Study to Determine the Effect of Moisture on Volatile Organic
Compound Recovery Rates for Through-The-Probe Audits
into Stainless Steel Canisters

This page last reviewed April 29, 2010

Alice Westerinen and Barry B. Reisman
Monitoring and Laboratory Division
California Air Resources Board
1309 T Street
Sacramento, CA 95814

ABSTRACT

Through-the-probe (TTP) audits of criteria pollutants have been a part of the California Air Resources Board's (ARB) quality assurance program since 1981.  The TTP audit allows us to check the integrity of the entire sampling and analysis phases of ambient air pollution monitoring.  By auditing through-the-probe we can detect not only problems associated with instrument calibration and operation, but also problems associated with the station probe line from leaks or contamination.   In 1989 we expanded the TTP audit program to include the ambient toxic monitoring program.  This field audit program joined an existing laboratory audit program for volatile organic compounds (VOCs).  Using this approach we were able to determine the field and lab biases and identify areas for improvement in the network.  Based on the TTP audits, the ARB conducted a two part canister comparison study to determine the effect of canister moisture on VOC recovery rates from stainless steel canisters.  The results of the study will be used to determine the best auditing method for VOC recovery.  In Part 1 of the study, we compared the results of standard gas administered into dry canisters with the results of wet injected canisters.  The results confirmed improved sample recovery utilizing wet injected canisters for VOC sampling.  Humidifying the sample canisters with high grade water improved recovery rates in eight out of 18 VOCs while using a pre-run zero air purge.  In Part 2 of the study, wet injected canisters were compared with canisters humidified by adding moisture to the diluent zero air stream.  Both audit procedures used a pre-run zero air purge.  The humidified diluent ensured all components of the collection system were exposed to moisture prior to being combined with the audit gas and administered to the probe.  This practice improved VOC recovery rates over the wet injected canister method in three out of 18 VOCs.  This paper is a report on the results of the two part canister humidification study.

INTRODUCTION

TTP audits of VOCs have been a part of the ARB's TTP audit program for over seven years.  Since 1989, the ARB has been refining its TTP program for the ambient air toxic monitoring program in California.  By introducing known quantities of VOCs at the probe inlet, the audits check the integrity of the field probe, the toxic sampler (leaks, contamination, etc.), the transport system (sample canister), and the laboratory analysis accuracy of the canister contents.  Until 1993, auditing was conducted using dry audit gas in dry, clean stainless steel canisters.  Recovery rates for several aromatics, in particular, were low (Table 1).  Literature investigations indicated that humidification improved VOC recovery rates in canisters.  A field study was conducted in two parts which compared:  1) dry canisters to canisters injected with high pressure liquid chromatography (HPLC) grade water (100 micro-liters (ul)), and 2) wet injected canisters versus in-line humidified canisters.  The goal of the study was to improve our ability to assess the data accuracy for the ambient toxics and photochemical assessment monitoring station (PAMS) programs.  Humidification by either method resulted in improved VOC recovery rates.  In theory, the improvement is achieved by coating the "active sites" of the audit train with water molecules to passivate potential surface reactions with VOCs.

EXPERIMENT

Part 1

The field study was conducted to determine the differences in VOC recovery rates between auditing through-the-probe into dry canisters versus humidified canisters.  The dry canister audit utilized a blend of air and toxic compounds in a National Institute of Standards and Technology (NIST) traceable cylinder, a zero air source, a dilution system, a tee with a vol-o-flow, an adapter probe line to connect the audit gas to the probe line of the field sampler, and a dry, clean stainless steel canister.  The set-up is shown in Figure 1.  The toxic blend cylinder was a high concentration (6-1,000 parts per billion (ppb)) NIST traceable blend of 26 VOCs.  The compounds are listed on Table 1.  The ARB lab analyzes and reports data for 18 VOCs.  The cylinder is recertified by NIST on a regular basis.  Using the zero air source and the dilution system, concentrations of audit gas were varied to approximate typical ambient concentrations.  The zero air source was an Aadco pure air generator combined with a methane reactor with a capacity of providing up to 10 liters per minute of zero air.  The methane reactor eliminated hydrocarbons in the ambient air used as the air source for the system.  The zero air quality is confirmed quarterly using the laboratory's limits of detection as the criteria for cleanliness (Table 2).  The dilution system consisted of a temperature controlled environment with mass flow meters (MFM), mass flow controllers (MFC), and a mixing chamber.  The MFMs measured the flow of the toxic cylinder and the zero air system.  Both MFCs can be independently controlled allowing a wide range of audit concentrations.  The dilution system was calibrated and certified quarterly.  All lines used were stainless steel or Teflon; all fittings were stainless steel.  A vol-o-flow and tee were used to measure the flow rate through the dilution system.  This arrangement permits venting the excess flow so as not to pressurize the sampling system while maintaining enough excess flow (i.e., 0.25 inch H2O) to prevent inward leakage of ambient air.  The audit probe line consisted of 100 feet of 1/4 inch TFE Teflon tubing connected to the tee and to the station probe inlet.  The dry sampling container is identical to those used for ambient sampling.  All canisters used in the network are batch cleaned, testing one out of eight for adherence to the cleanliness criteria (Table 3), and evacuated to minus 30 inches of mercury prior to field sampling.  To conduct a dry canister audit, staff hooked up the audit apparatus, tested the flow, and connected the audit probe to the station probe inlet.  Before connecting the audit canister to the toxic sampler, a two hour air purge was conducted in which zero diluent air is run through the toxic audit apparatus into the station probe.  The audit canister was filled using ambient sampling procedures.  The canister was a XonTech 910A toxic sampler.  The audits ran for 24 hours, the same as the ambient sampling, at a sample flow rate of 7.5 cubic centimeters per minute.  The final pressure of the audit canister was between eight and 16 PSIG.  The canister shipping and analysis were identical to that of an ambient sample.  Audit runs with wet injected canisters were conducted as above except they were pre-injected with 100 ul of HPLC grade water with a syringe using a septum adapter on the canister inlet.  Part 1 consisted of 52 field audits conducted between May 1993 and December 1995 using the standard toxics TTP apparatus diagrammed in Figure 1.

Part 2

This element of the field study was to determine the differences in VOC recovery rates between auditing with wet injected versus in-line humidified canisters.  Audit gas was introduced into the dry canisters with humidified zero air (using an in-line bubbler) to dilute the toxic gas blend and achieve approximately 60 percent relative humidity.  The in-line humidification audit apparatus consisted of the same dilution system used for dry canister audits with the addition of an in-line bubbler system.  The configuration of the bubbler is shown in Figure 2. In-line humidification was accomplished by inserting a bubbler consisting of a six liter SUMMA canister with center dip tube downstream of the dilution unit.  The canister was charged with three liters of Nanopure water that had been carbon filtered and boiled by staff to remove traces of dissolved organic compounds.  Any condensation from the bubbler was collected in an empty six liter SUMMA canister located after the bubbler.  After exiting the empty canister, the humidified zero air entered the dilution unit where it mixed with the incoming toxic gas blend.  The humidified, diluted audit gas passed through the audit line to the station inlet probe on to the toxic sampler, and finally to the dry audit canister.  Shipping and laboratory processing were conducted in the same manner as an ambient sample.  Humidifying the entire audit train and sampler probe line was expected to passivate the dry portions of the inlet probe and improve VOC recovery rates over those observed using the water injected canister method.  Part 2 of the study is still in progress.  Six field audits conducted since December 1995 are reported here and have been used in these comparisons to the wet injected canisters.

EXPERIMENT RESULTS

Part 1 - Dry Canister versus Wet Injected Canister

Eight out of 18 VOCs showed statistically improved recovery rates (confidence level = 95%) using the wet injected canister versus the dry canister method (Figures 3 and 4).  Most of the VOCs showing improvement were aromatic compounds.  The improvement ranged from 7% for 1,3-butadiene to 39% for styrene.  The average improvement was 25%.  The average improvement for aromatics was 29%.  Statistically, the remaining 10 VOCs were not affected by the audit technique.  The arithmetic means of the percent biases for each VOC are presented as an absolute value of the bias.  In most cases the biases from the audit results for the VOCs of concern, tended to be either always positive or always negative although there was some variability in the data.

Part 2 - Wet Injected Canister versus In-Line Humidified Canister

Based on six audits, three of 18 VOCs (two additional VOCs as well as further improvement of a third VOC) showed improved recovery rates using the in-line humidification method versus the wet injected method (Figures 5 and 6).  The improvement ranged from 5% for chlorobenzene to 11% for styrene and ortho-dichlorobenzene.  The average improvement was 9%.  Statistically, the remaining 15 VOCs were not affected by the audit technique.

CONCLUSIONS

Substantial improvement in VOC recovery rates was observed using the two forms of humidification included in this study.  Eight out of 18 VOCs improved with the introduction of the wet injected canister method.  The preliminary results of the in-line humidification versus the wet injected method TTP audits show that the in-line method may improve recovery rates for three of the 18 VOCs.  Two of these VOCs were in addition to the improvements observed with the wet injected canister method.  The third VOC showed increased improvement over results from the wet injected method.  In summary, changing from the dry canister to the in-line humidification TTP audit method improved recovery rates in nine out of 18 VOCs (Figures 7 and 8).  Overall recoveries improved 43%.  Improvements ranged from 22% for trichloroethylene to 94% for 1,1,1-trichloroethane.  Statistically, seven out of nine of the remaining VOCs were not affected one way or the other.  Staff are investigating the causes for the decreased recoveries for tetrachloroethylene and benzene using the in-line humidification method.  Three potential causes are being pursued:  limited sample size; site specific contamination; and laboratory operations.  Although the dry canister versus in-line humidification method improved results for nine compounds compared to eight compounds for dry canister versus wet injected canister, a final assessment of the in-line technique will require a larger sample size.  Many of the VOCs showed a "stair step" improvement based on the improved audit methods (Figures 9 and 10).  As changes were made from the dry canister, to the wet injected canister, to the in-line humidification TTP methods, the recovery rates continued to improve.  This trend is most noticeable for the aromatics.  For several VOCs, the in-line humidification audit results are approaching the laboratory accuracy estimates.  For dibromoethane and styrene the laboratory accuracy appears to exceed the in-line humidification results.  We have identified and are investigating two potential causes: limited sample size and laboratory operations.  Work on Part 2 of the TTP audit study will continue to obtain results from all sites.  Depending upon the final results of our continued study, we will consider changing the TTP audits for toxics and PAMS to the in-line humidification method.


Michael Miguel, Program Contact - mmiguel@arb.ca.gov



Toxics Audit
preload