1 Collection and Preparation of Gaseous Samples To collect such large volumes of air, ambient air sampling is frequently con ducted over many hours, usually 12 h or 24 h. These are termed integrated air sam ples. Air samplers frequently draw air through a filter to remove the particles, then through a sorbent material to trap the gaseous components. In this filter/sorbent geometry air sampler, it is now well known that artifacts are produced, the most serious of which is the volatilization of the particle adsorbed semivolatile com pounds due to the pressure drop across the filter. These volatilized compounds pass through the filter and are trapped on the adsorbent where they are analysed as though they were gaseous compounds. Temperature variations during sampling ill also influence the gas/particle partitioning of the SvOCs. The filter sorbent geometry should be restricted to determining the total combined gas and particle burden of a particular air sample but should not be used to determine gas/particle partition ratios of semivolatile compounds. If the gas/particle partition ratios are to be determined, it is preferable to remove the gas phase first, and then to remove the particles, as is done in the annular diffusion denuder samplers. The sample must subsequently be extracted from the trapping medium and pro- essed before it is analysed by the desired analytical method. Very often the pre essing requires a series of steps to isolate a particular compound or compound class for the analysis to be effective. If the work-up steps are ignored, the collected sample is usually too complex for even the most selective and sensitive analytical methods available today If the concentration of the analyte fluctuates with time, then the result of the sampling is a concentration averaged over the sampling time period. The shorter the sampling period, the greater will be the temporal resolution of the concentra tion variations of the analyte Many chemical species found in the atmosphere are chemically and/or photo- hemically reactive. If such compounds are present, inert sample inlets and sur- faces must be employed since there is a strong potential for the formation of arti- facts. This is not a simple problem to overcome. It is also necessary to prevent further chemical degradation of the compounds during and after the sampling been completed. The sample must not react with the sampling surfaces, filters or adsorbents. When adsorbents are used, the efficiency of extraction of the com- pounds from the adsorbents must be determined for each compound. If the gases and vapors pass through particle filters, the gases and vapors must not be adsorbed by, or react with the material from which the filters are made. Glass fiber and er filters, for example, are well known to adsorb organic compoun whereas Teflon-coated glass fiber filters are much less likely to adsorb organic vapors. On the other hand, if the primary objective of the sampling is to determine the elemental carbon content of the collected particles, then a Teflon-coated glass fiber filter would obviously be a poor choice since the Teflon coating would become part of the analyte during the high temperature heating of the sample. As stated previously, it is imperative that the sampling system suits the problem that is to be solved. If adsorbents or diffusion denuders are used, it is possible that gas phase material adsorbed on the surfaces of the adsorbents might break through the collector. This must be investigated and, if significant, then either a different
To collect such large volumes of air, ambient air sampling is frequently conducted over many hours, usually 12 h or 24 h. These are termed integrated air samples. Air samplers frequently draw air through a filter to remove the particles, then through a sorbent material to trap the gaseous components. In this filter/sorbent geometry air sampler, it is now well known that artifacts are produced, the most serious of which is the volatilization of the particle adsorbed semivolatile compounds due to the pressure drop across the filter. These volatilized compounds pass through the filter and are trapped on the adsorbent where they are analysed as though they were gaseous compounds. Temperature variations during sampling will also influence the gas/particle partitioning of the SVOCs. The filter sorbent geometry should be restricted to determining the total combined gas and particle burden of a particular air sample but should not be used to determine gas/particle partition ratios of semivolatile compounds. If the gas/particle partition ratios are to be determined, it is preferable to remove the gas phase first, and then to remove the particles, as is done in the annular diffusion denuder samplers. The sample must subsequently be extracted from the trapping medium and processed before it is analysed by the desired analytical method. Very often the processing requires a series of steps to isolate a particular compound or compound class for the analysis to be effective. If the work-up steps are ignored, the collected sample is usually too complex for even the most selective and sensitive analytical methods available today. If the concentration of the analyte fluctuates with time, then the result of the sampling is a concentration averaged over the sampling time period. The shorter the sampling period, the greater will be the temporal resolution of the concentration variations of the analyte. Many chemical species found in the atmosphere are chemically and/or photochemically reactive. If such compounds are present, inert sample inlets and surfaces must be employed since there is a strong potential for the formation of artifacts. This is not a simple problem to overcome. It is also necessary to prevent further chemical degradation of the compounds during and after the sampling has been completed. The sample must not react with the sampling surfaces, filters or adsorbents. When adsorbents are used, the efficiency of extraction of the compounds from the adsorbents must be determined for each compound. If the gases and vapors pass through particle filters, the gases and vapors must not be adsorbed by, or react with the material from which the filters are made. Glass fiber and quartz fiber filters, for example, are well known to adsorb organic compounds whereas Teflon-coated glass fiber filters are much less likely to adsorb organic vapors. On the other hand, if the primary objective of the sampling is to determine the elemental carbon content of the collected particles, then a Teflon-coated glass fiber filter would obviously be a poor choice since the Teflon coating would become part of the analyte during the high temperature heating of the sample. As stated previously, it is imperative that the sampling system suits the problem that is to be solved. If adsorbents or diffusion denuders are used, it is possible that gas phase material adsorbed on the surfaces of the adsorbents might break through the collector. This must be investigated and, if significant, then either a different 1 Collection and Preparation of Gaseous Samples 7
1. 3 Active vs. Passive Sampling adsorbing material must be used or a breakthrough factor must be statistically de- termined. All potential artifacts must be considered and minimized, if not elimi nated, through proper sampler design and analytical process control. This often ne- cessitates lengthy quality assurance(QA) and quality control (QC) programs Active vs Passive Sampling There are two basic means of collecting a gaseous sample: active sampling and pas- In active sampling procedures, air is drawn through an absorbing or adsorbing medium by a pump in order to trap the gas phase material. It is important that the sampler has an accurate, calibrated means to determine the total volume of the gaseous sample and the rate at which the gas is being sampled. This is most easil accomplished by the use of calibrated mass flow controllers In passive sampling devices, an adsorbing material is placed at a fixed distance away from the air being sampled. Gas phase molecules must pass through a mem- brane or filter and diffuse across this distance and be trapped on or react with the collecting medium. The principles of diffusion are utilized to calculate the concen tration of the specified contaminant in the ambient air. The diffusion rate across the passive sampler(and/or through the membrane) is analogous to the flow rate in an active sampler. 3. Active Air Collection Methods There are many active air sampling methods available. Each method utilizes differ- ent ways to collect gaseous samples and each sampling method has its own partic ular inherent artifacts, and, thus, each method has its own strengths and weak nesses. It cannot be stated too often that the appropriate sampling method must be selected to address a particular question or problem. There are several basic mechanisms whereby gases and vapors may be collected for subsequent analysis. Gaseous samples may be adsorbed on the surface of various substances, which have large surface areas and are specifically designed e gaseous chemical species desired. They may re some chemical adsorbed on the surface of the collection device or on particles in the collector. Gases may be collected in bags or canisters or trapped in bubblers, in mist chambers or cryogenically. Each of these methods will be described briefly
adsorbing material must be used or a breakthrough factor must be statistically determined. All potential artifacts must be considered and minimized, if not eliminated, through proper sampler design and analytical process control. This often necessitates lengthy quality assurance (QA) and quality control (QC) programs. 1.3 Active vs. Passive Sampling There are two basic means of collecting a gaseous sample: active sampling and passive sampling. In active sampling procedures, air is drawn through an absorbing or adsorbing medium by a pump in order to trap the gas phase material. It is important that the sampler has an accurate, calibrated means to determine the total volume of the gaseous sample and the rate at which the gas is being sampled. This is most easily accomplished by the use of calibrated mass flow controllers. In passive sampling devices, an adsorbing material is placed at a fixed distance away from the air being sampled. Gas phase molecules must pass through a membrane or filter and diffuse across this distance and be trapped on or react with the collecting medium. The principles of diffusion are utilized to calculate the concentration of the specified contaminant in the ambient air. The diffusion rate across the passive sampler (and/or through the membrane) is analogous to the flow rate in an active sampler. 1.3.1 Active Air Collection Methods There are many active air sampling methods available. Each method utilizes different ways to collect gaseous samples and each sampling method has its own particular inherent artifacts, and, thus, each method has its own strengths and weaknesses. It cannot be stated too often that the appropriate sampling method must be selected to address a particular question or problem. There are several basic mechanisms whereby gases and vapors may be collected for subsequent analysis. Gaseous samples may be adsorbed on the surface of various substances, which have large surface areas and are specifically designed to collect the gaseous chemical species desired. They may react chemically with some chemical adsorbed on the surface of the collection device or on particles in the collector. Gases may be collected in bags or canisters or trapped in bubblers, in mist chambers or cryogenically. Each of these methods will be described briefly below. 8 1.3 Active vs. Passive Sampling
1 Collection and Preparation of Gaseous Samples Sorbents come in many varieties and may be used as beds (packed in glass or metal tubes), surfaces( deposited on tubular or annular surfaces)or in chemically treated filters designed to trap compounds selectively. They may be organic poly- mers, inorganic materials or made from activated carbon. Each sorbent material has specific advantages and disadvantages in specific sampling situations. Some sorbents are chemically treated to react with a single component and are used pecific gas samplers. They indicate the presence of a gas by a color change and the concentration of the gas by the length of the color developed in the adsorbent column. Since these sorbents give a direct indication of the gas concentration, no further analysis is performed. As a consequence, they are outside the scope of this book and will not be considered further Other sorbents are not compound specific and, as a result, trap a wide range of mpounds. Unless specifically desired (see later in this section the discussion of DNPH-coated sorbents), it is vitally important that the compounds collected do not react with the sorbent. It is an unfortunate reality that the efficiency of recovery of most adsorbed compounds from the sorbents is less than 100%. For this reason, the efficiency of the sorbent for the desired compounds and the extraction or des ption efficiency of the compounds from the sorbent must be determined Organic polymers have proven themselves to be effective adsorbents for man ganic chemical species. They include materials such as TenaxTM(2, 4-diphenyl p-phenylene oxide), XAD(styrene-divinylbenzene copolymer) and polyurethane foam(PUF). Tenax and XAD are available as small beads (less than 1 mm in dia eter) and have large surface areas for effective adsorption of organic chemicals. These materials are hydrophobic which makes them suitable for the collection of organic vapors in gases that contain a significant relative humidity. As water moisture causes significant problems for gas chromatographic analysis, the use of hydrophobic adsorbents can be a significant advantage. These resins are particu- larly effective for neutral and aromatic organics but are less effective in the trap- ping of highly polar organics. Under extremely moist conditions, however, these adsorbents may lose their efficiency, particularly if water condenses on the sorbent. PUF is suitable for the retention of polychlorinated organics such as the PCBs but is ineffective in trapping low molecular weight organics. PUF is not effective for trapping aromatics such as naphthalene, acenaphthene and acenaphthylene Although Tenax has good thermal stability it has a major disadvantage in that it notoriously difficult to clean. The XAD resins also have good thermal stability but are still difficult to clean. If not manufactured from pure materials, solvent extrac tion or thermal desorption of the polymer will release the impurities present in the starting materials or produced during the manufacturing process. This problem is not easy to eliminate. For that reason, the sorbents must be exhaustively extracted or thermally desorbed prior to use. Tenax is cleaned by thermal treatment whereas the XAD resins are usually solvent extracted. Blanks are necessary to establish the level of interferences to the analytical procedure. In addition, the sorbents, although mely efficient for the trapping of the higher molecular weight organic less effective in trapping and retaining the lower molecular weight VOCs
1.3.1.1 Sorbents Sorbents come in many varieties and may be used as beds (packed in glass or metal tubes), surfaces (deposited on tubular or annular surfaces) or in chemically treated filters designed to trap compounds selectively. They may be organic polymers, inorganic materials or made from activated carbon. Each sorbent material has specific advantages and disadvantages in specific sampling situations. Some sorbents are chemically treated to react with a single component and are used in specific gas samplers. They indicate the presence of a gas by a color change and the concentration of the gas by the length of the color developed in the adsorbent column. Since these sorbents give a direct indication of the gas concentration, no further analysis is performed. As a consequence, they are outside the scope of this book and will not be considered further. Other sorbents are not compound specific and, as a result, trap a wide range of compounds. Unless specifically desired (see later in this section the discussion of DNPH-coated sorbents), it is vitally important that the compounds collected do not react with the sorbent. It is an unfortunate reality that the efficiency of recovery of most adsorbed compounds from the sorbents is less than 100 %. For this reason, the efficiency of the sorbent for the desired compounds and the extraction or desorption efficiency of the compounds from the sorbent must be determined. Organic polymers have proven themselves to be effective adsorbents for many organic chemical species. They include materials such as Tenax (2,4-diphenylp-phenylene oxide), XAD (styrene-divinylbenzene copolymer) and polyurethane foam (PUF). Tenax and XAD are available as small beads (less than 1 mm in diameter) and have large surface areas for effective adsorption of organic chemicals. These materials are hydrophobic which makes them suitable for the collection of organic vapors in gases that contain a significant relative humidity. As water moisture causes significant problems for gas chromatographic analysis, the use of hydrophobic adsorbents can be a significant advantage. These resins are particularly effective for neutral and aromatic organics but are less effective in the trapping of highly polar organics. Under extremely moist conditions, however, these adsorbents may lose their efficiency, particularly if water condenses on the sorbent. PUF is suitable for the retention of polychlorinated organics such as the PCBs but is ineffective in trapping low molecular weight organics. PUF is not effective for trapping aromatics such as naphthalene, acenaphthene and acenaphthylene. Although Tenax has good thermal stability it has a major disadvantage in that it is notoriously difficult to clean. The XAD resins also have good thermal stability but are still difficult to clean. If not manufactured from pure materials, solvent extraction or thermal desorption of the polymer will release the impurities present in the starting materials or produced during the manufacturing process. This problem is not easy to eliminate. For that reason, the sorbents must be exhaustively extracted or thermally desorbed prior to use. Tenax is cleaned by thermal treatment whereas the XAD resins are usually solvent extracted. Blanks are necessary to establish the level of interferences to the analytical procedure. In addition, the sorbents, although extremely efficient for the trapping of the higher molecular weight organics, are less effective in trapping and retaining the lower molecular weight VOCs. 1 Collection and Preparation of Gaseous Samples 9
1. 3 Active vs. Passive Sampling Carbon-based traps have a lower affinity for water than does Tenax, but they must be purged with ultra-pure helium while being heated to drive off adsorbed mpurities. Surrogates should be added before this clean-up procedure to deter mine the efficiency of the purge. After activation, sorbents must be handled with care as they may adsorb organic vapors from the air, thus resulting in adsorp- n artifacts Because of their physical structure and specific surface area, polymeric have a finite capacity for the collection of organic compounds. It is, therefore, ne cessary to determine the capacity of a particular sorbent under the particular sampling conditions desired. The minimum sampling duration will be defined by the concentration of the organics in the sampled gas and the sampling rate. If the capacity of a sorbent is exceeded, breakthrough of the sample will occur. In practice, it is wise, if not necessary, to use two sorbents in tandem. Any com- pounds which break through the first sorbent will be trapped by the second sor- bent. This will indicate the extent to which breakthrough is a problem. Adsorbents such as XAD can be ground to sub-micron sized particles thus greatly increasing their surface area and capacity. When these particles are properly applied to concentric glass tubes called annular diffusion denuders, they provide a large surface area for the collection of gases and vapors. Passing air through these devices will result in the gases and vapors being adsorbed to the walls of the denu- der. The particles, because of their greater mass and momentum, pass through the denuder and are trapped by a filter. This sorbent /filter geometry greatly reduces the artifacts inherent in the filter/sorbent geometry PUF has commonly been used downstream of a particle filter to collect the gas phase material that passes through the filter. PUF is reasonably effective in trap- ing the higher molecular weight organics, but, like the resins, it is much less ef- fective in trapping the low molecular weight organics. PUF is also notoriously dif- cult to clean and is well known to undergo chemical degradation when exposed Atmospheric oxidants. This can be seen over the course of a single 12 h sample by a yellowing of the foam that remains after solvent extraction Inorganic sorbents include silica gel, alumina and molecular sieves. Because they are polar substances, they are particularly effective in trapping polar vapors For these sorbents, it is the degree of polarity which determines how well a partic alar gas or vapor is retained. A source of potential error is that very polar gases may displace less polar compounds from the adsorbent For atmospheric sampling this presents a significant problem in that these sorbents are also very efficient in col- lecting water, which can cause serious deactivation of the sorbents. As a result, these compounds are not often used for the collection of organic vapors An adsorbent such as silica may be treated chemically with, for example, 2, 4-di nitrophenylhydrazine(DNPH). When an air sample is passed through this mate- rial, organic carbonyls react with the dNPh to form the dinitrophenylhydrazone which can then be extracted and easily detected In such a system, the sorbent simply acts as a large area substrate for the chemical reactant. Activated carbon in beds and impregnated in glass fiber mats has been used collect organic vapors. These sorbents are relatively non-polar and trap a wide
Carbon-based traps have a lower affinity for water than does Tenax, but they must be purged with ultra-pure helium while being heated to drive off adsorbed impurities. Surrogates should be added before this clean-up procedure to determine the efficiency of the purge. After activation, sorbents must be handled with care as they may adsorb organic vapors from the air, thus resulting in adsorption artifacts. Because of their physical structure and specific surface area, polymeric sorbents have a finite capacity for the collection of organic compounds. It is, therefore, necessary to determine the capacity of a particular sorbent under the particular sampling conditions desired. The minimum sampling duration will be defined by the concentration of the organics in the sampled gas and the sampling rate. If the capacity of a sorbent is exceeded, breakthrough of the sample will occur. In practice, it is wise, if not necessary, to use two sorbents in tandem. Any compounds which break through the first sorbent will be trapped by the second sorbent. This will indicate the extent to which breakthrough is a problem. Adsorbents such as XAD can be ground to sub-micron sized particles thus greatly increasing their surface area and capacity. When these particles are properly applied to concentric glass tubes called annular diffusion denuders, they provide a large surface area for the collection of gases and vapors. Passing air through these devices will result in the gases and vapors being adsorbed to the walls of the denuder. The particles, because of their greater mass and momentum, pass through the denuder and are trapped by a filter. This sorbent/filter geometry greatly reduces the artifacts inherent in the filter/sorbent geometry. PUF has commonly been used downstream of a particle filter to collect the gas phase material that passes through the filter. PUF is reasonably effective in trapping the higher molecular weight organics, but, like the resins, it is much less effective in trapping the low molecular weight organics. PUF is also notoriously difficult to clean and is well known to undergo chemical degradation when exposed to atmospheric oxidants. This can be seen over the course of a single 12 h sample by a yellowing of the foam that remains after solvent extraction. Inorganic sorbents include silica gel, alumina and molecular sieves. Because they are polar substances, they are particularly effective in trapping polar vapors. For these sorbents, it is the degree of polarity which determines how well a particular gas or vapor is retained. A source of potential error is that very polar gases may displace less polar compounds from the adsorbent. For atmospheric sampling this presents a significant problem in that these sorbents are also very efficient in collecting water, which can cause serious deactivation of the sorbents. As a result, these compounds are not often used for the collection of organic vapors. An adsorbent such as silica may be treated chemically with, for example, 2,4-dinitrophenylhydrazine (DNPH). When an air sample is passed through this material, organic carbonyls react with the DNPH to form the dinitrophenylhydrazone which can then be extracted and easily detected. In such a system, the sorbent simply acts as a large area substrate for the chemical reactant. Activated carbon in beds and impregnated in glass fiber mats has been used to collect organic vapors. These sorbents are relatively non-polar and trap a wide 10 1.3 Active vs. Passive Sampling
range of organics. It is extremely difficult to remove adsorbed organic commpound n from activated carbon sorbents. This limits their applicability in air sampling un- less the sole purpose of the adsorbent is to remove organics from an air stream. 1.3.1.2Bags Bags made of aluminum/plastic or of plastic laminates can be used to collect gas eous samples. They are filled either with inert surface pumps or indirectly by pla cing the bag in a non-flexible, closed container and evacuating the space betweer the bag and the container. When the space between the bag and the rigid container is evacuated, the bag will inflate, drawing in the air sample. Bags must be carefully cleaned and examined and tested for leaks prior to sampling. Loss of organics to the walls can be a significant problem. Diffusion of gases and vapors through the walls has been greatly reduced by the use of the laminated plastics. The us of bags allows a grab sample of air, usually less than a cubic meter in volume, to be collected. Because a relatively small volume of air is collected, the compound of interest must be in sufficiently high concentration that it can be detected and quantified. Bag collectors can be bulky and difficult to transport. If the collected sample is to be analysed by a technique such as gas chromatogra phy-mass spectrometry, the sample must be passed through an adsorbent to con- centrate the hydrocarbon gases and vapors. The sample must then be released from the adsorbent, either by solvent extraction or thermal desorption, prior to in- jection into the gas chromatograph. 1.3.1.3 Canis Air may be collected in glass or steel containers. Glass containers may be evacuated prior to sample collection or the air sample may only the collection of a grab be drawn through the container. Glass containers because of their small size. al sample. Steel canisters with electropolished or chemically deactivated interiors nay also be used to collect air samples. Inner surface treatment is necessary, as stainless steel is an adsorptive medium. Most canisters are designed to be evacu- ated in the laboratory then transported to the sampling site. However, prior to use, the canisters must be cleaned. This is a laborious procedure that requires that the canisters be evacuated to below 0.05 Torr and heated for several hours The procedure may have to be repeated several times if the canister was previous exposed to a"dirty"sample. Even this may not satisfactorily clean a"dirty"canister. The canister may have to be wet cleaned to remove some polar compounds in the canister is opened to allow the air sample to leak into the at a defined rate. The leak rate is fixed, often by a use of a suitable critical orifice. Depending on the leak rate, the size of the canister and the initial vacuum, the can- isters may collect short-term grab samples or may extend the collection process up to as long as 48 hours, although sampling times of 6 to 8 hours are more common. Canisters have been used primarily to collect vOCs. Water management is a major problem with canisters and, to reduce its effects, Nafion drying tubes may have to
range of organics. It is extremely difficult to remove adsorbed organic compounds from activated carbon sorbents. This limits their applicability in air sampling unless the sole purpose of the adsorbent is to remove organics from an air stream. 1.3.1.2 Bags Bags made of aluminum/plastic or of plastic laminates can be used to collect gaseous samples. They are filled either with inert surface pumps or indirectly by placing the bag in a non-flexible, closed container and evacuating the space between the bag and the container. When the space between the bag and the rigid container is evacuated, the bag will inflate, drawing in the air sample. Bags must be carefully cleaned and examined and tested for leaks prior to sampling. Loss of organics to the walls can be a significant problem. Diffusion of gases and vapors through the walls has been greatly reduced by the use of the laminated plastics. The use of bags allows a grab sample of air, usually less than a cubic meter in volume, to be collected. Because a relatively small volume of air is collected, the compound of interest must be in sufficiently high concentration that it can be detected and quantified. Bag collectors can be bulky and difficult to transport. If the collected sample is to be analysed by a technique such as gas chromatography-mass spectrometry, the sample must be passed through an adsorbent to concentrate the hydrocarbon gases and vapors. The sample must then be released from the adsorbent, either by solvent extraction or thermal desorption, prior to injection into the gas chromatograph. 1.3.1.3 Canisters Air may be collected in glass or steel containers. Glass containers may be evacuated prior to sample collection or the air sample may be drawn through the container. Glass containers, because of their small size, allow only the collection of a grab sample. Steel canisters with electropolished or chemically deactivated interiors may also be used to collect air samples. Inner surface treatment is necessary, as stainless steel is an adsorptive medium. Most canisters are designed to be evacuated in the laboratory then transported to the sampling site. However, prior to use, the canisters must be cleaned. This is a laborious procedure that requires that the canisters be evacuated to below 0.05 Torr and heated for several hours. The procedure may have to be repeated several times if the canister was previously exposed to a “dirty” sample. Even this may not satisfactorily clean a “dirty” canister. The canister may have to be wet cleaned to remove some polar compounds. A valve in the canister is opened to allow the air sample to leak into the canister at a defined rate. The leak rate is fixed, often by a use of a suitable critical orifice. Depending on the leak rate, the size of the canister and the initial vacuum, the canisters may collect short-term grab samples or may extend the collection process up to as long as 48 hours, although sampling times of 6 to 8 hours are more common. Canisters have been used primarily to collect VOCs. Water management is a major problem with canisters and, to reduce its effects, Nafion drying tubes may have to 1 Collection and Preparation of Gaseous Samples 11