Sampling and Laboratory Analysis for Solvent Stabilizers - Environmental Investigation and Remediation - page 219

sample concentration step in EPA Method 8270C are reduced in EPA Method 1624 by limiting the
degree of concentration (2 mL i nal volume versus 1 mL i nal volume by the standard procedure)
and using high-volume injection into the gas chromatograph (Linton and Alonso, 2006). The detec-
tion limit of EPA Method 1624 depends more on the level of interferences than instrumental limita-
tions (USEPA, 2001).
1,4-DIOXANE METHOD COMPARISON MINI-STUDY
A groundwater investigation to delineate 1,4-dioxane occurrence was conducted at a for-
mer industrial facility located in Tallevast, Florida. Analyses for 1,4-dioxane initially used a
modii ed Method 8270C. Beginning in 2006, samples were analyzed with both EPA Method
8270C and EPA Method 8260B in SIM mode. Results reported for the Method 8260B (SIM)
analysis were consistently and signii cantly higher than those reported for the Method 8270C
analysis, in some cases, as much as two orders of magnitude higher. Environmental scientists
with the consulting i rm conducting the investigation, Blasland Bouck and Lee Environmental
Services (now part of Arcadis), undertook a comparison study to identify differences in ana-
lytical methods for 1,4-dioxane (Linton and Alonso, 2006).
Samples were collected in triplicate in two separate sampling events from 13 monitoring
wells and two private wells by using low-l ow sampling techniques. Analysis was performed
by EPA Methods 8270C, 8260B, and 8270 with isotope dilution (“8270-ID”) at three dif-
ferent commercial laboratories. The purpose of the comparison study was to obtain a more
quantitative statement of a pattern of conl icting results for analyses by different parties using
different methods to analyze samples from domestic wells. The conl icting results led to con-
troversy in the affected community in Bradenton County, Florida, and brought the subtleties
of GC-MS coni gurations into the local newspaper. Data used in the method evaluation by
Linton and Alonso have not been fully validated, and the study was not intended to be a rig-
orous or fully validated review of these three analytical methods. Full validation of the data
would require an exhaustive review of laboratory quality control procedures, i eld methods,
and so on. Samples were analyzed by Method 8260B in standard scan mode initially and
subsequently reanalyzed in SIM mode if results were below the detection limit. To check for
analyte loss and potential interference, i eld samples were spiked at three different concentra-
tions and used as matrix spike-matrix spike duplicate (MS/MSD) samples for all three meth-
ods to determine method accuracy (percent recovery) and method precision (relative percent
difference, RPD). Recovery was also measured on laboratory control standards, to distinguish
instrument accuracy from possible matrix effects attributable to sample characteristics.
Results for samples analyzed by Method 8260B were highest among the three methods;
some results were more than an order of magnitude higher than the Method 8270C results.
Accuracy was consistent for the three spike concentration levels, and precision increased
somewhat with spike concentration. Method 8260B in scan and SIM modes had an average
recovery of 73% and an average RPD of 5%. Accuracy increased with increasing spike con-
centration, ranging from 49% recovery for the low spike level to 98% recovery for the high
spike level. Precision was relatively constant at all spike concentration levels.
As shown in Figure 4.7 , the results using Method 8260B paralleled results for the isotope-
dilution method. Method 8270C had the lowest recoveries. The average recovery was 65%,
and the average RPD for recoveries in MS/MSD analyses was 20%. Precision and accuracy of
Method 8270C decreased as the concentration of spike addition increased. Overall accuracy
for Method 8270ID averaged 100% recovery, and precision averaged 9% RPD in the i rst
sampling event and 12% in the second sampling event. Accuracy was independent of spike
concentrations, whereas precision increased somewhat with spike concentration. Figure 4.7
presents a comparison of the concentration trends for the results from the three methods.