps can be converted into methane and carbon dioxide. It is well recognized that monitoring the specific concentration of VFAs can give vital information on the process status [26, 58–68]. Often acetate and propionate will be dominating, which earlier led to the assumption that the ratio between propionate/acetate could be used as indication of process imbalance, if the ratio exceeded 1.4 for CSTR systems [58]. Test on immobilized biomass did, however, show that this ratio could not be generally used [61], while other studies have shown similar results for completely mixed reactors [69]. Other studies have shown that the iso-forms of butyrate and valerate are better indicators of changes in the process balance [59–62]. One of the main reasons for not being able to make general interpretations of the VFA concentrations in a stable process is the fact that the feed composition has great influence on the metabolic pathways used for the production of methane. The main pathways of organic matter, whether acetate alone, acetate and propionate or acetate, propionate and butyrate, will give different stable VFA concentrations. The complexity in the interpretation of the specific VFA concentrations has made it difficult to develop actual control systems based on VFAs alone. However, off-line measurements of VFAs to verify the state of the biomass have been conducted in numerous control studies based on other on-line parameters. An overloaded process is often accompanied by increased acetate concentration. However, if methane production increases with increasing acetate concentration, the anaerobic process can very well find a new balance between acetate production and conversion, at a higher concentration level [18, 62].An increased concentration of VFAs should therefore not be seen as a sure sign of process failure, but more as a process balance change that potentially could lead to failure.
Monitoring and Control of Anaerobic Reactors
151
Accumulation of propionate and butyrate is often seen as an imbalance between fermentation and acetogenesis. Propionate, butyrate and higher VFAs require very low hydrogen concentrations for degradation to be energetically favorable. Propionate degradation requires 5–6 times lower concentrations of hydrogen than butyrate degradation [68], and accumulation of propionate is often seen during step changes, pulses, and feed composition changes due to short-time increases in dissolved hydrogen levels [62, 65]. If the increase in hydrogen is more severe, an increase in butyrate will also be seen [68]. If the stress factor on the reactor is removed and the process can stabilize itself, it will often be observed that butyrate is removed before propionate as a consequence of the difference in hydrogen sensitivity [68]. The effect of temperature changes on VFAs will always depend on the actual situation. However, propionate degraders are generally more sensitive to temperature changes and it is not uncommon to see propionate accumulation due to sudden temperature changes [70]. Degradation of propionate and butyrate can also be inhibited by high concentrations of acetate, since their degradation rates are thermodynamically limited by the concentration of acetate and hydrogen [66, 71]. Butyrate exists in two different isomeric forms: n-butyrate and isobutyrate. n-butyrate is considered the main product of higher molecule degradation. The formation of iso-butyrate originates primarily from the degradation of the branched amino acid valine [72], so the actual concentration of iso-butyrate should normally be very low, but transformation between iso- and n-butyrate is well established [73–77], so iso-butyrate can act as a temporary storage for butyrate. Focusing on valerate, transformation between iso- and n-valerate has not been reported [75]. Due to the low concentrations of iso-butyrate and isovalerate in a stable process [60], the relative change in these components will be higher when imbalance occurs, and reactor studies have implied that the isoforms of butyrate and valerate are the best indicators of changes in the process balance [59–62]. Inhibitory levels of VFA have been found to be dependent on pH. At a pH higher than 7.5, the level of VFAs can exceed 4.5 g · L–1 or approximately 50 mM before any inhibition is observed [62, 78]. Furthermore, propionate and n-butyrate were found to be more inhibiting than n-valerate to the acetate-utilizing bacteria [78]. Batch tests on biomass from fluidized-bed reactors have shown that inhibition by VFA is clearly associated with the undissociated form [65]. Therefore, the inhibiting effect of VFAs will be much higher in systems with low pH ranging from 6.0–7.5, since increased VFA concentration will also cause pH lowering if the alkalinity is low. The pH dependency together with the different affinity to VFAs of the different methanogens [74, 79, 80] leads to the conclusion that no general assumptions on inhibitory levels of VFA are possible. Instead a more complex evaluation of the VFA concentration has to be conducted [20, 22, 23, 26, 62, 81]. Simple attempts to use propionate concentration for adaptive control have been conducted [23], and the use of VFA concentration as imbalance indicator has been tested [20]. VFAs were measured manually by the use of GC or HPLC. Linear input-output models have been used to control a fluidized-bed reactor by on-line measurements of VFA [9, 10]. However, the work has not been extended
152
P.F. Pind et al.
to other reactor systems, due to the lack of a simple and reliable method for online VFA measurements. The earliest attempts for using on-line VFA measurements were conducted by placing a filter on the recirculation line in a fluidizedbed reactor treating butyrate media, but no data were published [82]. Later, a 0.45 mm filter was used before acidifying and sending the sample to a GC using a flow-through cell [10], and others have attempted to use HPLC analysis on-line on UASB reactors [83], but the technique suffered from membrane fouling. Since specific measurements of VFAs are so difficult, the use of indirect methods for determination of the total VFA concentration have been well established. Powell and Archer describe in theory the measurement by titration [84]. This method requires that the waste does not contain particulate matter, since this would clog the used pipe system. The use of titration methods has been used for the control of CSTR systems [24, 25]. The interference of other acids and bases when measuring by titration can blur the results, and it is important to evaluate the reliability of the method by comparison with GC or HPLC. Rozzi et al. [63] suggested and tested the use of an off-line nitrate batch test, for performing indirect measurements of total VFAs as proportional to the measured nitrate reduction. Such biological methods for estimating readily degradable matter, such as VFAs, can be very informative. The use of lactate and alcohol as process variables has been very limited [85, 86], even though the degradation pathways are well established [64]. Although lactate may be an important intermediate from glucose degradation, it is quickly further fermented to VFA and often is only present in very low concentrations [87]. Costello et al. included lactate in their model of the anaerobic process [85, 86], and the comparison of actual concentration with the model simulation is used verbally but not for actual control [28]. Ethanol is normally only present in very low concentrations, except at low pH [88]. The fact that ethanol and lactate are only present in detectable levels when treating special types of waste, and the lack of on-line methods have limited the use of them for monitoring. When on-line procedures for measuring VFAs with GC, GCMS, HPLC, or MS are established in the future, lactate and alcohol, could easily be measured as well, provided that they are present at detectable levels. 3.3 Inorganic Chemical Components and Their Activity
Techniques for measuring specific inorganic components and their activity have been well established for process monitoring in general. The possible use of these methods for the anaerobic process should therefore be evaluated. 3.3.1 Redox Potential
One of the simplest measurements is the redox potential, which can reflect changes in oxidizing or reducing agents. Attempts have been made to follow the redox potential in a laboratory CSTR reactor to evaluate the information value of the parameter [21]. The monitoring could detect inhibition by oxygen, but could
153
Monitoring and Control of Anaerobic Reactors
not be used otherwise to give information about the system. The applicability of redox potentials is also very limited, since the anaerobic process is only energetically favorable at stable redox potentials below approximately 400 mV, so that only inhibition by oxidizing matter can be detected by this measurement. 3.3.2 Ammonia
Relative specific growth rate (d–1)
Focusing on ammonia is necessary if the waste treated has a high content of nitrogen, e.g., animal wastes and protein-rich waste.Ammonia or the ionized form ammonium (NH3/NH4+) serves as important nutrients for the cells, but also as degradation products, especially of proteins. It has been proven that concentrations of free ammonia ranging from 100–1100 mg-N · L–1 can cause inhibition of the methanogenic process [89–92]. However, adaptation of the process of up to 1900 mg-N · L–1 has also been shown [89]. The fraction of total soluble nitrogen in the form of free ammonia is dependent on temperature and pH. High temperature and pH both lead to a higher fraction of free ammonia. Furthermore, the addition of active matrices such as activated carbon can reduce the effect of ammonia inhibition [90]. Monitoring the content of total nitrogen and free ammonia would make sense if the actual concentration were close to a phase change in inhibition (as illustrated in Fig. 9). However, use of ammonia monitoring for control purposes is rare, since the actual concentration is often almost constant. However, indirect and unintentional influence on the free ammonia concentration is obtained by pH and temperature control in many cases. Most often the problem of ammonia inhibition is overcome by choosing mesophilic rather than thermophilic process conditions. Techniques based on electrodes or on colorimetric/photometric methods are available for ammonia measurement [93], but most techniques are optimized for wastewater with a low content of suspended solids and are difficult to use with, e.g., animal waste.
Exp. 1 Exp. 2 Mat. Descrip.
NH2 (g-N/l)
Fig. 9. Relative specific growth rate as a function of the free ammonia concentration [89]
154
P.F. Pind et al.
3.3.3 pH
The proton activity reflected as pH, reflects all cations and anions in the mixture, and is a function of many chemical components including organic acids and bases. So, pH can indicate changes in the chemical balance, when acids, bases, anions and cations are being produced or removed as a consequence of metabolic activity. Furthermore, biomass activity is highly dependent on pH, with an optimal pH around 7–7.5 for most microorganism activities, except for hydrolysis/fermentation, which works optimally at pH 5–7. Monitoring pH is, therefore, essential if the pH is not kept constant by the process itself. Interpreting the pH measurement is, however, complicated, since dominating buffers such as bicarbonate, ammonia, and VFAs can stabilize the pH, or at least affect it considerably. Therefore, pH measurements cannot be used as indirect measurements of specific components or activities alone (e.g., for inferential control), but should be used as a significant additional measurement that describes the state of the digester. Often, pH is measured as an extra process variable without using it as a process output for control purposes [15, 17, 21]. Measuring pH is quite easy, with the use of electrodes that are available in industrialized standards, although they always require continuous cleaning and calibration. This makes it more convenient to use pH measurements on waste with no suspended solids and on effluent rather than influent streams. Low-rate reactors treating animal waste or protein-rich waste often have a stable pH above 7.5–8.0, because of the high ammonia content, situations in which pH monitoring or control becomes uninteresting and too expensive. Therefore, pH monitoring is mostly used for high-rate reactors treating waste with low alkalinity. 3.3.4 Alkalinity
As mentioned earlier, pH changes do not necessarily reflect the metabolic activity, due to the presence of strong buffers such as bicarbonate, ammonia, and VFAs. To overcome this information gap, alkalinity measurements have become a common way of estimating the total buffer capacity in a liquid. Total alkalinity (TA) is measured by titrating the sample to pH 3.7 [94] and using the total number of moles added as the measurement. This measurement would normally include both bicarbonate and all VFAs. TA, therefore, reflects the liquid’s capability to withstand acidification without lowering the pH. Traditionally, TA is commonly expressed in units of mg CaCO3 · L–1 instead of mole. Powell and Archer made a detailed description of the theory behind this and developed an on-line method for TA [84]. It is the liquid’s capability to withstand increased VFA concentrations, without pH lowering that is interesting to measure, but TA also includes VFAs, and increasing VFA concentrations leads to increasing TA. In fact, it is the bicarbonate buffer capacity that is interesting, which is why bicarbonate alkalinity (BA) has gained more attention than TA. By changing the end-point of the alkalinity titration to 5.75 [94], it is possible to exclude VFAs from the alkalinity [81] measurement, and thereby truly characterize the
Monitoring and Control of Anaerobic Reactors
155
Fig. 10. Experimental data of gas production, CO2 percentage in gas and BA level in a fluidized-
bed. The organic loading is increased from 17.6 g COD · L–1 · d–1 to 75.0 g COD · L–1 · d–1 in the time span indicated by the arrows [13]. Dashed lines show model predicted change in each parameter 30 minutes ahead
presence of bicarbonate. Hawkes et al. [35] developed a method for on-line measurements of BA that is more suitable for reactor systems. A bypass of the effluent is saturated with CO2 , and afterwards pH is lowered below 4 with acid addition. The rate of CO2 evaporation from the effluent is proportional to the BA. The advantage of this method is that analytical instruments are not in contact with the waste; which makes it very stable. The response time is approximately 30 minutes, and has been tested in detail [16]. Access to on-line measurement of BA made it possible to generate linear models with capability of predicting BA levels 30 minutes ahead [13]. The measured levels of BA compared to the prediction made by the linear model are shown in Fig. 10. Notice the change in BA level when the organic loading is increased from 17.6 g COD · L–1 · d–1 to 75.0 g COD · L–1 · d–1. 3.4 Indirect Measurements of Organic Matter
The main issue of anaerobic digestion is conversion of organic matter into methane and carbon dioxide. Measuring the content of organic matter before and after the process can, therefore, reflect process efficiency. However, organic matter includes both solid and dissolved matter, different types of organic matter such as carbohydrates, proteins, lipids, acids, amino acids, etc., so no precise and simple method exists that can include all of this in one unit. Instead, a simple and indirect method called assessment of volatile solids (VS) is used [94]. Prior to VS measurement the content of total solids (TS) has to be determined. TS is measured as the dry matter content (as weight) after drying the sample for at least
156
P.F. Pind et al.
1 hour at approximately 103–105 °C, and includes both inorganic and organic matter with a boiling point above 105 °C. VS are measured as the amount lost when annealing the dry matter, from the TS measurement, at approximately 550 °C for 1 hour. The VS amount reflects the organic matter in the sample. If it is desirable to distinguish between dissolved and suspended matter, one can filter the sample, using a 1.2 mm pore-size filter. The matter left on the filter is characterized as the total suspended solid (TSS) or the volatile suspended solid (VSS) by the same drying/annealing process. Both TS and VS measurements are time consuming, and faster methods for indirect TS or VS measurement have been developed, but they must be calibrated with the standard methods when applied on new wastes. Often the same types of waste will show a more or less fixed ratio between TS and VS, which makes it possible to estimate the VS content, based on TS measurements only. When dealing with dissolved organics at low concentrations, total organic carbon (TOC) becomes a suitable method [94]. TOC determines the amount of organic matter by evaporating all water and then chemically oxidizing all carbon to CO2 , which is measured by an infrared analyzer. TOC is not suitable for suspended matter, and TOC is therefore mostly used in high-rate systems treating completely dissolved organics. Chemical oxygen demand (COD) is another way of estimating all organic matter, by chemical oxidation at destructive temperatures [94]. An advantage of COD is the fixed conversion factor of 350 L CH4 · (kg COD)–1 for all types of organic matter. However, the COD also includes organic matter, which is not biologically degradable in anaerobic processes. That is why biological degradation methods are preferable for estimating the organic matter. One method is the biological oxygen demand (BOD). BOD estimation is carried out by measuring the oxygen used during aerobic degradation of the sample over a fixed period, normally 5 days; then BOD is termed BOD5 [94]. However, aerobic degradation (BOD) does not necessary reflect all the possibly biodegradable organic compounds under anaerobic conditions. Therefore, a better evaluation of process performance can be achieved by comparing the actual yield with the yield established by anaerobic batch degradation experiments for each practical type of waste. The methane yield is then compared to the added VS to the batch, giving the Bo yield [mL CH4 · (g VS added)–1], or the methane yield is compared to the amount of degraded VS: B¢o [mL CH4 · (g VS degraded)–1] [95, 96]. Having measured the organic matter as VS, it is common to compare the methane yield of the waste as mL CH4 · (g VS)–1 or L CH4 · (kg VS)–1. This factor can be compared to the maximum theoretical yields of methane from carbohydrates [approximately 400 mL CH4 · (g VS)–1], proteins [approximately 500 mL CH4 · (g VS)–1] and lipids [approximately 1000 mL CH4 · (g VS)–1]. In practice the actual yield will always be lower, and in some cases considerably lower than the theoretical yield, due to effluent losses, VS spent for biomass synthesis and the presence of non-degradable complexes. VS is a good parameter for all types of organic matter, but the analysis time and method makes it unsuitable for on-line measurement, and the application is therefore mostly conducted manually for low-rate reactors. COD is faster, and
Monitoring and Control of Anaerobic Reactors
157
Fig. 11. Experimental data of TOC measurements compared with output signals for the influent pump in a fluidized-bed. The organic loading is increased from 16.1 g COD · L–1 · d–1 to 69.9 g COD · L–1 · d–1 in the intervals indicated by the pump output signal [13]. Dashed lines show predicted changes in TOC 30 minutes ahead
control-experiments on both UASB [2] and CSTR reactors [22, 23] have been made, but in all these cases the measurements were taken manually. On-line measurement of TOC has been applied by using a filter of a few microns on an effluent or bypass stream [10] in a fluidized-bed reactor treating whey powder wastewater. Even advanced model prediction of TOC has proven possible [13], as shown in Fig. 11. Indirect measurement of the organic matter can give vital information on the process yield and potential yield, but the more precise methods (VS, COD, and biogas potential) are so time-consuming and consequently are not suitable for on-line control, but only for off-line yield comparisons. 3.5 Metabolic Activity Measurements
The process measurements covered so far have been external parameters or products produced by cell activity. However, the actual activity of the microorganisms is not necessary proportional to these values. In fact, the usual purpose of anaerobic process monitoring is to optimize the metabolic activity. That is why the importance of being able to estimate metabolic activity has been recognized for some time, and several reviews on this topic have been published [18, 40, 97]. 3.5.1 Microbial Techniques
Traditional techniques such as morphological studies are found to be insufficient for characterizing the population of microorganisms [18, 97]. Other cultivationbased methods, such as most probable number (MPN) or cell-forming unit
158
P.F. Pind et al.
(CFU), are selective, due to choice of media and other growth conditions [18], and cannot be considered suitable for exact characterization or quantification of the populations [97]; however, they are still widely used. Specific methanogenic activity (SMA), a test specifically developed for anaerobic conditions, estimates the potential gas production rate in a reactor system, and SMA has been used to follow the start up of a membrane reactor [98] treating brewery waste. In the specific experiment [98], comparison of reactor performance with SMA on acetate did, however, reveal that SMA could not be compared with the gas production obtained from reactor treating the more complex brewery waste. Likewise, severe experimental difficulties have been experienced when dealing with samples from sludge systems [99]. Furthermore, SMA, MPN and CFU techniques all require several measurements over a time span of hours to days to obtain usable results. 3.5.2 Molecular Techniques
Today, molecular techniques for the identification and quantification of anaerobic microorganisms are being developed. Both immunological techniques and techniques based on RNA and DNA probing are used for identification [18, 97, 100–102]. These techniques can be used both for simple identification of specific microorganisms or even for quantitative determinations. Other possibilities of the highly advanced DNA and RNA techniques are detection of specific gene activity, either by incorporating reporter genes, by genetic manipulation or mRNA detection with traditional probing techniques or the newly developed in situ polymerase chain reaction (in situ PCR) [103]. However, all molecular techniques are only off-line methods and require tremendous amount of laboratory work. The techniques are, therefore, primarily used for obtaining better process understanding rather than actual monitoring. 3.5.3 Chemical Indicators
Cell-produced indicators, such as enzymes or phospholipid fatty acids, have also been examined intensively [18, 97]. Coenzyme F420 and NADH have gained interest because of their easy detectability by fluorescence monitoring. The first experiments with fluorescence monitoring in anaerobic reactors were carried out in low-loaded CSTR systems (actually sequence batch loaded) treating either glucose or VFAs [104]. The detection probe showed consistent calibration for both coenzyme F420 and NADH in the media.A clear transient response in NADH and coenzyme F420 was observed as a reflection of the daily loading. The levels of NADH and coenzyme F420 increased instantly when feeding took place, followed by slow decrease in the level. Pulse loading with formate, acetate, and propionate, either as salts or acids, showed very complicated responses in coenzyme F420 and NADH levels, probably depending on the prehistory of the process (Fig. 12). Similar tests were later carried out to investigate the use of adaptive control on laboratory CSTR systems [21]. Overloading, underloading and inhibition by phe-
Monitoring and Control of Anaerobic Reactors
159
Fig. 12. Different measurements obtained from a CSTR system subjected to batch feeding [104]. Additions: G – glucose; SF – sodium formate; AA – acetic acid; SP – sodium propionate; FA – formic acid; SA – sodium acetate; PA – propionic acid
160
P.F. Pind et al.
nol addition were tested. Detection of inhibition by oxygen was rapid, while detection of phenol inhibition was somewhat delayed. The response to overload and underload was, however, inconsistent and delayed [21]. The use of near-infrared spectroscopy (NIR) can be expanded to measure other compounds than coenzyme F420 and NADH. By analyzing the light scattering data from a broader range of the infrared area, it is possible to correlate known reactor behavior to the light scatter pattern. Recent laboratory studies [105, 106] showed good correlation between phospholipid fatty acids, acetate, propionate, glucose and the light scattering band obtained from a laboratory reactor loaded with glucose media. The NIR spectra of a digester are, however, very complex [105] and require intensive data filtering and analysis before they can be used as input to a controller. Unfortunately, particulate matter and fibers act as noise when measuring the NIR spectra, and therefore the method is not easily adaptable to systems treating complex wastes.
4 Manipulated Variables Manipulated variables are process parameters that can be varied directly, either automatically or by human interaction, and which are affecting the evolution of the process.A manipulated variable should have a range of operation, in order to facilitate refined control rather than simple on-off control. Most of the external physical conditions such as temperature, volume, and loading rate can be used as manipulated variables, but also internal physical conditions, e.g., pH, mixing, internal temperature, and concentrations of specific components can be included, provided that the reactor design allows it, and their setting at desired values may be accomplished almost instantaneously (in reality the true manipulated variable is something else, such as a heating element for manipulating the temperature, or rate of acid or base addition for manipulating the pH, etc.). Most of the manipulated variables used in anaerobic reactors are similar to those used in nonbiological, as well as in aerobic reactors. However, the application of the manipulated variables in anaerobic processes is quite different, due to the complexity of the process, the non-linearity and especially the substrate- and product-inhibition. 4.1 Hydraulic Retention Time, Sludge Retention Time and Organic Loading Rate
The most commonly used manipulated variable is the dilution rate (D), which is the reciprocal of the HRT (hydraulic retention time). D (time–1) is comparable to the actual growth rate of the biomass, which in CSTR systems should be higher than D in order to prevent wash-out. Otherwise, HRT is the most used term for expressing the volumetric loading of the reactor. Some low-rate reactors use recirculation of concentrated sludge from the reactor effluent to increase the retention time of the biomass (the biosolids) by increasing SRT (sludge retention time). In some cases treatment of waste with a high dry matter content makes it necessary to recycle process water, allowing for separate HRT and SRT control
Monitoring and Control of Anaerobic Reactors
161
[25]. Otherwise, almost all CSTR systems have the same HRT as SRT, and the manipulated variables are in principle restricted to a single variable. High-rate reactors all have higher SRT than HRT, but alteration of SRT is rarely a possibility, since the reactor system is optimized towards obtaining the maximum SRT possible. Generally speaking, the only readily manipulated variable available is HRT, but even the use of HRT is restricted. Most anaerobic reactor plants are treating a continuous production of waste, and storage capacities are often restricted to a few hours up to a couple of days. This restricts the variation of HRT to an emergency brake, which is only used in order to prevent total reactor failure. The use of HRT is therefore mostly restricted to experimental setup in pilot plant and laboratory studies to determine the optimal running conditions for specific reactor designs and waste types [2, 6, 7, 9–12, 14, 19–25, 27, 52]. The limiting levels of HRT for CSTR treating animal waste have been extensively studied in the past [95, 96, 107, 108], finding that HRT down to 4–6 days was feasible at thermophilic temperatures. More recent studies have shown that HRT should not be lower than 10–15 days when treating the waste at mesophilic temperatures [109]. Generally, HRT are kept somewhat higher than feasible to ensure a more stable process. OLR must also be taken into consideration when deciding on HRT, since the organic content of the waste is more or less fixed. In reactors using recycling of the effluent, it is possible to alter OLR independently of HRT, provided storage capacities allow it. High-rate reactors with recirculation of effluent can increase HRT while lowering OLR, making it possible to dilute influent concentrations. However, independent use of HRT, SRT, and OLR is often limited to a variation range of ±10%, if continuous operation is to be maintained. 4.2 pH
While interpretation of pH changes is difficult, controlling pH is easier, and the need for pH control has long been recognized as a way of controlling the degradation efficiency in high-rate systems, treating industrial waste with low buffer capacity [8, 110]. Strictly speaking, strong acid or base addition rate are the manipulated variables. Since, however, addition of strong base or acid can alter the pH quickly, we often consider pH as the manipulated variable. In high-rate reactors, it is common to use pH control on both preacidification reactors and the main reactor itself. Simple control of the pH at a desired setpoint can reduce the inhibitory effects of increased VFA concentration, when overloading the process, and thereby reduces the amount of non-degraded VFAs in the effluent [8]. When considering pH control, two process variables can be the aim of the control: influent pH and reactor pH, the latter being more difficult than the first. Waste with low alkalinity, often requires pH control by addition of base such as NaOH or Ca(OH)2 just to maintain pH above 6.5–7. One way of controlling the pH, is to use simple control of the inlet pH to the reactor [9, 27–29]. Because of the simplicity of pH control at the inlet, a simple pH controller is often used, as opposed to, e.g., more sophisticated control schemes for the reactor itself [23].
162
P.F. Pind et al.
Others have used pH measurements in a more complicated control of both inlet pH and the whole process pH [6, 8, 10–12, 14, 25]. Even advances in fault detection have been made, based on pH control [5, 27]. The use of pH control in the main reactor is much more difficult, due to the dynamic response of the process when altering the ion balance. This, together with the logarithmic scaling of pH, has turned the focus on alkalinity control (i.e., use of alkalinity as the manipulated variable for controlling the pH) rather than direct pH control by acid/base addition. 4.3 Bicarbonate Alkalinity
Provided that on-line measurements of BA are available, this manipulated variable becomes superior to pH manipulation in the reactor when the control objective is constant pH. The actual level of BA depends on the desired pH range and waste composition, but normally BA should exceed 1000 mg CaCO3 · L–1 and preferably be in the range of 2000–3000 mg CaCO3 · L–1 with high-rate systems [17]. In general, higher BA results in higher pH up to approximately a pH of 7.5. Further increase in the pH, requires a relatively high concentration of ammonia or other buffers with high pKa . Higher alkalinity implies better and more stable pH, but addition of chemicals to maintain a high BA and TA is costly. In fact, the largest operating cost is the addition of chemicals when treating carbohydrate-based wastes in high-rate systems [28, 29]. In order to reduce the costs of chemicals, many investigations have been carried out, in the effort to reduce the amount necessary to balance the pH [15–17, 28, 29]. One approach is to accept a low BA level, relying on a good BA controller being able to react quickly and correct disturbances. This approach was successfully implemented in an anaerobic filter treating simulated ice-cream wastewater as the influent. A trained neural network having the last 8 measurements of BA as input was capable of maintaining a stable system at a low BA level of 1000 – 1200 mg CaCO3 · L–1 despite heavy disturbances [17]. Later, the same control system was tested on a fluidized bed treating simulated baker’s yeast wastewater, with a stable BA level around 2500 mg CaCO3 · L–1 [15]. Another approach to BA control is to use the CO2 produced during biomethanation and recycle it to the influent, instead of adding bicarbonate salts. In UASB reactors, it is common to use effluent recycling to maintain a high up-flow velocity, and thereby increase considerably the retention time of dissolved CO2 . Increasing CO2 retention increases the reactor BA. This natural BA control is sufficient if the waste pH is higher than approximately 6.5. If the waste pH is below 6.5, however, direct pH control in the influent is necessary. However, addition of base to control the influent pH and thereby BA can be reduced, if one accepts a reactor pH lower as 6.5–6.8 [28, 29]. A side-effect of lowering reactor pH is of course stripping of CO2 , resulting in lower methane concentration in the gas.
Monitoring and Control of Anaerobic Reactors
163
4.4 Temperature
This parameter is very often kept constant or at least attempted to be kept constant and is therefore not used as a manipulated variable. Studies of the effects of temperature on inhibition [111] by ammonia have revealed the possibility of reducing inhibition by lowering the temperature. So temperature, as a manipulated variable, could in principle be used for handling rapid increases in ammonia. However, rapid temperature changes can cause dramatic drops in degradation over a short period [112], and only few attempts have been made to use temperature for control purposes [21]. Using temperature as a manipulated variable requires that the effect of changes can be evaluated in a simple manner. Since temperature affects both biological and chemical reactions, no simple output can be expected from a temperature change. Instead, temperature control must focus on finding an optimal temperature rather than changing it continuously. Using methane yield to evaluate temperature changes has been tested. The tests were carried out in laboratory reactor systems treating a mainly glucose medium, in which optimal cultivation conditions were found to be at 47.5 °C when starting out at 30 °C [21]. In this case, the waste was easily degraded and methanogenesis would be expected to be the bottleneck in the conversion to methane. Many anaerobic plants have been built with a limited specific heat capacity and cannot implement easily temperature control over a large range. Consequently, most temperature control schemes have been focused on steady temperature conditions, sometimes using feed forward of the feed temperature to control the reactor temperature [6]. 4.5 Waste-Management or Co-Digestion
Waste-management as a manipulated variable offers the opportunity to manage the composition of the waste. The major use is for making up for nutrient deficiencies and specifically for fixing the C/N ratio [113]. By combining wastes rich in carbohydrates and ammonia it is possible to obtain the optimal C/N ratio between 25–30 for high solids digester [113]. Another use when dealing with solid wastes is to control the OLR and the dry matter content in the reactor. The drawback of waste-management is the requirement for storage facilities of the waste. Therefore, most use of waste-management is found in CSTR systems treating solid waste. Here, the loading rate is sufficiently low to allow for inexpensive storage of the waste. Especially the use of energy-rich waste, such as lipids-rich waste, together with low potential waste, can make anaerobic treatment of low yielding animal sludge more economically feasible [114–117]. There is only very little documentation on the effect and use of waste-management for on-line control purposes. Instead, the use of waste-management has focused on obtaining high methane yields with fixed ratios, a process termed codigestion [116]. Co-digestion, as a control perspective, becomes more interesting when trying to control inhibition caused by ammonia or xenobiotic compounds from small waste streams being added to the reactor [111]. With the high-rate
164
P.F. Pind et al.
systems, co-digestion is the only method used to ensure sufficient nutrient levels for the treatment of carbohydrates-rich waste.
5 Control Approaches The complexity and sensitivity of the anaerobic digestion process necessitates the use of efficient control strategies. The intriguing point of anaerobic digestion control is to meet the requirement of process stability and the maximum yield. However, these two objectives are often conflicting and the controller must be highly sophisticated to accomplish this task. Generally, the complexity of the controller increases with the diversity of the objective(s) it is designed to meet. As a result, the works in this field of research that can be found in the literature vary from use of simple control laws, to more complicated adaptive controllers or artificial intelligence control schemes (i.e., fuzzy, knowledge-based, and neural network controllers). It is the intention of this section to introduce the reader to these different approaches of automatic control on anaerobic systems, without going into specific details on the theory involved. 5.1 Simple Controllers
The simplest controller is the on/off-type controller. The controller turns the manipulated input on or off so that the controlled output varies around a predetermined value (set point). This controller cannot decide on the size of the control action, taking into account the size of the error of the controlled output (which is the difference of the controlled output from its set-point). As a result, the controlled output may be highly oscillating. To maintain a fixed controlled output (set point), e.g., measured biogas yield, one can use proportional control; by choosing a manipulated input (e.g., OLR), a linear relationship is used between the controlled and the manipulated variable, e.g., biogas potential of 0.25 L gas · (g VS)–1. Deciding on the desired biogas yield (set point), one could set up a simple proportional controller (P) on the OLR using the proportionality coefficient (gain) of 0.25–1 g VS (L gas)–1. Nevertheless, P controllers introduce an offset in cases of step changes in the set-point or the disturbances. The most widely used controllers for set point control are the so-called PID (proportional integral and differential) controllers; they have an integral and a differential term apart from the proportional. The integral term makes the controller output change as long as there is an error, while the differential term determines the control action based on the time derivative of the error. The PID controller has one parameter for every term (proportional gain, integral time constant, and derivative time constant). The determination of the controller parameters is called tuning and can be accomplished by using various criteria, empirical rules, or a process model. In general, PID control is best applied for the non-biological parts of reactor systems, such as for pH control in the feed, temperature control, technical control of pumps, mixers and reactor levels. There are several works that have successfully used simple set-point controllers (examples in Table 2 and [118]).
On-off The on-off range was 0.13 pH units below the set point value. The set-point was higher (7.9) in normal operation than the start-up period (7.48). In both cases, the controller kept the conditions inside the reactor within close limits and a good quality effluent was produced, while possible organic overload and pH shocks could be eliminated (1) Increasing or decreasing the feed rate by 20% after a given time interval (5 h) if the dissolved hydrogen was lower or higher (respectively) than 6.5 Pa (set-point).
Dilution rate (the feed pump was set to provide an average OLR of 5 g COD · L –1 · d –1)
Dilution rate (the reactor was fed on a semi-continuous mode)
pH measured in a side stream, where, by regulating the pCO2 , the bicarbonate alkalinity effect on the pH would be minimal and a small change in the metabolic activity of the microorganisms would reflect on the pH
Dissolved hydrogen (by a technique described in [53])
To keep the organic loading rate high in a UASB reactor fed on a petrochemical effluent containing a mixture of short-chain fatty acids at a concentration of 17 g COD·L–1
To keep the dissolved hydrogen partial pressure below 6.5 in an L anaerobic digester fed on glucose (408 mM)
P. The change of feed rate was proportional to the deviation of the hydrogen concentration from the set-point. The sensitivity of the dissolved hydrogen permitted the laboratory anaerobic digester to operate at loading rates close to the maximum loading capacity
PID The determination of the controller parameters was done through the solution of an optimization problem, which used a simple model for anaerobic digestion as a constraint. The model simulated the dynamic response of the digester to shock conditions. The PID controller was tested with success over different values of organic loading shock
Addition of NaHCO3
BA by measuring the CO2 stripped from a liquid sidestream [35]
Controller type
To maintain the bicarbonate alkalinity (BA) on desired levels in an anaerobic filter with recycle (pilot scale) fed on easily biodegradable organic matter
Manipulated inputs
Measurements
Control objective – reactor
Table 2. Simple control laws for the anaerobic digestion process
[52]
[120]
[119]
Reference
Monitoring and Control of Anaerobic Reactors
165
PI, PI, PID, PI (reactively to the controlled outputs) The design of the controller was based on a complex process model This paper refers to a number of other simple control studies
Empirical The basic idea is to deliberately cause disturbances (overloading pulses), evaluate the excess biogas production (due to the overloading), compare it to the expected one, and increase, decrease or not change at all the dilution rate. The comparison allows an assessment of the activity of the microorganisms
Dilution rate
Dilution rate
pH, acetic acid, propionic acid, dissolved hydrogen (simulation) reactively to the controlled outputs
Biogas production rate, pH serves as an alarm and it inactivates the proposed strategy if it falls below a certain value
To maintain the controlled output (pH, acetic acid, propionic acid, dissolved hydrogen) around the set-point in single and two-step fluidized bed reactors fed on whey under a step change in the feed concentration
To maintain the organic loading rate high and keep the effluent concentration low in a labscale and a pilot scale fluidized bed reactor fed raw wine distillery wastewater (COD between 20 and 30 g · L–1)
PI and PID PI algorithm was tested while the set-point changed from I2 to I3 and PID was tested while the set point changed successively from I2 Æ I3 Æ I4 Æ I1 . The algorithms were successful except for the set-point I4 which was near the washout condition of the microorganisms
Dilution rate
Biogas production rate
To keep the biogas production at a set-point corresponding to optimal operation. Different set-points were defined according to a performance index: maximization of biogas production (I1), minimization of: the organic pollutant concentration (I2), the ratio of the pollutant concentration to biogas production (I3) and the organic effluent rate minus the biogas production rate (I4) Simulation results using a 3 step model (hydrolysis, acidogenesis, methanogenesis)
Manipulated inputs Controller type
Measurements
Control objective – system
Table 2 (continued)
[7]
[118]
[121]
Reference
166 P.F. Pind et al.
Dilution rate
Addition of NaHCO3
Alkali addition required to keep pH at 7
BA (measured as described in [35]
To maintain the effluent quality of a lab-scale fluidized bed with recycle between 100 and 400 mg acids · L–1 for easily biodegradable wastewater
To keep bicarbonate alkalinity (BA) around a set-point (2700 ± 100 mg CaCO3) in two lab-scale fluidized bed in series fed on a simulated baker’s yeast wastewater
On-off This type of control although it did not allow the BA to drop below the lower set-point in conditions of overloading, it caused an overshoot in BA concentration It is compared to a neural-network type of controller, which seemed to be superior since it did no induce any oscillation in the BA concentration. The authors refer to another work [122], where a PID and a neural type-network controller were applied on an anaerobic filter fed on ice-cream wastewater. The PID controller was not successful and the authors attributed its failure to: ∑ the use of linear models to determine the PID constants, ∑ the sensitivity of the process not reflected to the PID unchanged constants, and ∑ the long time constants of the process which tend to drive the integrator in the PID algorithm
Empirical Setting an upper and lower threshold for the consumption of alkali, the dilution rate: ∑ stops for a time period proportional to the surplus alkaline consumption if that exceeds the upper limit, ∑ remains the same if the alkaline consumption is between the two thresholds, and ∑ doubles for a time period proportional to the alkaline difference from the lower threshold if the alkaline consumption falls below it [15]
[8]
Monitoring and Control of Anaerobic Reactors
167
168
P.F. Pind et al.
5.2 Adaptive Controllers
The use of simple controllers, such as the PID controllers becomes limited due to the fixed parameters of, e.g., the proportional, integral, and differential part of the PID controller. The optimal values of these tuning parameters are very much dependent on the desired steady state of the process, because of the process nonlinearity.As long as the set point changes, the PID parameters should also change if an optimal response is desired. Moreover, the process itself may have characteristic parameters that change with time. This also affects the controller efficiency. Since the anaerobic digestion process is non-linear and highly dynamic, the use of a controller, which can adjust its parameters in a way that it compensates for the variations in the process characteristics in an optimal manner, would be highly superior to a simple PID, as far as performance (robustness and speed of response) is concerned. Such a controller is called adaptive. The theory, in general, is covered in Åström and Wittenmark [123]. Given a control objective (e.g., regulation at a particular set-point or optimization of a process performance measure), and a process model, the process outputs (measurements) are used to determine, on-line, the best values of all or some of the model parameters. The model is then used to determine the optimal values of the manipulated inputs. Table 3 contains some of the works in the field of the adaptive control of anaerobic digestion. These works span a wide range of control objectives (either regulation or optimization) as well as types of models used. Using an on-line identified kinetic model, a controller with all the same elements as a PID controller could be developed, except that the parameters of the controller will be expressed as functions of kinetic and physical constants, together with state variables. These functions can be either linear or non-linear. Since in most control systems the state variables of the process (simulated by the model) are more than the measured outputs, a model simplification is often required. The model simplification could be based on knowledge of the rate-limiting step of the process [124]. It is desirable that the model reduction is systematic and based on a rigorous mathematical line of reasoning [124]. Bastain and Dochain [125] have made a comprehensive description and evaluation of different model based adaptive non-linear control algorithms for bioreactors. However, the success of an adaptive control algorithm, does not necessarily depend on the complexity of the on-line identified model. It rather depends more decisively on the ability to predict the response of the measured outputs to changes in the manipulated inputs. To this end, the use of simple input-output models that describe this cause-effect relationship is often adequate. Premier and coworkers [13] were quite successful in making a simple adaptive linear model (a 2nd and 3rd order of a general family model structure called ARX model which stands for Auto Regressive with eXtra input), that could predict the behavior of CO2 percentage in the gas, gas production, BA and TOC in the effluent 30 minutes ahead. The linear model was deduced using data obtained from a fluidized-bed reactor treating simulated baker's yeast wastewater. The prediction of BA, gas rate, %CO2 and TOC are shown in Figs. 10 and 11, respectively. Similarly,
Measurements
COD concentration of the influent and effluent (every 2 h) with an offline method with reduced reflux heating time (from 2 h to 10 min). Methane production rate (on-line)
COD concentration of the influent, propionate concentration (every 1 or 2 days) and methane production rate (online)
Control objective
To regulate the effluent organic concentration at a prescribed level in a pilot-scale CSTR fed on wastewater from citric acid fermentation
To regulate the propionate concentration at a prescribed level in a pilot-scale CSTR fed on wastewater from citric acid fermentation
The controller results via external linearization te- [22] chniques, which “transfer” the non-linearity of the system into the control scheme so as for the closedloop system to obtain linear characteristics. The control law calculates the value of the manipulated variable as a function of all the measurements. This function includes a variable parameter (a yield coefficient of the methane produced per substrate consumed), which is updated on-line to track the long-term variations of the process parameters The adaptive control law was tested in two experiments (the one with a pulse and the other with a step increase in the influent COD) conducted in a pilot-scale CSTR fed on wastewater from citric acid fermentation. The controller succeeded, although the off-line measurement of the COD imposed a minimum time of controller action of 2 h [23] The same external linearization techniques were applied in this work too as in [22]. The controller contains a variable parameter (a conversion yield factor of the feed into propionate). The value of this parameter is adapted on-line by an estimation algorithm (least-squares). The control law was tested on a pilot-scale CSTR fed on wastewater from citric acid fermentation. The experiment involved a start-up period, during which the feed concentration was increased three times ant the normal operation (one more step increase). Totally the increase in the feed concentration was from 10 to 50 g COD . L–1 within 130 days. The controller application was successful
One step process (two mass balances for biomass and substrate respectively rearranged so as for the controlled variable to be expressed as a function the measurements)
Two-step process (four mass balances for the initial substrate and the intermediate VFAs expressed as propionate and the two microorganisms groups). The first step was assumed to be very fast (quasisteady state). After proper manipulation of the equations as previously [22], the controlled variable was expressed via the measurements
Dilution rate
Dilution rate
Reference
Control law
Models
Manipulated inputs
Table 3. Adaptive control laws for the anaerobic digestion process Monitoring and Control of Anaerobic Reactors
169
Measurements
Dissolved hydrogen, gaseous hydrogen production rate and glucose concentration in the feed
Substrate concentration (volatile acids, VA), methane production rate, substrate concentration in the feed (because of the high content of the volatile solids, VS, in the feed, the influent substrate concentration, as VA, was estimated through the VS destroyed)
Control objective
To regulate the dissolved hydrogen concentration at a prescribed level in a CSTR fed on glucose
To maintain the substrate concentration or the methane production rate close to a set-point, in two full-scale anaerobic digesters fed from the thickener of a municipal wastewater treatment plant
Table 3 (continued)
The dilution rate was expressed via the measurements by proper handling of the model equations. The parameter that was on-line adapted (by using a Lyapounov design estimation algorithm) was a proportional coefficient via which the specific growth rate of the hydrogen-utilizing microorganisms was related to the dissolved hydrogen concentration. No experiments were performed, so the controller was tested successfully on simulations, in which a five-step model was used. The controller proved to be robust with respect to the dynamics of some neglected steps of the process, and not to be affected by the deliberately wrong assignment of values (deviated 20% from the ones used by the model during the simulations) to some constants of the controller
Under the assumption that acidogenesis was a very fast step and the propionate could be neglected, the mass balance for the dissolved hydrogen contained only terms of the measurements or variables, which the change rate could be related with the measurements Methanogenic step
Dilution rate
Dilution rate
A control law was developed for each output to be regulated. The dilution rate was expressed as a function of the measured variables. The controller parameter was adapted on-line using a predictive control law. A combined control strategy was also developed so as to keep the substrate concentration low and the methane production rate high. This was accomplished by using the two controllers at each sampling interval depending on which one of the two objectives was important at that time. The two control laws were applied separately in two full-scale anaerobic digesters fed from the thickener of a municipal wastewater treatment plant. The measurements were performed off-line every 12 h and the control action was calculated. The controllers operated effectively, when the set point changed, although a time-delay in the reactor response to the controller was mentioned
Control law
Models
Manipulated inputs
[24]
[51]
Reference
170 P.F. Pind et al.
Dissolved hydrogen, gaseous hydrogen and methane production rate, glucose concentration in the feed
Propionate concentration, methane production rate, glucose concentration in the feed
COD concentration, methane production rate, glucose concentration in the feed
To regulate the dissolved hydrogen concentration in a CSTR fed on glucose
To regulate the propionate concentration in a CSTR fed on glucose
To regulate the COD concentration in a CSTR fed on glucose
It was assumed from a five-step process that all steps but the hydrogen consumption were very fast, and glucose, propionate and acetate concentrations were equal to zero. The one resultant mass balance was for hydrogen and contained only the measured variables It was assumed from a five-step process that all steps but the propionate consumption were very fast, and glucose, acetate and hydrogen concentrations were equal to zero. The one resultant mass balance was for propionate and contained only the measured variables The mass balances of glucose, propionate and acetate were multiplied by the proper COD conversion factors and were added. Then one equation resulted for the COD mass balance. Again, assuming all the steps of a five-step model but the glucose consumption were fast, and propionate, acetate and hydrogen concentrations were equal to zero, the COD equations could be transformed to contain only the measured variables
Dilution rate
Dilution rate
Dilution rate
The control law in each case consisted of an equation giving the dilution rate as a function of the measured variables. The control law contained a parameter for on-line adaptation using a Lyapounov approach, according to which, the parameter change of rate is proportional to the control error (i.e., deviation from the set-point). The other design parameters of the control law should change according to the measurements they were related to, or remain constant, so that the closed-loop dynamics is equivalent to a stable stationary system The efficiency of each controller was tested on a simulated process subjected to step changes in the organic load. The results were quite good. In the case where the measurement of the glucose concentration in the feed was not included, the controller was successful only in the dissolved hydrogen regulation, but not on the COD and propionate [127]
Monitoring and Control of Anaerobic Reactors
171
Organic acid concentration, methane production rate
Maximize the methane production rate without permitting the organic acid concentration to be high in one and two stage anaerobic fluidized bed reactors
Methane production rate
COD of the influent and effluent (off-line)
To regulate the effluent concentration around a set-point in a UASB reactor (lab-scale)
Maximize the methane production rate by selecting the proper temperature, while the anaerobic digester is run using the constant yield modus operandi on a lab-scale CSTR fed on glucose
Measurements
Control objective
Table 3 (continued)
Temperature, T
Dilution rate
Dilution rate
Manipulated inputs
Control law
Based on the adaptive control law of Renard et al. (1988) [22], instead of having an extra measurement (the methane production rate), the change rate of the substrate was calculated from the already existing measurements (without being necessary to correlate it with the methane production rate). The control law was applied in a UASB (lab-scale) with success). A comparison was made between the proposed control law and the one which it was based on [22]. The results were almost identical, indicating that the measurement of the methane production rate could be omitted The parameters of the model were adapted on-line to give a good fit at the current time interval. A new value for the dilution rate was then determined from the model towards the optimal operation. In case of unknown disturbances, the outputs would be away from the desired point, the model parameters would be re-calculated and the dilution rate would change again towards a new optimum. The controller was tested in one and two stage anaerobic fluidized bed reactors. The algorithm estimates the parameters of the model using a recursive least square method with variable forgetting factor. Once a new measurement had been obtained and the parameters values had been updated, the optimal temperature could be determined. The algorithm was tested with success on a lab-scale CSTR fed on glucose in a rate proportional to the methane production rate, corresponding to a constant yield modus operandi [128]
Models
Methanogenic step
A linear empirical model that relates the input directly to the output (the most recent output data)
A dynamic inputoutput model (methane production rate given as a 2nd degree polynomial in T)
[126]
[118]
[2]
Reference
172 P.F. Pind et al.
Monitoring and Control of Anaerobic Reactors
173
simple linear and quadratic input-output models have been used successfully in [118] and in [126], to determine the dilution rate and the temperature that maximize the methane production rate in CSTR glucose-fed digesters. In general, all adaptive control schemes can overcome relatively slow changes in the kinetic parameters such as yield coefficients. The main advantage of identifying a process model that adequately describes the anaerobic digestion process, by relating to the key microbial processes that take place inside the reactor, is that in addition to controlling the process, a better understanding of some changes in the state of the digester may be gained. 5.3 Other Control Schemes 5.3.1 Fuzzy Logic Controllers
Often, the knowledge basis of a specific anaerobic process/design is rather difficult to formulate by mathematical models. Instead, linguistically defined rules and observations could be made. This is often the basis for human intuition and intervention in manual control. Fuzzy regulation is a tool that can convert linguistically defined knowledge into automatic control. The formulation of fuzzy rules can be done in several ways [5, 6, 25, 27, 129, 130]. We will, however, only shortly illustrate the traditional way of setting up rules. The values of process variables are evaluated linguistically by dividing the values into subset values such as high, medium, low, etc. or good and bad. The membership of the process variable, e.g., pH, to each of the subset values is described by membership functions as illustrated in Fig. 13. The linguistically defined evaluation of pH is thereby converted into a mathematical membership function with a value between 0 and 1, a process called a fuzzification. Having made the fuzzification, the set of fuzzy rules can be formulated for certain actions to be made on the manipulated variable i.e. dilution rate: for example reduce 20%, maintain, or increase 20% as shown in Table 4. The actual action on dilution rate is formulated using a mathematical description of the rules based on the membership of each variable, a process called defuzzification. In this case, the suggested rules would probably lead to low loading if not a complete stop, since only one situation allows for an increased loading. The complexity of the rules can be increased by an increasing number of sub-
Fig. 13. Fuzzy sets of the process variables pH, showing the membership function (value [0–1]) of each set depending on the process value
174
P.F. Pind et al.
Table 4. An example of fuzzy rules formulation. Dilution rate is decreased, maintained or in-
creased as a function of pH and methane yield pH
low medium high
Methane yield bad
good
decrease 20% decrease 20% decrease 20%
decrease 20% maintain increase 20%
set divisions and increasing numbers of variables – leading to an increasing number of rules. Instead of formulating several hundred rules manually, specially designed algorithms or decision schemes can be used to perform this process automatically [129, 130]. Fuzzy control has been extensively developed for batch fermentation of pure cultures with very high success [129], and attempting to transfer this success to the kinetically more complex biogas process is natural. One of the first fuzzy controllers constructed for a CSTR system included HRT, OLR, pH, TA, total VFA, NH3 , CH4 percentage, gas flow and a Ks as process variables [25]. The Ks parameter reflected the state of the biomass based on observations of the rate and degree of digestion. This required intensive reactor monitoring, carried out manually once per day. The controller controlled the mixing of OFMSW with recycle water from the reactor, thereby controlling both the HRT and the OLR independently. To avoid a nervous control, a dynamic delay factor (filter) was added to the controller. The controller was tested successfully on the continuous running of a 3 m3 pilot plant. Time delay and response in anaerobic systems control are especially a problem, and using the time derivative of the process variables is sensible. In the design of a fuzzy controller the time derivative of pH and gas flow and their actual values were individually evaluated by fuzzification for a fluidized bed treating wine distillery wastewater [14]. Gas flow and its derivative were used as the primary process parameters, and pH drops were considered as an indication of acidification due to overloading. The fuzzy controller consisted of 500 rules. The controller was later modified to include the temperature and its derivative [6]. Since adding of two process variables would increase the rules by a factor of 25, it was decided to use a hierarchical architecture of the controller (Fig. 14). First, the gas flow and the derivative were evaluated resulting in a regulation output, then each of the four other variables were evaluated individually (pH, derivative of pH, temperature and derivative of temperature) together with the output value, with the possibility to lower the output or leave it unchanged. This procedure reduced the number of rules to 125, letting the last four variables act as process alarms only. Both controllers were successfully tested at a pilot scale [6, 14]. The advantage of these controllers was the use of simple process variables that could be measured on-line, without significant investment: gas flow, pH and temperature.
Monitoring and Control of Anaerobic Reactors
175
Fig. 14. Hierarchical architecture of a fuzzy controller used on a fluidized-bed reactor. Gas pro-
duction and the derivative of gas production is used to set the regulation output. pH, derivative of pH, temperature and the derivative of temperature each can reduce the regulating output or leave it unchanged. Partly recreated from Steyer et al. (1997) [6]
5.3.2 Knowledge-Based Expert Systems
The fuzzy control described previously has been used for specific manipulated variables of the process. However, a supervisory (or indirect) scheme is another way of using knowledge-based information for a control purpose [129]. This special supervision of the system performance can be directed towards conventional controllers but data evaluation and fault detection can also be included. Using fuzzy logic, it was possible to detect faults generated by foaming in the reactor and clogging of pipes and control of the feed flow rate successfully, in the same fluidized-bed test as previously mentioned [5]. The concept of fault detection was later applied to pH and temperature supervision with success [27]. The use of knowledge-based information is not limited to fuzzy control; other evaluation schemes can be designed based on the “if-then” formulation or as decision trees, resulting in the so-called expert systems (since they are based on the knowledge of experts). Although it is a rather tedious task to record an expert's knowledge to form rules which, in time, need refinement or change, the expert system is a powerful tool of integrating many control systems into one, making it applicable in most cases. An example of an expert system is the manipulation of the dilution rate proportionally to the production rate of methane, according to a constant yield control law (CYCL) [19, 34]. A conventional set-point control law was applied to the gas production, when the process was considered stable and normal. If measurements indicated inhibition or overloading, the constant yield control was
176
P.F. Pind et al.
used instead. This approach was later expanded to the combination of the expert system with a rather simple adaptive controller to maintain a desired methane production rate [21]. Once again, the expert system takes over the control, if the methane production exceeds certain limits due to inhibition or severe overloading. Finally, an expert system manipulated the dilution rate of a CSTR system in four different modes: conventional set-point control (with respect to the methane production rate), constant yield control law (CYCL), batch operation, and constant dilution rate. The choice of which controller should be used was decided upon statistical criteria (t-test) and future projections. The expert system was tested on a glucose-fed CSTR in conditions of severe overloading, mild to severe underloading, and moderate to severe inhibition (via phenol entry) with success [20]. The fluidized-bed system treating wine distillery wastewater that was previously covered with respect to fuzzy control has also been tested with decision tree structures. The expert system changed the dilution rate every half an hour based on the measurement of hydrogen in the gas phase, pH, and gas rates. Hydrogen was intended to act as an inhibition indicator if it exceeded a certain level. However, the hydrogen concentration returned to the “normal” level faster than the actual recovery of process stability [11, 12]. 5.3.3 Neural Networks
Neural networks have recently found a great acceptance in the field of process identification and control because, although they do not require a model, they succeed to map the non-linear relationship between the input-output pairs. Neural networks consist of neurons (computing elements which manage the information or impulses they receive in the way human brain cells do) placed in well-ordered layers that can be interconnected in many different ways. The multilayer structure of neural networks as well as the variety of the so-called training algorithms (necessary to determine the appropriate weights of the synapses between the neurons, a procedure called training) gives a diversity in the possible control strategies based on this concept [131–133]. In a neural network, the first and last layer are known as input and output layers, while the intermediate layers are called hidden. Wilcox et al. [17] developed a neural network to control the BA in a fluidized bed. The neural network structure is shown in Fig. 15. It consists of 8 inputs (the last 8 measurements of BA) going to the input layer of the network (8 input neurons). The input neurons balance the weight of the input before sending it to all the 8 neurons in a new hidden layer. The neurons in the hidden layer balance the input value of all 8 inputs sending output to a single output neuron in the output layer. The output neuron also balances all 8 inputs resulting in a degree of imbalance of the BA control (likelihood of being out of control). Using input data sets together with a predetermined optimal output, performs the tuning of balancing inputs to a single output in a neuron.An automatic mathematical procedure then performs the tuning by either weighting the whole data
Monitoring and Control of Anaerobic Reactors
177
Fig. 15. Schematic presentation of the neural network used by Wilcox et al. [17] to control the BA level in a fluidized-bed reactor. BA 1–8 represent eight consecutive data points of BA measurement
set equally, or by using one data set at a time and a disregarding factor of the previous tuning. The last method can be used for continuous tuning, while the network is in actual use. In a specific case [17], the network was trained using 80 hours of experimental data, distinguishing between good and bad control of the BA. The network control showed good behavior when disturbing the reactor by changing the OLR. The control was later used with good results on two fluidized-bed reactors running in parallel with different waste [15] without new training. The behavior of the control response was quite similar to that observed with a PID controller. The authors had previously tried to obtain the parameters for a PID controller, but found that the knowledge of the dynamics was inadequate to estimate any reliable parameters for PID control. The neural network controller of course also benefits from the knowledge of the last 8 measurements rather than only the last two as for a PID controller. The previously described fuzzy control system for the faults detection in a fluidized-bed reactor [5] was combined with a neural network. The neural network was designed to use the output of the fuzzy evaluation and translate it into fault detection or dangerous state.
178
P.F. Pind et al.
The successful use of a neural network control depends on the availability of a large set of initial training data that cover all aspects of system operation and give adequate information about process dynamics. Since the dynamics of anaerobic digestion (very complex with kinetic parameters that are slowly changing) render it difficult or even impossible to obtain a proper set of data for this purpose, an adaptive neural control network was proposed by Emmanouilides and Petrou [134]. In this control scheme, a neural network model was incorporated into a neural network controller to set up a combined adaptive controller with model predictive control. The neural network model was first trained on data obtained from a kinetic model simulating approximately 32 years of operation and set up to continuously adapt itself to actual process data, recently obtained from a running reactor. An adaptable neural network controller was then set up to control the process without prior training. By first letting the control action of the network controller to be tested by the network model, an optimal response could be selected (and adapted for actual control) within a test cycle of maximum 500 repetitions for each time cycle. In this way, the control would be trained on-line by the network model and the model would be adapted to the actual process. The model was, however, only tested on the same kinetic model used for training the network model. The possibility of using a model to test the controlled output on-line before employing it, it can hold many advantages, and it is the only way to prove its applicability.
6 Concluding Remarks The reactor design of an anaerobic system is very important for the choice of automatic control on both high-rate and low-rate systems. Furthermore, waste characteristics influence the possible monitoring of process parameters/outputs that could be included in the control. For automatic control the monitoring should preferably be accessible on-line rather then manually obtained. Until today the best techniques for on-line motoring are applied with gas measurements, focusing on gas flow and gas composition of CO2 and CH4 . Hydrogen measurements in the gas phase are possible, however the informative value of dissolved hydrogen is higher, and so interest should be focused on the newly developed laboratory techniques within that area. Monitoring of the content of organic matter, both intermediate compounds and total matter, is crucial for the evaluation of the process yield and stability. However, on-line techniques for measurement of COD, VS, TS, and VFA are still not fully developed. pH measurement and BA/TA monitoring for high-rate systems have been examined extensively, and are currently the best tools for controlling pH levels in high-rate systems, such as fluidized-bed reactors and anaerobic filters. The possible manipulated variables are usually restricted to HRT and OLR, though the use of separate SRT has been employed in few cases. Waste-management has not yet been used for control, but it possesses advantages that could be used. Temperature control is often limited to predetermined values, decided upon in the design and planning of the reactor. The use of temperature control must
Monitoring and Control of Anaerobic Reactors
179
be considered highly theoretical and complete understanding of the consequences of changes is still lacking. The uses of simple controllers such as PID and other set-point controllers are often limited to manipulated variables, such as temperature and influent pH. The dynamics of the anaerobic process often make it desirable to use control methods that allow for adaptation of process parameters. Adaptive control schemes, aiming at regulation or optimization of anaerobic digestion processes have been developed through time. Models based on knowledge of kinetics, thermodynamics, etc., have been expanded and improved tremendously, making it more difficult to condense this knowledge to an adaptive controller. However, simple input-output models have been found to be quite adequate in most instances. Knowledge-based controllers such as fuzzy control and expert systems are emerging alternative control approaches. Such knowledge-based control has been employed with success on both high-rate and low-rate systems. Knowledgebased controllers hold the advantage of implementing linguistic formulated knowledge on the process behavior. Furthermore, supervisory control makes it possible to combine many control objectives in the system. The more recent development has been the use of neural networks for the control of anaerobic digesters.
7 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Weiland P, Rozzi A (1991) Water Sci Technol 24:257 Monroy O, AlvarezRamirez J, Cuervo F, Femat R (1996) Ind Eng Chem Res 35:3442 Hickey RF, Wu WM, Veiga MC, Jones R (1991) Water Sci Technol 24:207 Lettinga G, Field J, van Lier J, Zeeman G, Pol LH (1997) Water Sci Technol 35:5 Steyer JP, Rolland D, Bouvier JC, Moletta R (1997) Water Sci Technol 36:209 Steyer JP, Estaben M, Polit M (1997) Proceedings of the Sixth IEEE International Conference on Fuzzy Systems. (Cat. No. 97CH36032 3 vol (xii+1730): 1245 Steyer JP, Buffiere P, Rolland D, Moletta R (1999) Water Res 33:2059 Denac M, Lee PL, Newell RB, Greenfield PF (1990) Water Res 24:583 Ryhiner G, Dunn IJ, Heinzle E, Rohani S (1992) J Biotechnol 22:89 Ryhiner GB, Heinzle E, Dunn IJ (1993) Biotechnol Prog 9:332 Ehlinger F, Escoffier Y, Couderc JP, Leyris JP, Moletta R (1994) Water Sci Technol 29:289 Moletta R, Escoffier Y, Ehlinger F, Couderc JP, Leyris JP (1994) Water Sci Technol 30:11 Premier GC, Dinsdale R, Guwy AJ, Hawkes FR, Hawkes DL, Wilcox SJ (1997) Water Sci Technol 36:229 Estaben M, Polit M, Steyer JP (1997) Control Eng Prac 5:1303 Guwy AJ, Hawkes FR, Wilcox SJ, Hawkes DL (1997) Water Res 31:2019 Hawkes FR, Guwy AJ, Hawkes DL, Rozzi A (1994) Water Sci Technol 30:1 Wilcox SJ, Hawkes DL, Hawkes FR, Guwy AJ (1995) Water Res 29:1465 Ahring BK,Angelidaki I (1997) Proceeding at the 8th international Conference on Anaerobic Digestion. 25–29 May 1997, Sendai, Japan 1:40 Pullammanappallil P, Svoronos SA, Lyberatos G, Chynoweth DP (1992) Proceedings of the 1992 American Control Conference. 3:1911 Pullammanappallil PC, Svoronos SA, Chynoweth DP, Lyberatos G (1998) Biotechnol Bioeng 58:13 Chynoweth DP, Svoronos SA, Lyberatos G, Harman JL, Pullammanappallil P, Owens JM, Peck MJ (1994) Water Sci Technol 30:21 Renard P, Dochain D, Bastin G, Naveau H, Nyns E-J (1988) Biotechnol Bioeng 31:287
180
P.F. Pind et al.
23. Renard P,Van Breusegem V, Nguyen M-T, Naveau H, Nyns E-J (1991) Biomass Bioenergy 38:805 24. Polihronakis M, Petrou L, Deligiannis A (1993) Comput Chem Eng 17:1167 25. Boscolo A, Mangiavacchi C, Drius F, Rongione F, Pavan P, Cecchi F (1993) Water Sci Technol 27:57 26. Pind FP, Angelidaki I, Ahring BK (1999) Proceeding at IAWQ II Symposium. Barcelona 15–17 June, 1999 II: 41 27. Genovesi A, Harmand J, Steyer JP (1999) Biochem Eng J 3:171 28. Romli M, Keller J, Lee PL, Greenfield PF (1994) Water Sci Technol 30:35 29. Romli M, Greenfield PF, Lee PL (1994) Water Res 28:475 30 Alvarez Ramirez J, Femat R, Monroy O, Cuervo F (1996) Proceedings of the 1996 IEEE International Conference on Control Applications. Dearborn, MI – September 15–18, 1996 1019 31. Guwy AJ, Hawkes DL, Hawkes FR (1995) Water Res 29:977 32. Angelidaki I, Ellegaard L, Ahring BK (1992) Biotechnol Bioeng 39:351 33. Veiga MC, SOTO M, Mendez R, Lema JM (1990) Water Res 24:1551 34. Pullammanappallil PC, Harmon J, Chynoweth DP, Lyberatos G, Svoronos SA (1991) Appl Biochem Biotechnol 28/29:33 35. Hawkes FR, Guwy AJ, Rozzi AG, Hawkes DL (1993) Water Res 27:167 36. Merkel W, Krauth K (1999) Water Res 33:2011 37. Spérandio M, Paul E (1996) Biotechnol Bioeng 53:243 38. Archer DB, Hilton MG, Adams P, Wieko H (1986) Biotechnol Lett 8:197 39. Pauss A, Samson R, Guiot SR, Beauchemin C (1990) Biotechnol Bioeng 35:492 40. Switzenbaum MS, Giraldogomez E, Hickey RF (1990) Enzyme Microb Technol 12:722 41. Kidby DW, Nedwell DB (1991) Water Res 25:1007 42. Guwy AJ, Hawkes FR, Hawkes DL, Rozzi AG (1997) Water Res 31:1291 43. Collins LJ, Paskins AR (1987) Water Res 21:1567 44. Björnsson L, Hörnsten EG, Mattiasson B (2001) Biotechnol Bioeng 73:35 45. Björnsson L, Murto M, Jantsch TG, Mattiasson B (2000) PhD Thesis. Intensification of the biogas process, Department of Biotechnol, Lund University 2000 1 46. Lundström I (1981) Sens Actuators 1:403 47. Lee MJ, Zinder SH (1988) Appl Environ Microbiol 54:1457 48. Pauss A, Guiot SR (1993) Water Environ Res 65:276 49. Pauss A, Andre G, Perrier M, Guiot SR (1990) Appl Environ Microbiol 56:1636 50. Frigon JC, Guiot SR (1995) Enzyme Microb Technol 17:1080 51. Dochain D, Perrier M, Pauss A (1991) Ind Eng Chem Res 30:129 52. CordRuwisch R, Mercz TI, Hoh CY, Strong GE (1997) Biotechnol Bioeng 56:626 53. Strong GE, CordRuwisch R (1995) Biotechnol Bioeng 63 54. Kuroda K, Silveira RG, Nishio N, Sunahara H, Nagai S (1991) J Ferment Bioeng 71:418 55. Meyer B, Heinzle E (1998) Biotechnol Bioeng 57:127 56. Hansson M, Nordberg Å, Mathisen B (1999) Proceeding at IAWQ II Symposium. Barcelona 15–17 June, 1999 II: 61 57. Brandgård J, Nordberg Å, Sundh I, Mandenius, C-F, Mathisen B (1999) Proceeding at IAWQ II Symposium. Barcelona 15–17 June, 1999 II: 57 58. Hill DT (1982) Trans ASAE 25:1374 59. Hill DT, Bolte JP (1989) Biol Wastes 28:33 60. Hill DT, Holmberg RD (1988) Biol Wastes 23:195 61. Cobb SA, Hill DT (1991) Trans ASAE 34:2564 62. Ahring BK, Sandberg M, Angelidaki I (1995) Appl Microbiol Biotechnol 43:559 63. Rozzi A, Massone A, Antonelli M (1997) Water Sci Technol 36:183 64. Sørensen AH, Winther-Nielsen M, Ahring BK (1991) Appl Microbiol Biotechnol 34:823 65. Mosche M, Jordening HJ (1999) Water Res 33:2545 66. Hickey RF, Switzenbaum MS (1991) Res J Water Pollut Control Fed 63:141 67. Alonso MS (1992) Biotechnol Lett 14:535 68. Öztürk M (1991) Water Res 25:1509
Monitoring and Control of Anaerobic Reactors
181
69. Pullammanappallil PC, Chynoweth DP, Lyberatos G, Svoronos SA (2001) Bioresource Technol 78:165 70. Öztürk M (1993) J Environ Eng 119:321 71. Ahring BK, Westermann P (1988) Appl Environ Microbiol 54:2393 72. Allison MJ (1978) Appl Environ Microbiol 35:872 73. Angelidaki I, Ahring BK (1995) Antonie Van Leeuwenhoek 68:285 74. Mösche M, Jördening H-J (1998) Appl Microbiol Biotechnol 49:793 75. Wang QH, Kuninobu M, Ogawa HI, Kato Y (1999) Biomass Bioenergy 16:407 76. Aguliar A, Casas C, Lafuente J, Lema JM (1990) J Ferment Bioeng 69:261 77. Matthies C, Schink B (1992) Appl Environ Microbiol 58:1435 78. Aguliar A, Casas C, Lema JM (1995) Water Res 29:505 79. Mladenovska Z, Ahring BK (2000) FEMS Mcrobiol Ecol 31:225 80. Vavilin VA, Lokshina LYa (1996) Bioresurce Technology 57:69 81. Björnsson L, Murto M, Henrysson T (1997) Proceedings at the 8th International Conference on Anaerobic Digestion 3:79 82. Slater WR, Merigh M, Ricker NL, Labib F, Ferguson JF, Benjamin MM (1990) Water Res 24:121 83. Zumbusch Pv, Meyer-Jens T, Brunner G, Märkl H (1994) Appl Microbiol Biotechnol 42:140 84. Powell GE, Archer DB (1989) Appl Microbiol Biotechnol 33:570 85. Costello DJ, Greenfield PF, Lee PL (1991) Water Res 25:859 86. Costello DJ, Greenfield PF, Lee PL (1991) Water Res 25:872 87. Romli M, Keller J, Lee PJ, Greenfield PF (1995) Process Saf Environ Protect 73:151 88. Ren N, Wan B, JuChang H (1996) Biotechnol Bioeng 54:428 89. Hansen KH, Angelidaki I, Ahring BK (1998) Water Res 32:5 90. Hansen KH, Angelidaki I, Ahring BK (1999) Water Res 33:1805 91. Angelidaki I, Ahring BK (1993) Appl Microbiol Biotechnol 38:560 92. Braun R, Huber P, Meyrath J (1981) Biotechnol Lett 3:159 93. Schlegel S, Baumann P (1996) Water Sci Technol 33:139 94. APHA-AWWA-WPCF (1975) Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington 95. Hashimoto AG (1983) Biotechnol Bioeng 25:185 96. Hashimoto AG (1982) Biotechnol Bioeng 24:2039 97. Amann R, Lemmer H, Wagner M (1998) FEMS Mcrobiol Ecol 25:205 98. Ince O, Anderson GK, Kasapgil B (1995) Water Res 29:349 99. Sørensen AH, Ahring BK (1993) Appl Microbiol Biotechnol 40:427 100. Macario AL, deMacario EC (1993) Biomass Bioenergy 5:203 101. Raskin L, Poulsen LK, Noguera DR, Rittmann BE, Stahl DA (1994) Appl Environ Microbiol 60:1241 102. Raskin L, Amann RI, Poulsen LK, Rittmann BE, Stahl DA (1995) Wat Sci Tech 31:261 103. Lange M, Tolker-Nielsen T, Molin S, Ahring BK (2000) Appl Environ Microbiol 66:1796 104. Peck MW, Chynoweth DP (1992) Biotechnol Bioeng 39:1151 105. Nordberg Å, Nordkvist E (1999) Proceeding at IAWQ II Symposium. Barcelona 15–17 June, 1999 II: 53 106. Nordberg Å, Hansson M, Sundh I, Nordkvist E, Carlsson H, Mathisen B (1999) Water Sci Technol 41:1 107. Varel VH, Isaacson HR, Bryant MP (1977) Appl Environ Microbiol 33:298 108. Varel VH, Hashimoto AG, Chen YR (1980) Appl Environ Microbiol 40:217 109. Linke B (1997) Environ Technol 18:849 110. Anderson GK, Yang G (1992) J Environ Eng 118:551 111. Angelidaki I, Ahring BK (1994) Water Res 28:727 112. Cha GC, Noike T (1997) Water Sci Technol 36:247 113. Kayhanian M, Tchobanoglous G (1992) BioCycle 33:58 114. Ahring BK, Angelidaki I, Johansen K (1992) Water Sci Technol 25:311 115. Angelidaki I, Ahring BK (1997) Biodegradation 8:221
182
P.F. Pind et al.: Monitoring and Control of Anaerobic Reactors
116. 117. 118. 119. 120. 121. 122.
Tafdrup S (1994) Water Sci Technol 30:133 Tafdrup S (1995) Biomass Bioenergy 9:303 Heinzle E, Dunn IJ, Ryhiner G (1993) Adv Biochem Eng/Biotechnol 48:79 Marsili-Libelli S, Beni S (1996) Ecol Model 84:215 Pretorius WA (1994) Wat Sci Tech 30:1 Simeonov I (1994) Chem Biochem Eng Q 8:45 Wilcox SJ, Guwy AJ, Hawkes DL, Hawkes FR (1994) Biogas Technology as an Environmental Solution to Pollution. REUR Technical Series. 46 Åström K, Wittenmark B (1995) Adaptive control. Addison-Wesley Publishing Company Dochain D, Perrier M (1993) Wat Sci Tech 28:283 Bastin G, Dochain D (1990) On-line Estimation and Adaptive Control of Bioreactors. Elsevier Harmon JL, Svoronos SA, Lyberatos G, Chynoweth D (1993) Biomass Bioenergy 5:279 Perrier M, Dochain D (1993) Int J Adaptive Cont Signal Proces 7:309 Harmon J, Pullammanappallil PC, Svoronos SA, Lyberatos G, Chynoweth DP (1990) American Autom Control Council 3:1540 Konstantinov KB, Yoshida T (1992) Biotechnol Bioeng 39:479 Guthke R, Schmidt-Heck W, Pfaff M (1998) J Biotechnol 65:37 Bhat N, McAcoy TJ (1990) Comput Chem Eng 14:573 Hertz J, Krogh A, Palmer RG (1991) Introduction to the Theory of Neural Computation. Addison-Wesley, New York Kosko B (1992) Networks and Fuzzy Systems. Prentice Hall, Englewood Cliffs Emmanouilides C, Petrou L (1997) Comput Chem Eng 21:113
123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134.
Received: May 2001
