METHOD TO-1 Revision 1.0April, 1984

METHOD FOR THE DETERMINATION OF VOLATILE ORGANIC COMPOUNDSIN AMBIENT AIR USING TENAX® ADSORPTION ANDGAS CHROMATOGRAPHY/MASS SPECTROMETRY (GC/MS)

1. Scope

1.1 The document describes a generalized protocol forcollection and determination of certain volatile organiccompounds which can be captured on Tenax® GC (poly(2,6-Diphenyl phenylene oxide)) and determined by thermaldesorption GC/MS techniques. Specific approaches usingthese techniques are described in the literature (1-3).

1.2 This protocol is designed to allow some flexibility inorder to accommodate procedures currently in use.However, such flexibility also results in placement ofconsiderable responsibility with the user to documentthat such procedures give acceptable results (i.e.,documentation of method performance within eachlaboratory situation is required). Types ofdocumentation required are described elsewhere in thismethod.

1.3 Compounds which can be determined by this method arenonpolar organics having boiling points in the range ofapproximately 80E - 200EC. However, not all compoundsfalling into this category can be determined. Table 1gives a listing of compounds for which the method hasbeen used. Other compounds may yield satisfactoryresults but validation by the individual user isrequired.

2. Applicable Documents

2.1 ASTM Standards:

D1356 Definitions of Terms Related to AtmosphericSampling and Analysis.

E355 Recommended Practice for Gas ChromatographyTerms and Relationships.

2.2 Other documents:

Existing procedures (1-3).

U. S. EPA Technical Assistance Document (4).

3. Summary of Protocol

3.1 Ambient air is drawn through a cartridge containing -1-2grams of Tenax and certain volatile organic compounds aretrapped on the resin while highly volatile organiccompounds and most inorganic atmospheric constituentspass through the cartridge. The cartridge is thentransferred to the laboratory and analyzed.

3.2 For analysis the cartridge is placed in a heated chamberand purged with an inert gas. The inert gas transfersthe volatile organic compounds from the cartridge onto acold trap and subsequently onto the front of the GCcolumn which is held at low temperature (e.g., -70EC).the GC column temperature is then increased (temperatureprogrammed) and the components eluting from the columnare identified and quantified by mass spectrometry.Component identification is normally accomplished, usinga library search routine, on the basis of the GCretention time and mass spectral characteristics. Lesssophisticated detectors (e.g., electron capture or flameionization) may be used for certain applications buttheir suitability for a given application must beverified by the user.

3.3 Due to the complexity of ambient air samples only highresolution (i.e., capillary) GC techniques are consideredto be acceptable in this protocol.

4. Significance

4.1 Volatile organic compounds are emitted into theatmosphere from a variety of sources including industrialand commercial facilities, hazardous waste storagefacilities, etc. Many of these compounds are toxic;hence knowledge of the levels of such materials in theambient atmosphere is required in order to determinehuman health impacts.

4.2 Conventional air monitoring methods (e.g., for workspacemonitoring) have relied on carbon adsorption approacheswith subsequent solvent desorption. Such techniquesallow subsequent injection of only a small portion,typically 1-5% of the sample onto the GC system.However, typical ambient air concentrations of thesecompounds require a more sensitive approach. The thermal

desorption process, wherein the entire sample isintroduced into the analytical (GC/MS) system fulfillsthis need for enhanced sensitivity.

5. Definitions

Definitions used in this document and any user prepared SOPsshould be consistent with ASTM D1356(6). All abbreviationsand symbols are defined with this document at the point ofuse.

6. Interferences

6.1 Only compounds having a similar mass spectrum and GCretention time compared to the compound of interest willinterface in the method. The most commonly encounteredinterferences are structural isomers.

6.2 Contamination of the Tenax cartridge with the compound(s)of interest is a commonly encountered problem in themethod. The user must be extremely careful in thepreparation, storage, and handling of the cartridgesthroughout the entire sampling and analysis process tominimize this problem.

7. Apparatus

7.1 Gas Chromatograph/Mass Spectrometry system - should becapable of subambient temperature programming. Unit massresolution or better up to 800 amu. Capable of scanning30-400 amu region every 0.5-1 second. Equipped with datasystem for instrument control as well as dataacquisition, processing and storage.

7.2 Thermal Desorption Unit - Designed to accommodate Tenaxcartridges in use. See Figure 2a or b.

7.3 Sampling System - Capable of accurately and preciselydrawing an air flow of 10-500 ml/minute through the Tenaxcartridge. (See Figure 3a or b.)

7.4 Vacuum oven - connected to water aspirator vacuum supply.

7.5 Stopwatch.

7.6 Pyrex disks - for drying Tenax.

7.7 Glass jar - Capped with Teflon-lined screw cap. Forstorage of purified Tenax.

7.8 Powder funnel - for delivery of Tenax into cartridges.

7.9 Culture tubes - to hold individual glass Tenaxcartridges.

7.10 Friction top can (paint can) - to hold clean Tenaxcartridges.

7.11 Filter holder - stainless steel or aluminum (toaccommodate 1 inch diameter filter). Other sizes may beused if desired. (optional)

7.12 Thermometer - to record ambient temperature.

7.13 Barometer (optional).

7.14 Dilution bottle - Two-liter with septum cap for standardspreparation.

7.15 Teflon stirbar - 1 inch long.

7.16 Gas-tight glass syringes with stainless steel needles -10-500 µ1 for standard injection onto GC/MS system.

7.17 Liquid microliter syringes - 5.50 µL for injecting neatliquid standards into dilution bottle.

7.18 Oven - 60 + 5EC for equilibrating dilution flasks.

7.19 Magnetic stirrer.

7.20 Heating mantel.

7.21 Variac

7.22 Soxhlet extraction apparatus and glass thimbles - forpurifying Tenax.

7.23 Infrared lamp - for drying Tenax.

7.24 GC column - SE-30 or alternative coating, glass capillaryor fused silica.

7.25 Psychrometer - to determine ambient relative humidity.(optional)

8. Reagents and Materials

8.1 Empty Tenax cartridges - glass or stainless steel (seeFigure 1a or b).

8.2 Tenax 60/80 mesh (2,6-diphenylphenylene oxide polymer).

8.3 Glasswool - silanized.

8.4 Acetone - Pesticide quality or equivalent.

8.5 Methanol - Pesticide quality or equivalent.

8.6 Pentane - Pesticide quality or equivalent.

8.7 Helium - Ultra pure, compressed gas. (99.9999%)

8.8 Nitrogen - Ultra pure, compressed gas. (99.9999%)

8.9 Liquid nitrogen.

8.10 Polyester gloves - for handling glass Tenax cartridges.

8.11 Glass Fiber Filter - one inch diameter, to fit in filterholder. (optional)

8.12 Perfluorotributylamine (FC-43).

8.13 Chemical Standards - Neat compounds of interest. Highestpurity available.

8.14 Granular activated charcoal - for preventingcontamination of Tenax cartridges during storage.

9. Cartridge Construction and Preparation

9.1 Cartridge Design

9.1.1 Several cartridge designs have been reportedin the literature (1-3). The most common (1)is shown in Figure 1a. This design minimizescontact of the sample with metal surfaces,which can lead to decomposition in certaincases. However, a disadvantage of this designis the need to rigorously avoid contaminationof the outside portion of the cartridge sincethe entire surface is subjected to the purgegas stream during the desorption process.Clean polyester gloves must be worn at all

times when handling such cartridges andexposure of the open cartridge to ambient airmust be minimized.

9.1.2 A second common type of design (3) is shown inFigure 1b. While this design uses a metal(stainless steel) construction, it eliminatesthe need to avoid direct contact with theexterior surface since only the interior ofthe cartridge is purged.

9.1.3 The thermal desorption module and samplingsystem must be selected to be compatible withthe particular cartridge design chosen.Typical module designs are shown in Figure 2aand b. These designs are suitable for thecartridge designs shown in Figures 1a and b,respectively.

9.2 Tenax Purification

9.2.1 Prior to use the Tenax resin is subjected to aseries of solvent extraction and thermaltreatment steps. The operation should beconducted in an area where levels of volatileorganic compounds (other than the extractionsolvents used) are minimized.

9.2.2 All glassware used in Tenax purification aswell as cartridge materials should bethoroughly cleaned by water rinsing followedby an acetone rinse and dried in an oven at250EC.

9.2.3 Bulk Tenax is placed in a glass extractionthimble and held in place with a plug of cleanglasswool. The resin is then placed in thesoxhlet extraction apparatus and extractedsequentially with methanol and then pentanefor 16-24 hours (each solvent) atapproximately 6 cycles/hour. Glasswool forcartridge preparation should be cleaned in thesame manner as Tenax.

9.2.4 The extracted Tenax is immediately placed inan open glass dish and heated under aninfrared lamp for two hours in a hood. Caremust be exercised to avoid over heating of theTenax by the infrared lamp. The Tenax is thenplaced in a vacuum oven (evacuated using a

water aspirator) without heating for one hour.An inert gas (helium or nitrogen) purge of 2-3ml/minute is used to aid in the removal ofsolvent vapors. The oven temperature is thenincreased to 110EC, maintaining inert gas flowand held for one hour. The oven temperaturecontrol is then shut off and the oven isallowed to cool to room temperature. Prior toopening the oven, the oven is slightlypressurized with nitrogen to preventcontamination with ambient air. The Tenax isremoved from the oven and sieved through a40/60 mesh sieve (acetone rinsed and ovendried) into a clean glass vessel. If theTenax is not to be used immediately forcartridge preparation it should be stored in aclean glass jar having a Teflon-lined screwcap and placed in a desiccator.

9.3 Cartridge Preparation and Pretreatment

9.3.1 All cartridge materials are pre-cleaned asdescribed in Section 9.2.2. If the glasscartridge design shown in Figure 1a isemployed all handling should be conductedwearing polyester gloves.

9.3.2 The cartridge is packed by placing a 0.5-lcmglasswool plug in the base of the cartridgeand then filling the cartridge to withinapproximately 1 cm of the top. A 0.5-1cmglasswool plug is placed in the top of thecartridge.

9.3.3 The cartridges are then thermally conditionedby heating for four hours at 270EC under aninert gas (helium) purge (100 - 200 ml/min).

9.3.4 After the four hour heating period thecartridges are allowed to cool. Cartridges ofthe type shown in Figure 1a are immediatelyplaced (without cooling) in clean culturetubes having Teflon-lined screw caps with aglasswool cushion at both the top and thebottom. Each tube should be shaken to ensurethat the cartridge is held firmly in place.Cartridges of the type shown in Figure 1b areallowed to cool to room temperature underinert gas purge and are then closed withstainless steel plugs.

VMAX'VbxW

1.5

9.3.5 The cartridges are labeled and placed in atightly sealed metal can (e.g., paint can orsimilar friction top container). Forcartridges of the type shown in Figure 1a theculture tube, not the cartridge, is labeled.

9.3.6 Cartridges should be used for sampling within2 weeks after preparation and analyzed withintwo weeks after sampling. If possible thecartridges should be stored at -20EC in aclean freezer (i.e., no solvent extracts orother sources of volatile organics containedin the freezer).

10. Sampling

10.1 Flow Rate and Total Volume Selection

10.1.1 Each compound has a characteristic retentionvolume (liters of air per gram of adsorbent)which must not be exceeded. Since theretention volume is a function of temperature,and possibly other sampling variables, onemust include an adequate margin of safety toensure good collection efficiency. Someconsiderations and guidance in this regardareprovided in a recent report (5).Approximate breakthrough volumes at 38EC(100EF) in liters/gram of Tenax are providedin Table 1. These retention volume data aresupplied only as rough guidance and aresubject to considerable variability, dependingon cartridge design as well as samplingparameters and atmospheric conditions.

10.1.2 To calculate the maximum total volume of airwhich can be sampled use the followingequation:

where

V is the calculated maximum total volume inMAXliters.

V is the breakthrough volume for the leastbretained compound of interest (Table 1)in liters per gram of Tenax.

W is the weight of Tenax in the cartridge,in grams.

QMAX'VMAXt

x1000

B'QMAXBr 2

1.5 is a dimensionless safety factor to allowfor variability in atmospheric conditions.This factor is appropriate for temperatures inthe range of 25-30EC. If higher temperaturesare encountered the factor should be increased(i.e. maximum total volume decreased).

10.1.3 To calculate maximum flow rate use thefollowing equation:

where

Q is the calculated maximum flow rate inMAXmilliliters per minute.

t is the desired sampling time in minutes.Times greater than 24 hours (1440minutes) generally are unsuitable becausethe flow rate required is too low to beaccurately maintained.

10.1.4 The maximum flow rate Q should yield aMAXlinear flow velocity of 50-500 cm/minute.Calculate the linear velocity corresponding tothe maximum flow rate using the followingequation:

where

B is the calculated linear flow velocity incentimeters per minute.

r is the internal radius of the cartridgein centimeters.

If B is greater then 500 centimeters perminute either the total sample flow rate (V )MAXshould be reduced or the sample flow rate(Q ) should be reduced by increasing theMAXcollection time. If B is less then 50centimeters per minute the sampling rate (Q )MAXshould be increased by reducing the samplingtime. The total sample value (V ) cannot beMAXincreased due to component breakthrough.

10.1.5 The flow rate calculated as described abovedefines the maximum flow rate allowed. Ingeneral, one should collect additional samplesin parallel, for the same time period but atlower flow rates. This practice yields ameasure of quality control and is furtherdiscussed in the literature (5). In general,flow rates 2 to 4 fold lower than the maximumflow rate should be employed for the parallelsamples. In all cases a constant flow rateshould be achieved for each cartridge sinceaccurate integration of the analyteconcentration requires that the flow beconstant over the sampling period.

10.2 Sample Collection

10.2.1 Collection of an accurately known volume ofair is critical to the accuracy of theresults. For this reason the use of mass flowcontrollers, rather than conventional needlevalves or orifices is highly recommended,especially at low flow velocities (e.g., lessthan 100 milliliters/minute). Figure 3aillustrates a sampling system utilizing massflow controllers. This system readily allowsfor collection of parallel samples. Figure 3bshows a commercially available system based onneedle valve flow controllers.

10.2.2 Prior to sample collection insure that thesampling flow rate has been calibrated over arange including the rate to be used forsampling, with a "dummy" Tenax cartridge inplace. Generally calibration is accomplishedusing a soap bubble flow meter or calibratedwet test meter. The flow calibration deviceis connected to the flow exit, assuming theentire flow system is sealed. ASTM MethodD3686 describes an appropriate calibrationscheme, not requiring a sealed flow systemdownstream of the pump.

10.2.3 The flow rate should be checked before andafter each sample collection. If the samplinginterval exceeds four hours the flow rateshould be checked at an intermediate pointduring sampling as well. In general, arotameter should be included, as shown inFigure 3b, to allow observation of thesampling flow rate without disrupting thesampling process.

10.2.4 To collect an air sample the cartridges areremoved from the sealed container just priorto initiation of the collection process. Ifglass cartridges (Figure 1a) are employed theymust be handled only with polyester gloves andshould not contact any other surfaces.

10.2.5 A particulate filter and holder are placed onthe inlet to the cartridges and the exit endof the cartridge is connected to the samplingapparatus. In many sampling situations theuse of a filter is not necessary if only thetotal concentration of a component is desired.Glass cartridges of the type shown in Figure1a are connected using teflon ferrules andSwagelok (stainless steel or teflon) fittings.Start the pump and record the followingparameters on an appropriate data sheet(Figure 4): data, sampling location, time,ambient temperature, barometric pressure,relative humidity, dry gas meter reading (ifapplicable), flow rate, rotameter reading (ifapplicable), cartridge number and dry gasmeter serial number.

10.2.6 Allow the sampler to operate for the desiredtime, periodically recording the variableslisted above. Check flow rate at the midpointof the sampling interval if longer than fourhours. At the end of the sampling periodrecord the parameters listed in 10.2.5 andcheck the flow rate and record the value. Ifthe flows at the beginning and end of thesampling period differ by more than 10% thecartridge should be marked as suspect.

10.2.7 Remove the cartridges (one at a time) andplace in the original container (use glovesfor glass cartridges). Seal the cartridges orculture tubes in the friction-top cancontaining a layer of charcoal and package forimmediate shipment to the laboratory foranalysis. Store cartridges at reducedtemperature (e.g., -20EC) before analysis ifpossible to maximize storage stability.

QA'Q1%Q

2%...QNN

Vm'T×QA1000

Vs'Vm×PA760

×298

273%tA

10.2.8 Calculate and record the average sample ratefor each cartridge according to the followingequation:

where

Q = Average flow rate in ml/minute.AQ , Q , ....Q = Flow rates determined at1 2 Nbeginning, end, and intermediate points duringsampling.N = Number of points averaged.

10.2.9 Calculate and record the total volumetric flowfor each cartridge using the followingequation:

where

V = Total volume sampled in liters atmmeasured temperature and pressure.

T = Stop time.2T = Start time.1T = Sampling time = T = T , minutes2 1

10.2.10 The total volume (V ) at standard conditions,s25EC and 760 mmHg, is calculated from thefollowing equation:

where

P = Average barometric pressure, mmHgAt = Average ambient temperature, EC.A

11. GC/MS Analysis

11.1 Instrument Set-up

11.1.1 Considerable variation from one laboratory toanother is expected in terms of instrumentconfiguration. Therefore each laboratory mustbe responsible for verifying that theirparticular system yields satisfactory results.Section 14 discusses specific performancecriteria which should be met.

11.1.2 A block diagram of the typical GC/MS systemrequired for analysis of Tenax cartridges isdepicted in Figure 5. The operation of suchdevices is described in 11.2.4. The thermaldesorption module must be designed toaccommodate the particular cartridgeconfiguration. Exposure of the sample tometal surfaces should be minimized and onlystainless steel, or nickel metal surfacesshould be employed. The volume of tubing andfittings leading from the cartridge to the GCcolumn must be minimized and all areas must bewell-swept by helium carrier gas.

11.1.3 The GC column inlet should be capable of beingcooled to -70EC and subsequently increasedrapidly to approximately 30EC. This can bemost readily accomplished using a GC equippedwith subambient cooling capability (liquidnitrogen) although other approaches such asmanually cooling the inlet of the column inliquid nitrogen may be acceptable.

11.1.4 The specific GC column and temperature programemployed will be dependent on the specificcompounds of interest. Appropriate conditionsare described in the literature (1-3). Ingeneral a nonpolar stationary phase (e.g., SE-30, OV-1) temperature programmed from 30EC to200EC at 8E/minute will be suitable. Fusedsilica bonded phase columns are preferable toglass columns since they are more rugged andcan be inserted directly into the MS ionsource, thereby eliminating the need for aGC/MS transfer line.

11.1.5 Capillary column dimensions of 0.3 mm ID and50 meters long are generally appropriatealthough shorter lengths may be sufficient inmany cases.

11.1.6 Prior to instrument calibration or sampleanalysis the GC/MS system is assembled asshown in Figure 5. Helium purge flows(through the cartridge) and carrier flow areset at approximately 10 ml/minute and 1-2ml/minute respectively. If applicable, theinjector sweep flow is set at 2-4 ml/minute.

11.1.7 Once the column and other system componentsare assembled and the various flowsestablished the column temperature isincreased to 250EC for approximately fourhours (or overnight if desired) to conditionthe column.

11.1.8 The MS and data system are set according tothe manufacturer's instructions. Electronimpact ionization (70eV) and an electronmultiplier gain of approximately 5 x 10 should4

be employed. Once the entire GC/MS system hasbeen setup the system is calibrated asdescribed in Section 11.2. The user shouldprepare a detailed standard operatingprocedure (SOP) describing this process forthe particular instrument being used.

11.2 Instrument Calibration

11.2.1 Tuning and mass standardization of the MSsystem is performed according tomanufacturer's instructions and relevantinformation from the user prepared SOP.Perfluorotributylamine should generally beemployed for this purpose. The material isintroduced directly into the ion source thougha molecular leak. The instrumental parameters(e.g., lens voltages, resolution, etc.) shouldbe adjusted to give the relative ionabundances shown in Table 2 as well asacceptable resolution and peak shape. Ifthese approximate relative abundances cannotbe achieved, the ion source may requirecleaning according to manufacturer'sinstructions. In the event that the user'sinstrument cannot achieve these relative ionabundances, but is otherwise operatingproperly, the user may adopt another set ofrelative abundances as performance criteria.However, these alternate values must berepeatable on a day-to-day basis.

11.2.2 After the mass standardization and tuningprocess has been completed and the appropriatevalues entered into the data system the usershould then calibrate the entire system byintroducing known quantities of the standardcomponents of interest into the system. Threealternate procedures may be employed for thecalibration process including 1) directsyringe injection of dilute vapor phasestandards, prepared in a dilution bottle, ontothe GC column, 2) injection of dilute vaporphase standards into a carrier gas streamdirected through the Tenax cartridge, and 3)introduction of permeation or diffusion tubestandards onto a Tenax cartridge. Thestandards preparation procedures for each ofthese approaches are described in Section 13.The following paragraphs describe theinstrument calibration process for each ofthese approaches.

11.2.3 If the instrument is to be calibrated bydirect injection of a gaseous standard, astandard is prepared in a dilution bottle asdescribed in Section 13.1. The GC column iscooled to -70EC (or, alternately, a portion ofthe column inlet is manually cooled withliquid nitrogen). The MS and data system isset up for acquisition as described in therelevant user SOP. The ionization filamentshould be turned off during the initial 2-3minutes of the run to allow oxygen and otherhighly volatile components to elute. Anappropriate volume (less than 1 ml) of thegaseous standard is injected onto the GCsystem using an accurately calibrated gastight syringe. The system clock is startedand the column is maintained at -70EC (orliquid nitrogen inlet cooling) for 2 minutes.The column temperature is rapidly increased tothe desired initial temperature (e.g., 30EC).The temperature program is started at aconsistent time (e.g., four minutes) afterinjection. Simultaneously the ionizationfilament is turned on and data acquisition isinitiated. After the last component ofinterest has eluted acquisition is terminatedand the data is processed as described inSection 11.2.5. The standard injectionprocess is repeated using different standardvolumes as desired.

11.2.4 If the system is to be calibrated by analysisof spiked Tenax cartridges a set of cartridgesis prepared as described in Sections 13.2 or13.3. Prior to analysis the cartridges arestored as described in Section 9.3. If glasscartridges (Figure 1a) are employed care mustbe taken to avoid direct contact, as describedearlier. The GC column is cooled to -70EC,the collection loop is immersed in liquidnitrogen and the desorption module ismaintained at 250EC. The inlet valve isplaced in the desorb mode and the standardcartridge is placed in the desorption module,making certain that no leakage or purge gasoccurs. The cartridge is purged for 10minutes and then the inlet valve is placed inthe inject mode and the liquid nitrogen sourceremoved from the collection trap. The GCcolumn is maintained at -70EC for two minutesand subsequent steps are as described in11.2.3. After the process is complete thecartridge is removed from the desorptionmodule and stored for subsequent use asdescribed in Section 9.3.

11.2.5 Data processing for instrument calibrationinvolves determining retention times, andintegrated characteristic ion intensities foreach of the compounds of interest. Inaddition, for at least one chromatographicrun, the individual mass spectra should beinspected and compared to reference spectra toensure proper instrumental performance. Sincethe steps involved in data processing arehighly instrument specific, the user shouldprepare a SOP describing the process forindividual use. Overall performance criteriafor instrument calibration are provided inSection 14. If these criteria are notachieved the user should refine theinstrumental parameters and/or operatingprocedures to meet these criteria.

11.3 Sample Analysis

11.3.1 The sample analysis process is identical tothat described in Section 11.2.4 for theanalysis of standard Tenax cartridges.

11.3.2 Data processing for sample data generallyinvolves 1) qualitatively determining thepresence or absence of each component ofinterest on the basis of a set ofcharacteristic ions and the retention timeusing a reverse-search software routine, 2)quantification of each identified component byintegrating the intensity of a characteristicion and comparing the value to that of thecalibration standard, and 3) tentativeidentification of other components observedusing a forward (library) search softwareroutine. As for other user specificprocesses, a SOP should be prepared describingthe specific operations for each individuallaboratory.

12. Calculations

12.1 Calibration Response Factors

12.1.1 Data from calibration standards is used tocalculate a response factor for each componentof interest. Ideally the process involvesanalysis of at least three calibration levelsof each component during a given day anddetermination of the response factor(area/nanogram injected) from the linear leastsquares fit of a plot of nanograms injectedversus area (for the characteristic ion). Ingeneral quantities of component greater than1000 nanograms should not be injected becauseof column overloading and/or MS responsenonlinearity.

12.1.2 In practice the daily routine may not alwaysallow analysis of three such calibrationstandards. In this situation calibration datafrom consecutive days may be pooled to yield aresponse factor, provided that analysis ofreplicate standards of the same concentrationare shown to agree within 20% on theconsecutive days. One standard concentration,near the midpoint of the analytical range ofinterest, should be chosen for injection everyday to determine day-to-day responsereproducibility.

Y'A%BX%CX 2

YA'A%BXA%CXA

CA'XAVS

12.1.3 If substantial nonlinearity is present in thecalibration curve a nonlinear least squaresfit (e.g., quadratic) should be employed.This process involves fitting the data to thefollowing equation:

where

Y = peak areaX = quantity of component, nanogramsA, B, and C are coefficients in the equation

12.2 Analyte Concentrations

12.2.1 Analyte quantities on a sample cartridge arecalculated from the following equation:

where

Y is the area of the analyte characteristic ionAfor the sample cartridge.

X is the calculated quantity of analyte on theAsample cartridge, in nanograms.

A, B, and C are the coefficients calculated fromthe calibration curve described in Section 12.1.3.

12.2.2 If instrumental response is essentially linearover the concentration range of interest alinear equation (C=O in the equation above)can be employed.

12.2.3 Concentration of analyte in the original airsample is calculated from the followingequation:

where

C is the calculated concentration of analyte inAnanograms per liter.

V and X are as previously defined in SectionS A10.2.10 and 12.2.1, respectively.

WT'WIVI×VB

VT'WTd

13. Standard Preparation

13.1 Direct Injection

13.1.1 This process involves preparation of adilution bottle containing the desiredconcentrations of compounds of interest fordirect injection onto the GC/MS system.

13.1.2 Fifteen three-millimeter diameter glass beadsand a one-inch Teflon stirbar are placed in aclean two-liter glass septum capped bottle andthe exact volume is determined by weighing thebottle before and after filling with deionizedwater. The bottle is then rinsed with acetoneand dried at 200EC.

13.1.3 The amount of each standard to be injectedinto the vessel is calculated from the desiredinjection quantity and volume using thefollowing equation:

where

W is the total quantity of analyte to beTinjected into the bottle in milligrams

W is the desired weight of analyte to beIinjected onto the GC/MS system or spikedcartridge in nanograms

V is the desired GC/MS or cartridge injectionIvolume (should not exceed 500) in microliters.

V is total volume of dilution bottle determinedBin 13.1.1, in liters.

13.1.4 The volume of the neat standard to be injectedinto the dilution bottle is determined usingthe following equation:

where

V is the total volume of neat liquid to beTinjected in microliters.

d is the density of the neat standard in gramsper milliliter.

13.1.6 The bottle is placed in a 60EC oven for atleast 30 minutes prior to removal of a vaporphase standard.

13.1.7 To withdraw a standard for GC/MS injection thebottle is removed from the oven and stirredfor 10-15 seconds. A suitable gas-tightmicrober syringe, warmed to 60EC, is insertedthrough the septum cap and pumped three timesslowly. The appropriate volume of sample(approximately 25% larger than the desiredinjection volume) is drawn into the syringeand the volume is adjusted to the exact valuedesired and then immediately injected over a5-10 seconds period onto the GC/MS system asdescribed in Section 11.2.3.

13.2 Preparation of Spiked Cartridges by Vapor Phase Injection

13.2.1 This process involves preparation of adilution bottle containing the desiredconcentrations of the compound(s) of interestas described in 13.1 and injecting the desiredvolume of vapor into a flowing inert gasstream directed through a clean Tenaxcartridge.

13.2.2 A helium purge system is assembled wherein thehelium flow 20-30 mL/minute is passed througha stainless steel Tee fitted with a septuminjector. The clean Tenax cartridge isconnected downstream of the tee usingappropriate Swagelok fittings. Once thecartridge is placed in the flowing gas streamthe appropriate volume vapor standard, in thedilution bottle, is injected through theseptum as described in 13.1.6. The syringe isflushed several times by alternately fillingthe syringe with carrier gas and displacingthe contents into the flow stream, withoutremoving the syringe from the septum. Carrierflow is maintained through the cartridge forapproximately 5 minutes after injection.

13.3 Preparation of Spiked Traps Using Permeation or DiffusionTubes

13.3.1 A flowing stream of inert gas containing knownamounts of each compound of interest isgenerated according to ASTM Method D3609(6).

Note that a method of accuracy maintainingtemperature within + 0.1EC is required and thesystem generally must be equilibrated for atleast 48 hours before use.

13.3.2 An accurately known volume of the standard gasstream (usually 0.1-1 liter) is drawn througha clean Tenax cartridge using the samplingsystem described in Section 10.2.1, or asimilar system. However, if mass flowcontrollers are employed they must becalibrated for the carrier gas used in Section

13.3.1 (usually nitrogen). Use of air as the carriergas for permeation systems is not recommended,unless the compounds of interest are known tobe highly stable in air.

13.3.3 The spiked cartridges are then stored orimmediately analyzed as in Section 11.2.4.

14. Performance Criteria and Quality Assurance

This section summarizes quality assurance (QA) measures andprovides guidance concerning performance criteria which shouldbe achieved within each laboratory. In many cases thespecific QA procedures have been described within theappropriate section describing the particular activity (e.g.,parallel sampling).

14.1 Standard Operating Procedures (SOPs)

14.1.1 Each user should generate SOPs describing thefollowing activities as they are performed intheir laboratory:1) assembly, calibration, and operation of

the sampling system,2) preparation, handling and storage of

Tenax cartridges,3) assembly and operation of GC/MS system

including the thermal desorptionapparatus and data system, and

4) all aspects of data recording andprocessing.

14.1.2 SOPs should provide specific stepwiseinstructions and should be readily availableto, and understood by, the laboratorypersonnel conducting the work.

14.2 Tenax Cartridges Preparation

14.2.1 Each batch of Tenax cartridges prepared (asdescribed in Section 9) should be checked forcontamination by analyzing one cartridgeimmediately after preparation. While analysiscan be accomplished by GC/MS, manylaboratories may choose to use GC/FID due tologistical and cost considerations.

14.2.2. Analysis by GC/FID is accomplished asdescribed for GC/MS (Section 11) except foruse of FID detection.

14.2.3 While acceptance criteria can vary dependingon the components of interest, at a minimumthe clean cartridge should be demonstrated tocontain less than one fourth of the minimumlevel of interest for each component. Formost compounds the blank level should be lessthan 10 nanograms per cartridge in order to beacceptable. More rigid criteria may beadopted, if necessary, within a specificlaboratory. If a cartridge does not meetthese acceptance criteria the entire lotshould be rejected.

14.3 Sample Collection

14.3.1 During each sampling event at least one cleancartridge will accompany the samples to thefield and back to the laboratory, withoutbeing used for sampling, to serve as a fieldblank. The average amount of material foundon the field blank cartridge may be subtractedfrom the amount found on the actual samples.However, if the blank level is greater than25% of the sample amount, data for thatcomponent must be identified as suspect.

14.3.2 During each sampling event at least one set ofparallel samples (two or more samplescollected simultaneously) will be collected,preferably at different flow rates asdescribed in Section 10.1. If agreementbetween parallel samples is not generallywithin + 25% the user should collect parallelsamples on a much more frequent basis (perhapsfor all sampling points). If a trend of lowerapparent concentrations with increasing flow

rate is observed for a set of parallel samplesone should consider using a reduced flow rateand longer sampling interval if possible. Ifthis practice does not improve thereproducibility further evaluation of themethod performance for the compound ofinterest may be required.

14.3.3 Backup cartridges (two cartridges in series)should be collected with each sampling event.Backup cartridges should contain less than 20%of the amount of components of interest foundin the front cartridges, or be equivalent tothe blank cartridge level, whichever isgreater. The frequency of use of backupcartridges should be increased if increasedflow rate is shown to yield reduced componentlevels for parallel sampling. This practicewill help to identify problems arising frombreakthrough of the component of interestduring sampling.

14.4 GC/MS Analysis

14.4.1 Performance criteria for MS tuning and masscalibration have been discussed in Section11.2 and Table 2. Additional criteria may beused by the laboratory if desired. Thefollowing sections provide performanceguidance and suggested criteria fordetermining the acceptability of the GC/MSsystem.

14.4.2 Chromatographic efficiency should be evaluatedusing spiked Tenax cartridges since thispractice tests the entire system. In generala reference compound such as perfluorotolueneshould be spiked onto a cartridge at the 100nanogram level as described in Section 13.2 or13.3. The cartridge is then analyzed by GC/MSas described in Section 11.4. Theperfluorotoluene (or other reference compound)peak is then plotted on an expanded time scaleso that its width at 10% of the peak can becalculated, as shown in Figure 6. The widthof the peak at 10% height should not exceed 10seconds. More stringent criteria may berequired for certain applications. Theasymmetry factor (see Figure 6) should bebetween 0.8 and 2.0. The asymmetry factor for

DL'A%3.3S

any polar or reactive compounds should bedetermined using the process described above.If peaks are observed that exceed the peakwidth or asymmetry factor criteria above, oneshould inspect the entire system to determineif unswept zones or cold spots are present inany of the fittings and are necessary. Somelaboratories may choose to evaluate columnperformance separately by direct injection ofa test mixture onto the GC column. Suitableschemes for column evaluation have beenreported in the literature (7). Such schemescannot be conducted by placing the substancesonto Tenax because many of the compounds(e.g., acids, bases, alcohols) contained inthe test mix are not retained, or degrade, onTenax.

14.4.3 The system detection limit for each componentis calculated from the data obtained forcalibration standards. The detection limit isdefined as

where

DL is the calculated detection limit innanograms injected.

A is the intercept calculated in Section12.1.1 or 12.1.3.

S is the standard deviation of replicatedeterminations of the lowest levelstandard (at least three suchdeterminations are required).

In general the detection limit should be 20nanograms or less and for many applicationsdetection limits of 1-5 nanograms may berequired. The lowest level standard shouldyield a signal to noise ratio, from the totalion current response, of approximately 5.

14.4.4 The relative standard deviation for replicateanalyses of cartridges spiked at approximately10 times the detection limit should be 20% orless. Day to day relative standard deviationshould be 25% or less.

14.4.5 A useful performance evaluation step is theuse of an internal standard to track systemperformance. This is accomplished by spikingeach cartridge, including blank, sample, andcalibration cartridges with approximately 100nanograms of a compound not generally presentin ambient air (e.g., perfluorotoluene). Theintegrated ion intensity for this compoundhelps to identify problems with a specificsample. In general the user should calculatethe standard deviation of the internalstandard response for a given set of samplesanalyzed under identical tuning andcalibration conditions. Any sample giving avalue greater than + 2 standard deviationsfrom the mean (calculated excluding thatparticular sample) should be identified assuspect. Any marked change in internalstandard response may indicate a need forinstrument recalibration.

REFERENCES

1. Krost, K. J., Pellizzari, E. D., Walburn, S. G., and Hubbard,S. A., "Collection and Analysis of Hazardous OrganicEmissions", Analytical Chemistry, 54, 810-817, 1982.

2. Pellizzari, E. O. and Bunch, J. E., "Ambient Air CarcinogenicVapors-Improved Sampling and Analytical Techniques and FieldStudies", EPA-600/2-79-081, U.S. Environmental ProtectionAgency, Research Triangle Park, North Carolina, 1979.

3. Kebbekus, B. B. and Bozzelli, J. W., "Collection and Analysisof Selected Volatile Organic Compounds in Ambient Air", Proc.Air Pollution Control Assoc., Paper No. 82-65.2. Air Poll.Control Assoc., Pittsburgh, Pennsylvania, 1982.

4. Riggin, R. M., "Technical Assistance Document for Sampling andAnalysis of Toxic Organic Compounds in Ambient Air", EPA-600/4-83-027, U.S. Environmental Protection Agency, ResearchTriangle Park, North Carolina, 1983.

5. Walling, J. F., Berkley, R. E., Swanson, D. H., and Toth, F.J. "Sampling Air for Gaseous Organic Chemical-Applications toTenax", EPA-600/7-54-82-059, U.S. Environmental ProtectionAgency, Research Triangle Park, North Carolina, 1982.

6. Annual Book of ASTM Standards, Part 11.03, "AtmosphericAnalysis", American Society for Testing and Materials,Philadelphia, Pennsylvania.

7. Grob, K., Jr., Grob, G., and Grob, K., "ComprehensiveStandardized Quality Test for Glass Capillary Columns", J.Chromatog., 156, 1-20, 1978.

TABLE 1. RETENTION VOLUME ESTIMATES FOR COMPOUNDS ON TENAX

ESTIMATED RETENTION VOLUME ATCOMPOUND 100EF (38EC)-LITERS/GRAM

Benzene 19Toluene 97Ethyl Benzene 200Xylene(s) -200Cumene 440n-Heptane 20l-Heptene 40

Chloroform 8Carbon Tetrachloride 81,2-Dichloroethane 101,1,1-Trichloroethane 6Tetrachloroethylene 80Trichloroethylene 201,2-Dichloropropane 301,3-Dichloropropane 90Chlorobenzene 150Bromoform 100Ethylene Dibromide 60Bromobenzene 300

TABLE 2. SUGGESTED PERFORMANCE CRITERIA FOR RELATIVE ION ABUNDANCES FROM FC-43 MASS CALIBRATION

% RELATIVEM/E ABUNDANCE

51 1.8 + 0.5 69 100100 12.0 + 1.5119 12.0 + 1.5131 35.0 + 3.5169 3.0 + 0.4219 24.0 + 2.5264 3.7 + 0.4314 0.25 + 0.1

Tenax

~ 1.5 Grams (6 cm Bed Depth)

Glass Wool Plugs

(0.5 cm Long)

(a) Glass Cartridge

Glass Cartridge

(13.5 mm ODx

100 mm Long)

1/2" to

1/8"

Reducing

UnionGlass Wool

Plugs

(0.5 cm Long)

1/8" End Cap

Metal Cartridge

(12.7 mm OD x

100 mm Long)

Tenax

~ 1.5 Grams (7 cm Bed Depth)

(b) Metal Cartridge

1/2"

Swagelok

Fitting

FIGURE 1. TENAX CARTRIDGE DESIGNS

Teflon

Compression

Seal

Purge

Gas

Cavity for

Tenax

Cartridge

Latch for

Compression

Seal

Effluent to

6-Port Valve

To GC/MS

Vent

Freeze-Out

Loop

Liquid

Nitrogen

Coolant

Carrier

Gas

(a) Glass Cartridges (Compression Fit)

Purge

Gas

Swagelok

End Fittings

Heated Block

Tenax

Trap

Effluent to

6-Port Valve

To GC/MS

Vent

Liquid

Nitrogen

Coolant

Carrier

Gas

(b) Metal Cartridges (Swagelok Fittings)

Figure 2. Tenax Cartridge Desorption Modules

FIGURE 2. TENAX CARTRIDGE DESORPTION MODULES

Couplings

to Connect

Tenax

Cartridges

Mass Flow

ControllersOilless

Pump

Vent(a) Mass Flow Control

Vent

Dry

Test

Meter

Rotameter

Needle

Valve

Pump

Coupling to

Connect Tenax

Cartridge(b) Needle Valve ControlFigure 3. Typical Sampling System Configurations

FIGURE 3. TYPICAL SAMPLING SYSTEM CONFIGURATIONS

SAMPLING DATA SHEET(One Sample Per Data Sheet)

PROJECT: DATE(S) SAMPLED:

SITE: TIME PERIOD SAMPLED:

LOCATION: OPERATOR:

INSTRUMENT MODEL NO: CALIBRATED BY:

PUMP SERIAL NO:

SAMPLING DATA

Sample Number: Start Time: Stop Time: *Dry Gas* * Flow *Ambient*Barometric* * * Meter *Rotameter*Rate,*Q* Temp. *Pressure, *Relative * Time*Reading* Reading *ml/Min * EC * mmHg *Humidity,%* Comments))))))3)))))))3)))))))))3)))))))3)))))))3))))))))))3))))))))))3))))))))))1. * * * * * * *))))))3)))))))3)))))))))3)))))))3)))))))3))))))))))3))))))))))3))))))))))2. * * * * * * *))))))3)))))))3)))))))))3)))))))3)))))))3))))))))))3))))))))))3))))))))))3. * * * * * * *))))))3)))))))3)))))))))3)))))))3)))))))3))))))))))3))))))))))3))))))))))4. * * * * * * *))))))3)))))))3)))))))))3)))))))3)))))))3))))))))))3))))))))))3))))))))))N. * * * * * * *))))))2)))))))2)))))))))2)))))))2)))))))2))))))))))2))))))))))2))))))))))

Total Volume Data**

V = (Final - Initial) Dry Gas Meter Reading, or = Litersm

= Q + Q + Q ...Q x 1 = Liters1 2 3 N N 1000 x (Sampling Time in Minutes)

* Flowrate from rotameter or soap bubble calibrator (specify which).** Use data from dry gas meter if available.

FIGURE 4. EXAMPLE SAMPLING DATA SHEET

Purge

Gas

Thermal

Desorption

Chamber

6-Port High-Temperature

Valve

Heated

Blocks

Carrier

Gas Liquid

Nitrogen

Coolant

Freeze-Out Loop

Vent

Capillary

Gas

Chromatograph

Mass

Spectrometer

Data

System

Figure 5. Block Diagram of Analytical System

FIGURE 5. BLOCK DIAGRAM OF ANALYTICAL SYSTEM

E

A B C

D

Asymmetry Factor =

BC

AB

Example Calculation:

Peak Height = DE = 100 mm

10% Peak Height = BD = 10 mm

Peak Width at 10% Peak Height = AC = 23 mm

Therefore: Asymmetry Factor = 12

11

= 1.1

AB = 11 mm

BC = 12 mm

FIGURE 6. PEAK ASYMMETRY CALCULATION

METHOD TO-2Revision 1.0April, 1984

METHOD FOR THE DETERMINATION OF VOLATILE ORGANIC COMPOUNDS INAMBIENT AIR BY CARBON MOLECULAR SIEVE ADSORPTION AND GASCHROMATOGRAPHY/MASS SPECTROMETRY (GC/MS)

1. Scope

1.1 This document describes a procedure for collection anddetermination of selected volatile organic compoundswhich can be captured on carbon molecular sieve (CMS)adsorbents and determined by thermal desorption GC/MStechniques.

1.2 Compounds which can be determined by this method arenonpolar and nonreactive organics having boiling pointsin the range -15 to +120EC. However, not all compoundsmeeting these criteria can be determined. Compounds forwhich the performance of the method has been documentedare listed in Table 1. The method may be extended toother compounds but additional validation by the user isrequired. This method has been extensively used in asingle laboratory. Consequently, its generalapplicability has not been thoroughly documented.

2. Applicable Documents

2.1 ASTM StandardsD 1356 Definitions of Terms Related to AtmosphericSampling and Analysis.E 355 Recommended Practice for Gas Chromatography Termsand Relationships.

2.2 Other DocumentsAmbient Air Studies (1,2).U.S. EPA Technical AssistanceDocument (3).

3. Summary of Method

3.1 Ambient air is drawn through a cartridge containing -0.4of a carbon molecular sieve (CMS) adsorbent. Volatileorganic compounds are captured on the adsorbent whilemajor inorganic atmospheric constituents pass through (orare only partially retained). After sampling, thecartridge is returned to the laboratory for analysis.

3.2 Prior to analysis the cartridge is purged with 2-3 litersof pure, dry air (in the same direction as sample flow)to remove adsorbed moisture.

3.3 For analysis the cartridge is heated to 350E-400EC, underhelium purge and the desorbed organic compounds arecollected in a specially designed cryogenic trap. Thecollected organics are then flash evaporated onto acapillary column GC/MS system (held at -70EC). Theindividual components are identified and quantifiedduring a temperature programmed chromatographic run.

3.4 Due to the complexity of ambient air samples, only highresolution (capillary column) GC techniques areacceptable for most applications of the method.

4. Significance

4.1 Volatile organic compounds are emitted into theatmosphere from a variety of sources including industrialand commercial facilities, hazardous waste storage andtreatment facilities, etc. Many of these compounds aretoxic; hence knowledge of the concentration of suchmaterials in the ambient atmosphere is required in orderto determine human health impacts.

4.2 Traditionally air monitoring methods for volatile organiccompounds have relied on carbon adsorption followed bysolvent desorption and GC analysis. Unfortunately, suchmethods are not sufficiently sensitive for ambient airmonitoring, in most cases, because only a small portionof the sample is injected onto the GC system. Recentlyon-line thermal desorption methods, using organicpolymeric adsorbents such as Tenax® GC, have been usedfor ambient air monitoring. The current method uses CMSadsorbents (e.g. Spherocarb®) to capture highly volatileorganics (e.g., vinyl chloride) which are not collectedon Tenax®. The use of on-line thermal desorption GS/MSyields a sensitive, specific analysis procedure.

5. Definitions

Definitions used in this document and any user prepared SOPsshould be consistent with ASTM D1356 (4). All abbreviationsand symbols are defined with this document at the point ofuse.

6. Interferences

6.1 Only compounds having a mass spectrum and GC retentiontime similar to the compound of interest will interferein the method. The most commonly encounteredinterferences are structural isomers.

6.2 Contamination of the CMS cartridge with the compound(s)of interest can be a problem in the method. The usermust be careful in the preparation, storage, and handlingof the cartridges through the entire process to minimizecontamination.

7. Apparatus

7.1 Gas Chromatograph/Mass Spectrometry system - must becapable of subambient temperature programming. Unit massresolution to 800 amu. Capable of scanning 30-300 amuregion every 0.5-0.8 seconds. Equipped with data systemfor instrument control as well as data acquisition,processing and storage.

7.2 Thermal Desorption Injection Unit - Designed toaccommodate CMS cartridges in use (See Figure 3) andincluding cryogenic trap (Figure 5) and injection valve(Carle Model 5621 or equivalent).

7.3 Sampling System - Capable of accurately and preciselydrawing an air flow of 10-500 ml/minute through the CMScartridge. (See Figure 2a or b.)

7.4 Dewar flasks - 500 mL and 5 liter.

7.5 Stopwatches.

7.6 Various pressure regulators and valves - for connectingcompressed gas cylinders to GC/MS system.

7.7 Calibration gas - In aluminum cylinder. Prepared by useror vendor. For GC/MS calibration.

7.8 High pressure apparatus for preparing calibration gascylinders (if conducted by user). Alternatively, customprepared gas mixtures can be purchased from gas supplyvendors.

7.9 Friction top can (e.g. one-gallon paint-can) - With layerof activated charcoal to hold clean CMS cartridges.

7.10 Thermometer - to record ambient temperature.

7.11 Barometer (optional).

7.12 Dilution bottle - Two-liter with septum cap for standardpreparation.

7.13 Teflon stirbar - 1 inch long.

7.14 Gas tight syringes - 10-500 Fl for standard injectiononto GC/MS system and CMS cartridges.

7.15 Liquid microliter syringes - 5-50 FL for injecting neatliquid standards into dilution bottle.

7.16 Oven - 60 + 5EC for equilibrating dilution bottle.

7.17 Magnetic stirrer.

7.18 Variable voltage transformers - (120 V and 1000 VA) andelectrical connectors (or temperature controllers) toheat cartridge and cryogenic loop.

7.19 Digital pyrometer - 30 to 500EC range.

7.20 Soap bubble flow meter - 1, 10 and 100 mL calibrationpoints.

7.21 Copper tubing (1/8 inch) and fittings for gas inletlines.

7.22 GC column - SE-30 or alternative coating, glass capillaryor fused silica.

7.23 Psychrometer (optional).

7.24 Filter holder - stainless steel or aluminum (toaccommodate 1 inch diameter filter). Other sizes may beused if desired. (optional)

8. Reagents and Materials

8.1 Empty CMS cartridges - Nickel or stainless steel (SeeFigure 1).

8.2 CMS Adsorbent, 60/80 mesh-Spherocarb® from Analabs Inc.,or equivalent.

8.3 Glasswool - silanized.

8.4 Methylene chloride - pesticide quality, or equivalent.

8.5 Gas purifier cartridge for purge and GC carrier gascontaining charcoal, molecular sieves, and a dryingagent. Available from various chromatography supplyhouses.

8.6 Helium - Ultra pure, (99.9999%) compressed gas.

8.7 Nitrogen - Ultra pure, (99.9999%) compressed gas.

8.8 Liquid nitrogen or argon (50 liter dewar).

8.9 Compressed air, if required - for operation of GC ovendoor.

8.10 Perfluorotributylamine (FC-43) for GC/MS calibration.

8.11 Chemical Standards - Neat compounds of interest. Highestpurity available.

9. Cartridge Construction and Preparation

9.1 A suitable cartridge design is shown in Figure 1.Alternate designs have been reported (1) and areacceptable, provided the user documents theirperformance. The design shown in Figure 1 has a built-inheater assembly. Many users may choose to replace thisheater design with a suitable separate heating block oroven to simplify the cartridge design.

9.2 The cartridge is assembled as shown in Figure 1 usingstandard 0.25 inch O.D. tubing (stainless steel ornickel), 1/4 inch to 1/8 inch reducing unions, 1/8 inchnuts, ferrules, and endcaps. These parts are rinsed withmethylene chloride and heated at 250EC for 1 hour priorto assembly.

9.3 The thermocouple bead is fixed to the cartridge body, andinsulated with a layer of Teflon tape. The heater wire(constructed from a length of thermocouple wire) is woundaround the length of the cartridge and wrapped withTeflon tape to secure the wire in place. The cartridgeis then wrapped with woven silica fiber insulation (Zetexor equivalent). Finally the entire assembly is wrappedwith fiber glass tape.

9.4 After assembly one end of the cartridge is marked with aserial number to designate the cartridge inlet duringsample collection.

9.5 The cartridges are then packed with -0.4 grams of CMSadsorbent. Glass wool plugs (-0.5 inches long) areplaced at each end of the cartridge to hold the adsorbentfirmly in place. Care must be taken to insure that nostrands of glasswool extend outside the tubing, thuscausing leakage in the compression endfittings. Afterloading the endfittings (reducing unions and end caps)are tightened onto the cartridge.

9.6 The cartridges are conditioned for initial use by heatingat 400EC overnight (at least 16 hours) with a 100mL/minute purge of pure nitrogen. Reused cartridges needonly to be heated for 4 hours and should be reanalyzedbefore use to ensure complete desorption of impurities.

9.7 For cartridge conditioning ultra-pure nitrogen gas ispassed through a gas purifier to remove oxygen, moistureand organic contaminants. The nitrogen supply isconnected to the unmarked end of the cartridge and theflow adjusted to -50 mL/minute using a needle valve. Thegas flow from the inlet (marked) end of the cartridge isvented to the atmosphere.

9.8 The cartridge thermocouple lead is connected to apyrometer and the heater lead is connected to a variablevoltage transformer (Variac) set at 0 V. The voltage onthe Variac is increased to -15 V and adjusted over a 3-4minute period to stabilize the cartridge temperature at380-400EC.

9.9 After 10-16 hours of heating (for new cartridges) theVariac is turned off and the cartridge is allowed to coolto -30EC, under continuing nitrogen flow.

9.10 The exit end of the cartridge is capped and then theentire cartridge is removed from the flow line and theother endcap immediately installed. The cartridges arethen placed in a metal friction top (paint) cancontaining -2 inches of granulated activated charcoal (toprevent contamination of the cartridges during storage)in the bottom, beneath a retaining screen. Clean papertissues (e.g., Kimwipes) are placed in can to avoiddamage to the cartridges during shipment.

9.11 Cartridges are stored in the metal can at all timesexcept when in use. Adhesives initially present in thecartridge insulating materials are "burnt off" duringinitial conditioning. Therefore, unconditionedcartridges should not be placed in the metal can sincethey may contaminate the other cartridges.

QMAX'VMAXt

×1000

9.12 Cartridges are conditioned within two weeks of use. Ablank from each set of cartridges is analyzed prior touse in field sampling. If an acceptable blank level isachieved, that batch of cartridges (including thecartridge serving as the blank) can be used for fieldsampling.

10. Sampling

10.1 Flow Rate and Total Volume Selection

10.1.1 Each compound has a characteristic retentionvolume (liters of air per unit weight ofadsorbent). However, all of the compoundslisted in Table 1 have retention volumes (at37EC) in excess of 100 liters/cartridge (0.4gram CMS cartridge) except vinyl chloride forwhich the value is -30 liters/cartridge.Consequently, if vinyl chloride or similarlyvolatile compounds are of concern the maximumallowable sampling volume is approximately 20liters. If such highly volatile compounds arenot of concern, samples as large as 100 literscan be collected.

10.1.2 To calculate the maximum allowable samplingflow rate the following equation can be used:

whereQ is the calculated maximum sampling rateMAX

in mL/minute.t is the desired sampling time in minutes.V is the maximum allowable total volumeMAX

based on the discussion in 10.1.1.

10.1.3 For the cartridge design shown in Figure 1 QMAXshould be between 20 and 500 mL/minute. IfQ lies outside this range the sampling timeMAXor total sampling volume must be adjusted sothat this criterion is achieved.

10.1.4 The flow rate calculated in 10.1.3 defines themaximum allowable flow rate. In general, theuser should collect additional samples inparallel, at successive 2- to 4-fold lowerflow rates. This practice serves as a quality

control procedure to check on componentbreakthrough and related sampling andadsorption problems, and is further discussedin the literature (5).

10.2 Sample Collection

10.2.1 Collection of an accurately known volume ofair is critical to the accuracy of theresults. For this reason the use of mass flowcontrollers, rather than conventional needlevalves or orifices is highly recommended,especially at low flow rates (e.g., less than100 milliliters/minute). Figure 2aillustrates a sampling system based on massflow controllers which readily allows forcollection of parallel samples. Figure 2bshows a commercially available sampling systembased on needle valve flow controllers.

10.2.2 Prior to sample collection the sampling flowrate is calibrated near the value used forsampling, with a "dummy" CMS cartridge inplace. Generally calibration is accomplishedusing a soap bubble flow meter or calibratedwet test meter connected to the flow exit,assuming the entire flow system is sealed.ASTM Method D 3686 (4) describes anappropriate calibration scheme, not requiringa sealed flow system downstream of the pump.

10.2.3 The flow rate should be checked before andafter each sample collection. Ideally, arotometer or mass flow meter should beincluded in the sampling system to allowperiodic observation of the flow rate withoutdisrupting the sampling process.

10.2.4 To collect an air sample the cartridges areremoved from the sealed container just priorto initiation of the collection process.

10.2.5 The exit (unmarked) end of the cartridge isconnected to the sampling apparatus. Theendcap is left on the sample inlet and theentire system is leak checked by activatingthe sampling pump and observing that no flowis obtained over a 1 minute period. Thesampling pump is then shut off.

QA'Q1%Q

2%...QNN

10.2.6 The endcap is removed from the cartridge, aparticulate filter and holder are placed onthe inlet end of the cartridge, and thesampling pump is started. In many situationsa particulate filter is not necessary sincethe compounds of interest are in the vaporstate. However, if large amounts ofparticulate matter are encountered, the filtermay be useful to prevent contamination of thecartridge. The following parameters arerecorded on an appropriate data sheet (Figure4): date, sampling location, time, ambienttemperature, barometric pressure, relativehumidity, dry gas meter reading (ifapplicable), flow rate, rotometer reading (ifapplicable), cartridge number, pump, and drygas meter serial number.

10.2.7 The samples are collected for the desiredtime, periodically recording the variableslisted above. At the end of the samplingperiod the parameters listed in 10.2.6 arerecorded and the flow rate is checked. If theflows at the beginning and end of the samplingperiod differ by more than 10%, the cartridgeshould be marked as suspect.

10.2.8 The cartridges are removed (one at a time),the endcaps are replaced, and the cartridgesare placed into the original container. Thefriction top can is sealed and packaged forimmediate shipment to the analyticallaboratory.

10.2.9 The average sample rate is calculated andrecorded for each cartridge according to thefollowing equation:

whereQ = Average flow rate is ml/minuteAQ , Q ....Q = Flow rates determined at1 2 Nbeginning, end, and immediate points duringsampling.

N = Number of points averaged.

Vm'T×QA1000

Vs'Vm×Pa760

×298

273%ta

10.2.10 The total volumetric flow is obtained directlyfrom the dry gas meter or calculated andrecorded for each cartridge using thefollowing equation:

whereV = Total volume sampled in liters atm

measured temperature and pressure.T = Sampling time = T -T , minutes.2 1

10.2.11 The total volume sampled (V ) at standardsconditions, 760 mm Hg and 25EC, is calculatedfrom the following equation:

wherePa = Average barometric pressure, mm Hgta = Average ambient temperature, EC.

11. Sample Analysis

11.1 Sample Purging

11.1.1 Prior to analysis all samples are purged atroom temperature with pure, dry air ornitrogen to remove water vapor. Purging isaccomplished as described in 9.7 except thatthe gas flow is in the same direction assample flow (i.e. marked end of cartridge isconnected to the flow system).

11.1.2 The sample is purged at 500 mL/minute for 5minutes. After purging the endcaps areimmediately replaced. The cartridges arereturned to the metal can or analyzedimmediately.

11.1.3 If very humid air is being sampled the purgetime may be increased to more efficientlyremove water vapor. However, the sum ofsample volume and purge volume must be lessthan 75% of the retention volume for the mostvolatile component of interest.

11.2 GC/MS Setup

11.2.1 Considerable variation from one laboratory toanother is expected in terms of instrumentconfiguration. Therefore, each laboratorymust be responsible for verifying that theirparticular system yields satisfactory results.Section 14 discusses specific performancecriteria which should be met.

11.2.2 A block diagram of the analytical systemrequired for analysis of CMS cartridges isdepicted in Figure 3. The thermal desorptionsystem must be designed to accommodate theparticular cartridge configuration. For theCMS cartridge design shown in Figure 1, thecartridge heating is accomplished as describedin 9.8. The use of a desorption oven, inconjunction with a simpler cartridge design isalso acceptable. Exposure of the sample tometal surfaces should be minimized and onlystainless steel or nickel should be employed.The volume of tubing leading from thecartridge to the GC column must be minimizedand all areas must be well-swept by heliumcarrier gas.

11.2.3 The GC column oven must be capable of beingcooled to -70EC and subsequently temperatureprogrammed to 150EC.

11.2.4 The specific GC column and temperature programemployed will be dependent on the compounds ofinterest. Appropriate conditions aredescribed in the literature (2). In general,a nonpolar stationary phase (e.g., SE-30, OV-1) temperature programmed from -70 to 150EC at8E/minute will be suitable. Fused silica,bonded-phase columns are preferable to glasscolumns since they are more rugged and can beinserted directly into the MS ion source,thereby eliminating the need for a GC/MStransfer line. Fused silica columns are alsomore readily connected to the GC injectionvalve (Figure 3). A drawback of fused silica,bonded-phase columns is the lower capacitycompared to coated, glass capillary columns.In most cases the column capacity will be lessthan 1 microgram injected for fused silicacolumns.

11.2.5 Capillary column dimensions of 0.3mm ID and 50meters long are generally appropriate althoughshorter lengths may be sufficient in manycases.

11.2.6 Prior to instrument calibration or sampleanalysis the GC/MS system is assembled asshown in Figure 3. Helium purge flow (throughthe cartridge) and carrier flow are set atapproximately 50 mL/minute and 2-3 mL/minuterespectively. When a cartridge is not inplace a union is placed in the helium purgeline to ensure a continuous inert gas flowthrough the injection loop.

11.2.7 Once the column and other system componentsare assembled and the various flowsestablished the column temperature isincreased to 250EC for approximately fourhours (or overnight if desired) to conditionthe column.

11.2.8 The MS and data system are set up according tothe manufacturer's instructions. Electronimpact ionization (70eV) and an electronmultiplier gain of approximately 5 x 10 should4

be employed. Once the entire GC/MS system hasbeen setup the system is calibrated asdescribed in Section 11.3. The user shouldprepare a detailed standard operatingprocedure (SOP) describing this process forthe particular instrument being used.

11.3 GC/MS Calibration

11.3.1 Tuning and mass standardization of the MSsystem is performed according tomanufacturer's instructions and relevant userprepared SOPs. Perfluorotributylamine (FC-43)should generally be employed as the referencecompound. The material is introduced directlyinto the ion source through a molecular leak.The instrumental parameters (e.g., lensvoltages, resolution, etc.) should be adjustedto give the relative ion abundances shown inTable 2, as well as acceptable resolution andpeak shape. If these approximate relativeabundances cannot be achieved, the ion sourcemay require cleaning according tomanufacturer's instructions. In the event

that the user's instrument cannot achievethese relative ion abundances, but isotherwise operating properly, the user mayadopt another set of relative abundances asperformance criteria. However, thesealternate values must be repeatable on a day-to-day basis.

11.3.2 After the mass standardization and tuningprocess has been completed and the appropriatevalues entered into the data system, the usershould then calibrate the entire GC/MS systemby introducing known quantities of thecomponents of interest into the system. Threealternate procedures may be employed for thecalibration process including 1) directinjection of dilute vapor phase standards,prepared in a dilution bottle or compressedgas cylinder, onto the GC column, 2) injectionor dilute vapor phase standards into a flowinginert gas stream directed onto a CMScartridge, and 3) introduction of permeationor diffusion tube standards onto a CMScartridge. Direct injection of a compressedgas cylinder (aluminum) standard containingtrace levels of the compounds of interest hasbeen found to be the most convenient practicesince such standards are stable over a severalmonth period. The standards preparationprocesses for the various approaches aredescribed in Section 13. The followingparagraphs describe the instrument calibrationprocess for these approaches.

11.3.3 If the system is to be calibrated by directinjection of a vapor phase standard, thestandard, in either a compressed gas cylinderor dilution flask, is obtained as described inSection 13. The MS and data system are setupfor acquisition, but the ionizer filament isshut off. The GC column oven is cooled to-70EC, the injection valve is placed in theload mode, and the cryogenic loop is immersedin liquid nitrogen or liquid argon. Liquidargon is required for standards prepared innitrogen or air, but not for standardsprepared in helium. A known volume of thestandard (10-1000 mL) is injected through thecryogenic loop at a rate of 10-100 mL/minute.

11.3.4 Immediately after loading the vapor phasestandard, the injection valve is placed in theinject mode, the GC program and system clockare started, and the cryogenic loop is heatedto 60EC by applying voltage (15-20 volts) tothe thermocouple wire heater surrounding theloop. The voltage is adjusted to maintain aloop temperature of 60EC. An automatictemperature controller can be used in place ofthe manual control system. After elution ofunretained components (-3 minutes afterinjection) the ionizer filament is turned onand data acquisition is initiated. The heliumpurge line (set at 50 mL/minute) is connectedto the injection valve and the valve isreturned to the load mode. The looptemperature is increased to 150EC, with heliumpurge, and held at this temperature until thenext sample is to be loaded.

11.3.5 After the last component of interest haseluted, acquisition is terminated and the datais processed as described in Section 11.3.8.The standard injection process is repeatedusing different standard concentrations and/orvolumes to cover the analytical range ofinterest.

11.3.6 If the system is to be calibrated by analysisof standard CMS cartridges, a series ofcartridges is prepared as described inSections 13.2 or 13.3. Prior to analysis thecartridges are stored (no longer than 48hours) as described in Section 9.10. Foranalysis the injection valve is placed in theload mode and the cryogenic loop is immersedin liquid nitrogen (or liquid argon ifdesired). The CMS cartridge is installed inthe helium purge line (set at 50 mL/minute) sothat the helium flow through the cartridge isopposite to the direction of sample flow andthe purge gas is directed through thecryogenic loop and vented to the atmosphere.The CMS cartridge is heated to 370-400EC andmaintained at this temperature for 10 minutes(using the temperature control processdescribed in Section 9.8). During thedesorption period, the GC column oven iscooled to -70EC and the MS and data system aresetup for acquisition, but the ionizerfilament is turned off.

11.3.7 At the end of the 10 minute desorption period,the analytical process described in Sections11.3.4 and 11.3.5 is conducted. During theGC/MS analysis heating of the CMS cartridge isdiscontinued. Helium flow is maintainedthrough the CMS cartridge and cryogenic loopuntil the cartridge has cooled to roomtemperature. At that time, the cryogenic loopis allowed to cool to room temperature and thesystem is ready for further cartridgeanalysis. Helium flow is maintained throughthe cryogenic loop at all times, except duringthe installation or removal of a CMScartridge, to minimize contamination of theloop.

11.3.8 Data processing for instrument calibrationinvolves determining retention times, andintegrated characteristic ion intensities foreach of the compounds of interest. Inaddition, for at least one chromatographicrun, the individual mass spectra should beinspected and compared to reference spectra toensure proper instrumental performance. Sincethe steps involved in data processing arehighly instrument specific, the user shouldprepare a SOP describing the process forindividual use. Overall performance criteriafor instrument calibration are provided inSection 14. If these criteria are notachieved, the user should refine theinstrumental parameters and/or operatingprocedures to meet these criteria.

11.4 Sample Analysis

11.4.1 The sample analysis is identical to thatdescribed in Sections 11.3.6 and 11.3.7 forthe analysis of standard CMS cartridges.

11.4.2 Data processing for sample data generallyinvolves 1) qualitatively determining thepresence or absence of each component ofinterest on the basis of a set ofcharacteristic ions and the retention timeusing a reversed-search software routine, 2)quantification of each identified component byintegrating the intensity of a characteristicion and comparing the value to that of thecalibration standard, and 3) tentative

Y'A%BX%CX 2

identification of other components observedusing a forward (library) search softwareroutine. As for other user specificprocesses, a SOP should be prepared describingthe specific operations for each individuallaboratory.

12. Calculations

12.1 Calibration Response Factors

12.1.1 Data from calibration standards is used tocalculate a response factor for each componentof interest. Ideally the process involvesanalysis of at least three calibration levelsof each component during a given day anddetermination of the response factor(area/nanogram injected) from the linear leastsquares fit of a plot of nanograms injectedversus area (for the characteristic ion). Ingeneral, quantities of components greater than1,000 nanograms should not be injected becauseof column overloading and/or MS responsenonlinearity.

12.1.2 In practice the daily routine may not alwaysallow analysis of three such calibrationstandards. In this situation calibration datafrom consecutive days may be pooled to yield aresponse factor, provided that analysis ofreplicate standards of the same concentrationare shown to agree within 20% on theconsecutive days. In all cases one givenstandard concentration, near the midpoint ofthe analytical range of interest, should beinjected at least once each day to determineday-to-day precision of response factors.

12.1.3 Since substantial nonlinearity may be presentin the calibration curve, a nonlinear leastsquares fit (e.g. quadratic) should beemployed. This process involves fitting thedata to the following equation:

whereY = peak areaX = quantity of component injected nanogramsA, B, and C are coefficients in the equation.

YA'A%BXA%CX2

CA'XAVs

12.2 Analyte Concentrations

12.2.1 Analyte quantities on a sample cartridge arecalculated from the following equation:

whereY is the area of the analyteA

characteristics ion for the samplecartridge.

X is the calculated quantity of analyte onAthe sample cartridge, in nanograms.

A, B, and C are the coefficients calculatedfrom the calibration curve described inSection 12.1.3.

12.2.2 If instrumental response is essentially linearover the concentration range of interest, alinear equation (C=O in the equation above)can be employed.

12.2.3 Concentration of analyte in the original airsample is calculated from the followingequation:

whereC is the calculated concentration of analyteAin ng/L.V and X are as previously defined in Sections A10.2.11 and 12.2.1, respectively.

13. Standard Preparation

13.1 Standards for Direct Injection

13.1.1 Standards for direct injection can be preparedin compressed gas cylinders or in dilutionvessels. The dilution flask protocol has beendescribed in detail in another method and isnot repeated here (6). For the CMS methodwhere only volatile compounds (boiling point

CT'VI×d

VC×

14.7PC%14.7

×24.4×1000

compressed gas cylinders has been found to bemost convenient. These standards aregenerally stable over at least a 3-4 monthperiod and in some cases can be purchased fromcommercial suppliers on a custom preparedbasis.

13.1.2 Preparation of compressed gas cylindersrequires working with high pressure tubing andfittings, thus requiring a user prepared SOPwhich ensures that adequate safety precautionsare taken. Basically, the preparation processinvolves injecting a predetermined amount ofneat liquid or gas into an empty high pressurecylinder of known volume, using gas flow intothe cylinder to complete the transfer. Thecylinder is then pressurized to a given value(500-1000 psi). The final cylinder pressuremust be determined using a high precisiongauge after the cylinder has thermallyequilibrated for a 1-2 hour period afterfilling.

13.1.3 The concentration of components in thecylinger standard should be determined bycomparison with National Bureau of Standardsreference standards (e.g., SRM 1805-benzene innitrogen) when available.

13.1.4 The theoretical concentration (at 25EC and 760mm pressure) for preparation of cylinderstandards can be calculated using thefollowing equation:

whereC is the component concentration, in ng/mLT

at 25EC and 760 mm Hg pressure.V is the volume of neat liquid componentI

injected in FL.V is the internal volume of the cylinder,c

in L.d is the density of the neat liquid

component, in g/mL.P is the final pressure of the cylinderc

standards, in pounds per square inchgauge (psig).

13.2 Preparation of Spiked Traps by Vapor Phase Injection

This process involves preparation of dilution flask orcompressed gas cylinder containing the desiredconcentrations of the compound(s) of interest andinjecting the desired volume of vapor into a flowing gasstream which is directed onto a clean CMS cartridge. Theprocedure is described in detail in another method withinthe Compendium (6) and will not be repeated here.

13.3 Preparation of Spiked Traps Using Permeation or DiffusionTubes

13.3.1 A flowing stream of inert gas containing knownamounts of each compound of interest isgenerated according to ASTM Method D3609 (4).Note that a method of accurately maintainingtemperature within + 0.1EC is required and thesystem generally must be equilibrated for atleast 48 hours before use.

13.3.2 An accurately known volume of the standard gasstream (usually 0.1-1 liters) is drawn througha clean CMS cartridge using the samplingsystem described in Section 10.2.1, or asimilar system. However, if mass flowcontrollers are employed, they must becalibrated for the carrier gas used in Section13.3.1 (usually nitrogen). Use of air as thecarrier gas for permeation systems is notrecommended, unless the compounds of interestare known to be highly stable in air.

13.3.3 The spiked traps are then stored orimmediately analyzed as in Sections 11.3.6 and11.3.7.

14. Performance Criteria and Quality Assurance

This section summarizes the quality assurance (QA) measuresand provides guidance concerning performance criteria whichshould be achieved within each laboratory. In many cases thespecific QA procedures have been described within theappropriate section describing the particular activity (e.g.parallel sampling).

14.1 Standard Operating Procedures (SOPs)

14.1.1 Each user should generate SOPs describing thefollowing activities as accomplished in theirlaboratory: 1) assembly, calibration andoperation of the sampling system, (2)preparation, handling and storage of CMScartridges, 3) assembly and operation of GC/MSsystem including the thermal desorptionapparatus and data system, and 4) all aspectsof data recording and processing.

14.1.2 SOPs should provide specific stepwiseinstructions and should be readily availableto, and understood by the laboratory personnelconducting the work.

14.2 CMS Cartridge Preparation

14.2.1 Each batch of CMS cartridges, prepared asdescribed in Section 9, should be checked forcontamination by analyzing one cartridge,immediately after preparation. While analysiscan be accomplished by GC/MS, manylaboratories may choose to use GC/FID due tologistical and cost considerations.

14.2.2 Analysis by GC/FID is accomplished asdescribed for GC/MS (Section 11) except foruse of FID detection.

14.2.3 While acceptance criteria can vary dependingon the components of interest, at a minimumthe clean cartridge should be demonstrated tocontain less than one-fourth of the minimumlevel of interest for each component. Formost compounds the blank level should be lessthen 10 nanograms per cartridge in order to beacceptable. More rigid criteria may beadopted, if necessary, within a specificlaboratory. If a cartridge does not meetthese acceptance criteria, the entire lotshould be rejected.

14.3 Sample Collection

14.3.1 During each sampling event at least one cleancartridge will accompany the samples to thefield and back to the laboratory, having beenplaced in the sampler but without sampling

air, to serve as field blank. The averageamount of material found on the field blankcartridges may be subtracted from the amountfound on the actual samples. However, if theblank level is greater than 25% of the sampleamount, data for that component must beidentified as suspect.

14.3.2 During each sampling event at least one set ofparallel samples (two or more samplescollected simultaneously) should be collected,preferably at different flow rates asdescribed in Section 10.1.4. If agreementbetween parallel samples is not generallywithin +25% the user should collect parallelsamples on a much more frequent basis (perhapsfor all sampling points). If a trend of lowerapparent concentrations with increasing flowrate is observed for a set of parallel samplesone should consider using a reduced samplingrate and longer sampling interval, ifpossible. If this practice does not improvethe reproducibility further evaluation of themethod performance for the compound ofinterest might be required.

14.3.3 Backup cartridges (two cartridges in series)should be collected with each sampling event.Backup cartridges should contain less than 10%of the amount of components of interest foundin the front cartridges, or be equivalent tothe blank cartridge level, whichever isgreater.

14.4 GC/MS Analysis

14.4.1 Performance criteria for MS tuning and massstandardization have been discussed in Section11.2 and Table 2. Additional criteria can beused by the laboratory, if desired. Thefollowing sections provide performanceguidance and suggested criteria fordetermining the acceptability of the GC/MSsystem.

14.4.2 Chromatographic efficiency should be evaluateddaily by the injection of calibrationstandards. A reference compound(s) should bechosen from the calibration standard andplotted on an expanded time scale so that its

DL'A%3.3S

width at 10% of the peak height can becalculated, as shown in Figure 6. The widthof the peak at 10% height should not exceed 10seconds. More stringent criteria may berequired for certain applications. Theasymmetry factor (see Figure 6) should bebetween 0.8 and 2.0. The user should alsoevaluate chromatographic performance for anypolar or reactive compounds of interest, usingthe process described above. If peaks areobserved that exceed the peak width orasymmetry factor criteria above, one shouldinspect the entire system to determine ifunswept zones or cold spots are present in anyof the fittings or tubing and/or ifreplacement of the GC column is required.Some laboratories may choose to evaluatecolumn performance separate by directinjection of a test mixture onto the GCcolumn. Suitable schemes for columnevaluation have been reported in theliterature (7).

14.4.3 The detection limit for each component iscalculated from the data obtained forcalibration standards. The detection limit isdefined as

whereDL is the calculated detection limit in

nanograms injected.A is the intercept calculated in Section

12.1.3.S is the standard deviation of replicate

determinations of the lowest levelstandard (at least three suchdeterminations are required). The lowestlevel standard should yield a signal tonoise ratio (from the total ion currentresponse) of approximately 5.

14.4.4 The relative standard deviation for replicateanalyses of cartridges spiked at approximately10 times the detection limit should be 20% orless. Day to day relative standard deviationfor replicate cartridges should be 25% orless.

14.4.5 A useful performance evaluation step is theuse of an internal standard to track systemperformance. This is accomplished by spikingeach cartridge, including blank, sample, andcalibration cartridges with approximately 100nanograms of a compound not generally presentin ambient air (e.g., perfluorotoluene).Spiking is readily accomplished using theprocedure outlined in Section 13.2, using acompressed gas standard. The integrated ionintensity for this compound helps to identifyproblems with a specific sample. In generalthe user should calculate the standarddeviation of the internal standard responsefor a given set of samples analyzed underidentical tuning and calibration conditions.Any sample giving a value greater than + 2standard deviations from the mean (calculatedexcluding that particular sample) should beidentified as suspect. Any marked change ininternal standard response may indicate a needfor instrument recalibration.

14.5 Method Precision and Recovery

14.5.1 Recovery and precision data for selectedvolatile organic compounds are presented inTable 1. These data were obtained usingambient air, spiked with known amounts of thecompounds in a dynamic mixing system(2).

14.5.2 The data in Table 1 indicate that in generalrecoveries better than 75% and precision(relative standard deviations) of 15-20% canbe obtained. However, selected compounds(e.g. carbon tetrachloride and benzene) willhave poorer precision and/or recovery. Theuser must check recovery and precision for anycompounds for which quantitative data areneeded.

References

1. Kebbekus, B. B. and J. W. Bozzelli. Collection and Analysisof Selected Volatile Organic Compounds in Ambient Air.Proceedings of Air Pollution Control Association, Paper No.82-65.2, Air Pollution Control Association, Pittsburgh,Pennsylvania, 1982.

2. Riggin, R. M. and L. E. Slivon. Determination of VolatileOrganic Compounds in Ambient Air Using Carbon Molecular SieveAdsorbents, Special Report on Contract 68-02-3745 (WA-7), U.S.Environmental Protection Age