SITE INFORMATION AND CORRESPONDENCE

Home Browse Search

October 2002 • LL(STLine Bulletin 42 tivel' .As a result, ethanol and TBA these oxygenates can be determined than 5 pg/L.However,if alcohols(or are not ionized and cannot be for the same samples and in some of acetone)are analytes of concern, the detected by these lamps.The poten- the same analytical runs used to water sample must be heated to tial required to ionize MTBE is 10 eV, determine BTEX and MTBE, pro- attain adequate recovery of analytes. which is right at the maximum vided that the concentrations of all If the sample is not heated, the effec- potential of these lamps. Although target compounds fall within the tive limit of quantitation for TBA the PID may respond to MTBE when operational calibration ranges of the using Method 5030 is near 100 pg/L; the lamp is new, the response detectors used. when the water sample is heated to becomes weaker as the lamp ages 80°C the limit of quantitation is near with use. If the calibration curve for 10 pg/L. Method 8021 is not current, the procedure to calibrate for all o/the: In response to problems identi- method can return false negatives for pfied with current analytical practice, MTBE when MTBE is present at con- oxygenates and, until now, nosing EPA conducted a study to determine centrations above regulatory action method with this capability has the optimum conditions for purge- levels. and-trap sample preparation of undergone a rigorous demonstratio" Method 8021 (PID) may also be MTBE and the other fuel oxygenates subject to coelution interferences of applicability. Conventions/ in river water samples both with and and generate false positive results analytical procedures designed/o without BTEX interferences in the when real-world samples contain sig- petroleum hydrocarbons(i.e., BTE form of gasoline spiked at 600 jig/L. nificant concentrations of other The compounds included in the can also detect MTBE and the othe contaminants such as petroleum study were MTBE,TBA,RIPE,ETBE, hydrocarbons. Halden et al. (2001) 'hers when properly calibrated to TAME,TAEE, and acetone.The tar- found that when a sample contains 'them, but they have very poor '' get sensitivity was 5 pg/L(U.S.EPA, petroleum contamination (as total sensitivity/ar TBA and the 2002). petroleum hydrocarbon, TPH) of The study was performed over a greater than about 1,000 pg/L(1 part other alcohols. five-point calibration range of 2 pg/L per million),Method 8021 is subject to 40 pg/L for each target analyte. to false positive results for MTBE.He Another important concern is the The analytes were purged at 80°C for also found that the effect is concen- method detection limit or the report- seven minutes and trapped on a tration-dependent (i.e., the effect ing limits of current analytical proto- Supelco H trapl, held at 35 C, dry increases as the concentration of cols for the alcohols, and TBA in purged, desorbed and baked for other contaminants in the sample particular. Analysis of the alcohol three minutes each,and analyzed on increases).Most laboratory QA/QC oxygenates is a more difficult chal- a standard VOA column and a wax procedures for MTBE are not set up lenge than analyzing for BTEX (or column. Water samples were run to identify circumstances in which even MTBE).Many commercial labo- both with and without BTEX present coelution and concentration effects ratories set reporting limits for TBA in the samples.An additional evalua- compromise the reliability of the that are much higher than reporting tion using purge-and-trap conditions method.Without this information the limits for BTEX and MTBE. Typical at ambient temperature (20° C) and analyst may have the mistaken reporting limits for TBA may be as the standard VOA column was also impression that the analytical results high as 100 or 1,000 pg/L. These performed. are accurate, when in fact they are reporting limits are higher than the The results of EPA's study erroneous. concentrations of TBA that are of reg- demonstrate that the recoveries of A more important concern in- ulatory interest to many states. low levels of MTBE and related oxy- olves the unequivocal determination genates can be improved over cur- of the presence of oxygenates.UsingOvercoming Analytical rent practice. The most consistent either GC/MS (Method 8260) or Obstacles oxygenate recoveries were obtained GC/FID (Method 8015) with an using the following combination of appropriate GC column and an Methods 8015 (GC/FID) and 8260 methods: sample preparation using appropriate sample-preparation tech- (GC/ C) MS)are appropriate for deter- Method 5030 with a heated (80° nique, it is possible to detect oxy- mining the presence and concentra- purge-and-trap, then analysis by genates at concentrations of 5 pg/L tion of fuel oxygenates and BTEX. Method 8260 using a DB-Wax capil- or less. However, GC/MS provides Appropriate sample-preparation lary column as the determinative positive confirmation of the chemical methods include Methods 5021 (sta- method.Use of an RTX-Volatile cap- identity of the analyte that is tic headspace), 5030 (purge-and- illary column with a heated purge detected,while GC/FID does not. trap), or 5032 (vacuum distillation). did not significantly improve the It is not necessary to modify TBA can also be recovered for analy- t existingconventional practice for sis using the azeotropic distillation Performance with other brands t traps P technique Method 5031). If ethers may vary from that of the present study. chromatography to obtain data for all q ( If a different trap is used,its performance of the oxygenates; only the sample are the only target analytes of inter- must be demonstrated,not merely preparation and method calibration est,then using Method 5030 at ambi- assumed to be comparable to the Supelco steps need to be modified. If calibra- ent temperature (rather than heated) H trap.Silica gel is needed as a trapping tion curves are nun for all of the other is adequate to determine concentra- material for the trap to perform properly. ethers, then concentrations of all of tions of oxygenates that are greater ■continued on page 4 3