Document 4Jj1OKkBLJpZn4K9eJKLewo5a

TO Lloyd Stakes AT Lake Charles FROM D* B. Smith AT New Haven SUBJECT Guide To The Use Charcoal Tubes For Monitoring Airborne Organic Vapors date September 9,'1980 COPY TO L. A. Krause N. Patel C. Self The use of glass tubes packed with activated carbon for the quantitative collection of airborne organic vapors is well documented throughout the industrial hygiene literature. Generalization of parameters influencing collection and desorption efficiencies have also been reported. An excellent review by Melcher et al appearing in the American Industrial Hygiene Association Journal (39) 5/78 page 3^9 is attached. Complete sampling and analytical methods for a number of single compounds and mixtures are abundently available. Many of these methods have been rigorously validated following the use of vapor generation systems and exhaustive field testing. When monitoring for a new compound or a unique mixture it is imperative that the desorption efficiency be determined from tubes spiked with known amounts of.compound. Collection efficiency can only be determined from the use a vapor generation system which produces known concentration of the compound(s) or by closely monitoring the breakthrough or backup section of a charcoal tube from an actual field sample. Generally, if the amount of analyte in the breakthrough section exceeds 255? of the main section the sample is declared invalid. Data from such a sample is often reported as "minimum concentration present". A reduction of sampling time and/or sampling rate will reduce breakthrough to within exceptable limits in most cases.~ As we recently discussed there is a need to standardize how a charcoal tube is treated relative to our two laboratories to Insure that desorption efficiencies are being determined in the same manner. Information contained in the following sections highlights some important aspects of charcoal tube technology. The Charcoal Tube Three sizes of charcoal tube are readily available commerically for the routine monitoring of a wide range of organic compounds. The tubes are generally packed with a coconut based charcoal and differ only by the amount of charcoal and tube dimensions. The volume of desorption solvent relative to the amount of charcoal can drastically effect desorption efficiency. The-threS types of common charcoal tubes and recommended desorption solvent volumes are described in Table I. < ----- OLI 7729 . . *' # L. Stakes September 9> 1980 Page 2 ' TABLE X TUBE TYPE Standard Large Jumbo MILLIGRAMS FRONT/REAR 100/50 *100/200 800/200 MILLILITERS OF DESORPTION SOLVENT 1/1 3/1 5/2 Desorption Solvent The most commonly used desorption solvent used in conduction with the charcoal tube.and the PID detector is carbon disulfide (CS2) owing to its minimal FID response. It is also a eommon practice to utilize solvent mixtures such as 95% CS2 - 5% isopropanol to improve the desorption of moderately polar compounds. Other solvents such as hexane and dichloromethane are also used. The need to use a solvent mixture or an alternate to CS2 is generally dictated by poor desorption efficiency (<J 50$7 for a particular compound. The use of mixtures and alternate solvents can lead to analytical problems on the GC particularly if a large solvent peelc masks components from the sample. An increase In desorption solvent volume may sometimes improve recovery with an attendent loss in sensitivity due to dilution of the sample. Quality Control The most valuable aid to establishing credibility relative to charcoal tube analysis is the routine use of spiked tubas and the attendent percent recovery of the analyte. Failure to achieve reproducible recovery may Indicate an instrument malfunction or a calibration standard problem. As a matter of routine spiked tubes prepared by a competent analyst or chemist must be analyzed along with each batch of field samples. Data from a routine analysis can be tabulated and evaluated statistically as it accumulates resulting in a wealth of Information concerning the precision and accuracy of the analytical method. DBS/gmm Attachments qlI 7730 uu I *4 .A smalt tube containing a solid sorbent is convenient to use for collecting and ` concentrating trace organics in ambient air and workerbreathing zones. Many techniques for air sampling with solid sorbents, especially for the long term eight hour sample, are relatively new. It is necessary to have criteria to judge methods utilizing these techniques and to establish guidelines which are accurate and practical. Parameters which affect collection and recovery and information necessary to define, develop and evaluate a sampling method are discussed. Progressive steps for the development and validation of air sampling methods are recommended. Criteria for the evaluation of methods for the collection of organic pollutants in air using solid sorbents* i RICHARD G. MELCHER,* RALPH R. LANGNER* and RONALD 0. KAGEL' `Michigan Division Analytical Laboratory, Dow Chemical U.S.A., Midland, Michigan 48640; `industrial Hygiene Laboratory, Dow Chemical U.S.A., Midland, Michigan 48640; `Environmental Services, Dow Chemical U.S.A., Midland, Michigan 48640. introduction In the early development of solid sorbent sampling techniques, a ten minute sample period was adequate to establish the concentration of a compound in air in an area. Since the sampling period was short, a large number of variables could be tested for the method validation in a relatively short time. Secondly, parameters such as humidity were not considered critical for short samples. Recently, the emphasis has been toward the development of portable personnel sampling systems which could be used up to eight hours to determine "time weighted average" (TWA) exposures. Validation of methods for long term sampling becomes quite time consuming and more difficult since parameters such as humidity and flowrate are more critical. It is therefore also important to define the data necessary to validate a method in the most efficient manner. The main emphasis in this paper is on the collection of organic pollutants with solid sorbents followed by desorption with a suitable solvent and analysis. This includes a majority ofthe samples presently being collected and new methods being developed. Basic principles and specific `Dow Chemical Company CRI No. B 600-134-77 parameters which can cause errors if not considered, are discussed. Although the procedures discussed will be directed toward ambient air monitoring, either for industrial hygiene or air quality, adaptation can be made for monitoring stationary sources if particular attention is given to examining the extreme conditions often found in these situations. With this scope in mind, the general areas discussed will be: (1) Some of the parameters which affect collection and recovery, how the parameters can be evaluated, and what extrapolations can be made. (2) What information is necessary to define and develop a tentative collection and recovery procedure. (3) What criteria are necessary to evaluate a detailed validation of a tentative procedure which has been developed through an in-depth study or by extrapolation of a method for a similar compound. .analytical procedure The total monitoring method can be divided into ^merton InduMriil Hygiene Associetion JOURNAL (33) S/7! 349 OLI 7731 V'. % g~ length 01 Sorttni Bed 3 ft /3. y-- Effluent Monitored Collection Tube Figure 1 -- Concentration distribution in a solid sorbent and vapor breakthrough. the Analytical Procedure and the Sampling Procedure. If we define the Analytical Procedure as the determination of specific compounds in the desorption solvent, then either procedure can be changed independently as long as the interfacing parameters are considered. However, an adequate analytical procedure must be available before collection and recovery' data can be obtained. Gas chromatography is the most widely used analytical technique because ofits selectivity and sensitivity. Gas chromatographic procedures are available or can be developed for a wide range of organic compounds by the proper selection of columns and detectors. The range can be broadened by using chemical derivatives and gas TABLE I Interfacing Analytical and Sampling Procedures 1. Sensitivity 2. Specificity 3. ReproduCibiBty of analysis 4. Reproducibitty of apparatus 6. Adaptability 6. Simplicity 7. Automated sensitivity range compatible with amount of compound to be collected interferences in the area and on a broader scale investigated over concentration range to be studied operating conditions wall defined and columns, reagents etc. widelyavailable compatible with desorption solvents end other reagents rapid, uncomplicated procedure automatic injection and data handling systems chromatographic/mass spcctromctric tech niques. It is beyond the scope of this paper to discuss the complexities of the gas chromatographic or other analytical techniques; however, Table I lists analytical parameters which should be determined before and during the Sampling Procedure development. A general guide for industrial hygiene laboratory quality control has been prepared.<" sampling procedure The sampling procedure can be subdivided into collection and recovery, each ofwhich involve a number of parameters which must be determined. In many cases, extrapolation of existing information serves as a guide to the selection of a suitable collection sorbent and desorption solvent. Although many of the parameters involved hav&been studied in detail, the number of variables encountered in field sampling are often too great to rely solely on these extrapolations for untested compounds. Verification of the collection and recovery in the laboratory before the samples arc taken saves time and reduces errors. Deteriorative effects on collection efficiency will be detected, after the fact, by high concentrations on the "back-up" section. However, if the factors are recognized before or during sampling, a greater safety factor can be applied to the sample volume which will reduce 350 Am. M. Hyt. Assoc. I (33) My, 1978 OLI 7732 f. respirator cartridges. the number of invalid samples and the need to resample. Factors which cause changes in recovery are more insidious. Unless these are understood and can be related to the chemical and physical properties of the compounds collected, errors can go undetected. collection sorption process Without getting too deeply in adsorption theory, we can look at what happens when air, containing an organic compound at concen tration Ci, is pulled through a tube containing a solid sorbent. Figure 1 illustrates the distribution curve of the compound through the sorbent bed and the breakthrough curve of the compound in the effluent. At the start of collection the compound is distributed through a section of the sorbent as shown by curve 1. After continued collection this section becomes "saturated" that is, an equilibrium is established with incoming concentration Ci where the vapor sorption and desorption rates are equal. This is shown by curve 2. The sorption front has moved through the front section of the collection tube (L 2/3) and is being collected on the backup section. If collection continues, the equilibrium zone lengthens and the compound begins to breakthrough the collection tube at time Tb, as shown by curve 3. The breakthrough curve 4 is generated by monitoring concentrations of the compound in the effluent, and when all the sorbent is "saturated" the effluent concentration approaches the input concentration Ci. Breakthrough is defined as the detection ofthe compound in the effluent as a percentage of Ci. As a practical guide in air sampling this is often set at 5%."' The breakthrough volume, which is more indicative ofthe collection efficiency, is the product of the sampling rate and the sampling time. A more detailed discussion of the collection mechanism can be found in the literature for sorbents such as charcoal/5-7' silica gel<,,l and gas chromatographic packings."0'"' A general guideline can be proposed for actual field samples. If less- than 10% of the total amount collected is found on the 1/3 backup section,. significant loss of the compound has probably not occurred, and ifgreater than 25% is detected loss has probably occurred and results should be reported as "minimum amount present". More specific values can be obtained for this general statement by a detailed study of the breakthrough profiles for each specific sorbent and sorbate. Factors which affect collection efficiency are discussed below. The degree to which each factor Americtn Industrial Hygiene Assodition JOURNAL (39) S/78 3S1 contributes depends on the specific sorbent and mixtures. By decreasing the sorbent size, the sorbate. efficiency increases because of the increase in ij properties of the compound surface area;'7' however, the pressure drop also increases and limitations of the sampling pump A number of equations have been developed to must be considered. Prior treatment and relate the properties of the compound to activation of the sorbent will affect efficiencies. collection efficiencies.'3'4'** Some equations For example, not all types of charcoal are the become quite involved and often, require same and not all batches of the same type information not readily available for many charcoal necessarily have the same collection compounds. It is usually more efficient to obtain and recovery properties."3-15' experimental breakthrough data on a specific collection system for the compounds of interest. size of collection tube Correlation graphs can then be prepared for a In general, if the size (amount of sorbent) is group of compounds by plotting the doubled the breakthrough volume is doubled. A breakthrough volume (or time) against a somewhat greater increase in sampling time may physical property. For example, experimental be observed depending on the shape of the breakthrough data, or data from the literature,'41 sorption zone front, as shown in Figure 1, since can be plotted as shown in Figure 2. Up to a the sorbent is not at full capacity at the point of boiling point of 120C, a very definite trend is breakthrough. observed. By relating the boiling point of an untested compound, some predictions can be made if the other factors discussed below are considered. flow rate The effect of flow rate varies with the sorbent. In some cases once the optimum flow rate is reached no increase in breakthrough volume is nature of sorbent observed with reduction in flow rate.'7' In other In selecting a sorbent for high collection cases the efficiency continues to increase,'13'131 as efficiency, the recovery of the compound must shown in Figure 3. The coconut base carbon also be considered. Often a mutual compromise reached a maximum efficiency at approximately may be necessary to obtain an acceptable 100 ml/min. while Dow developmental system. Table II lists some of the most used adsorbent XF-4175L(Saran Carbon) continued sorbents and the types of compounds collected. to increase in efficiency. The linear velocity at Overlap allows some flexibility when sampling optimum conditions should be an important i I l I i: i 352 TABLE II General Sorption~desorption Systems for Organic Compounds Sorbent Desorption Solvent Types of Compounds Activated carbon Silica get Activated alumina Porous polymers Chemically bonded and other GC packings Carbon disulfide. Methylene chloride, Ether (1 % Methanol or 5% Isopropanol sometimes added) Methanol. Ethanol Diethyl ether, water Water, Diethyl ether. Methanol Ether, Hexane. Carbon disulfide. Alcohols Ether, Hexane. Methanol Misc. volatile organics: Methyl chloride vinyl chloride and other chlorinated ^ aliphatic*, aliphatic and aromatic solvents, acetates, ketones, alcohol* etc. Polar compound*: alcohol*, phenols, chlorophenols. chlorobenzenes, aliphatic and aromatic amines, etc. Polar compounds: alcohols, glycols, ketones, aldehydes, etc. Wide range of compounds: phenols, acidic and basic organics, multi-functional organics, etc. Specialized: High boiling compounds pesticides, herbicides, polynuclear aromatics, etc. I Am. Ind. Hyt- Assoc. J. (39) May, 1978 OLI 7734 Figure 3 -- Collection efficiency of activated coconut base carbon and Saran carbon for vinyl chloride. consideration when extrapolating data to different size tubes.*1*1 concentration Breakthrough occurs sooner for higher concentrations. For charcoal the empirical Freundlich isotherm appears to apply.,t',J) This equation takes the form: log T* = log a + b log c (1) A straight line results if the log of the breakthrough time Tb is plotted against the log of the concentration, C. The line will have the slope b and intercept of log a, and is shown in Figure 4. Figure 4 -- Breakthrough time as a function of concentration. American Industrial Hjrgiene Association JOURNAL (3$) S/78 353 Length Of Sorbent Bed Figure S -- Effect of Coadsorption on breakthrough. Equations developed for the collection of amines on silica gel predict the amount of a compound collected is independent of the inlet concentration and flow rate.'9' A theoretical approach to the collection of compounds on gas chromatographic packings has also been reported.09' humidity In general, an increase in humidity will result in an increase in breakthrough. The magnitude of this effect depends on the properties of the sorbent and sorbate. Although only a small amount ofwater is collected on charcoal the high ratio of water molecules to the compound of interest affects the sorption-desorption equilibrium and increases breakthrough as much as 50%.n41 A greater effect is observed for low concentrations ofsorbate as indicated by the change in the slope for high humidities in Figure 4. Polar sorbents such as silica gel and alumina, adsorb water more strongly than most organic compounds. This not only continues to reduce the effective length ofthe collection tube but also displaces compounds already collected. The extent of displacement is dependent on the polarity of the collected compounds. This effect can be treated in terms of coadsorption discussed next. collected, the compound most strongly held will displace the other compounds, in order, down the length of sorbent bed Figure 5. For a polar sorbent, compounds with the largest dielectric constant and dipole moment are most strongly held. For non-polar compounds, preferences are usually for the compounds of a higher boiling point or a larger molecular volume.01 temperature An increase in temperature will result in an increase in breakthrough. There is little specific information available for the various sorbents. For charcoal a general guideline has been suggested that for every 108C rise in temperature the breakthrough time will be reduced by 1I0%.7' migration A false indication of breakthrough may be caused by migration of the compounds collected on the front section to the backup section over an extended storage period. This equilibration approaches 33.3% on the backup section for the more volatile compounds.01'141 This problem can be reduced by refrigerating samples as soon as possible, or eliminated by separating the front and backup sections. coadsorption Figure 1 .deals with the collection of one compound. When two or more compounds are recovery desorption efficiency A number of factors influence recovery 354 km. Ind. H/i Assoc. I (39) Miy, 1978 TZ t- OLI 7736 Mg. Styrene Monomer/5 ml. CSj/0.5 gm. Carbon Figure 6 -- Variation of desorption efficiency ofstyrene monomer with concentration end desorption time. efficiencies (that fraction recovered of total compound collected when no breakthrough has occurred). The desorption efficiency is the most significant ofthe factors in defining the sorptiondesorption system. Although desorption efficiency cannot always be isolated, it can be determined experimentally and is one ofthe first indicators of potential in a suggested method. Some of the sorption-desorption systems used are listed in Table I. Although tbe desorption efficiency is generally treated empirically, some insights are obtained from a theoretical approach. A study has shown,'14* that for several types of compounds collected on charcoal, the system can be treated as a phase equilibrium. Preliminary data also indicates the same treatment can be applied to silica gel and porous polymers.120'2" The organic compounds are partitioned between the solid adsorbent and the desorption solvent according to the equation: J___me sorbent , . D K mg solvent T 1 (2) where D is the desorption efficiency expressed as a fraction and K is an equilibrium constant. This concept is important since it indicates that the solvent - sorbent ratio cannot be changed without affecting the desorption efficiency. This property can be used to advantage for improving recovery. For example, Amcrictn Industriil Hygiene Auo&tion JOURNAL (39) 5/78 if a desorption efficiency of 50% has been obtained by desorbing 100 mg of charcoal with 1/2 ml of carbon disulfide, by using 5 ml of solvent, an efficiency of91%should be obtained. It also indicates that if the backup section containing one-half the amount of sorbent is desorbed in the same volume, a 5 to 15% positive error could be introduced. temperature The temperature effect on recovery can be significant. Ifhigh temperatures are encountered while sampling; decomposition, polymerization or other chemical reactions may occur on the activated surface of the sorbent It is a common practice to cool the solvent before desorbing samples to reduce the heating effects. However, the equilibrium constant will change with temperature and the final desorption temperature may be critical. An efficiency change has been observed for MEK from 89.5% to 66.8% at 258C and 0C respectively.1141 The magnitude of the effect depends on the compound. Vinyl chloride shows no effect with desorption temperature in this range.'1" humidity High humidity during collection may produce low recoveries for compounds which are easily hydrolyzed. If a large amount of water is collected it may prevent good contact with a 1 i' 3SS .4 non-polar desorption solvent, or change the equilibrium if it dissolves in the desorption solvent. This effect is more pronounced with polar sorbents and can sometimes be counteracted by using longer desorption times and stronger (more polar) desorption solvents. coadsorption Collection of other compounds can also affect recovery. The chemistry of all the compounds collected must be considered and their relative effects tested in the laboratory. 0 4 8 12 16 Min. Figure 7 -- Formation ofinterferences from methanolin a carbon disulfide and charcoal mixture. TABLE III Desorption Efficiency of Aromatic Compounds Sorbent: Cocontfi 6m Carbon Q.5 gram* Solvent; Carbon Ditulfidt 6 ml Concentration: 02-0.3 mg/6 ml Beniene Deeorption Efficiency After 1 hr. After 16 hr*. 98 68 Toluene 88 68 Nephthelene y Biphenyl (J.o{ ^ Diphenyl Oxkto Qc-O Diphenyl Methene ^ ^ ^-CH> Per* Methyl Biphenyl 43 65 93 93 84 40 S3 98 67 desorption time One-half to one hour with good agitation is usually sufficient desorption time for most systems. Once optimum desorption has taken place the system is generally stable; however, some exceptions have been observed.<19' Figure 6 shows the decrease in recovery for styrene monomer over a 16 hour period. This effect is more pronounced for lower concentrations. When this situation is observed, the extract should be removed from the sorbent if samples cannot be analyzed during the optimum desorption period. This type of decrease was not observed for the compounds listed in Table III. It is also necessary to establish the stability of the solutions and reagents because of the possible catalytic effect of the activated sorbent surface. In some procedures using charcoal, 1% methanol is added to increase desorption efficiency, particularly for polar compounds. If the mixture is allowed to stand in contact with the charcoal longer than four hours a reaction takes place forming a number of peaks shown in Figure 7. Similar chromatograms were observed with Durapak - Carbowax 400/ Porasil F, 5% QFI or 5% OV 210 gas chromatographic column using a temperature program. Tentative identification by gas chromatography/ mass spectrometry indicates dimethyl di-, tri-, and tetrasulfides, mercaptans and polyether thioether compounds.,,9, Ethanol gave one major peak (6), but no reaction peaks were observed with isopropanoL storage In most cases samples are not analyzed the same day they are taken, and some compounds may show low recoveries after a one or two week storage period. Refrigeration ofthe samples may 356 Am. tnd. Hut Assoc. X (33) May, 1978 help or immediate desorption for later analysis can be used if the solutions are stable. concentration The phase equilibrium model predicts the desorption efficiency should not vary with concentration. For most methods developed for organics on charcoal, this concept is generally true (within ~ 5%) for concentrations of interest in industrial hygiene samples. Some compounds have a tendency toward lower recoveries as illustrated by styrene monomer in Figure 6. As the concentration decreases a decrease in desorption efficiency is observed. Any factors which affect the recovery will usually have a greater percentage effect op the lower concentrations. recovery prediction Some extrapolations can be made from existing information but more data is necessary before accurate prediction can be made from physical and chemical properties. An example is shown in Table III which illustrates high attraction (low desorption efficiencies) for biphenyl and naphthalene amd lesser attraction for similar compounds possibly because of bond angles or steric effects.'20' experimental procedures desorption efficiency direct injection (static) Desorption efficiency is usually determined by placing the scwbent from the front section of a collection tube in a small vial and injecting a known amount (a few /d) of the compound or solution of tbe compound directly into the sorbent at several points. The vial is capped and analyzed subsequently using the exact procedure to be used for the samples. Periodic analyses of this same solution over a longer time period (4-8 hours) can be vised to optimize desorption time and determine a time limitation. If several preparations are made a week or more in advance, an estimate of the storage effects can also be made- phase equilibrium The phase equilibrium technique for determining desorption efficiency approaches the equilibrium from the other direction.11*' Known concentrations ofthe compound(s) to be studied are prepared in the desorption solvent and analyzed. An aliquot of this solution, equal to the volume used in desorbing samples, is cooled and the front section of sorbent from a sampling tube is added. The solution is analyzed periodically until a steady state is obtained. preliminary bias and breakthrough studies Although method bias and breakthrough will be determined more accurately during the validation procedure, early identification of excessive breakthrough or large differences between the desorption efficiency and total recovery of a simulated air sample may save considerable time. If the desorption efficiencies as determined by the direct injection are considerably lower than phase equilibrium values, interaction or reaction on the sorbent surface is indicated. If the total recovery from a simulated air collection is lower than the direct injection efficiency, even though no break through has occurred, hydrolysis, oxidation, or other reaction may be indicated. breakthrough and recovery The factors which affect each of these have been discussed. The extent of the effects can be determined in the laboratory with simulated air sampling experiments. A number ofsystems for producing a known atmosphere have been suggested. These include air bags, permeation tubes, diluters, etc., and are discussed in detail in the literature.'22' A device, adapted from the literature,'23-' has been used with success for a wide range of liquids and solids.'241 Air, which is pulled through a saturated salt solution to create a known humidity, is swept through a U-tube, and carries the compound vapors into-the collection tube. Breakthrough for volatile compounds can be determined by monitoring the effluent by gas chromatography using a gas syringe or gas sampling valve. For compounds which cannot be monitored this way, tubes containing four sections of sorbent, each 1/2 the size of a field sampling tube can be used (the cross-section should be the same). Section no. 3 would represent the backup section and section no. 4 could be used to determine breakthrough. The different parameters arc examined in subsequent runs by varying the concentration, humidity, American Industrial Hygiene Association JOURNAL (33) 5/79 (357 OLI 7739 r- temperature of the collection tube, etc., and analyzing each section separately to determine sorption profiles. See Figure 5. After each run, the U-tube is rinsed with an appropriate solvent, which is analyzed to obtain a mass balance for each component. A modification of this technique can be made for very low concentrations of high boiling materials such as pesticides,*151 An extra small glass wool plug is placed in the front end of the divided tube. A small amount ofa dilute solution is injected into this plug at the start of a collection test. Analysis of the plug separately from the sections will give a mass balance. criteria for a validated method The factors which affect air sampling with solid sorbents can be quite complex. It is necessary to define some general criteria as a standard to evaluate acceptable methods and a guide to determine the data necessary to support the evaluation. The various steps in method development and validation can be divided into several sections as outlined in Table IV. TABLE IV Progressive Development and Validation Steps Suggest procedure (little or no experimental work) 1. Chemical, physical, and toxicological properties listed. Z. Concentration range end required sensitivity estimated. 3. Sorbent-desorption eystem(s) extrapolated from similar compounds. 4. Analytical procedure suggested. 5. Possible interferences indicated. Tentative procedure (limited experimental work) 1. Analytical procedure tested. 2. Desorption efficiency determined. 3. Optimized desorption time. 4. Preliminary breakthrough and bias study. 5. Preliminary reU samples. 6. Effect of interferences, coadsorption and temperature (H indicated). Validated method (systematic study) 1. Determination of breakthrough volume at concentrations of 2 times TIV and relative humidity of 85% or greater. 2. Five samples at each concentration, 0.1.0.6,1 and 2 times TLV for relative humidities less than 50% and greater than CSV (40 samples). 3. Six samples at the TLV concentration (three collected at each humxSey) and stored for at least 14 days. 4. Statistical evataation of data. 5. Field validation by comparison to accepted method or by field spiking. . Collaborative studies with other laboratories. TABLE V Criteria for Succeasful Validation 1. Accuracy and Precision of Combined Analytical and Collection Procedures (or concentration range of l/10TLVto2TLV -- 116% relative at the 95% confidence level. 2. Total Recovery efficiency of at least 75%. 3. Bias between Total Recovery end desorption efficiency less than 110% relative. 4. Capable of taking 10-15 minute samples at ceiling and excursion levels. 5. Minimum sampling time one hour, preferred 4 to 8 hours for TWA. 6. Storage samples should compare within 110% relative to initial samples after being stored for 14 days. 7. The flow rate of the sampling pump should be known with the accuracy of 5%. The amount of testing necessary depends on the extent of the sampling project Methods for limited surveys, where the sampling conditions and concentrations are defined and relatively constant, can be validated for the specific situation. The minimum amount of data necessary will-vary with the project, but a guideline such as the 10-10-10 principle,*2*1 or a recently suggested 7-7-7 principle for water effluent analysis, may be adapted. For a specific situation, this would involve seven determin ations on controls (blanks and interferences), seven determinations on various concentrations for recovery data (spiked tubes) and seven determinations in a controlled or simulated air sampling situation to determine the precision of the procedure (specified humidity, temperature etc.). If additional sampling variables are subsequently introduced or if the procedure is proposed for general sampling, additional data should be obtained. Specific guidelines, in Jablc V and discussed below, have been developed from the literature and study of a wide range of organic compounds in industrial areas. They are intended to indicate the degree of reliability which should be obtained from an acceptable method using diligent but practical application of develop mental and sampling techniques. If a meth d fails to meet some of the criteria, it could still be used as long as its limitations are known and dealt with effectively. It is important to point out that although a method is validated for a specific ppm v/v 351 Am lad. Hyg. Assoc, i. (39) May, 1978 V- . ' - ' V.i*' >. v r%-,. ,, 'i- OLI 7740 jvi-V,11 V917 J.K' : j concentration range (1/10 to 2 TLV) we are actually dealing with a specific weight of chemical. If much smaller volumes of air are \Rn taken for actual field samples, poor precision Statistical treatment of data obtained as will reflect the low milligram amount collected suggested in Table IV will establish an overall even though the calculated ppm v/v precision of the total recovery and help identify ft concentration is within the range validated. any bias. Data should be obtained for the lower \ accuracy and precision concentration of 1/10 TLV since many areas Accuracy, how close the observed value x is to contain mixtures of several compounds. If the true value x, can be expressed in terms of methods are validated for concentrations only relative error. down to 1/2 TLV, it would be difficult to -Mrelative error 100 support data for mixtures of more than two (3) compounds as indicated by the ACGIH additive The relative error can be separated into rule.*1*1 "determinate or systematic" errors and The suggested precision guidelines of 16% "indeterminate or random" errors. The were developed from several groups of systematic errors, which are unidirectional and information. Data has been obtained in our arc inherent in the method, are due to laboratories for the collection and analysis of incomplete desorption, inefficient extractions, etc., and once determined can be expressed in terms of percent recovery, R. The deviations due to the random errors can be treated by an expression of precision. over 50 compounds for sampling periods of4 to 8 hours. The average relative precision was 6.5% with a range of 3% to 10% at the 95% confidence level (not including sampling pump precision). After the series of recovery experiments the Out of the 98 NIOSH methods published for data are treated by calculating the standard solid sorbent-gas chromatographic tech deviation a and applying it in equations (4) (5) niques,(J,) 87 fall within 16% relative precision and (6). level with less than 12.9% relative precision at The relative precision for a single determination, RPs, at the statistical confidence the 95% confidence level. In another study/29' where fifteen investigators sampled seven level of 95%, is given by equation (4), with a determined average percent recovery, Rn, for n number of determinations. different compounds, each over a range of5% to 2 times the TLV (a total of 300 samples by each person), the relative error in the range of 60% k RPs = x 100 TLV to 2 times the TLV was 12.7%., At a concentration of 5% of the TLV, the relative Rn (4) error was double or approximately 25%. All The relative precision for the determination of the samples were taken for 10 minutes. average percent recovery, RPn, is RPs bias, desorption efficiency and total recovery RPn = Vn" (5) By treating the various groups of data If the validation data has been obtained using a mass balance generation technique such as direct injection or U-tube, the contribution of pump precision, RPp must be considered. The total relative precision, RP-r. is then given by equation (6). RPt =V(Rps)1 + (RPn)1 + (RPp)1 (6) separately, a bias may be detected when the recovery varies as a function ofa parameter such as concentration, humidity, storage, flow rate etc., even though no breakthrough has occurred. When such a bias, (the difference between the desorption efficiency and total recovery,) is observed to be greater than 10%, it should be plotted as a function of the variable and a The true concentration, C, forany experimen correction for it included in the method. When tally determined concentration Z, is given by bias is observed for more than one parameterthe equation (7) when Rn and RPt are given as a collection and recovery procedures should be percent: ;l reconsidered. ,f. Hv Americin Industrie! Hygiene Attociitiofl JOURNAL (39) 5/78 OL1 7741 359 sam fpfling time A one hour sampling period may be sufficient to determine a time weighted average (TWA). A more accurate value can be obtained with a longer sample especially in areas where the atmosphere varies greatly or when personnel spend a large part of their time in a number of different areas. In most cases the procedures used for lone term samples will work for short term samples when determining ceiling and excursion values. If it is necessary to increase the flow rate to obtain the required sensitivity, additional breakthrough data will be necessary. factor from the spiked tube by difference. Direct injection spiking can be used in field samples if no bias has been observed between this technique and dynamic spiking techniques. Several unused tubes, to serve as blanks, and several spiked tubes containing appropriate amounts of the expected compounds should be included with the samples as a check on storage, transportation factors and analytical procedures. Widely varying recoveries will signal difficulties and the necessity to reevaluate the data but should not be used for correcting the results. sampling pump One of the largest errors which can he introduced into a validated method is inaccurate sample volume measurement. The flow rate of a pump can he measured within 5%; however, the flow rate may change as the batteries are depleted. Digitized pumps will measure an accurate volume but may introduce a bias since the strokes per minute may decrease. This effect communications The development, sampling and analytical sections of a survey are strongly interdependent and it is important that the specialists working on each section have a basic understanding of the total effort. Good communications before, during, and after are essential factors in a successful validation and sampling project. becomes more important with long sampling periods. A periodic check of the performance curve of each battery operated pump should be references , made. Sampling equipment is being continually improved, and good accuracy can be obtained 1. Linch, B. E.: Quality Control for Sampling and Laboratory Analysis. The Industrial Environment Its evaluation and Control. NIOSH, Government with careful calibration and use. Printing. Washington, D.C. 2. Documentation of NIOSH Validation Tests. field testing Although many of the collection and recovery parameters can be tested in the laboratory, it is not often possible or feasible to simulate actual NIOSH Contract No. CDC-99-74-45. 3. MSA Research Corporation: Package Sorption Device System Study. EPA Contract EHSD 71-2, National Technical Information Service. PB-221 138. field sampling conditions. Judicious field testing during method development and after a successful validation will either uncover unpredictable circumstances or increase confidence. 4. Jonas. t_ A. and J. A. flehrmenn: Predictive Equations in Gas Adsorption Kinetics. Carbon 11:59 (1973). 5. Jonas. L. A. and J. A. Rehrmann: The Rate of Gas Adsorption by Activated Carbon. Carbon 12:95 (19741. , One technique used in field testing is on-site 6. Nelson, G. O. and C. A. Harder Respirator comparison to an established (but possibly inconvenient) alternate method. The alternate Cartridge Efficiency Studies: VI; Effect of Concentration. Am. Ind. Hyg. Assoc. J. 37:205 (1976). procedure may be a time averaging technique 7. Fraust, C. L and F. R. Hermann: The Adsorption of . such as a portable infrared, or use of periodic Aliphatic Acetate Vapors onto Activated Carbon. short term or instantaneous grab samples. Am.-Ind. Hyg. Assoc. J. 30:494 (1969)- 8. Saltzman. B.E.: BasicTheoryofGas IndicatorTube | Field spiking is also a useful technique where a Calibration. Am..Ind. Hyg. Assoc. J. 23\\ 12 (1962). duplicate sampling system is set up in an on-site 9. Wood, G. O. and R. G. Anders n: Personal area. One of the collection tubes is then spiked Sampling for Vapors of Aniline Compounds. Am. with an appropriate amount of the test compound. Identical treatment and analysis of the duplicate tubes will indicate the unknown concentration in the area as well as the recovery Ind. Hyg. Assoc. J. 36:638(1976). 10. Novak, J., V. Vasak and J. Janak: Chromato graphic Method for the Concentration of Trace Impurities in theAtmosphere and other gases. Anal. Chem. 35:735 (1963). i . 360 Am. M. H/l Assoc. I (39) My. 1978 OL1 7742 y:T~ 11. Pesek. J. J., and J. E. Daniels: Investigation of the Retention Mechanism of Chemically Bonded Stationary Phases in Gas Chromatography. J. Chrom. Sci. 14:288(1976). 12. Severs. L. W. and L. K. Skory: Personnel Sampling and Analysis of Vinyl Chloride. Am. tnd. Hyg. Assoc. J. 36:669 (1975). 13. Hill. Jr..'R. H.. C. S. McCammon. A. T. Saalwaechter. A. W. Teass and IN. J. Woodfin: Gas-Chromatographic Determination of Vinyl Chloride in Air.Samples Collected on Charcoal. Anal. Chem. 43:1395 (1976). 14. Teass. A, W.. C. S. McCammon. C. P. Roper and A. T. Saalwaechter: Sampling Organic Vapors Using Activated Charcoal. 2ndN10SHSolidSorbent Roundtable. Cincinnati. Ohio. December 1973. HEW Publication No. (NIOSH) 76-193 (1976). 15. Ballou. E. V.: An Overview of the Use of Solid Sorbents for Air Analysis. 2nd NIOSH SolidSorbent Roundtable. Cincinnati, Ohio. December 1973. HEW Publication No(NIOSH) 76-193 (1976). 16. Fraust. C. L and F. R. Hermann: Charcoal Sampling Tubes For Organic Vapor Analysis By Gas Chromatography. Am. tnd. Hyg. Assoc. J.. 27:68 (1966). 17. Nelson. G. O.. A. N. Correia and C. A. Harder: Respirator Cartridge Efficiency Studies: W; Effect of Relative Humidity and Temperature. Am. tnd. Hyg. Assoc. J. 37:280 (1976). 18. Dommer, R. A. and R. G. Matcher: Phase Equilibrium Method for Determination of Desorption Efficiencies. Am. tnd. Hyg. Assoc. J. 39:240(1978). 19. Irwin. M. M. and W. Kracht: Unpublished Data, The Dow Chemical Company, Midland, Michigan. 20. Melcher. R. G.: Unpublished Data, The Dow Chemical Company, Midland, Michigan. 21. Kerr. M. L: Unpublished Data. The Dow Chemical Company. Midland Michigan. 22. Saltzman. B. E.: Preparation of Known Concentrations of Air Contaminants. The Industrial Environment-Its Evaluation and Control. NIOSH, Government Printing, Washington. D.C. 23. Whitman. N. E. and A. E. Johnston: Sampling and Analysis of Aromatic Hydrocarbon Vapors in Air: A Gas-Liquid Chromatographic Method. Am. tnd. Hyg. Assoc. J. 25:464 (1964). 24. Severs. L. W., R. G. Melcher and M. J. Kocsis: Dynamic U-Tube System for Solid Adsorbent Air Sampling Method Development. Am. tnd. Hyg. Assoc. J. In Press. 25. Melcher. R. G., L W. Severs. M. J. Kocsis and W. L. Garner; Specialized Solid Sorbents for Sampling Organic Compound of High and LowVolatility in Air. Presented at 172nd ACS National Meeting. San Francisco (1976). 26. Kagel, R. 0.: Analytical Problems in Effluent Analysis. Presented at Manufacturing Chemist Association Symposium on the Approaching Toxic ra(1976). 27. American Conference of Governmental Industrial Hygienists: TLVs. Threshold Limit Values for Chemical Substances in Workroom Air Adopted by ACGIH for 1976. 28. NIOSH Analytical Methods for Set A-J Standards Completion Program. National Technical Information Service, Springfield, Virginia. 29. NIOSH Contract HSM-99-72-98, Scott Research Laboratories, Inc.: Collaborative Testing ofActivated Charcoal Sampling Tubes for Seven Organic Compounds HEW Publication (NIOSH) 76-173 (1976). Accepted August 12,1977 Comments now wanted . . . revision on occupational exposure to ozone standard now in progress The ASTM task force on Occupational Exposure to Ozone is revising the ASTM standard practice (E591) for Safety and Health Requirements Relating to Occupational Exposure toOzone.The task force is under the aegis of Committee E-34 on Occupational Health and Safety Aspects of Materials, Physical and Biological Agents. These voluntary consensus guidelines establish recommendations in industrial hygiene, industrial medicine, safety engineering, engineering controls and monitoring, with appropriate toxicological and epidemiological rationales, relating to occupational exposure to ozone. The current revision to the standard incorporates significant changes in the appendix (rationale) detailing Biological Effects of Exposure to Ozone on Humans. Due to the extensive nature of this standards development activity all^reas of expertise are encouraged to participate. Upon request, therefore, a copy of the standard will be provided for comment. The current revision will be discussed at the upcoming meeting of E-34, May 21 -23,1978, Radisson Muehtebach, Kansas City, Missouri. For further information and/or a copy of the working draft standard, contact the staff manager, Donald A. Tobias, ASTM, 1916 Race St, Phila., PA 19103(215)299-5546. American Industrial Hygiene Association JOURNAL (35) S/78 OLI 7743 Ml