WO2001007118A1

WO2001007118A1 - Filtration system

Info

Publication number: WO2001007118A1
Application number: PCT/GB2000/002811
Authority: WO
Inventors: Richard James Holmes
Original assignee: UK Secretary of State for Defence
Current assignee: UK Secretary of State for Defence
Priority date: 1999-07-22
Filing date: 2000-07-20
Publication date: 2001-02-01
Anticipated expiration: 2002-01-22
Also published as: AU6005400A; GB9917105D0

Abstract

A filter suitable for use in the high temperature and pressure section of an aircraft Environmental Control System (ECS). The filter is able to remove, and optionally to destroy, harmful gases or vapours of nerve, mustard and lewisite agents from the air passing through the filter. Advantageously, the filter may be associated with the engine bleed air supplies without impeding the performance of the engine by decreasing the flow rate or supply of air from the bleeds to the ECS. The filter includes an adsorbent composition comprising a zeolite molecular sieve.

Description

FILTRATION SYSTEM

This invention relates to filtration systems. More particularly the present invention relates to filtration systems for the removal and possible denaturing of noxious or toxic substances or other pollutants in a fluid.

In the description which follows the present invention will be described with particular reference to its preferred use in the removal of harmful gases, such as chemical warfare agents, in the air inlet to an aeroplane environmental control system and the conditioning air supplies derived from it. However, it is not intended that the invention be limited to such use since the invention has equal utility in other applications, such as the removal of other noxious gases or vapours in the intake air to a building air conditioning system, particularly where constraints of size and/or weight are of prime importance.

Aircraft electronics and systems require conditioned air supplies to maintain cabin and equipment bay temperatures. A loss of this supply will render the aircraft inoperable. In addition, conditioned air supplies are passed into the cockpit or cabin spaces containing personnel and mission systems/avionics. The ait* is obtained from the auxiliary power unit or main engine bleed, and is conditioned within an environment control system (ECS), which is generally present on modern jet aircraft (including rotary wing types). The air is then directed to the required locations through a series of pipes and ducts. Many of these distribution pipes and ducts are manufactured from, or contain, silicon rubber or similar compounds which can be attacked or degraded by chemicals present in the ambient surrounding fluid. These pipes and ducts are usually inaccessible.

It is well known that chemical warfare (CW) agents (as both vapour and liquid/aerosol) can be absorbed by surfaces and by components (particularly silicone rubber and similar compounds) and, once absorbed, they can be released giving rise to a desorption hazard. This hazard may persist for a considerable period of time, and would require users to wear respiratory protection equipment. Additionally, these CW agents may physically damage or render inoperable certain types of equipment, including unprotected microelectronics.

In the event of CW exposure, agents would be ingested through the engines and into the ECS. Once within the ECS, agents would then be absorbed and may be distributed throughout the cabin and bays. As a consequence of such contamination the aircraft could be rendered inoperable.

The CW agents known to be most damaging to aircraft and operations are nerve, mustard and lewisites (e.g. GB (sarin), GD (soman), HD (mustard), VX, L (Lewisite)). CW agents which are gaseous at normal conditions of temperature {e.g. AC (hydrogen cyanide), CK) are not damaging in this respect.

It is therefore desirable to prevent these agents from entering the ECS. For various technical reasons the only feasible location for a filter pack is generally considered to be in the high pressure section of the ECS (upstream of the air conditioning packs). This part of the ECS contains fresh air at high temperature and pressure (generally in excess of 100^QC, and at >1 bar). By virtue of these conditions, the use of conventional carbon based filtration is not possible, or at the least there would be poor absorption and a risk of combustion of the carbon. In this respect, other than means for removing condensed water vapour, no currently available ECS has an integrated filtration system capable of removing harmful or noxious gases and vapours.

The present invention provides a filter for removing noxious or otherwise harmful gases, particularly but not exclusively for an aircraft, the filter including an adsorbent composition comprising a zeolite molecular sieve.

Optionally, the zeolite molecular sieve is impregnated with an active metal or metallic species. Preferably, the metal or metallic species is a transition metal or a salt thereof, more preferably, the metal or metallic species comprises copper or chromium. Ideally the metal or metallic species comprises copper (II) or chromium (VII) species.

Advantageously, the filter is located upstream of the air conditioning packs in each aircraft environmental control system (ECS). The filter may be located in the high pressure (HP) and high temperature section of an aircraft ECS.

Surprisingly, the present inventors observed that when using the filter pack of the present invention, air flow to systems downstream of the pack is not reduced to an extent where the operation of the system is impeded by excessive pressure losses.

Advantageously, such a solution allows for a collective protection (colpro) liner to be employed in the aircraft, to allow crew to enter the cabin after decontamination of any liquid agent in the colpro itself. The use of the filter would preferably prevent ingress of CW agents (and the like) into the aircraft and would allow for correct operation of the colpro (which should be supplied with clean air - in the present case meaning air free from CW and biological warfare agents and nuclear hazards in the form of radioactive dust).

Where the filter is in the low pressure section of the ECS it is preferably located downstream of the air conditioning system, ideally immediately downstream and is at a reduced temperature to minimise the number of filter elements required.

Where the filter is provided in the HP section of the ECS it is preferred that the pressure loss across the filter is not great. Preferably, the pressure loss is not more than 0.07 bar, more preferably not more than 0.05 bar and ideally not more than 0.03 bar. The tolerable pressure loss will depend on the specific ECS design and the operational characteristics of the aircraft.

Preferably, the filter located in the HP section is able to operate at a temperature of between 100 ° C - 250 ° C. Advantageously, the filter is able to operate at pressures up to 4.6 bar.

The filter advantageously comprises a two layer bed wherein one layer comprises an impregnated zeolite molecular sieve and one layer comprises an unimpregnated molecular sieve. This provides the advantageous features of both the unimpregnated and impregnated layers. The first layer presented to the airflow preferably comprises the impregnated molecular sieve.

Preferably, the filter is associated with means for deactivating the CW agent. Such means preferably comprise destructive adsorption of the agent on the filter. The agent is adsorbed onto the filter and then broken down by the means for deactivating the agent. In this respect, the present inventors have found that a filter containing an impregnated or unimpregated zeolite molecular sieve in accordance with the present invention is effective in the removal of CW agents from an airstream.

The present invention also provides a filtration system comprising a plurality of filters, at least one of which is a filter according to the present invention, wherein one or more of the filters are used cyclically between an active condition aad a recovery condition. This allows filtered air to be supplied to the crew of an aircraft, for example, whilst another filter bed is being desorbed. Such a filter may be operated in the same manner as those used as part of a pressure swing adsorption system.

Advantageously, elevated pressure has a beneficial effect on filtration performance, since adsorption efficiency increases with pressure. A further benefit of operation at increased pressure is that the residence time of a fixed mass of air on a given volume of filter media is longer, resulting in a greater likelihood of adsorption taking place (meaning that a smaller volume of adsorbent media is likely to be required).

Embodiments of the invention will now be described in detail, by way of example only, with reference to Figure 1 which shows in schematic form a test apparatus for measuring the performance of filters in accordance with the present invention. Experimental procedure

A 13X molecular sieve (1 mm bead size) was obtained from UOP Ltd.. A test apparatus was constructed as shown schematically in Figure 1. The adsorption column 9 comprises the 13X molecular sieve and is attached via insulated 1/4" supply pipework. A flame ionisation detector (FID) 6 and mass spectrometer 11 were used to monitor the inlet and product gas, in addition to an adsorbent containing sampling tube 14 or solvent containing bubbler device (not shown) which provided a means to detect very low quantities of CW agent or reaction products (specifically Lewisite (L), Lewisite oxide, or AC). The duration of each agent challenge measurement is the sum of each of the individual tube or bubbler sampling times (the tubes / bubblers being used consecutively). GD, GB, L and HD were obtained from internal stock. The sieve was used as supplied or after impregnation with copper or chromium species.

Compressed air containing water (and oil) vapour was passed to a pressure regulator fitted with a water separator 2. The use of an oil based compressor 1 reflected turbine (jet) engine bleed air, which is known to contain some oil vapour. This approach enabled the composition of the ECS air to be mimicked in the laboratory. Note that the outlet air from the compressor possessed an d vious odour of oil vapour. Operation of the two way selectable valve 3 determined whether this air was supplied either to a pressure vessel (vapour generator) 4 containing CW agent(s) or through a bypass and then to the test line (fabricated from 1/4" stainless steel (1/4" copper for the heat exchanger)). The test pressure was set using the regulator 2 by reference to the pressure gauge 5. The presence of agent in the input airstream was monitored prior to passage of the airflow through the heat exchanger 8 located in a tube furnace using a flame ionisation detector (FID) 6 and / or a chemical agent monitor (CAM ) 7. For convenience of representation, two CAMs 7 and two FIDs 6 are shown in Figure 1. In reality, only one CAM 7 and one FID 6 were used. Each instrument was used to monitor both the inlet and product gases. It would, of course, be perfectly reasonable to use two CAMs and two FIDs. The CAM 7 is a sensitive time of flight mass spectrometer which is used to detect for the presence of G and H agents in the field. The instrument is operated in either G or H mode. After passage through the heat exchanger 8, the airflow was supplied to the adsorption column 9 via insulated 1/4" copper and stainless steel pipe. The inlet temperature to the adsorption column 9 was monitored using a K type thermocouple which protruded slightly into the filter bed. The inlet temperature was set by reference to the thermocouple by adjusting the set point of the tube furnace. Air exiting the adsorption column 9 was supplied to the FID 6 and / or the mass spectrometer (MS) 11 , to the CAM 7 and to an adsorbent sampling tube 14 (containing Poropak porous polymer) or a bubbler device (containing solvent) for the collection of any agent or reaction products present in the product gas from the adsorption column 9. The bubbler devices are used in experiments involving L. The flow through the sample tubes or bubblers is typically 1 litre per minute (Ipm). The major flow through the adsorption column 9 was passed through a rotameter 15 via a needle valve 13 which was used to adjust the major flow. The CAM 7 was also used to monitor the product gas exiting this rotameter.

Experimental technique specific to control experiments

Two control experiments were carried out to establish the fate of agent-laden air passing through the system in the absence of any adsorbent, the objective being to determine whether any degradation of HD or GD takes place. Whilst full scale aircraft trials using simulants had shown that little or no agent degradation would take place, the control measurements were carried out to demonstrate this. In each case, the adsorption column 9 contained only the packing materials used to retain the molecular sieve adsorbent (glass beads and glass wool). The first control experiment was carried out using HD at an adsorption column inlet temperature of ca. 110°C at 2.9 bar, the total flow rate being 55 Ipm. To attain this inlet temperature, the furnace (and hence heat exchanger 8) was operated at 360°C. The second control experiment was carried out using a combined challenge of GD and HD at an adsorption column inlet temperature of ca. 220°C at 4.6 bar, the total flow rate being 75 Ipm. To attain this inlet temperature, the furnace (and hence heat exchanger 8) was operated at 750°C. The control experiments are representative of the lower and upper limits respectively associated with typical HP ECS. Results from these experiments and additional experimental information can be found in Example 1. Experimental techniques specific to assessment of filters versus CW agents

The following identifies in outline some of the measurements of filtration performance which were carried out:

HD run (0.89 g and 1.13 g HD, loaded into the generator separately), pressure of 2.9 bar, 100°C adsorption column inlet temperature, 95 Ipm total continuous flow (initial set point), 35 g zeolite used in the adsorption column.

GD run (1.12 g and 0.92 g GD, loaded into the generator separately), 2.9 bar, 100°C adsorption column inlet temperature, 95 Ipm total continuous flow (initial set point), 35 g zeolite.

Combined GD (0.98 g) and HD (1.0 g) run, 4.6 bar, 230°C adsorption column inlet temperature, 75 Ipm total continuous flow (initial set point), 35 g zeolite, assessment of performance of filter in accordance with the present invention at a temperature and pressure which are representative of the upper limits of ECS operation, hereinafter known as 'upper limit assessment'.

Combined GD (0.7 g) and HD (1.1 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 35 g zeolite, assessment of performance of filter in accordance with the present invention at a temperature and pressure which are representative of the lower limits of ECS operation, hereinafter known as 'lower limit assessment'.

AC (1.0 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 35 g zeolite. Lower limit assessment.

AC (1.4 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 42.7 g impregnated zeolite. Lower limit assessment.

AC (0.5 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 42.6 g impregnated zeolite. Lower limit assessment. AC (0.6 g), 42.7 g impregnated zeolite, Upper limit assessment, 65 Ipm total continuous flow.

L (0.93 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 35 g zeolite. Lower limit assessment.

L (1.0 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 42.7 g impregnated zeolite. Lower limit assessment.

GB (0.93 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 35 g zeolite. Lower limit assessment

VX (1.02g), upper and lower limit assessment.

Examples 2-11 give further experimental details.

In each of the above experiments, a fresh charge of molecular sieve was used.

Several filters were exposed to a series of challenges to determine whether or not the efficiency of filters in accordance with the present invention is affected by previous challenges. Examples 12 to 15 give further experimental details concerning these serial challenges.

The general experimental procedure that was used in the experiments listed above is now discussed.

During each measurement, the inlet temperature and pressure were recorded at regular intervals, as was the output of the CAM detector 7 (for G and H agents only), which was used to monitor both the adsorption column inlet supply and the product gas for the presence of agent. The response of the FID detector 6, sampling either the inlet or product gas supply, was also recorded. The mass spectrometer 11 was set to continuously record ions present in the product gas between 10 and 200 amu (VG gas Gaslab capillary leak instrument, electron multiplier detector, operating at 1350 V, operating on continuous scan mode). Poropak containing adsorbent sampling tubes 14 or solvent containing bubblers were supplied continuously with 0.5 or 1 Ipm of product flow, and were changed at intervals between 10 and 300 minutes. Exposed tubes were sealed in individual sample bags for analysis. Bubbler inlets were covered with laboratory film to prevent contamination. The output of the CAM detector 7 is a bar scale (maximum eight bars). The significance of CAM output in terms of the hazard is that a 1 bar response or more requires the use of respiratory protection. As the number of bars displayed increases, the time allowed for respite from the use of a respirator (or collective protection) decreases significantly. At higher agent levels, removal of respiratory protection is forbidden.

The Poropak (porous polymer) adsorbent sampling tubes 14 were analysed using gas chromatography (GC), the tube being eluted with solvent. Detection was by a flame photometric detector (FPD) operating in sulphur or phosphorous mode. The method also enabled any reaction products containing sulphur or phosphorous to be detected. In some cases, tubes were examined by thermal desorption to establish if any other reaction products were present. The analyses are highly specific, meaning that the technique allows for the unambiguous identification of agents And reactions products. The bubbler devices were analysed using specific measurements to determine the presence of CW agent or reaction products (primarily being used to determine the presence of AC, L or its oxide, and for total arsenic content).

In addition to the product gas analyses, the zeolite adsorbent was removed for examination after exposure to agents. The zeolite samples were checked with CAM 7 for any residual vapour hazard and then extracted using d-chloroform and d- methanol (deuterated solvents) for examination using nuclear magnetic resonance spectroscopy (NMR). ¹H (proton), ³¹P (phosphorous) and ¹⁹F (fluorine) spectra were recorded using Jeol Lambda 300 and 500 MHz instruments. The purpose of these analyses was to determine whether any agent was present on the zeolite on completion of the filtration experiment, or whether there were any extractable reaction products. This, combined with the use of adsorbent sampling tubes, was necessary to allow the nature of the filtration process to be established. The use of two different solvents enabled non-polar and polar chemicals, if present, to be extracted. Comparison was made against GD and HD agent standards made up in each solvent.

Example 1 - Comparative Examples

Control experiments

The HD control experiment carried out at an adsorption column inlet temperature of ca. 110°C at 2.9 bar (furnace temperature of 360°C) was continued for one and a half hours, after which the air supply was selected to bypass the vapour generator 4. Within minutes of selecting bypass air, no HD was detectable in either the supply or product gas. During this time period, nine Poropak adsorbent tubes were used to sample the product gas. The inlet agent concentration to the adsorption column 9 was measured using the FID 6. CAM 7 was used to monitor both the inlet and product for HD.

The FID signal rose and fell during this period as agent was evaporated. The adsorption column product gas agent concentrations measured using CAM 7 remained high throughout the measurement period, CAM 7 usually being overloaded due to the high HD concentrations present. The levels detected throughout were therefore dangerously high for unprotected personnel, and the measurement confirmed that heating the agent to the temperatures used did not lead to any obvious degradation. This was confirmed through analysis of the adsorbent sampling tubes (table 1 , tubes 9902408-9902414): in each case they were overloaded with HD, and there were no reaction products detectable in the extract from the tubes. Note the absence of agent on the control tubes (9902415-6), which sampled clean air on completion of the agent run.

The combined GD and HD control experiment carried out at an adsorption column inlet temperature of ca. 220°C at 4.6 bar (furnace temperature of 750°C) was continued for one hour, after which the air supply was selected to bypass the vapour generator 4. Within minutes of selecting bypass air, no GD or HD was detectable in either the supply or product gas. During this time period, three Poropak adsorbent tubes 14 were used to sample the product gas. The inlet agent concentration to the adsorption column was measured using the FID 6. CAM 7 was used to monitor both the inlet and product for HD and GD.

The FID signal rose and fell during this period as agent was evaporated. The adsorption column product gas agent concentrations measured using CAM 7 remained high throughout the measurement period, CAM 7 usually being overloaded in both G and H mode due to the high HD and GD concentrations present. The levels detected throughout were therefore dangerously high for unprotected personnel, and the measurement confirmed that heating the agent to the temperatures used did not lead to any obvious degradation. This was confirmed through analysis of the adsorbent sampling tubes (table 2): in each case they were overloaded with GD and HD. Note the absence of agent on the control tube (tube 9902562), which was sampling clean air prior to switching to agent laden air. There were no reaction products (containing sulphur or phosphorous) detectable in the extract from the tubes.

It should be noted that the highly sensitive nature of the analysis for HO present on the adsorbent tubes is such that tube overload occurs when the amount of agent present is relatively low (in term of the mass adsorbed, noting the high toxicity of these chemicals). This means that the amount of agent detected, when the tube is overloaded, is reported as greater than 37μg which is the maximum HD concentration used for calibration, or greater than 35 μg in the case of GD. Amounts reported as less than 1 μg of HD means that there was none (or only a trace amount) detectable when compared to the peak obtained for the lowest HD calibration standard (less than 0.25 μg for GD). This was the pass criteria in this study (this is also applicable to the G agent results, noting that the maximum and minimum calibration standards differed). L agent detection (bubblers) provided similarly high sensitivity. During each measurement, the vapour generator 4 was heated using a fan heater to maximise the challenge concentration. No liquid agent remained in the generator on completion of the measurements.

Table 1; HD (Control Measurement) Adsorbent Tube results

Table 2; HD and GD Adsorbent Tube results (Control)

The control experiments demonstrated that heating the agents to very high temperatures (up to 750°C during passage through the heat exchanger 8), did not result in any measurable degradation. The measurements also confirm that the copper based heat exchanger 8, the glass beads and the glass wool (the latter being adsorption column packing materials) do not promote agent degradation. Heating alone is therefore not a viable option for agent removal. It is unlikely that increasing the residence time in the HP section of an ECS, for example, by markedly increasing the length of pipework, whilst providing a heat source to maintain the air temperature, would result in any significant degradation of these agents. During each measurement, CAM 7 (G and H agents only) and FID response as a function of time was consistent with the delivery of a continuous but variable concentration of CW agent vapour to the test filter. Delivery of the whole dosage of agent to the filter was confirmed by checking the vapour generator on completion of each measurement. The CAM measurements of product gas contamination were consistent with the adsorbent tube data.

Initial measurements were used to identify the minimum quantity of zeolite required to remove CW agent (with no agent detectable in the product gas from the filter) using the maximum possible air flow throughput. From these measurements, a zeolite weight of 35 g and a corresponding flow rate of 65 Ipm were identified. For a high flow ECS, this would be equivalent to a filter of 700 mm length and 300 mm diameter processing a continuous airflow of 66,000 Ipm. Each 1 g agent challenge is therefore equivalent to a vapour challenge of 1.1 kg of CW agent to the full scale filter.

Analysis of the extracts from the zeolite after exposure to CW agents using NMR spectroscopy demonstrated that no CW agent remained.

Filtration experiments

Example 2 - HD

35 g of zeolite was challenged with HD vapour. At no time during the measurement was HD detected in the product gas from the filter bed using CAM 7 (0 bars displayed), and the total flow rate through the bed was maintained at 95 Ipm throughout. Apart from one case, no HD was found on any of the adsorbent sampling material, confirming the CAM measurements (Table 3). The CAM 7 itself was regularly checked using the CAM confidence test kit. This was after challenging the filter with 2.02 g of HD. The amount of HD detected on the adsorbent material where agent was present was low, being equivalent to a total of 0.13 mg in 6,650 I of product gas (representing 99.994% HD removal) (table 3). When measured using CAM 7, the concentration of HD in the inlet air to the adsorption column was high, usually resulting in overload of the instrument. Toward the end of the measurement (5 1/2 hours duration) the inlet concentration decreased (three / fours bars on CAM, still a hazardous level). No further HD was added on the basis of the CAM measurements made which indicated that the filter possesses an inherently high capacity for HD under the test conditions.

On completion of the measurement, the bed was cooled and the zeolite removed. It was not discoloured, and remained free flowing. No HD was detected when the sample was checked with CAM 7, indicating there was no residual vapour hazard associated with the filter media after exposure to HD. The sample weight had increased to 42.9 g (+22.5 w/w%). NMR results obtained for the solvent extracts from the zeolite sample indicate that the increases in weight was due to adsorbed water vapour and to the adsorption of reaction products of HD.

Table 3; HD Run Adsorbent Tube results

No HD was present in the solvent extract from the zeolite (¹H NMR analysis (d- chloroform and d-methanol), three samples) indicating that this CW agent is decomposed during passage through the molecular sieve. The only identifiable substances were 1 ,4 dithiane and 1 ,4 oxythiane and hydrocarbon species which are thought to be non-toxic. It was apparent that there was sufficient molecular sieve to provide reliable filtration at realistic flow rates applicable to aircraft ECS requirements. The results confirm the principle of using a filter containing this adsorbent to remove HD. On the basis that the filter efficiently removed HD, it was likely that nerve agent would also be removed effectively. This is because HD is a more stable molecule than many nerve agents, which are more susceptible to decomposition through thermal degradation. Therefore, an assessment of filtration performance using GD was carried out using substantially the same test conditions and substantially the same amount of zeolite adsorbent (a fresh charge of zeolite being used).

Example 3 - GD

During the initial part of the measurement (the first twelve minutes), GD was detectable in the product gas supply (95 Ipm) from the adsorption column 9 (CAM, 0 / 1 bar). The presence of a very high concentration of GD in the inlet air to the adsorption column was apparent from the FID trace (CAM 7 immediately overloading when brought near to the inlet air supply bleed). As a result, the flow rate was reduced in two 10 Ipm steps to 75 Ipm ( a 0 / 1 bar response being observed at 85 Ipm). After reducing the flow, no further CAM response was observed when sampling the product gas. On this basis, a second charge of GD (0.92 g, total delivered 2.04 g) was added. After 2.75 hours, the run was terminated. Whilst GD vapour was still present in the inlet air up to this time (5 - 6 bars (CAM 7)), there was no liquid agent present in the vapour generator 4 when it was dismantled, indicating that essentially all of the GD added was delivered to the adsorbent bed. Analysis of the adsorbent tubes 14 confirmed that no GD was present in the product gas during the run, all the tubes being clear of agent. Therefore, the CAM response, whilst real (the instrument was checked using the confidence test kit), probably represents a transient effect resulting from the very high challenge observed during the first part of the run. On the basis that no GD was detected in any of the adsorbent sampling tubes (table 4), removal efficiency was very close to 100% (an allowance being necessary for the transient CAM responses observed during the initial part of the run). No further GD was added on the basis of the CAM measurements made which indicated that the filter possesses an inherently high capacity for GD under the test conditions.

Table 4; GD Run Adsorbent Tube results

On completion of the measurement, the bed was cooled and the zeolite removed. It was not discoloured, and remained free flowing. No GD was detected when the sample was checked with CAM 7, indicating there is no residual vapour hazard associated with the filter media after exposure to GD. The sample weight had increased to 43.4 g (+24 w/w%, similar to the weight gain observed after HD challenge (noting that a similar quantity of agent was delivered in both cases)).

Examination of the NMR spectra of the solvent extracts showed that no GD was present. This is on the basis that there are no similarities between the ¹H and ³¹P NMR spectra for the GD samples and those of the extracts. Methyl phosphonic acid was detected in small amounts (a non-volatile reaction product which is a characteristic breakdown product of GD) in addition to a phosphorous-fluorine containing species (a breakdown product of GD).

During this run, it was apparent that there was sufficient molecular sieve to provide reliable filtration at realistic flow rates applicable to an aircraft ECS requirements, even though the flow rate through the bed was reduced during the initial part of the measurement when a CAM response was observed whilst sampling the product gas. The result confirmed the principle of using a filter containing this adsorbent to remove nerve agents, as well as mustard. On the basis of the measurements described above in Examples 2 and 3, further assessments of filter performance were made at the two extremes of typical ECS operation (upper and lower limits of temperature and pressure - ground and flight cases).

Example 4 - GD and HD combined challenge - upper limit assessment

A zeolite bed (35 g) was subjected to a dual agent (simultaneous) GD (0.98g) and HD (1.0g) challenge. The temperature and pressure of the air passing through the adsorption column 9 was 230°C and 4.6bar respectively. These parameters represent the upper limits of the operating range of an ECS.

During the initial part of the measurement (the first ten minutes), GD (but not HD) was detectable in the product gas supply (75 I min^"1) from the adsorption column 9 (CAM 7, 0 / 1 bar). The presence of a very high concentration of GD and HD in the inlet air to the adsorption column was apparent from the use of CAM 7, which immediately overloaded in G and H mode when brought near to the inlet air supply bleed. As a result, the flow rate was reduced in two steps (to 65 Ipm) (a 0 / 1 bar response also being observed at 70 Ipm). After reducing the flow, no*further CAM response was observed when sampling the product gas. During this measurement, the FID 6 was used to monitor the product gas. Whilst no G or H agent was detectable in the product gas (see Table 5), it was apparent that there were detectable species present. These appeared to be present in lower concentrations on the basis of a direct comparison of the FID plots obtained during this and the prior measurements. As part of the analytical procedure developed for the adsorbent tubes, the materials contained by the adsorbent tubes are extracted with isopropyl alcohol (IPA), and then analysed using specific sulphur (HD) and phosphorous (GD) detectors. This is necessary to allow the presence of any agent to be revealed (particularly at very low concentrations). As a result, the technique will not allow for the detection and quantification of species which contain neither sulphur nor phosphorous atoms or functional groups. Therefore, in an attempt to identify the components present in the product gas which were giving rise to the FID signal, one of the adsorbent tubes (which sampled the air during the peak FID response (12 - 24 minutes elapsed time)) was analysed via thermal desorption. The adsorbent used in the tubes (Poropak) is highly porous, and so is capable of collecting relatively volatile components present in an air sample. The use of thermal desorption does not involve the use of any solvents, which could interfere with the analysis on the basis that they may be present in the air sample. For example, I PA is a potential byproduct of the degradation of GD. A major peak was found in the chromatogram so obtained, and analysis using a mass selective detector indicated this was due to hydrocarbon species containing 6 carbon atoms (which may be hexane, iso butanes and/or iso pentanes, for example). The product gas was safe to breathe, given the observed concentrations of C6 hydrocarbons. Furthermore, these levels do not represent an explosion hazard since the amounts are too low, and they would not be retained in the ECS or cabin. The product gas was also monitored using a continuous sampling mass spectrometer 11 (up to 200 amu). The spectrum did not change throughout the period of the measurement, supporting the theory that the reaction products present in the air sampled after the adsorption column were not present in significant quantities. Note that a FID is very much more sensitive to hydrocarbons compared to a mass spectrometer and that the thermal desorption technique allows for much higher detection sensitivity because the material is concentrated on the GC column prior to being passed to the FID ^ and mass selective detector.

On the basis that no GD or HD was detected on or in any of the adsorbent sampling tubes, removal efficiency was very close to 100% (see Table 5). An allowance is necessary for the transient CAM responses observed during the initial part of the run. No further agent was added on the basis of the CAM measurements made which indicated that the filter possesses an inherently high capacity for GD and HD under test conditions which represent the upper limits of pressure and temperature within a typical aircraft ECS.

On completion of the measurement, the bed was cooled and the zeolite removed. It was discoloured (slightly brown, initially white), but remained free flowing. No GD or HD was detected when the sample was checked with CAM 7, indicating there is no residual vapour hazard associated with the filter media after exposure to GD and HD. The sample weight had increased to 44.2 g (+26.3 w/w%, similar to the weight gain observed after challenge with only one of GD and HD (noting that a similar quantity of agent was delivered compared to the previous single agent runs - about g)).

No GD or HD was present in the solvent extract from the zeolite according to ¹H and ³¹ P NMR analysis. There were no identifiable substances in any of the extracts and no phosphorus species were observed in either the d-methanol or d-chloroform extracts. This indicates that the extent of agent degradation is greater at the higher temperature used in this run. This is consistent with the discolouration of the molecular sieve, which is probably due to the presence of reaction products. Note that heating the zeolite to ca. 600°C does not result in discolouration. The result further confirms the principle of using a filter containing this adsorbent to simultaneously remove both mustard and nerve agents when the ECS is operating at high temperature and pressure.

Tube 9902651 was subjected to thermal desorption analysis

Table 5; GD / HD Run Adsorbent Tube results

Example 5 - GD and HD combined challenge - lower limit assessment

A further assessment of filter performance was made at the lower extreme of ECS operational conditions. This assessment was carried out by subjecting the bed (35g) to a dual agent (simultaneous) GD (0.7g) and HD (1.1g) challenge. A continuous air flow rate through the filter of 65 Ipm was used. Throughout the experiment the temperature and pressure were set at 100°C and 2.5bar respectively to simulate the lower operating limits of an aircraft ECS.

At no time during the measurement was GD or HD detectable in the product gas supply (65 Ipm) from the adsorption column using CAM 7. The presence of a very high concentration of GD and HD in the inlet air to the adsorption column was apparent when using CAM 7 (it immediately overloaded in G and H mode when brought near to the inlet air supply bleed). During this measurement, the FID 6 was used to monitor the product gas. Whilst no G or H agent was detectable in the product gas using CAM 7, it was apparent that, as in example 4, there were detectable species present. Again, the concentrations were low on the basis of a direct comparison of the FID plots obtained during this and the prior measurements. The results of thermal desorption analysis indicated that the product gas contained only C6 hydrocarbons.

On the basis of the previous results, which demonstrated that when CAM 7 indicated that the product gas was clear of agent there was no agent detected on the adsorbent sampling tubes during subsequent analysis, the odour of the* product gas was checked during the first 20 minutes of the run immediately after completing a CAM check. There was a detectable (but not strong or unpleasant) odour of a butane or butane like component, an observation consistent with the presence of C6 hydrocarbons in the product gas (probably butane isomers). The air was not foul, and it should be noted that once agent is no longer present in the inlet air, the odour should disappear (i.e. the FID signal reduces to baseline). As before, the product gas was also monitored using a continuous sampling mass spectrometer 11 (up to 200 amu). The spectrum did not change throughout the period of the measurement, supporting the contention that the reaction products present in the air sampled after the adsorption column were not present in significant quantities.

Small amounts of GD and HD were detected on two of the adsorbent tubes (Table 6). The amount of HD detected on the two adsorbent tubes where agent was present was low, being equivalent to a total of 0.09 mg in 3,900 I and 0.3 mg in 6,825 I of product gas respectively (representing 99.97% HD removal). The corresponding values for GD were 0.02 mg in 3,900 I of product gas (representing 99.998% GD removal).

When measured using CAM 7, the concentration of GD and HD in the inlet air to the adsorption column was high, normally resulting in overload of the instrument. Toward the end of the measurement (3 1/2 hours) the inlet concentration decreased (three / fours bars on CAM 7 in both G and H mode, still a hazardous level). No further GD or HD was added on the basis of the CAM measurements made which indicated that the filter possesses an inherently high capacity for these agents.

Table 6; GD/HD Run Adsorbent Tube results

On completion of the measurement, the bed was cooled and the zeolite removed. It was not discoloured and was free flowing. No GD or HD was detected when the sample was checked with CAM 7, indicating there is no residual vapour hazard associated with the filter media after exposure to GD and HD. The sample weight had increased to 43.58 g (+24.5 w/w%, similar to the weight gain observed after the previous challenge runs). As before, no agent or reaction products was detected in the solvent extracts from the zeolite using NMR spectroscopy.

The result further confirms the principle of using a filter containing this adsorbent to simultaneously remove both mustard and nerve agents when the ECS is operating at low temperature and pressure. Example 6 - AC Challenge

Further assessments of filter performance were made by subjecting the bed to a challenge with hydrogen cyanide (AC). Summaries of the experimental conditions are found below.

AC (0.5 g) run, 2.5 bar, 100°C adsorption column inlet temperature, 65 Ipm total continuous flow, 42.6 g impregnated zeolite. Lower limit assessment.

AC (0.6 g), 42.7 g impregnated zeolite, 4.6 bar, 230°C adsorption column inlet temperature. Upper limit assessment.

A new filter was used for each challenge.

During evaporation of AC into the inlet of the unimpregnated zeolite filter, the FID signal increased. Monitoring the product gas using a specific detector (Monitox Ltd.) demonstrated that AC was penetrating the filter. Potassium hydroxide containing bubblers were also used to collect any AC present in the product gas, being analysed for the presence of cyanide ions. Two bubblers were used in each case, a lead bubbler and a slip bubbler, the slip bubbler being used to collect any AC passing through the lead bubbler. The results for these bubblers can be found in Table 7.

Due to the failure of the unimpregnated filter, a further assessment was made using an impregnated zeolite (aqueous impregnation, using copper (II) and chromium (VI) species). Impregnation was by soaking zeolite in an aqueous solution of copper (II) chloride and chromium (III) or chromium (VI) oxide for at least 60 minutes followed by draining and drying (-16 hours, air oven, 100°C).

This filter substantially reduced, but did not eliminate, AC penetration. Typical product concentrations generated by a -0.5 g challenge were -20 ppm, which is a concentration known to be safe to breathe. Various impregnant loadings on the zeolite were assessed, but none resulted in complete removal of AC (Table 7).

The concentrations of AC detected when the filter was challenged under the upper limit conditions of temperature and pressure were significantly lower (Table 7), indicating that AC destruction is more efficient under these conditions.

It is recognised that agents such as AC provide a lower threat to an aircraft, and do not give rise to any post attack desorption hazards or damage effects.

Table 7; AC Bubbler Results Example 7 - L Challenge (unimpregnated zeolite)

An assessment of filter performance was made at conditions representing the lower limits of ECS operational conditions. Air entering the filter bed was at a pressure of 2.5bar and a temperature of 100°C, with a 65 Ipm flow rate. This assessment was carried out by subjecting the zeolite bed (35 g) to a challenge with Lewisite (L) (0.93g).

Analysis of the bubbler devices used to sample for L demonstrated that no agent was detectable in the product gas supply from the adsorption column. During the measurement, the FID 6 was used to monitor the product gas. Whilst no L was detectable in the product gas, it was apparent that there were detectable species present. These products possessed a non-distinct odour.

On the basis that no agent was detected, removal efficiency was essentially 100%. On completion of the measurement, the bed was cooled and the zeolite removed. It was not discoloured and remained free flowing. Extraction of the sieve demonstrated that a significant quantity of Lewisite Oxide (L oxide) (a solid) was present on the adsorbent. Whilst L oxide does not possess the vesicant properties of L, it is an arsenical and as such presents a residual hazard, albeit of substantially less significance (noting it is a solid).

The results shown in Table 8 confirm the principle of using a zeolite based filter to remove L agents when the ECS is operating at or near the lower limits of temperature and pressure.

The difference in the minimum reported values (Table 8) is due to use of different standards. If present, L would be found in greater quantities in the lead bubblers.

Table 8; L Bubbler Results

Example 8 - L Challenge (impregnated zeolite)

An assessment of filter performance was made at the lower limit of ECS operational conditions using an impregnated filter. Air entering the filter bed was at a pressure of 2.5bar and a temperature of 100°C, with a 65 Ipm flow rate. This assessment was carried out by subjecting the bed (35 g) to a challenge with Lewisite (L) (1.0g). Impregnated zeolite (copper, chromium) manufactured according to Example 6 was used. There was approximately 20% w/w impregnant loading, including water adsorbed during impregnation not subsequently removed during drying.

Analysis of the bubbler devices (Table 9) demonstrated that no agent was detectable in the product gas supply (65 Ipm) from the adsorption column. During the measurement, the FID 6 was used to monitor the product gas. Whilst no L was detectable in the product gas, it was apparent that there were detectable species present. These products possessed a non-distinct odour. They did not contain arsenic on the basis of acid digestion and analysis of a selected number of the bubblers (Table 10).

Table 9; L Bubbler Results

Table 10; Total Arsenic Results

On the basis that no agent was detected, removal efficiency was essentially 100%. On completion of the filtration experiment, the bed was cooled and the zeolite removed. It was not discoloured and remained free flowing. Extraction of the sieve demonstrated that no L or Lewisite Oxide was present on the adsorbent, indicating that the impregnated form of the zeolite chemically decomposes L.

The result demonstrates the advantage of using the impregnated form of the zeolite to remove and detoxify L agents when the ECS is operating at low temperature and pressure.

It is notable that when unimpregnated zeolite is exposed to L at the lower temperature and pressure and then the conditions are adjusted to the upper temperature and pressure, no L or L oxide is found on the adsorbent, indicating that the higher temperature results in destruction of the agent.

Example 9 - H and G challenges of impregnated filters

Performance assessments of the impregnated form(s) of the zeolite using G and H agents demonstrated that the capacity of the sieve was reduced for these agents compared to the unimpregnated form.

Discussion

The most effective combination for the removal and detoxification of G, H and L agents from a hot (of the order of 100°C) pressurised supply of air is therefore a two layer filter, the first layer consisting of a 13X zeolite or similar impregnated with copper and chromium species (either or both species, in various proportions) and the second layer consisting of an unimpregnated zeolite (type 13X or similar). The second layer must be sized such that this part of the bed alone is sufficient to remove G and H agents, the impregnated layer being primarily for the full detoxification of L agents and the partial detoxification of acid gases such as AC.

Whilst the observations described here relate primarily to single pass filters, the use of the adsorbent systems described are equally applicable to two or more filter bed systems where one or more of the filter beds is providing filtered gas whilst others are being desorbed at elevated temperature and reduced pressure. This design might be referred to as a partially regenerable self decontaminating filtration system.

It is anticipated that the method of the invention in suit may be adapted to ensure a low pressure drop across the filter, this being desirable in the case of an aircraft. Two or more of the filters used in the method of the present invention may be placed in parallel so that the pressure drop across each filter is small.

Example 10 - GB Challenge

A further assessment of filter performance was made at or near the lower limit of ECS operational conditions. Air entering the filter bed was at a pressure of 2.5bar and a temperature of 100°C, with a 65 Ipm flow rate. The assessment was carried out using GB challenge. GB is a lower boiling nerve agent of the same family as GD, and is expected to be more susceptible to degradation on the basis of the observations made with GD and the chemistry of this agent.

Prior to GB challenge, the filter was operated for a cumulative period of 67 hours at 2.5 bar and 100°C. This was carried out to determine the effect of pre-exposure of the filter to oil and water vapour. The zeolite was removed and weighed prior to GB challenge, the weight increase being 6.46 g (+18.5 w/w %). This weight increase was due almost entirely to the adsorption of water vapour.

At no time during the measurement was GB detectable in the product gas supply (65 Ipm) from the adsorption column using CAM 7. During this measurement, the FID 6 was used to monitor the product gas. Whilst no G agent was detectable in the product gas using CAM 7, it was apparent that, as before, there were detectable species present. Again, the concentrations were low on the basis of a direct comparison of the FID plots obtained during this and the prior measurements. On the basis of the previous results, which demonstrated that when CAM 7 indicated that the product gas was clear of agent there was no agent detected on the adsorbent sampling tubes during subsequent analysis, the odour of the product gas was checked during the first 60 minutes of the run immediately after completing a CAM check. There was a detectable (but not strong or unpleasant) odour of butane or butane-like components in the product gas (probably butane isomers). The air was not foul, and it should be noted that once agent is no longer present in the inlet air, the odour will disappear (i.e. the FID signal reduces to baseline).

The absence of any GB on the adsorbent sampling tubes indicates that removal efficiency was essentially 100% (see Table 11).

Table 11; GB Run Adsorbent Tube results

On completion of the measurement, the bed was cooled and the zeolite removed. It was not discoloured and was free flowing. No GB was detected wheg the sample was checked with CAM 7, indicating there is no residual vapour hazard associated with the filter media after exposure to GB.

The result confirms the principle of using a filter containing this adsorbent to remove other nerve agents when the ECS is operating at the lowest temperatures and pressures.

Example 11 - VX challenge

The run was performed using a fresh filter (35.22g). The bed was challenged with 1.02g of VX. During the challenge, the vapour generator 4 was heated to maximise the challenge concentration. The filter was challenged 2 hours after initiating the air flow at a temperature and pressure associated with the lower limits of operation of the ECS (100°C and 1.5bar respectively). The FID (product gas) signal remained at baseline throughout this period (3.75 hours after initiation of air flow). The inlet conditions to the filter were then adjusted to reflect the maximum operating temperature and pressure in the ECS (250°C and 3.5bar respectively). During this heating period, the FID signal rose, but not substantially before falling. The run was terminated 6 hours after initiation of air flow.

On completion of the run, the vapour generator 4 was examined to ensure that none of the liquid agent remained in the generator.

The odour of the product gas was not checked throughout the run because VX is super-toxic. No VX vapour detection method was available.

During the run, sampling tubes 14 containing adsorbent materials were used to collect sample gas. The contents of these tubes were analysed to determine whether VX was present in the product gas at any time. The tubes were changed at regular intervals, and the results from those tubes are shown below in Table 12.

Table 12 - VX run adsorbent tube results

No VX was detectable in the extracts from any of the tubes. On completion of the run, the zeolite filter was removed and its weight was found to have increased by 2.1% to 35.97g. Extractions were performed on the zeolite using d-chloroform and d-methanol and the ¹H and ³¹P NMR spectra of the extracted materials measured and compared with the spectra of VX. No peaks in the extract were attributable to either VX or a phosphonic acid (non-volatile reaction product which is a characteristic breakdown product of VX). Hence, the residual zeolite adsorbent does not present a residual hazard.

The results above were substantially reproduced using another fresh filter.

Example 12 - Serial challenge 1 (unimpregnated zeolite)

An assessment of an unimpregnated zeolite filter (35 g) was carried out to define the capability of the filter to sustain repeated challenge with CW agents. The bulk of each challenge was delivered under low temperature and pressure to simulate ECS conditions associated with ground operation of the aircraft (2.5 bar, 100°C) over a period of about 4 hours (the vapour generator 4 being heated to promote evaporation of agent). Thereafter the conditions were altered to reflect those associated with flight (high pressure and temperature limits of ECS, about 4.6 bar, ~250°C) with any residual agent being delivered over the remaining period of the measurement - about three hours. On completion of each challenge, the filter was cooled and subjected to a further challenge after about 16 hours or a time period compatible with the next opportunity to use live CW agent. The same filter was used throughout. The filter was operated at 2.5 bar, 100°C and 4.6 bar, 230°C in the absence of agents (prior to any agent challenges) for a number of hours (7 and 8.5 respectively) to ensure that thermal cycling of the adsorbent did not affect its function with respect to degrading CW agents.

The filter was challenged with the following agents, in the order shown: • GD (1.02 g)

• L (1.15 g)

• HD (1.22 g) • GB (1.1 g)

Sampling was by CAM 7, FID 6, adsorbent tubes 14 and bubblers as before. At no time during the measurements was agent detectable in the product gas supply (65 Ipm) from the adsorption column 9 using the various detection methods. During the measurements, the FID 6 was used to monitor the product gas. During the low temperature and pressure phase of each challenge, there were detectable odours present in the product gas, but these were not unpleasant. Gas analysis showed that the product gas was not toxic. During the high temperature and pressure phase of each challenge, the odours became stronger during about the first 30 minutes, declining thereafter. This is associated with further degradation of the detoxified agents at the higher temperature. In the case of L, the odour became similar to that of baked onions during the hot phase. No L or L oxide was detected at any stage.

The absence of any agents in the product gas from the filter indicates that removal efficiency was essentially 100% in each case when the filter was operated under conditions representative of ground operations and then flight.

The result further confirms the principle of using a zeolite filter to remote a range of agents without the need to change the filter when the aircraft is operating under ground and then flight conditions (lower and then higher temperatures and pressures).

This sequence is equivalent to a total CW agent vapour challenge to the full scale (66,000 Ipm) filter of approximately 5.3 kg. It is apparent that the capacity of the filter exceeds this value.

Example 13 - Serial challenge 2

In order to simulate a realistic exposure sequence and to determine whether filter efficiency depends on prior challenges, an unimpregnated filter was subjected to the following challenges in sequence. The sequence also included periods of filter operation without challenge to realistically simulate longer periods of use. The sequence is presented in the order that it was carried out:

- VX (0.85g) - Lower limit of ECS operating conditions (100°C, 1.5bar) for 8 hours - GD (1.05g)

- Lower limit of ECS operating conditions (100°C, 1.5bar) for 16 hours

- Lower limit of ECS operating conditions (100°C, 1.5bar) for 3.5 hours, followed by upper limit of ECS operating conditions (250°C, 3.5bar) - repeated 3 times - HD (1.4g) (hereinafter HD1 ) - HD (1.5g) (hereinafter HD2)

- L (0.7g)

The performance of the filter was monitored throughout each challenge.

VX challenge and GD challenge

Both of these challenges were defeated by the filter, resulting in effectively nil penetration as measured by the CAM 7 and by analysis of contents of adsorbent tubes.

HD1 challenge

The bed was challenged with 1.4g of HD. During the challenge, the vapour generator 4 was heated to maximise the challenge concentration.

The filter was challenged approximately 3 hours after initiating the air flow at temperature and pressure associated with the lower limits of operation of the ECS (100°C and 1.5bar). The FID (product gas) signal remained at low levels throughout this period (3.25 hours after initiation of air flow). 3.25 hours after initiation of air flow, the inlet conditions to the filter were then adjusted to reflect the maximum operating temperature and pressure in the ECS (250°C and 3.5bar respectively). During this heating period, the FID signal rose, but significantly before falling. The run was terminated 5.75 hours after initiation of air flow. On completion of the run, the vapour generator 4 was examined to ensure that none of the liquid agent remained in the generator.

The odour of the product gas was checked throughout the run. Initially, the gas was odourless or slightly hydrocarbon in nature. Then, the odour changed to 'oniony' for the remainder of the run at the lower temperature and pressure. The odour remained oniony throughout the run, but increased in intensity during the temperature rise, before declining to become faint.

Sampling tubes containing adsorbent material were used to collect product gas throughout the run. A summary of the results from the sampling tubes is given below in Table 13. Note that a CAM 7 was also used in 'H' mode to sample the product gas, with zero bars displayed throughout the run.

Table 13 - HD1 sampler results

The results of Table 13 indicate that HD is present in the product gas, thus indicating partial failure of the filter. However, the filter was 99.89% efficient, thus giving excellent, if not faultless, performance.

HD2 challenge

The bed was challenged with 1.5g of HD. During the challenge, the vapour generator 4 was heated to maximise the challenge concentration.

The filter was challenged approximately 45 minutes after initiating the air flow at temperature and pressure associated with the lower limits of operation of the ECS (100°C and 1.5bar). The FID (product gas) signal remained at low levels throughout the lower ECS period. Several hours after initiation of air flow, the inlet conditions to the filter were then adjusted to reflect the maximum operating temperature and pressure in the ECS (250°C and 3.5bar respectively). During this heating period, the FID signal rose, but significantly before falling. The run was terminated 6 hours after initiation of air flow.

Sampling tubes containing adsorbent material were used to collect product gas throughout the run. A summary of the results from the sampling tubes is given below in Table 14. Note that a CAM 7 was also used in Η' mode to sample the product gas, with zero bars displayed throughout the run.

Table 14 - HD2 sampler results

The results of Table 14 indicate that HD is present in the product gas, thus indicating partial failure of the system. However, the filter was 98.93% efficient, thus giving excellent, if not faultless, performance. The data from this example indicate that the filter's ability to remove HD from a gas depends on previous challenges.

L challenge The bed was challenged with 1.4g of L. During the challenge, the vapour generator 4 was heated to maximise the challenge concentration.

The filter was challenged approximately 1 hour after initiating the air flow at temperature and pressure associated with the lower limits of operation of the ECS (100°C and 1.5bar). The FID (product gas) signal remained at low levels throughout this period (5.5 hours after initiation of air flow). 5.5 hours after initiation of air flow, the inlet conditions to the filter were then adjusted to reflect the maximum operating temperature and pressure in the ECS (250°C and 3.6bar respectively). During this heating period, the FID signal rose significantly before falling rapidly after the maximum temperature and pressure conditions were attained. The run was terminated several hours after initiation of air flow.

The odour of the gas was monitored throughout the run. During operation at the lower ECS temperature and pressure limits, the product gas possessed a metallic odour, which intermittently disappeared. During the first ten minutes after the setting of the temperature and pressure of associated with the upper operating limits of the ECS, the odour was strong, thereafter becoming faint again.

20 samples of the product gas were taken using bubblers, operated at O.δlitres per minute gas flow rate. Poropak-containing sampling tubes were used to sample the product gas to try to identify gas species giving rise to the metallic odour (if these are not L).

18 of the bubblers, which acquired product gas for periods of between 7 and 25 minutes, were clear of L, with <0.7μg being collected in each of these tubes. Two of the bubblers showed traces of L. One bubbler collected product gas for 16 minutes, and yielded 0.83μg of L, whereas the other bubbler was exposed to product gas for 18 minutes and yielded 0.73 μg of L.

Hence, traces of L penetrated the filter and thus the filter partially failed. However, the amounts penetrated were close to the detection limit and the filter was determined to be 99.99% efficient.

Samples extracted from the adsorbent material from the sampling tubes gave a number of GC-FID peaks, none of which were clearly identifiable against the toxic substances which may be generated from L. No HD was detected.

This series of experiments has demonstrated that a filter used in accordance with the present invention performs very well when exposed to a series of high dosage challenges of chemical warfare agents. The performance of the filter is seen to degrade very slightly on exposure to serial challenges. However, the inventors have demonstrated that there is no need to regenerate or desorb the material on the filters, whether by heating or other means.

Example 14- serial challenge 3

Example 13 illustrated that, whilst performing almost faultlessly against many chemical warfare agents, the filters used in the invention in suit do not provide faultless protection against HD when the filters are exposed to a series of challenges. In order to investigate this phenomenon further, a filter was exposed to the following series of challenges:

- Run at lower limit of ECS operating conditions (100°C, 1.5bar)

- Run at upper limit of ECS operating conditions (250°C, 3.5bar) - HD, 1.7g

- HD, 1.9g - HD, 1.0g The efficiency of the filter was measured using methodology previously described for HD. The efficiency of the filter during the three HD runs was 99.96%, 98.9% and 99.1 % respectively. This suggests that, although the filter partially failed, the performance of the filter was very good, especially given the high doses associated with the first two challenges.

Example 15 - serial challenge 4

A filter was subjected to a small series of challenges of high dose in order to determine how the filter would perform if exposed to higher than expected doses of chemical warfare agent. A filter was exposed to the following challenges in sequence:

- HD, 2g - L, 2g

The efficiency of the filter was measured throughout each run in accordance with procedures described previously. The efficiency of the filter was 99.94% and 99.1% for the HD and L challenges respectively. Hence, the filter partially failed to stop the challenge. However, the performance of the filter was very good, especially given the high doses involved. Failure may well be attributable to the heating of the agent which causes the filter to be subjected to maximum challenge concentration.

The inventors have demonstrated that filters in accordance with the present invention can remove GD, GB, L, VX and HD from an airstream which is at temperatures and pressures representative of an aircraft ECS.

The filters in accordance with the present invention were found to possess an inherently high capacity for these types of CW agent, and as a result, protection can be achieved using a relatively small filter bed. Within the limitations of the tests (which are associated with the amounts of agents that may be used) it is likely that a single filter containing 35.9 kg of zeolite can continuously process a 66.6 m³min^"1 (4,000 m³hr^"1) airflow and sustain multiple attacks with G, H and L agents without significant degradation of performance.

Furthermore, 13X molecular sieve is a low cost adsorbent which is commercially available in a range of bead sizes.

It is worth noting that the filters in accordance with the present invention are compatible with the proposed use of a collective protection liner (where appropriate) to allow crew to enter a cabin or cockpit free of any liquid agent whilst the liner (as appropriate) is being purged with air free of these agents. Use of such a filter may result in the relaxation of some of the BC hardening requirements for components supplied with ECS air which would, in the absence of active filtration, be subject to exposure to G and H agents in the event of an agent release during operation of the aircraft.

Claims

1 . A filter for removing noxious or otherwise harmful gases, particularly but not exclusively for an aircraft, the filter including an adsorbent composition comprising a zeolite molecular sieve

2. A filter according to claim 1 wherein the filter is impregnated with an active metal or metallic species.

3. A filter according to claim 2, in which the metal or metallic species is a transition metal or a salt thereof.

4. A filter according to either of claims 2 or 3, in which the metal or metallic species comprises copper or chromium.

5. A filter according to any one of claims 2 to 4, in which the metal or metallic species comprises copper (II), chromium (III) or chromium (VI) species.

6. A filter according to any preceding claim, in which the filter is located in the high pressure and temperature section of an aircraft Environmental Control System

(ECS).

7. A filter according to any preceding claim in which the air intake pressure loss across the filter is not more than 0.07 bar.

8. A filter according to any preceding claim in which the filter operates at a temperature of between 100⁹C - 250⁹C.

9. A filter according to any preceding claim in which the filter operates at pressures up to 4.6 bar.

10. A filter according to any preceding claim comprising a two layer bed wherein one layer comprises an impregnated zeolite molecular sieve and one layer comprises an unimpregnated zeolite molecular sieve.

11. A filter according to claim 10 wherein the first layer presented to the airflow is of the impregnated type.

12. A filter according to any preceding claim further comprising means for deactivating the CW agents of the general types G, H and L.

13. A filter substantially as hereinbefore described with reference to the examples.

14. A filtration system comprising a plurality of filters, at least one of which is a filter according to any preceding claim, wherein one or more of the filters are used cyclically between an active condition and a recovery condition.