WO1995026806A1

WO1995026806A1 - Permeation and biological treatment of waste gas

Info

Publication number: WO1995026806A1
Application number: PCT/GB1995/000744
Authority: WO
Inventors: Andrew Livingston; Luisa Maria Freitas Dos Santos
Original assignee: Imperial College Of Science, Technology & Medicine
Priority date: 1994-03-30
Filing date: 1995-03-30
Publication date: 1995-10-12
Also published as: EP0752909A1; AU2080795A; GB9406323D0; JPH09510917A

Abstract

A method of reducing the concentration of at least one gaseous or vapour phase organic compound present in a gaseous stream, wherein said gaseous stream is caused to flow in contact with an interior surface of a substantially water-insoluble selectively permeable sheet, tubular or envelope-type polymeric membrane whose permeability to the said at least one organic compound exceeds its permeability to chloride ion, whilst simultaneously maintaining in contact with an external surface of said membrane an aqueous reaction medium containing biologically active reaction means such as biomass bacteria or algae capable of reacting with said at least one compound after it permeates from the gaseous stream through said membrane from said interior surface thereof to said external surface.

Description

PERMEATION AND BIOLOGICAL TREATMENT OF WASTE GAS

Many industrial waste gas streams contain toxic, volatile compounds which we have found can be separated from the gas stream through preferential membrane permeation. The preferential permeation of solute compounds across a membrane appears to be related to the high intrinsic solubility of the solute compound in the membrane phase. For instance, we have found that many volatile organic compounds (VOC's) such as chlorinated solvents (DCE- dichloroethane) and aromatics (benzene, toluene) have high solubilities in silicone rubber membranes, and will diffuse across these membranes under the influence of a concentration driving force. These are examples of organic compounds which can be treated by the present method.

This preferential permeation can be combined with subsequent biodegradation of the permeating species. Biodegradation renders the toxic VOC harmless by converting it into harmless components such as carbon dioxide, water and mineral constituents such as chloride ion. Thus a membrane bioreactor can be used to extract toxic VOC's from a waste gas stream, and then biodegrade them. The advantages of this process over other biological treatment technologies such as bioscrubbers (shown schematically in Figure 2) and biofilters are that:

(1) A large gas-liquid interface and high gas velocities can be obtained without the significant pressure drops observed in biofilters and bioscrubbers.

(2) The membrane acts as a barrier between the biomedium and the contaminated gas stream (see figure 1) thus avoiding the entry into the biomedium of compounds (such as HC1) , which would hinder biological degradationT

(3) Ready control of pH and removal of bioxidation products is possible.

(4) The solutes may be transferred at a higher rate than in a conventional system since mass transfer through silicone rubber may be faster than mass transfer through a water film.

(5) The microorganisms can be maintained in a favourable growth environment, as they only receive the specific compound they need as carbon source from the gas stream.

(6) The ageing and reguired renewal of packing material is avoided.

(7) The control of the process parameters for bacterial growth in the system is more easily feasible.

(8) The treatment of compounds with low water solubility is possible.

(9) The dimensions of the installed equipment are more compact.

(10) . No disposal of sludge (as in bioscrubbers) is necessary.

According to this invention there is provided a method of reducing the concentration of at least one gaseous or vapour phase organic compound present in a gaseous stream, wherein said gaseous stream is caused to flow in contact with an interior surface of a substantially water-insoluble selectively permeable sheet, tubular or envelope-type polymeric membrane whose permeability to the said at least one organic compound exceeds its permeability to chloride ion, whilst simultaneously maintaining in contact with an external surface of said membrane an aqueous reaction medium containing biologically active reaction means such as biomass bacteria or algae or other microorganisms or immobilised cells capable of reacting with said at least one compound after it permeates from the gaseous stream through said membrane from said interior surface thereof to said external surface. Further embodiments are defined in the sub claims.

Any dense phase polymer through which the volatile organic compounds diffuse can be used for this process. Membrane modules may be constructed in any suitable configuration, e.g. spiral wound flat sheets, layered flat sheets, shell and tube.

Diagrams showing the operating principle are given in Figure 1, a system configuration including a remote bioreactor and membrane module is given in Figure 3, and operating data from determinations using the system including the remote bioreactor and the membrane module to biodegrade DCE containing waste gas are given in Figure 8.

A preferred feature of the present invention resides in use of a dense phase, hollow membrane such as a plurality of tubular synthetic elastomers like silicone rubber to create an interface between the gas and aqueous biological phases.

The present method is particularly useful in gas and vapour permeation, where problems of membrane-fouling or -scaling can be minimised, and parallel reactions may be avoided. The selective mass transport allows the recovery of single compounds without an additional process step. Thus, a membrane process can be described as a special type of extraction process which is more suitable for applications where the organic vapour content is too high to make carbon adsorption feasible, but too low to permit incineration.

One step further in the use of membranes is the combination of membrane extraction and a subsequent bioconversion. Microporous membranes have limited selectivities, which under certain conditions prevent complete VOC extraction. Additionally, the transmembrane pressure difference that destroys the membrane is much lower for microporous membranes, and the pores of the membrane can be blocked by biomass. Therefore, the use of dense membranes to treat VOCs is preferable.

A strain of microorganisms, for example Xanthobacter. autotrophicus GJ10 described by Janssen et al in 1985 is capable of utilising DCE and other haloalkanes that are structurally related to DCE (e.g. 1-chlorobutane, 1-chloropropane, 1, 3-dichloropropane, bromoethane) as sole sources of carbon and energy.

Without wishing to be bound by any theoretical considerations it appears the mechanism (shown in Figure 4) is as follows:

The first step is the conversion of DCE to 2- chloroethanol by a constitutively produced haloalkane dehalogenaεe. Following this, chloroethanol is converted, via chloroacetaldehyde to 2-chloroacetic acid by a quinoprotein (PQQ) alcohol dehydrogenaεe and a coenzyme (NAD) dependent aldehyde dehydrogenase. The second dehalogenase dechlorinates the 2-chloroacetic acid to glycolic acid which is finally degraded to biomass, water, carbon dioxide and hydrochloric acid. As the whole degradation is aerobic, the overall stoichiometric equation can be expressed by

C₂H₄Cl₂+0₂+NH₃ —> biomass (e.g. CH₂0_{Q 5}N_{Q 2}) +C0₂+H₂0+HC1

The overall balance reveals that 0.72 mg of chloride ions should be released into the biomedium if 1 mg of DCE is completely degraded. The formation of hydrochloric acid and the subsequent neutralisation with NaOH results in the accumulation of sodium chloride. Sodium chloride at a concentration of 100 mM already slows the growth rate significantly down. Thus, preferred bioreactors facilitate regulation of the amount of chloride ions present in the biomedium by means of proper pH control and nutrient salts flow.

In order that the invention may be illustrated, more easily appreciated and readily carried into effect by one skilled in the art, reference will now be made to the following examples purely by way of non-limiting embodiments, and to the accompanying drawings, wherein:

Figure 1 shows the principle of operation of the present membrane bioreactor system,

Figure 2 illustrates mass transfer and biodegradation in a bioscrubber,

Figure 3 shows a membrane module and remote bioreactor configuration,

Figure 4 illustrates the metabolic pathway for dichloroethane (DCE) degradation by Xanthobacter autotrophicus GJ10,

Figure 5 shows apparatus for investigating transmembrane transfer of DCE from a waste gas stream into a receiving solution,

Figure 6 shows an apparatus for determination of mass transfer coefficients,

Figure 7 shows a more detailed arrangement of membrane module and remote bioreactor apparatus for quantitative measurements of DCE extraction and biodegradation,

Figure 8 demonstrates removal efficiency of DCE in the membrane bioreactor,

Figure 9 show the results of transfer test using silicone rubber membrane. DCE concentration curves in the gas vessel are identified thus -A in the receiving solution -♦ and in the gas -O

Figure 10 shows mass transfer results with membrane module (gas flow rate: 750 mL min^-1, receiving solution flow rate: 1050 mL min^-1) :

DCE concentration curves in the gas production vessel (A) , gas inlet (o) , gas outlet (•) and the receiving solution

(♦),

Figure 11 shows mass transfer results with membrane module (gas flow rate: 1500 mL min , receiving solution flow rate: 1050 mL min^-1) : DCE concentration curves in the gas production vessel (A) , gas inlet (o) , gas outlet (•) and the receiving solution (♦),

Figure 12 shows^" mass transfer results with membrane module (gas flow rate: 750 mL min , receiving solution flow rate: 2300 mL min^-1) : DCE concentration curves in the gas production vessel (A) , gas inlet (o) , gas outlet (•) and the receiving solution (♦) ,

Figure 13 shows mass transfer results with membrane module (gas flow rate: 1500 L min , receiving solution flow rate: 2300 mL min^-1) : DCE concentration curves in the gas production vessel (A) , gas inlet (o) , gas outlet (•) and the receiving solution (♦) ,

Figure 14 shows the gas flow rate of DCE waste gas,

Figure 15 shows the pressure drop on the shell-side of the membrane,

Figure 16 shows removal efficiency of DCE in the membrane bioreactor,

Figure 17 shows the flux of DCE through the membrane,

Figure 18 shows carbon evolution: expected values assuming complete mineralisation of DCE to carbon dioxide (A) ; values from the analysis of the gas exhaust of the reactor (♦) ; carbon leaving the bioreactor as biomass (o) ,

Figure 19 shows chloride evolution in the biological medium: expected values assuming complete mineralisation of DCE (A) and measured concentrations (♦) . EXAMPLE 1

In order to obtain Xanthobacter autotrophicus GJ10 cells to inoculate the bioreactor, several shakeflasks with different volumes (100 mL, 250 L, 500 mL) were filled with nutrient salts solution (the composition of these nutrient salts and the vitamins medium is given in Table 1 and dichloroethane DCE at a concentration of 1000 mg L^-1 DCE. These shakeflasks were inoculated with Xanthobacter autotrophicus GJ10 and placed in an orbitant incubator at 30°C for 4 days. Control flasks were employed in parallel with any other microbial activity precluded by the addition of sodium azide (NaN₃) . The sealed flasks were subsequently analysed for DCE and Cl- concentrations and turbidity. Bacterial growth was observed with increasing turbidity which is proportional to the number of cells in the liquid, and decreasing DCE concentration showing that the microorganisms were able to utilise the provided DCE as a sole carbon source . Samples of Xanthobacter autotrophicus GJI O cells were stored in f lasks containing glycerol and placed in a freezer compartment .

Ta_ble 1 : Nutrient salts and vitamins medium for Xanthobacter autotrophicus GJ10

Assays Liquid Phase DCE Concentration

Gas chromatography was employed to determine concentrations of DCE in the waste gas, the gas vessel, the exhaust gas and in the biomedium. A gas chromatograph (GC) with a Flame lonisation Detector (FID) was fitted to a 25 m long 0.23 mm i.d. megabore column with BP-101 (SGE (Australia) Ltd) as the stationary phase. 1,1,2- Trichloroethane was mixed with the DCE sample for use as the internal standard. A lμL sample of the DCE and the TCE mixture (in equal parts) was injected into the column and the GC run with the following temperature program:

initial column oven temperature: 40°C initial time (period in which temperature stays constant): 2 min rate of column over temperature rise: 20C° per min final oven temperature: 100°C final time (period in which temperature is held before the oven is cooled down) 0.5 min

The concentration of DCE in the sample was calculated as follows:

The peak area generated by the GC is proportional to the mass of the detected compound. With the same volume of sample and standard, the ratio between the amount of DCE and TCE in sample and in the standard can be expressed by

and

^M2s ^S2s ^A2s

where 1 = DCE,

2 = TCE, s = sample, sa = standard. With the use of the same amount of TCE in the sample and the standard, the DCE concentration is obtained by _S = ^Als ^A2sa ς ^lΞ A ^A2s ^rtAlεa ^lεa

Where A is the value for the peak area in %.

The limit of sensitivity of this assay was approximately 2.5 mg L with an average error of 1%. Gas Phase DCE Concentration

DCE concentration in the waste gas and the exhaust gas was measured in the same GC and using the same temperature program. A 1 mL gas sample was injected into the column and the concentration was calculated against an external liquid standard.

Since different sample volumes were injected (1 mL gas and 1 μL liquid) , the concentrations were determined from

A_G V_L

^SG - ^Cl ^C2 A _V^ ^SL

Where A is the absolute peak value. The factors C-, and C₂ take into account different attenuations and the presence of other compounds in the standard. For the gas measurements, liquid DCE solution was injected and analysed at the same attenuation as the samples, so that these factors became 1.

The limit of sensitivity of this assay was approximately 0.05 mg L^-1 with an average error of 5%. Chloride Assay

Chloride was assayed colorimetrically according to a known technique. To 2.5 mL of aqueous sample solution 3.0 L of mercuric thiocyanate solution (1.5 g mercuric thiocyanate in 500 mL of 98% ethanol) and 1.0 mL of iron alumn solution (8g ferric ammonium sulphate in 100 mL 6M nitric acid) were added. The mixture was shaken and the colour allowed to develop for 10 min after which the absorption was measured at 460 nm. The obtained value was compared with a standard calibration curve and the chloride concentration was calculated by linear interpolation. The uncertainty of this assay (quoted as the standard deviation of 5 separate determinations) was 5% at the 10 mg L level.

Prior to analysis, samples were centrifuged in order to remove cell material which might cause interferences. Carbon dioxide assay

Carbon dioxide was measured in an infra-red gas analyser (Servo ex PA 404). The principle is based on the fact that many gases and liquids absorb energy in the infra¬ red region of the spectrum because of resonant molecular vibrations. These resonances produce highly characterised spectra which can be used to identify substances.

In order to quantify the concentration of carbon dioxide in the exhaust gas of the bioreactor, the gas stream was diverted through a cooler to condense the remaining moisture which might damage the infra-red analyser. The analyser read the carbon dioxide concentration in volume percent (volume carbon dioxide per volume air) which was converted in mg L^-1. In order to get the amount of carbon dioxide generated by the bioconversion, the concentration in the air entering the bioreactor (0.03% - 0.04%) was subtracted. Example 2

Membrane Bioreactor with Remote Fermenter or Bioreactor Configuration for DCE Removal

In order to determine mass transfer and extraction, an apparatus was constructed corresponding to Figure 5 using a tubular silicone rubber membrane (1 m length, 2 mm inner diameter, 0.5 mm wall thickness), being a commercially available poly(di ethylsiloxane) and inert filler for strength.

The gaseous waste stream was created by bubbling air into a vessel containing 10L of DCE solution with an initial concentration of 1000 mg L . Due to air stripping, DCE was transferred from the solution to the gaseous phase. This synthetic gas was pumped through the membrane at a flow rate of 0.5 L min . Due to this method used for gas generation, the gas produced also contained water vapour and therefore a humidity trap with silica gel was used to prevent humidity in the gas pump or on the gas side of the membrane. In order to keep constant atmospheric pressure in the gas production vessel, an outlet with an activated carbon trap was installed (see Fig. 5)

The synthetic DCE gas stream passed through the silicone tubing which was placed in a shakeflask containing the receiving solution. This solution was 530 L distilled water at pH 7.0.- The shakeflask was kept at a constant temperature of 30°C using a.waterbath.

A steady flow rate of 3.7 mL min^-1 of receiving solution was pumped through the receiving solution reservoir using a peristaltic pump. Therefore, it could be assumed that the concentration of pollutants in the receiving solution would achieve a steady-state value over the period of time.

All the tubing in the system other than the silicone rubber membrane was of polytetrafluoroethylene (PTFE) which had been chosen to minimise adsorption of DCE from the waste gas and the receiving solution.

Measurements of the DCE concentrations in the gas production vessel, the receiving solution and the gaseous waste stream were performed at regular intervals until the whole system reached equilibrium. Example 3 Membrane Bioreactor Operation Mass Transfer Results

In order to quantify mass transfer through the membrane module described in Example 1, the mass transfer coefficients were determined. Additionally, the influence of different gas and water flow rates were examined.

The apparatus used for such determinations is shown in Fig. 6:

The membrane module employed was a spiral wound silicone rubber tubing module with a transfer surface area of 2.5 m . With the approximate volume of the silicone membrane of 2.10~³m , the specific surface area a_M = A_H/V_M was 1667 m^-1.

The DCE containing gas was circulated on the tube side at flow rates of 750 and 1500 mL min^-1. The gas generation unit used in the previous examples was modified, as the transfer rates were expected to be higher and the concentration in the gas production vessel would consequently decrease rapidly preventing a constant gas flow. Thus, a high concentration DCE solution (2800-2950 mg L^-1) was added to the gas production vessel at flow rates that varied between 0.18 and 0.24 mL min^-1, and the same amount of solution was removed from the system via a multichannel peristaltic pump.

Water at a rate of 5.7 mL min- was pumped into a receiving solution flask containing 3.7 L water (including membrane hold-up) . The receiving solution was continuously recirculated on the shell side of the module at flow rates of 1050 and 2300 mL min^-1.

The DCE concentrations in the gas production vessel and in the receiving solution as well as the gas concentrations in the inlet and outlet of the membrane module were measured. Furthermore, the change in concentration of the highly concentrated DCE solution due to volatilisation was monitored.

In order to treat VOC contaminated air streams, this bioreactor system was run for 11 days to demonstrate system performance.

The membrane bioreactor system used (see Fig. 7) was similar to the previously described configuration for the determination of mass transfer coefficients. In this setup the receiving solution vessel was replaced by a continuously stirred tank reactor (CSTR) used as a fermenter. The CSTR was inoculated with Xanthobacter autotrophicus GJ10 cells grown on DCE under non-sterile conditions.

The membrane module employed was the same as described above.

The total biomedium volume was approximately 2 L including volume of the CSTR and membrane hold-up. Gas flow was set at a rate of 100 mL min- entering at the base of the CSTR through a stainless steel sparger. In order to achieve complete mixing and homogeneous aeration, the stirrer speed was set at 500 rpm. The temperature of the biomedium was held constantly at 30°C by a temperature controller with a heating element. A pH controller held the pH at 7.0 in the biomedium via the adding of acid (IM sulphuric acid solution) or base (IM sodium hydroxide solution) as necessary. Nutrient salt solution was added to the CSTR at a flow rate of 0.95 L min^-1 through a multichannel peristaltic pump. A water cooled condenser was fitted to the gas outlet to minimise the loss of DCE due to air stripping caused by the aeration of the bioreactor. Gas exiting the reactor was piped either to a fume cupboard or into the sample loop of the C0₂ analyser as required for C0₂ analysis.

The biomedium was recirculated at an initial flow rate of 2000 L min- on the shell side of the membrane. As the gas flow rate on the tube side, which was initially set at 1500 L min^-1, started to decrease due to membrane expansion, the biomedium flow rate was reduced to 1000 L min . The pressure drop on the biomedium side was measured regularly using a mercury manometer.

In order to perform mass balances for the compounds entering and exiting the system both DCE removal and product formation as a result of biodegradation were monitored.

The DCE concentration in the gas entering and exiting the membrane was measured. In order to maintain a constant gas flow, the concentration in the gas production vessel was monitored and the flow rate of the peristaltic pump adding the highly concentrated DCE solution was adjusted to between 0.176 and 0.32 mL min^-1.

Samples of the biomedium were analysed for DCE concentration and chloride ions. Gas exiting the reactor was analysed for both DCE and C0₂. Additionally, measurements of the biomass exiting the system were performed. EXAMPLE 4 - Mass Transfer through Silicone Rubber

The results of the transfer test using silicone rubber tubing is shown in Fig. 9.

Equilibrium DCE concentrations between the gas phase and the receiving solution were established after approximately 330 min. The DCE concentration in the gas phase remained at a constant value of 22 mg L , whereas the liquid concentration in the gas production vessel was 640 mg L^-1, and the concentration in the receiving solution reached a value of 295 mg L . Example 5 - Extraction and Destruction of DCE in Membrane Bioreactor System

The system reached steady state almost after 390-500 min of operation dependent on the amount of DCE crossing the membrane. The graphs in Figures 10 to 13 show the results of the mass transfer tests carried out with the membrane module at different flow rates on tube and shell side. The final DCE concentrations at steady state were determined.

The gas flow rate, which was initially set at 1500 mL min^-1, decreased during the operating period due to swelling of the silicone rubber and growth of biomass on the shell- side which caused expansion of the membrane. The change of these gas flow rates is shown in Fig. 14.

In order to obtain more information on the effects of biomass growth on the shell-side, the pressure drop was measured. The results show, that after a period of 8 days a constant value for the pressure drop of approximately 600 bar was achieved.

Therefore, it could be concluded that the biofil had reached a steady state thickness.

The development of the pressure drop is given in Fig. 15.

The average DCE gas concentration entering the membrane was 0.65 mg L^-1 and decreased to a concentration of 0.06 mg L- at the membrane outlet. The average removal efficiency shown in Fig. 16 was calculated as the ratio between outlet and inlet concentration and was approximately 91%.

The flux of DCE across the membrane calculated as the product of the gas flow rate and difference in DCE concentration between membrane inlet and outlet increased within the first two days of running the bioreactor apparatus. The concentration in the biomedium increased and subsequently decreased within this time span, as the microorganisms acclimatised to the new growth conditions. After 48 hours, no DCE was detected in the biomedium, and therefore the driving force between gas and liquid phases increased leading to a maximum DCE flux of approximately 55 mg h .

After the fourth day of operation, the DCE flux across the membrane decreased due to the intensive biofilm formation. A constant flux of 25 mg h was reached after 7 days as shown in Fig. 17.

When dealing with a volatile compound such as DCE, it is of major importance to verify product formation as a result of biodegradation instead of only relying on the measurements of substrate disappearance.

Hence, the formation of the products due to microbial metabolism such as carbon dioxide, chloride and biomass was evaluated in addition to the determination of the DCE removal. Carbon balance

The amount of carbon exiting the system was measured as C0₂ exiting through the gas outlet of the bioreactor, and as biomass leaving via the biomedium overflow.

The carbon evolved as C0₂ was calculated as the product of the air flow and the C0₂ concentration measured in the C0₂ analyser. The amount of carbon leaving as biomass was obtained from the product of the biomedium flow rate of the overflow and the amount of suspended biomass which stayed constant at a value of 20 mg L^-1. The amount of DCE lost due to air stripping was found to be less than 0.8% of the total amount of DCE entering the membrane bioreactor system and could therefore be neglected.

The results of the calculations for the carbon balance are given in Fig. 18. Chloride balance

Fig. 19 shows the measured evolution of chloride ions calculated from the biomedium overflow flow rate and the chloride concentration in the biomedium, and the theoretical evolution of chloride assuming 100% conversion of DCE to chloride.

The comparison of the two curves reveals a good agreement between the theoretical and the actual evolution.

In summary, a membrane bioreactor system has been used to remove an organic pollutant 1, 2-dichloroethane (DCE) from a gaseous waste stream. In comparison to conventional technologies such as thermal and catalytic oxidation, absorption, adsorption and condensation, this process allows the treatment of contaminated air streams at low flow rates and with low pollutant concentrations. In this configuration, a plurality of elongate, tubular hollow silicone rubber membranes separate the contaminated gas stream from the biomedium. This allows selective transfer of the organic compound through the membrane to the biomedium where biodegradation occurs.

The conditions in the biomedium can be controlled independently from the waste gas side. At a gas flow rate of 770 mL min^-1 and an average DCE concentration of 0.65 mg L^-1, 91% of the DCE present in the synthetically produced gaseous stream was biodegraded. During operation, extensive growth of biofilm attached to the membrane surface on the biomedium side was observed, which caused a pressure drop on the biomedium side of the membrane of approximately 600 bar when steady state conditions were achieved. The release of chloride ions due to the mineralisation of DCE was found to be 99%, while conversion of DCE to C0₂ was found to be 60% of the amount of carbon crossing the membrane. This discrepancy is thought to be due to the extensive biofilm formation. The mass transfer of DCE across the membrane in absence of microbial activity was also determined. With a gas flow rate of 770 mL min^-1 and a water flow rate of 1050 mL min , the overall mass transfer coefficient was determined to be 0.83.10^-3 m s^-1. The comparison of the DCE fluxes across the membrane in the presence and absence of microbial activity revealed that the biofilm lowers significantly the amount of DCE crossing the membrane.

Claims

1. A method of reducing the concentration of at least one gaseous or vapour phase organic compound present in a gaseous stream, wherein said gaseous stream is caused to flow in contact with an interior surface of a substantially water-insoluble selectively permeable sheet, tubular or envelope-type polymeric membrane whose permeability to the said at least one organic compound exceeds its permeability to chloride ion, whilst simultaneously maintaining in contact with an external surface of said membrane an aqueous reaction medium containing biologically active reaction means such as biomass bacteria or algae or other microorganisms or immobilised cells capable of reacting with said at least one compound after it permeates from the gaseous stream through said membrane from said interior surface thereof to said external surface.

2. A method as claimed in claim 1, wherein the membrane is generally tubular or in the shape of a hollow envelope, or in the form of a flat sheet or is spirally wound.

3. A method as claimed in claim 1 or 2, wherein the gaseous stream comprises waste gas from an industrial or municipal process which contains a plurality of pollutant organic compounds.

4. A method as claimed in any preceding claim, wherein at least one of the organic compounds is volatile.

5. A method as claimed in any preceding claim, wherein a biologically active film of microorganisms forms at the external surface of said membrane in contact with the reaction medium.

6. A method as claimed in any preceding claims, wherein the gaseous stream contains one or more compounds which impede development or growth of said reaction means.

7. A method as claimed in any preceding claim, wherein the membrane is formed of an elastomer such as synthetic or natural rubber in which the said at least one organic compound exhibits solubility.

8. A method as claimed in claim 7, wherein the elastomer comprises a synthetic silicone elastomer.

9. A method as claimed in claim 8, wherein the silicone comprises an alkylsiloxane such as dimethylpolysiloxane silicone.

10. A method as claimed in any one of claims 7 to 9 wherein the elastomer is a dense phase heterogeneous elastomer containing at least 5%, preferably 10-20% of an inert filler.

11. A method as claimed in any preceding claim wherein the organic compound comprises a halogenated hydrocarbon.

12. A method as claimed in claim 11 wherein the compound is chlorinated.

13. A method as claimed in claim 11 or 12 wherein the compound is an optionally substituted alkane.