SITE INFORMATION AND CORRESPONDENCE

Home Browse Search

,44 ICGC NoRiN AMEND VOLUME 11 NUMBER I JANUARY 2003 V NW41 w .chromatographyonlinexom because we anticipated our sampling pro- nology Evaluation program for MTBE at vide adequate results for oxygenates depend- ung upon the required sensitivity and that gram to require analysis of 1000 or more Port Hueneme CBC (Ventura, California). how As explained above, the oof method performance in aqueous matrices groundwater samples for MTBE,Automated method applicability was designed to assess should be verified by the laboratory per- to use a readily available, forming the analysis (14). Halden and co- method. For this reason, we chose to avoid the proficiency of purge-and-trap GGMS workers(4)also have suggested that purging investigating other separation techniques methods in analyzing groundwater samples and this compound using large sample volumes such as azeotropic distillation (EPA Method fiom the demonstration site for MTBE at elevated temperatures(40'Q could result 5031) or direct aqueous injection (6), two its degradation products — tert-butyl alco- in method detection limits as low as 1 µg/L. techniques that have shown promise for oxy- hol, 2-propanol, and acetone. The method Our objective in this study was to deter- genate analysis,if purge-and-trap procedures applicability study was designed to demon- mine if accurate quantitation of MTBE would satisfy our objectives. Of the dozens strate that purge-and-trap GC-MS methods could be obtained for compound concentm- of laboratories investigated, we found that could generate satisfactory data in the matrix tions of less than 5 lig/L,which is the cur- most claimed they could perform this analy- of interest for the treatment technology rent secondary maximum contaminanr level sis using purge-and-trap methodology fol- demonstrations. Based upon discussions for California. Although Halden and col- lowed by GC-MS (Method SW-846 with EPA's National Risk Management and leagues suggested that method detection Gm- 5030B/8260B).We eliminated GC analysis Research Laboratory quality assurance and its as low as 1 µg/L could be obtained using without MS as an option because of the project management staff, we defined the purge-and-trap methodology, for the pur- potential interferences expected in our sam- data requirements for the MTBE demon- poses of our demonstration it was important Ales. In addition, many commercial labora- strations as the ability to meet project accu- to achieve a practical quantitation limit of torim prefer GC-MS and commercial labo- racy and precision goals of 75-125%recov- 1 lig/L. ratory GC-MS analysis usually is no more ery and 25% relative percent difference, EPA Guidance on Data Quality Indicators expensive than GC analysis. respectively, at concentrations that were (EPA QA/G51) explains in detail the differ- For the purpose of our study, we nar- 20-25%of the lowest applicable regulatory ence between practical quantitation limits rowed our choice of laboratories to three, level.The action levels and practical quinti- and method detection limits (5). Method which we chose based upon their ability to ration limit requirements for the target com- detection limits are theoretical or statistical accommodate the potential workload,satisfy pounds are as follows: lower limits achievable by a particular EPA quality assurance requirements for the • MTBE: 5 µg/L (California secondary method and usually represent the minimum Superfund Innovative Technology Evalua- maximum contaminant level); target concentration of an analyte that can be dif- tion program,and their previous experienceprati cal q a oalcohoL 12 µg/Ltitatn limit: 2pg/(California ferentiated from zero with a specified level of in the analysis of MTBE at similar ground- • confidence.Practical quantitation limits, on water sites. In addition to MTBE, EPA was action level); target practical quantitation the other hand, represent the concentration interested in analyzing for other compounds limit•.6-8µg/L of an analyte that can be quantified within at low quantitation limits, including test- • 2-Propanol: no California regulatory the precision and accuracy constraints of the butyl alcohol, 2-propanol, and acetone. As level; target practical quantitation limit: analytical method. Many programs and lab- potential degradation products of MTBE, 10 µg/L oratories use the definition of a practical each of these compounds presents similar or • Acetone: no California regulatory level; quantitation limit as the lowest quantitation greater difficulties in terms of purging efft- target practical quantitation limit: standard used when developing a calibration ciency; therefore, we included them as pan 10 µg/L curve for the specified contaminant.Because of this study. Of the many potential laboratories inves- the practical quantitation limit incorporates The data from this method applicability Ligated,we chose three commercial laborato- the precision and accuracy requirements of study indicated that purge-and-trap meth- ries for this study through prequalification the method, we used it as the basis for ods attain acceptable levels of data quality audits based upon National Risk Manage- method sensitivity requirements in this and sensitivity for use in the referenced ment and Research Laboratory quality assur- study.The practical quantitation limit often MTBE demonstration program and should ance requirements.For purposes of this arti- is 5-10 times higher in concentration than continue to be considered as a method of cle, we will call them laboratory X, the method detection limit, and detection choice for analyzing MTBE. As described laboratory Y,and laboratory Z.Each labora- responses between the practical quantitation below, specific instrumentation or purge- tory conducted the demonstration of limit and the method detection limit often and-trap techniques used by one laboratory method applicability study in accordance are reported with J qualifiers as estimated compared with another proved better at with EPA methods with slight variations in concentrations. obtaining required quantitation limits with the purge-and-trap procedure. Laboratory X Another important consideration in acceptable recovery of spiked compounds. used a purge volume of 5 mL and purged at method selection for environmental pro- As pan of the study,we compiled raw ana- room temperature. (This same laboratory grams is the availability and cost of the lyrical data packages from the analytical lab- was the only laboratory to use a Varian methods of interest. Commercial labomto- oratories for review and validation to idm- Archon-EST autosampler [Palo Alto, Cali- ries, for example, work best if they can use tify potential method performance issues. fomial, and, as will be explained lacer, this automated methods for performing com- instrument proved to be an important factor Pound analysis. In fact, laboratory automa- Study Background affecting method performance.) Laboratory tion is one of the major reasons analytical This demonstration of method applicability Y purged at an elevated temperature of costs have declined significantly during the was conducted in support of the treatment 40 °C and used a 10-mL purge volume. past 15 years.Because analytical costs always technology demonstration that was initiated Laboratory Z purged at room temperature are a significant concern and in particular under the EPA Superfund Innovative Tech- using a 25-mL purge volume.