DISCUSSION - continued

DISCUSSION - TR00-200014

SUMMARY - TR00-200016

TR00-20
Page Navigation
11

Table 11. Relative percent differences (RPD)

Table 10. Percent recoveries and relative percent differ-

for the sample duplicates.

ences established for the matrix spike duplicates.

Technology Developer

Techology developer

Sample

CRREL

HM 2000

CRREL

Visual

HM 2000

No./ID

(%)

Sample

% Recovery

% Recovery % Recovery

No./ID

(% RPD)*

(% RPD)

1. Gasoline

1. Gasoline

2. Diesel

3. Motor oil

2. Diesel

failures, lack of existing calibration models, and

Soil, SDM-6 & 7

180

310

design problems delayed and confounded several

(250 mg TPH/kg)

(10)

of the laboratory trials.

Soil, SDM-21 & 22

250

160

It has been stated that both the HM 2000 and

(1000 mg TPH/kg)

(38)

visual method of estimating TPH in environmental

Water, WDM-1 & 2

matrices are capable of producing concentration

estimates within 10% of the concentration found

(0.98 mg TPH/L)

OR††

Water, WDM-16 & 17

170

by accepted methods of analysis, when the specific

(24 mg/kg TPH/L)

contaminant of concern is known (Hanby 1998).

However, this citation provided no experimental evi-

3. Motor oil

dence for this claim. This level of accuracy (10%),

Soil, M1 & M9

(1000 mg TPH/kg)

(12)

which is better than what has been reported else-

where on the basis of interlaboratory studies (Parr

* Relative percent difference.

†

Instrument failure.

et al. 1994), can be achieved by a single laboratory,

** Sample lost.

as demonstrated by the reference method (14 of 23

†† Over range, no dilution made.

cases had values within 10% for the analysis of

certified or spiked samples). In contrast, the HM

2000 only achieved this level of agreement for a certi-

occurs that quickly causes a very nauseating, chlorine-

fied or target value in 5 out of 34 cases (12%), and the

smelling gas to form. Therefore, cuvettes should only

visual method only achieved it in 2 out of 36 cases (5%).

be rinsed in an exhaust hood. A cleaning chamber, con-

In all cases where the HM 2000 or visual method gave

sisting of a washing station and reaction trap contain-

an average value within 10% of the certified or tar-

ing charcoal and sodium bisulfate, was fabricated by

geted concentration, the samples were spiked soils. Fig-

the technology developer to assist with this operation.

ure 3 shows the %D ranges of the values (33 visual, 31

However, in its current design, it failed to remove all

HM 2000, excluding less than or greater than values)

of the by-products, and additional rinses were neces-

found for both methods of analysis for all of the PE

sary once the cuvettes were removed from the cleaning

and spiked samples used in both the laboratory and field

chamber. Furthermore, we observed that these cuvettes

trials. This figure shows that roughly 60% of the visual

needed to be cleaned within a couple of hours of intro-

values for the spiked soils (12 of 21) and waters (7 of

ducing the catalyst, or the bottom window became per-

12), and the HM 2000 values for spiked soils (15 of

manently clouded. A combination of the time constraints

25), were within 50% of the certified or targeted val-

and need for an exhaust hood with the present cleaning

ues. However, only 1 out of 6 values (17%) attained by

chamber may limit the reuse of these cuvettes during

the HM 2000 for spiked water samples fell within this

field operations.

range. Clearly, neither method is capable of routinely

The instrumental problems experienced by the HM

achieving a 10% accuracy. With the exception of HM

2000 during the field exercise indicate that the present

2000 analysis of water matrices, perhaps they can

system is not robust, even when operated by the devel-

achieve a 50% level of accuracy, in the majority of

oper. Although not mentioned previously, instrumental

cases. It should be noted that, in comparison to the ref-

problems similar to those experienced in the field also

erence method for the analysis of field samples, this

plagued the laboratory trials. Intermittent instrument

To Contents

Index

Privacy Statement
Press Release
Contact