Table 8. Percent recoveries and relative standard deviations esti-
mated by the off-site laboratories and on-site by the technology
developer for the performance evaluation materials.
Hanby
CRREL
Ref. Lab
HM 2000
Sample
% Recov.
% Recov.
Visual
% Recov
no./ID
(% RSD)
(% RSD)
% Recov.
(% RSD)
GRO
SG-1, 2, 3, 4
99
112
160
310
(200 mg TPH/kg)
(8.6)
(4.4)
(38)
WG-1, 2, 3, 4
100
102
430
480
(1.0 mg TPH/L)
(7.0)
(10.2)
DRO
SDM-1, 2, 3, 4
NA
134
300
270
(401 mg TPH/kg)
(7.0)
(56)
SDM-25, 26, 27, 28
NA
177
270
290
(2480 mg TPH/kg)
(22.7)*
(35)
* Only two of the four replicates were analyzed.
(Tables 9 and 10) and were on average 35% for the HM
day was continued. Visual and HM 2010 measurements
were made on 34 of the 36 samples distributed.
2000 method. The precision of the visual method was
On the second day of the field exercise, the HM
not assessed because estimates arise from a discontinu-
2000 experienced another software failure after the 16
ous scale and are subjective.
water samples contaminated with DRO/bunker C had
The sampling team distributed 90 samples for on-
been analyzed. This could not be corrected during the
site analysis during the 2.5-day exercise. On the first
technology demonstration. At about the same time the
day, 20 soil and 16 water samples contaminated with
HM 2000 failed, the HM 2010 also failed because of a
GRO compounds were distributed. On the second day,
low battery charge. The combination of these two prob-
36 soil samples and 16 water samples were distributed.
lems forced the technology developer to treat 38
Of the soils distributed on day two, 27 were contami-
samples (36 soil samples and 2 water matrix spikes) by
nated by DRO/bunker C and 9 were contaminated by
taking them through the solvent extraction step, then
RRO compounds. All 16 water samples were contami-
shipping them off-site prior to adding the catalyst and
nated with DRO/bunker C. On the morning of the last
completing the analysis.
day, two water samples with DRO contamination were
The large amount of time spent to address these
distributed. The technology developer agreed to ana-
problems limited the time available to prepare a data
lyze the large number of samples on the second day of
report. As a result, no TPH values were reported dur-
the field exercise.
ing the technology demonstration; however, a prelimi-
The technology developer analyzed all of the 20 soil
nary data report was made available on the following
samples distributed on the first day; however, while
Monday (the technology demonstration finished on a
analyzing the 16 water samples, the HM 2000 instru-
Friday). This initial data report showed that of the 90
ment developed a software problem that required off-
samples distributed, 52 samples (58%) were analyzed
site assistance. When it became clear that the HM 2000
on-site by the visual comparison and the HM 2010
would be unable to continue, the technology developer
methods, and 47 samples (53%) had been analyzed by
chose to hold some of the sample extracts overnight,
the HM 2000. Preliminary TPH values were reported
prior to adding the catalyst and forming the Friedel-
at this time for the visual and HM 2000 methods of
Crafts reaction products. Because of this interruption,
analysis; however, none were reported for the HM 2010.
five water samples were not analyzed by the HM 2000
system. On the morning of the second day, the HM 2000
During this field exercise, the HM 2010 was only
system was brought back on-line and the analysis of
capable of producing voltage responses because of an
the water samples that had been distributed on the first
integrated circuit failure, and calibration models for the
13
Previous Page
