ivEPA
United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
P.O. Box 93478
Las Vegas NV 89193-3478
EPA/600/R-93/056
April 1993
Research and Development
Interlaboratory Evaluation
of an Off-Line SFE/IR
Method for Determination
of Petroleum Hydrocarbons
in Solid Matrices
Project Report
5181GR93QADCOV
-------
EVTERLABORATORY EVALUATION OF AN
OFF-LINE SFE/IR METHOD FOR
DETERMINATION OF PETROLEUM
HYDROCARBONS IN SOLID MATRICES
By
Viorica Lopez-Avila, Richard Young, and Robert Kim
Midwest Research Institute
California Operations
Mountain View, California 94043
Contract Number 68-C1-0029
Project Officer; Werner F. Beckert -
Quality Assurance and Methods Development Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89119
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89119
-------
NOTICE
The information in this document is the result of work funded wholly or in part by the United
States Environmental Protection Agency under Contract #68-C 1-0029 to Midwest Research Institute.
It has been subjected to the Agency's peer and administrative review, and it has been approved for
publication as an EPA document.
-------
PREFACE
This is the final report for Work Assignment 2-7, EPA Contract 68-C1-0029, conducted at
Midwest Research Institute, California Operations. The project was directed by Dr. Viorica Lopez-
Avila.
This report was written by Dr. Viorica Lopez-Avila. Technical support for the project was
provided by Richard Young and Robert Kim.
The following collaborators/laboratories (listed in alphabetical order) participated in this study:
Mark Bruce, Wadsworth/Alert Laboratories, North Canton, OH
Paul David, Amoco Corporation, Naperville, IL
Sally Eckert-Tilotta, Energy and Environmental Research Center, University of North
Dakota, Grand Forks, ND
Anna P. Emergy, NIST, Gaithersburg, MD
Chuck Hecht, Chemical Waste Management, Riverdale, IL
Joe Hedrick, Hewlett-Packard Company, Avondale, PA
Joe Levy, Suprex Corporation, Pittsburgh, PA
Shirley Liebman, CCS Instruments, West Grove, PA
Mary Ellen McNally/Connie Deardorff, DuPont Agricultural Products, Wilmington, DE
Midwest Research Institute, California Operations, Mountain View, CA
Nathan Porter, Dionex-Lee Scientific, Salt Lake City, UT
Steve Pyle, EPA, EMSL-LV, Las Vegas, NV
John L. Snyder, Lancaster Laboratories, Lancaster, PA
Joe Tehrani, Isco, Inc., Lincoln, NE
We gratefully acknowledge their contributions.
iii
-------
ABSTRACT
A collaborative study was conducted, with Midwest Research Institute, California Operations
(MRI-CO), as the lead laboratory, and with 13 additional volunteer laboratories participating, to
determine the method accuracy and precision of EPA SW-846 draft Method 3560. This method
describes the extraction of petroleum hydrocarbons from solid matrices with supercritical carbon
dioxide at 340 atm and 80°C for 30 min (dynamic; flow rate 1 to 2 mL/min as liquid) and collection
of the extracted materials in 3 mL tetrachloroethylene. The extracts generated by the participating
laboratories were shipped to MRI-CO for analysis by infrared spectrometry (draft Method 8440).
The collaborative study was based on the AOAC International blind replicate design with balanced
replicates. Four soil samples were to be extracted in triplicate. Three of these samples were standard
reference materials with TPH contents of 614, 2050, and 32,600 mg/kg, respectively. The fourth
sample was a clay soil spiked with motor oil at 10,000 mg/kg. Each of the four samples was
extracted by the laboratories in triplicate with supercritical carbon dioxide, and the extracts were
analyzed by MRI-CO on two different infrared spectrometers. In addition, each of the participating
laboratories extracted a sample of the unspiked clay soil, clay soil spiked with both corn oil and
reference oil (a mixture of n-hexadecane, isooctane, and chlorobenzene) at 1,000 mg/kg each, and clay
soil spiked with motor oil at 10,000 mg/kg and brought to a water content of 30 percent (the
individual laboratories then had to add anhydrous sodium sulfate). These latter three samples were
extracted only once. They were included in the study to address possible cross-contamination in the
supercritical fluid extraction system (the laboratories were asked to extract the unspiked clay soil right
after extraction of a clay soil sample spiked at 10,000 mg/kg with motor oil), the presence of
interferences such as oily materials, and the effect of water content on extraction efficiency.
After outlier removal (using both the Cachran and the Grubbs tests), we calculated the mean
concentration, repeatability standard deviation, reproducibility standard deviation, repeatability relative
standard deviation, and reproducibility relative standard deviation for each of the four matrices
extracted in triplicate. In addition, the relationship between the measured petroleum hydrocarbon
contents of these samples and the true concentrations was established.
The results from the triplicate analyses of the four materials show that the method accuracy
(percent recovery) for petroleum hydrocarbon concentrations ranging from 614 to 32,600 mg/kg was
83 percent; the mean recoveries of petroleum hydrocarbons from each of the four matrices ranged
from 78 to 107 percent for the analyses performed with a Perkin-Elmer FTIR spectrometer, and from
76 to 101 percent for the analyses performed with a Buck-Scientific IR spectrometer. The differences
of the results from the two instruments on a sample-by-sample basis were less than 17 percent for the
total petroleum hydrocarbon determinations. The interlaboratory method precisions ranged from 17
to 45 percent for analyses on the Perkin-Elmer FTIR system and from 17 to 48 percent for analyses
on the Buck-Scientific IR system; the intralaboratory method precisions ranged from 12 to 17 percent
for analyses on the Perkin-Elmer FTIR system and from 11 to 18 percent for analyses on the Buck-
Scientific IR system. Method accuracy and precision data are presented for the five laboratories that
used the Isco supercritical fluid extraction systems and for the seven laboratories that used different
supercritical fluid extraction systems, all having extraction vessels of 3.5-mL volume or less.
IV
-------
Analysis of variance (ANOVA) was used to separate (mathematically) the total variation of the
experimental measurements into an intralaboratory portion and an interlaboratory portion, with a
corresponding split of the total number of degrees of freedom. The ANOVA results indicate that, as
expected, the variation from laboratory to laboratory was greater than that attributed to the analytical
error displayed within laboratories. The matrix and operational parameters, such as flow rate, size
of the extraction vessel, extraction vessel design and orientation, mode of collection of the extracted
material, and temperature of the collection solvent/trap all seemed to be important.
The recoveries from the wet clay soil samples mixed with sodium sulfate were above 30 percent
for only two laboratories, and eight laboratories had recoveries below 8 percent. Separate experiments
conducted at MRI-CO showed that addition of anhydrous magnesium sulfate and Hydromatrix
(diatomaceous earth) raised the recoveries to 72 percent or higher for samples containing not more
than 20 percent water. When samples contained more than 20 percent water, they had to be mixed
with anhydrous magnesium sulfate and allowed to equilibrate for several hours (preferably overnight
in sealed containers and at 4°C to minimize losses of volatile petroleum hydrocarbons) before
extraction with supercritical carbon dioxide.
-------
CONTENTS
Notice ii
Preface iii
Abstract iv
Tables vii
1. Introduction 1
2. Conclusions 2
3. Recommendations 3
4. Experimental 4
Collaborative study . 4
Treatment of data 17
Quality assurance/quality control 18
5. Results and discussion 20
Interlaboratory method performance 20
Quality assurance/quality control 55
References 72
Appendices
A. Proposed draft protocol for supercritical fluid extraction of petroleum
hydrocarbons (Method 3560) A-l
B. Proposed draft protocol for determination of total recoverable petroleum
hydrocarbons by infrared spectrophotometry (Method 8440) B-l
C. List of instructions for collaborators C-l
VI
-------
TABLES
Number Page
1 Samples used in the interlaboratory study 5
2 Identification of the SFE systems used in the collaborative study 6
3 SFE operating conditions for the Isco SFE systems 7
4 SFE operating conditions for the Dionex-Lee Scientific SFE-703 systems 8
5 SFE operating conditions for the Suprex Prepmaster SFE systems 9
6 SFE operating conditions for the Suprex SFE-50 system 10
7 SFE operating conditions for the CCS Instruments SFE system 11
8 SFE operating conditions for the HP 7680A SFE systems 12
9 Flow rate of carbon dioxide used by the participating laboratories 13
10 Exact weights of the samples extracted in the study 15
11 Final volumes of the extracts before dilution 16
12 Concentrations of TPHs determined in TPH-1 soil extracts using the PE-FTIR
spectrometer 21
13 Concentrations of TPHs determined in TPH-1 soil extracts using the BSci-IR
spectrometer 22
14 Concentrations of TPHs determined in TPH-2 soil extracts using the PE-FTIR
spectrometer 23
15 Concentrations of TPHs determined in TPH-2 soil extracts using the BSci-IR
spectrometer 24
16 Concentrations of TPHs determined in SRS103-100 soil extracts using the
PE-FTIR spectrometer 25
VII
-------
TABLES (continued)
Number Page
17 Concentrations of TPHs determined in SRS103-100 soil extracts using the
BSci-IR spectrometer 26
18 Concentrations of TPHs determined in spiked clay soil extracts using the
PE-FTIR spectrometer 27
19 Concentrations of TPHs determined in spiked clay soil extracts using the BSci-IR
spectrometer 28
20 Summary of results from all laboratories (before outlier removal) 29
21 Summary of results from all laboratories (after outlier removal) 30
22 Summary of results from laboratories using Isco systems (before outlier
removal) 31
23 Summary of results from laboratories using Isco systems (after outlier
removal) 32
24 Summary of results from laboratories using extraction vessels 3.5 mL or less in
volume (before outlier removal) 33
25 Summary of results from laboratories using extraction vessels 3.5 mL or less in
volume (after outlier removal) 34
26 Outlier testing—the Cochran statistic, the single Grubbs statistic, and the double
Grubbs statistic 35
27 One-way ANOVA for the TPH-1 soil samples extracted by all laboratories 38
28 One-way ANOVA for the TPH-2 soil samples extracted by all laboratories 39
29 One-way ANOVA for the SRS 103-100 soil samples extracted by all
laboratories 40
30 One-way ANOVA for the spiked clay soil samples extracted by all laboratories . . 41
31 One-way ANOVA for the TPH-1 soil samples extracted by laboratories 05, 13,
14, 15, and 17 . . . 42
32 One-way ANOVA for the TPH-2 soil samples extracted by laboratories 05, 13,
14, 15, and 17 43
via
-------
TABLES (continued)
Number Page
33 One-way ANOVA for the SRS103-100 soil samples extracted by laboratories 05,
13, 14, 15, and 17 44
34 One-way ANOVA for the spiked clay soil samples extracted by laboratories 05,
13, 14, 15, and 17 45
35 One-way ANOVA for the TPH-1 soil samples extracted by laboratories 01, 04,
06, 08, 13, 14, and 17 46
36 One-way ANOVA for the TPH-2 soil samples extracted by laboratories 01, 04,
06, 08, 13, 14, and 17 . 47
37 One-way ANOVA for the SRS 103-100 soil samples extracted by laboratories 01,
04, 06, 08, 13, and 14 48
38 One-way ANOVA for the spiked clay soil samples extracted by laboratories 01,
04, 06, 08, 13, 14, and 17 49
39 Recoveries of oil/grease and TPHs determined in extracts of clay soil spiked with
corn oil and reference oil using the PE-FTIR spectrometer 51
40 Recoveries of oil/grease and TPHs determined in extracts of clay soil spiked with
corn oil and reference oil using the BSci-IR spectrometer 52
41 Concentrations and percent recoveries of TPHs determined in extracts of wet clay
soil spiked with motor oil using the PE-FTIR spectrometer 53
42 Concentrations and percent recoveries of TPHs determined in extracts of wet clay
soil spiked with motor oil using the BSci-IR spectrometer 54
43 Percent recoveries of TPHs determined in extracts of wet clay soil spiked with
motor oil using the PE-FTIR spectrometer . 56
44 Percent recoveries of TPHs determined in extracts of wet clay soil spiked with
motor oil or diesel oil using die PE-FTIR spectrometer 56
45 Correlation between the PE-FTIR data and the BSci-IR data 57
46 Calibration data for the PE-FTIR spectrometer using reference oil in PCE and the
10-mm path-length IR cell 58
IX
-------
TABLES (concluded)
/
Number Page
47 Calibration data for the BSci-IR spectrometer using reference oil in PCE and the
10-mm path-length IR cell 60
48 Calibration data for the PE-FTIR spectrometer using motor oil in PCE and the
10-mm path-length IR cell 62
49 Calibration data for the BSci-IR spectrometer using motor oil in PCE and the
10-mm path-length IR cell ; . . 63
50 Calibration data for the PE-FTIR spectrometer using reference oil in PCE and the
10-mm path-length IR cell (no silica gel cleanup) 64
51 Calibration data for the BSci-IR spectrometer using reference oil in PCE and the
10-mm path-length IR cell (no silica gel cleanup) 65
52 Concentrations of daily standards analyzed to verify system reproducibility 66
53 Concentrations of TPHs in the extracts from the unspiked clay soil samples 67
54 Concentrations of TPHs in the PCE blanks analyzed during this study 68
55 Concentrations of TPHs in the extracts submitted as system blanks 69
56 Percent recoveries of TPHs from the spiked clay soil samples, stored at 4°C in
the dark, as a function of time 69
-------
SECTION 1
INTRODUCTION
Widespread use of chlorofluorocarbons and other harmful organic solvents in environmental
testing laboratories is of major concern to the U.S. Environmental Protection Agency (EPA).
Therefore, EPA's Environmental Monitoring Systems Laboratory in Las Vegas, NV, started the
development of supercritical fluid extraction (SFE) methods that replace those harmful organic solvents
with supercritical carbon dioxide or carbon dioxide modified with small amounts of methanol or other
innocuous organic solvents as extractants. The first SFE method, SW-846 Method 3560, was recently
proposed (1, 2). The proposed method describes extraction of petroleum hydrocarbons with
supercritical carbon dioxide at 340 atm and 80°C for 30 min (dynamic), with a carbon dioxide flow
rate of 1 to 2 mL/min as liquid, or 500 to 1,000 mL/min as decompressed gas. Depending on the
extraction system used, the stream of carbon dioxide containing the extracted materials passes either
through a collection vial containing 3 mL tetrachloroethylene (also known as perchloroethylene, PCE),
or through an adsorbent trap. The technique is simple, requiring approximately 30 min for the
extraction of a 2- to 5-g solid sample, and 10 min for extract cleanup (to remove polar organic
compounds) and analysis by infrared (IR) spectrometry. The results of a single-laboratory evaluation
of this SFE method indicated that its performance is equivalent to the Soxhlet extraction method using
Freon-113 as extractant, its accuracy is 80 percent or better, and its precision is + 20 percent (1, 2).
A collaborative study of Method 3560 was conducted by 14 laboratories, with Midwest
Research Institute, California Operations (MRI-CO), as the lead laboratory, and with 13 additional
volunteer laboratories participating. The goal of this study was to determine the method accuracy and
precision of draft Method 3560 (Supercritical Fluid Extraction of Petroleum Hydrocarbons) when used
at different laboratories using instrumentation from different manufacturers.
The criteria for selecting the laboratories included availability to the laboratories of
commercial SFE systems that could be clearly described by the participants, willingness to perform
the extractions within a month after sample receipt, previous experience of the prospective
collaborators with the SFE technique in general, and willingness to participate as a volunteer.
Subsequent sections of this report present the conclusions and recommendations from this
study, the experimental details, and the results. The proposed draft protocols for SFE (Method 3560)
and for the total recoverable hydrocarbon determination by IR (Method 8440) are included as
Appendices A and B, respectively. The list of instructions sent to the collaborating laboratories is
included as Appendix C.
-------
SECTION 2
CONCLUSIONS
The results of this collaborative study, conducted with 14 laboratories, indicate that the
proposed EPA SW-846 Method 3560 (when used with EPA Method 8440) has an accuracy of
82.9 percent (this is the overall method recovery of petroleum hydrocarbons from samples containing
from 614 to 32,600 mg/kg TPHs). The interlaboratory method precisions ranged from 17.3 to
45.4 percent relative standard deviation for PE-FTIR analyses and from 16.7 to 47.9 percent for BSci-
IR analyses; the intralaboratory method precisions ranged from 11.5 to 17 percent for PE-FTIR
analyses and from 11.1 to 18.2 percent for BSci-IR analyses.
The matrices used in this study were three standard reference soils and clay soil spiked with
TPHs; the TPH recoveries for these samples, which were homogeneous and dry (with the exception
of the clay soil that had a water content of 10.6 percent), were greater than 75 percent. However,
when the clay soil was mixed with additional water to bring its water content to 30 percent, followed
by extraction by SFE, less than 38 percent recovery was achieved. Results from additional
experiments performed with anhydrous magnesium sulfate and diatomaceous earth as drying agents
indicated that samples containing 20 percent water or more need to be mixed with anhydrous
magnesium sulfate and allowed to equilibrate for several hours (preferably overnight in sealed
containers and at 4°C to minimize losses of volatile petroleum hydrocarbons) before extraction by
SFE.
Contamination of the SFE system is likely, especially when high-contamination samples are
extracted; the analyst must take precautions to minimize coss-contamination of extracts.
The participating laboratories generally stayed within the guidelines of the instructions
provided by MRI-CO; however, variations in the carbon dioxide flow rate, the extraction vessel
dimensions, design, and orientation, the mode of collection of the extracted material, and the
temperature of the collection solvent/trap were unavoidable because the various SFE systems were not
identical in design. Therefore, it is not surprising that the variation from laboratory to laboratory was
greater than that attributed to the analytical error displayed within laboratories. Nonetheless, the
participating laboratories did a good job; when the data were subjected to a software program from
the Association of Official Analytical Chemists (AOAC), data from only one laboratory were rejected
on two of the matrices.
-------
SECTION 3
RECOMMENDATIONS
On the basis of this interlaboratory method validation study, the Office of Solid Waste may
want to consider Method 3560 for incorporation in the SW-846 methods manual.
The specific recommendations from the participating laboratories are listed below:
• The temperature of the collection solvent should be kept at 0 to 5°C to minimize losses
of volatile organics.
• Laboratory 02 reported that all glass wool in their laboratory was contaminated with
petroleum hydrocarbons. Therefore, they recommended that glass wool be washed and
dried in a muffle furnace before use.
• Laboratory 02 recommended that glass wool be replaced with a drying agent
(magnesium sulfate, diatomaceous earth) to "protect" the frits and fill the void volume
of the extraction vessel.
• Laboratory 13 recommended that precleaned sand be used to fill the void volume of
the extraction vessel.
• Laboratories 05 and 13 recommended that the direction of carbon dioxide flow be
recorded. Furthermore, they stated that the carbon dioxide flow from top to bottom
of the extraction vessel is superior to flow from bottom to top when the extraction cell
is not full.
• The forms used by the laboratories in reporting the operating conditions (Appendix C)
were very helpful, and we^recommend that these forms be used to record the SFE
conditions.
The method presented here is not suitable for the extraction of hydrocarbons from gasoline-
contaminated soil samples because of poor collection efficiencies of the volatile hydrocarbons present
in gasoline. Improvements in the collection method, such as trapping onto an adsorbent trap held at
-10°C, followed by rinsing of the trap with an organic solvent (PCE) and analysis by gas
chromatography with infrared detection, is recommended for future studies.
-------
SECTION 4
EXPERIMENTAL
COLLABORATIVE STUDY
Study Design
The study was based on the AOAC's blind replicate design with balanced replicates for the
collaborative evaluation of precision and accuracy of an analytical method (3,4). In our study, four
soil samples were to be extracted in triplicate. Samples 1, 2, and 3 were standard reference materials
that had been certified for TPHs by the modified Method 418.1(5); their TPH contents were 614,
2050, and 32,600 mg/kg, respectively. The fourth sample was a clay soil that we spiked with motor
oil at 10,000 mg/kg. The samples, including their code numbers and TPH levels, are described in
Table 1.
Each laboratory received 10-g portions from matrices 1, 2, and 3, and three 3-g portions of
clay soil spiked with motor oil at 10,000 mg/kg. The spiked clay soil samples were identified as
samples 4, 6, and 7. In addition, they received the unspiked clay soil, identified as sample 5, and clay
soil samples spiked with corn oil/reference oil and motor oil (identified as samples 8 and 9
respectively). The corn oil was used to simulate lipids and similar materials; the reference oil is a
mixture of n-hexadecane, isooctane, and chlorobenzene. To sample 9 we added water (after spiking
with the motor oil) to bring its water content to 30 percent. A total of 15 extractions were to be
performed by each laboratory. The laboratories were instructed to extract three 3-g portions from
each of matrices 1, 2, and 3, and all of the material they received for samples 4 through 9.
Sets of test samples were sent to 17 laboratories; 14 laboratories submitted extracts within
the time frame specified by MRI-CO. The laboratories were to perform the extractions according to
the instructions given in Appendix C and then ship the extracts to MRI-CO for IR analysis. To
minimize errors due to reagent contamination, each laboratory was given tetrachloroethylene (PCE)
for use as collection solvent, and anhydrous sodium sulfate that was to be added to the water-
containing spiked clay sample. In addition, we made arrangements to provide each laboratory with
the SFE-grade carbon dioxide from Scott Specialty Gases. However, due to delays in the arrival of
the carbon dioxide, laboratories 05, 12, 15, and 17 used their own SFE-grade carbon dioxide (Air
Products).
Apparatus
The SFE systems used by the laboratories in this study are identified in Table 2. The actual
operating conditions, as reported by the laboratories, are given in Tables 3 through 8. The flow rates
of the carbon dioxide reported by the participating laboratories are given in Table 9. The exact
-------
TABLE 1. SAMPLES USED IN THE INTERLABORATORY STUDY'
Sample
no.
1A, B, C
2A, B, C
3A, B, C
4
Sample
identification
TPH-1 soil (Lot 91017)
without fatty acids
TPH-2soil(Lot91017)
with fatty acids
SRS 103-100 soil
Clay soil spiked with
TPH
concentration
(mg/kg)
614"
2,050"
32,600°
10,000d
Source
Environmental Research
Associates, Arvada, CO
Environmental Research
Associates, Arvada, CO
Fisher Scientific,
Pittsburgh, PA
MRI-CO6
5
6
motor oil
Clay soil (unspiked)
Clay soil spiked with 10,000d
motor oil
Clay soil spiked with 10,000d
motor oil
Clay soil spiked with
corn oil and reference oil l,000d
Clay soil spiked with 10,000d
motor oil and water
(approximately 30 percent
by weight)
MRI-CO
MRI-CO
MRI-CO
MRI-CO
MRI-CO
a Samples 1, 2, and 3 were extracted in triplicate by each laboratory. The three replicates were
identified as A, B, and C. Only one extraction was performed for samples 4 through 9.
Approximately 3 g of sample was extracted in each case. The exact amounts extracted by each
laboratory are given in Table 10. The final volumes of the extracts before dilution are given in
Table 11.
b Certified value.
c Determined by Soxhlet extraction with Freon-113 and analysis by IR spectrometry. Average of
duplicate determinations.
d Spike value.
° Midwest Research Institute, California Operations.
-------
TABLE 2. IDENTIFICATION OF THE SFE SYSTEMS USED IN THE COLLABORATIVE
STUDY
Laboratory code
SFE system identification
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Suprex SFE-50
Suprex Prepmaster
CCS Instruments SFE
Dionex-Lee Scientific SFE 703
Isco SFE System 1200
Suprex Prepmaster
Dionex-Lee Scientific SFE 703
HP 7680A
HP 7680A
HP 7680A
Isco SFE System 1200
Isco SFE System 1200
Isco SFE System 1200
Isco SFE System 1200
-------
TABLE 3. SFE OPERATING CONDITIONS FOR THE ISCO SFE SYSTEMS
Parameter
CO2 Pressure (atm)
CO2 Density (g/mL)
CO2 Flow rate (mL/min)
(see Table 9)
Oven temperature (°C)
Extraction time (min)
Extraction vessel volume (mL)
Extraction vessel dimensions
56-mm length
Extraction vessel orientation
Res trie tor dimensions
35-cm length
Restrictor temperature (°C)
Collection solvent
Volume of solvent (mL)
Temperature of collection
05
340
0.785
Varied
(see Table 9)
80
30
10
15-mm ID x
56-mm length
• Vertical, down flow
50-^m ID x
1 1-cm length
Not heated
PCE
3
Ambient
temperature
13
340
0.785
Varied
(see Table 9)
80
30
2.5
7.5-mm ID x
56-mm length
Vertical, down flow
32-/tm ID x
50-cm length
80
PCE
3
Ambient
temperature
14
340
0.785
Varied
(see Table 9)
80
30
2.5
7.5-mm ID x
56-mm length
Vertical, down flow
50-/tm ID x
37.5-cm length
Not heated
PCE
3
Ambient
temperature
15
340
0.785
Varied
(see Table 9)
80
30
10
15-mm ID x
56-mm length
Vertical, down flow
50-nm ID x
60/70-cm length"
Not heated
PCE
3
Ambient
temperature
17
340
0.785
Varied
80
30
2.5 and
7.5/15-mm ID
10"
x
Vertical, down flow
50-jim ID x
Not heated
PCE
3
Ambient
temperature'
" The 10-mL vessel was used for samples 3A, 3B, 3C, and 9.
b A 60-cm length restrictor was used with the 10-mL extraction vessel; a 70-cm length restrictor was used with the 2.5-mL extraction vessel.
0 Initial temperature was room temperature; however, no attempt was made to control the collection solvent temperature during SFE.
-------
TABLE 4. SFE OPERATING CONDITIONS FOR THE DIONEX-LEE SCIENTIFIC
SFE-703 SYSTEMS
Parameter
04
08
C02 Pressure (atm)
C02 Density (g/mL)
CO, Flow rate (mL/min)
Oven temperature (°C)
Extraction time (min)
Extraction vessel volume (mL)
Extraction vessel dimensions
Extraction vessel orientation
Restrictor dimensions
Restrictor temperature (°C)
Collection solvent
Volume of solvent (mL)
Temperature of collection vial (°C)
340
0.785
Varied (see
Table 9)
80
30
3.5
9.4-mm ID x
50-mm length
Horizontal
a
150
PCE
5
2
340
0.785
Varied (see
Table 9)
80
30
3.5
9.4-mm ID x
50-mm length
Horizontal
a
150
PCE
5
2
a Restrictor identified as "250" restrictor.
-------
TABLE 5. SFE OPERATING CONDITIONS FOR THE SUPREX PREPMASTER SFE
SYSTEMS
Parameter 02 06
CO2 Pressure (atm)
CO2 Density (g/mL)
C02 Flow rate (mL/min)
Oven temperature (°C)
Extraction time (min)
Extraction vessel volume (mL)
Extraction vessel dimensions
Extraction vessel orientation
Restrictor dimensions
Restrictor temperature (°C)
Collection solvent
Volume of solvent (mL)
Temperature of collection vial (°C)
340
0.785
1.2
80
30
5
9.5-mm ID x
65-mm length
Vertical, up flow
a
100
PCE
3
Ambient
temperature1"
340
0.785
Varied (see
Table 9)
80
30
3
10-mm ID x
38-mm length
Vertical, up flow
50-jim ID x
10-cm length
Not heated
PCE
3
Ambient
temperature
a A prototype VariFlow restrictor (variable) was used.
b Initial temperature was ambient temperature; however, no attempt was
made to control the collection solvent temperature during SFE.
-------
TABLE 6. SFE OPERATING CONDITIONS FOR THE SUPREX SFE-50 SYSTEM
Parameter
01
C02 Pressure (atm)
C02 Density (g/mL)
C02 Flow rate (mL/min)
Oven temperature (°C)
Extraction time (min)
Extraction vessel volume (mL)
Extraction vessel dimensions
Extraction vessel orientation
Restrictor dimensions
Restrictor temperature (°C)
Collection solvent
Volume of solvent (mL)
Temperature of collection vial (°C)
340
0.785
Varied (see
Table 9)
80
30
3.5
9.4-mm ID x
50-mm length
Horizontal
50-/itn ID x
60-cm length
Not heated
PCE
3
Ambient
temperature3
Initial temperature was ambient temperature; however,
no attempt was made to control the collection solvent
temperature during SFE.
10
-------
TABLE 7. SFE OPERATING CONDITIONS FOR THE CCS INSTRUMENTS SFE
SYSTEM
Parameter 03
C02 Pressure (atm) 340
C02 Density (g/mL) . 0.785
C02 Flow rate (mL/min) 0.2
Oven temperature (°C) 80
Extraction time (min) 30
Extraction vessel volume (mL) 6.0
Extraction vessel dimensions 9.5-mm ID x
127-mm length
Extraction vessel orientation Vertical, up flow
Restrictor dimensions 20-/*m ID x
5-cm length
Restrictor temperature (°C) 80 to 100
Collection solvent PCE
Volume of solvent (mL) 3
Temperature of collection vial (°C) Ambient
temperature3
3 Initial temperature was ambient temperature; however,
no attempt was made to control the collection solvent
temperature during SFE.
11
-------
TABLE 8. SFE OPERATING CONDITIONS FOR THE HP 7680A SFE SYSTEMS
Parameter
CO, Pressure (atm)
CO2 Density (g/mL)
C02 Flow rate (mL/min)
Oven temperature (°C)
Extraction time (min)
Extraction vessel volume (mL)
Extraction vessel dimensions
Volumes swept (mL)
Extraction vessel orientation
Nozzle temperature (°C)
Trap temperature (°C)
Trap packing material
Rinse solvent
Volume
Rate (mL/min)
Nozzle, temperature
during rinse (°C)
Trap temperature
during rinse (°C)
Number of rinses
10
340
0.78
2.0
80
30
7.0
10-mm ID x
90-mm length
10.2
Vertical, up flow
50
-15
Stainless-
steel beads
PCE
1.2
1.0
40
40
2
11
370
0.80
2.0
80
30
7.0
10-mm ID x
90-mm length
9.9
Vertical, up flow
45
10
Stainless-
steel beads
PCE
1.5
1.0
45
60
2
12
370
0.80
2.0
80
20
7.0
10-mm ID x
90-mm length
6.6
Vertical, up flow
45
10
ODSa
PCE
1.5
1.0
45
40
2
ODS - octadecylsilyl-bonded silica.
12
-------
TABLE 9. FLOW RATE (mL/min) OF CARBON DIOXIDE (AS LIQUID) USED BY THE PARTICIPATING
LABORATORIES
Sample
ID
1A
IB
1C
2A
2B
2C
3A
3B
3C
4
5
6
7
8
9
01
.2
.2
.1
.0
.2
.0
0.6
0.5
0.9
0.9
0.8
1.1
1.0
0.6
0.9
02
.2
.2
.2
.2
.2
.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
03
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
04 (
0.7
0.7
0.6 :
0.7
0.6
0.6
0.5
0.6
0.6
0.6
0.7
0.6
0.5
0.7
0.6
)5
1.9
1.7
>.o
1.9
1.6
.9
.5
.5
.7
.7
.7
.2
.5
.3
.4
06
0.9
0.7
0.5
1.2
1.2
0.6
2.2
1.2
0.9
1.5
1.5
1.0
2.2
2.4
2.4
08
0.8
0.7
0.6
0.7
0.7
0.6
0.7
0.6
0.6
0.4
0.3
0.6
0.8
0.6
0.3
10
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
11
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
12
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
13 14
0.8
0.8
0.9
0.6
0.5
0.6
0.6
0.6
0.6
0.7
0.6
0.7
0.9
0.8
0.8
15
2.4
2.3
2.3
2.2
2.2
2.0
2.0
2.0
.1.5
2.1
2.0
2.2
2.4
2.1
2.2
17
1.6
1.6
1.9
1.6
1.6
1.2
0.8
0.8
0.8
1.2
1.8
1.8
1.0
1.0
0.8
Reported as 1 to 2 mL/min but not measured for each extraction.
-------
weights of the samples extracted by the individual laboratories are listed in Table 10, and the volumes
of the extracts delivered to MRI-CO are given in Table 11.
Two infrared spectrometers were used: a Perkin-Elmer Corporation (Norwalk, CT) Model
1605 FTIR, interfaced with a Digital Equipment Corporation (Cupertino, CA) DEC 386SX computer
(scanning from 3200 to 2700 cm'1), and a Buck-Scientific IR filter spectrometer Model 404, centered
on 2924 cm'1 with a band of 15 cm'1 on each side (E. Norwalk, CT).
Materials
The reference oil standards used in the IR determination of TPHs were prepared as follows:
Pipet 15 mL n-hexadecane, 15 mL isooctane, and 10 mL chlorobenzene into a 50-mL glass-stoppered
flask. Mix by swirling the contents. Pipet 0.5 mL of this mixture into a tared 100-mL volumetric
flask, weigh to the nearest milligram, and dilute to volume with PCE. Pipet appropriate volumes of
this stock standard solution into 100-mL volumetric flasks and dilute to volume with PCE to make
10-/xg/mL, 25-/ig/mL, 50-jtg/mL, 100-/ig/mL, 250-jig/mL, and 500-/,ig/mL calibration standards. The
motor oil calibration standards at 10 jtg/mL, 25 /xg/mL, 50 ng/mL, 100 /xg/mL, 250 //.g/mL, and
500 /xg/mL were prepared by serial dilution of a stock solution made at 200 /ig//*L in PCE. Corn oil
at 40 mg/mL in Freon-113 was used to spike the clay soil.
The following materials were used in the study:
n-Hexadecane, isooctane, chlorobenzene (Aldrich Chemical, Milwaukee, WI).
Corn oil (local supermarket).
Motor oil (local Shell gas station).
PCE, glass-distilled, HPLC grade, lot no. 06626TX (Aldrich Chemical).
Silica gel, 70 to 230 mesh, ASTM (Baxter Scientific Products, McGaw Park, IL).
Anhydrous sodium sulfate, analytical reagent (Mallinckrodt, St. Louis, MO).
Anhydrous magnesium sulfate, analytical reagent (Mallinckrodt).
Hydromatrix (diatomaceous earth) (Isco, Inc., Lincoln, NB).
Clay soil (Sandoz Experimental Station, San Jose, CA) (33.6% sand, 35.4% silt, 31% clay;
organic carbon 1.8%, water 10.6%).
Carbon dioxide, SFC/SFE-grade (Air Products, Allentown, PA).
Carbon dioxide, SFE-grade (Scott Specialty Gases, Inc., Plumsteadville, PA).
The standard reference materials used are listed in Table 1.
14
-------
TABLE 10. WEIGHTS (g) OF THE SAMPLES EXTRACTED IN THE STUDY
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
1A
3.03
3.00
3.00
3.00
3.03
3.00
3.09
3.12
3.00
3.09
3.01
3.00
3.00
3.00
IB
3.00
3.00
3.00
3.00
3.00
3.00
3.02
3.07
3.00
3.07
2.98
3.00
3.00
3.00
1C
3.03
3.00
3.00
3.00
3.01
3.00
3.03
2.99
3.00
3.10
3.00
3.00
3.00
3.00
2A
3.00
3.00
3.00
3.00
3.00
3.01
3.02
3.01
3.00
3.03
3.01
3.00
3.00
3.01
2B
3.03
3.00
3.00
3.00
3.01
3.00
3.01
3.00
3.00
3.10
3.00
3.00
3.00
3.00
2C
3.04
3.00
3.00
3.00
3.01
3.01
3.07
3.00
3.00
3.06
3.00
3.00
3.00
3.00
3A
2.98
3.00
3.00
3.00
3.06
3.00
3.11
3.02
3.00
3.03
2.99
3.00
3.00
3.02
3B
3.07
3.00
3.00
3.00
3.01
3.00
2.98
3.00
3.00
3.10
3.03
3.00
3.00
3.03
3C
3.05
3.00
a
3.00
3.05
3.00
3.02
3.03
3.00
b
3.07
1.24
3.00
3.01
4
3.02
3.00
3.00
3.02
3.01
3.00
3.01
2.99
2.99
3.01
3.01
3.00
3.00
3.00
5
3.00
3.00
3.00
2.98
2.99
2.99
3.00
2.99
3.00
2.98
2.99
3.00
3.00
2.99
6
3.00
3.00
3.00
3.00
3.01
3.00
2.99
2.98
3.00
3.00
3.01
2.70
3.00
3.00
7
3.02
3.00
3.00
3.01
3.01
3.00
3.00
3.00
3.01
3.00
3.02
3.00
3.00
3.00
8
3.01
3.00
3.00
2.99
3.00
3.07
2.99
2.99
3.03
2.99
3.00
2.50
3.00
3.04
9
1.62
2.40
3.00
4.29
3.01
2.91
2.78
2.52
2.34
2.48
2.86
3.00
2.55
2.95
a The extraction was not performed.
b The extract was not submitted because the SFE system developed a leak.
-------
TABLE 11. FINAL VOLUMES OF THE EXTRACTS (mL) BEFORE DILUTION
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
1A
3.8
3.0
3.0
4.5
3.0
3.0
4.6
3.0
3.4
3.2
3.0
3.0
3.0
3.0
IB
3.0
3.0
3.0
4.5
3.0
3.0
4.7
3.0
3.4
3.2
3.0
3.0
3.0
3.0
1C
3.0
3.0
3.0
4.5
3.0
3.0
4.7
3.0
3.4
3.2
3.0
3.0
3.0
3.0
2A
3.0
3.0
3.0
4.5
3.0
3.0
4.7
3.0
3.3
3.2
3.0
3.0
3.3
3.0
2B
3.0
3.0
3.0
4.5
3.0
3.0
4.7
3.0
3.3
3.2
3.0
3.0
3.0
3.0
2C
3.0
3.0
3.0
4.5
3.0
3.0
4.7
3.0
3.4
3.2
3.0
3.0
3.2
3.0
3A
3.0
3.0
3.0
4.5
3.0
3.0
4.8
3.9
3.6
3.3
3.0
3.0
3.0
3.0
3B
3.0
3.0
3.0
4.6
3.0
3.0
4.7
3.8
3.6
3.4
3.0
3.0
3.0
3.0
3C
3.0
3.0
a
4.6
3.0
3.0
4.7
3.0
3.6
a
3.0
3.0
3.0
3.0
4
3.0
3.0
3.0
4.5
3.1
3.0
4.6
3.0
3.5
3.3
3.0
3.0
3.0
3.0
5
3.0
3.0
3.0
4.4
3.0
3.0
a
3.0
3.0
3.3
3.0
3.0
3.0
3.0
6
3.0
3.0
3.0
4.5
3.0
3.0
4.6
3.2
3.5
3.3
3.0
3.0
3.0
3.0
7
3.0
3.0
3.0
4.6
3.0
3.0
4.6
3.1
3.5
3.4
3.0
3.0
3.0
3.0
8
3.0
3.0
3.0
4.5
3.4
3.0
4.6
3.0
3.3
3.3
3.0
3.0
3.0
3.0
9
3.0
3.0
3.1
4.4
3.0
3.0
4.8
3.0
3.4
3.4
3.0
3.0
3.0
3.0
The extract was not submitted for analysis.
-------
Sample Spiking Procedure
For spiking of the clay soil samples with motor oil, 3.0-g portions of the soil were weighed
into 7-mL glass vials, and portions (150 /*L) of concentrated stock solution containing motor oil in
Freon-113 at 200 /xg//*L were added with a syringe while ensuring .that the solution did not contact
the walls of the vials. Mixing was performed with the tip of a disposable glass pipet. After the
solvent had completely evaporated (approximately 15 min), the vials with the spiked samples were
sealed with Teflon-lined caps and stored at 4°C for 5 days prior to shipping to the laboratories.
Spiking with corn oil and reference oil was done as described above for the motor oil, except that the
concentrations of the stock solutions were 20 /*g//xL. The clay soil samples that required adjustment
of their water content to 30 percent were also spiked individually; a 2.4-g portion of the clay soil
sample was weighed into a glass vial, spiked, and then 0.6 mL water was added.
Sample Extraction
The methods used to extract the samples and to analyze the extracts are included in
Appendices A and B to this report. Each laboratory used only one analyst to perform sample
extractions and allowed one block of time for all sample extractions.
TREATMENT OF DATA
Outlier Testing
Outlier testing was done using both the Cochran test and the Grubbs test (3, 4). The former
is used for the removal of results from laboratories that show significantly greater variability among
replicate (intralaboratory) analyses than the other laboratories for a given material. To calculate the
Cochran test statistic, the intralaboratory variance for each laboratory was computed, and the largest
of these values was divided by the sum of all these variances. When a laboratory was rejected on the
basis of these tests, its results were removed from the set of data for a particular matrix, and the test
was repeated using the remaining data in the subset.
Grubbs tests were performed to remove results from laboratories with extreme averages. The
single-value test (2-tail; P = 0.01) was run first; then, when no outlier was found, the pair value test
(two values at the highest end, two values at the lowest end, and two values, one at each end, at an
overall P = 0.01) was applied.
To calculate the single Grubbs test statistic, the average for each of the L laboratories was
computed, and the standard deviation of these L averages (designated as the original s) was calculated.
The standard deviation of the set of averages with the highest average removed (SH) was calculated,
then the standard deviation of the set of averages with the lowest average removed (sj was calculated.
Finally, the percentage decrease in standard deviation was calculated as follows:
100 x [1 - (SL/S)] and 100 x [l-(sH/s)]
The higher of these two percentage decreases was the single Grubbs statistic, which signals the
presence of an outlier to be omitted if it exceeds the critical value listed in the single Grubbs tables,
at the P = 0.01 level, 2-tail, for L laboratories.
17
-------
To calculate the Grubbs pair statistic, we proceeded in an analogous fashion, except that we
calculated the standard deviations, from the original set of averages. The smallest of these three
standard deviation values was taken and the corresponding percentage in standard deviation from the
original s was calculated. A Grubbs outlier pair was present if the selected value for the percentage
decrease from the original s exceeded the critical value listed in the Grubbs pair value table, at the
P = 0.01 level, for L laboratories.
Statistical Analysis
Statistical analysis was performed using the AOAC Lotus spreadsheet program developed for
the analysis of data from interlaboratory studies (4).
The summary statistics that are reported for each matrix include the mean concentration of
TPHs and the method precision (percent relative standard deviation). The repeatability relative
standard deviation (RSDr), which was determined from the repeatability standard deviation (sr) and
the mean concentration of a particular matrix, is an indication of the intralaboratory precision. The
reproducibility relative standard deviation (RSDR), which was determined from the reproducibility
standard deviation (SR) and the mean concentration of a particular matrix, is an indication of the
interlaboratory method precision.
QUALITY ASSURANCE/QUALITY CONTROL
The extractions of the study samples were performed according to the instructions provided
by MRI-CO to all collaborators (Appendix C). To minimize errors due to reagent contamination, we
provided each laboratory with the collection solvent, the extractant (carbon dioxide), and the drying
agent (anhydrous sodium sulfate). We made arrangements with Scott Specialty Gases to supply carbon
dioxide from the same lot to all participating laboratories. All but four laboratories used SFE-grade
carbon dioxide from Scott Specialty Gases to extract the test samples. Laboratories 05,12,15, and 17
experienced delay in delivery of the carbon dioxide and therefore used SFC/SFE-grade carbon dioxide
from Air Products. Each laboratory was instructed to report the SFE operating conditions on special
forms (provided by MRI-CO), and to record the exact mass of the study sample extracted by SFE.
The IR analyses were performed on two different IR systems according to the proposed
Method 8440 included in Appendix B. The following quality control procedures were implemented:
• A six-level calibration (at 10, 25, 50, 100, 250, and 500 /ig/mL) was performed every
day, on each instrument, during the period the IR analyses were performed. The
multilevel calibration was then verified after every 10 analyses by analyzing a reference
oil standard or a motor oil standard at 100 /ig/mL. The reference oil standard was used
to quantify TPHs in samples 1, 2, 3, 5, and 8. The motor oil standard was used to
quantify TPHs in samples 4, 6, 7, and 9. Corn oil was quantified against the reference
oil standard.
• PCE blanks were analyzed daily on each instrument during the period the IR analyses
were performed.
• All system blanks received from the participating laboratories were analyzed for TPHs.
• Sample 5 was a blind QC sample (unspiked clay soil).
18
-------
At the MRI-CO, six clay soil samples were spiked with motor oil at 10,000 mg/kg and stored
at 4°C in the dark; two were extracted with supercritical carbon dioxide after 22 days, two after
34 days, and two after 40 days of storage. These six samples were part of the batch of samples
prepared for the collaborative study and were expressly set aside for extraction at a later time but still
within the time frame in which the collaborating laboratories would carry out their extractions (six
laboratories submitted extracts within 30 days of sample spiking, five within 40 days, and three within
60 days).
19
-------
SECTION 5
RESULTS AND DISCUSSION
INTERLABORATORY METHOD PERFORMANCE
The results of the interlaboratory study have been summarized by matrix and by the IR
spectrometer used in the analysis. As mentioned in Section 1, all IR determinations were performed
at MRI-CO on two different IR spectrometers. Tables 12 through 19 present the concentration data
for each of the four matrices that were extracted in triplicate (TPH-1 soil, TPH-2 soil, SRS103-100
soil, and the clay soil spiked with motor oil). The mean concentrations, the repeatability standard
deviations (sr), the reproducibility standard deviations (sj, die repeatability relative standard deviations
(RSDr), the reproducibility relative standard deviations (RSDR) and the mean recoveries have been
summarized in Tables 20 through 25.
Rejection of Outliers
For the entire study, the AOAC software program rejected only one laboratory
(laboratory 03) on two matrices. This laboratory achieved very low recoveries on all test samples
because the carbon dioxide flow rate was too low (0.2 mL/min); however, the outlier test only
rejected values on the TPH-2 soil extract and on the extract from the clay soil spiked with motor oil.
Table 26 summarizes the outlier testing results. When data from all laboratories were pooled, and
the outlier testing was performed by matrix, then laboratory 03 data for TPH-2 and spiked clay soil
samples were rejected on the basis of the single Grubbs test (lowest average). When data from five
laboratories using Isco systems were pooled, and outlier testing was performed by matrix, then
laboratory 14 data were rejected on the basis of the Cochran test (maximum intralaboratory variance).
When data from laboratories using extraction vessels of 3.5 mL or less in volume were pooled, and
outlier testing was performed by matrix, then data from laboratory 17 were rejected on the basis of
the Cochran test (maximum intralaboratory variance).
Method Recovery
The summary statistics for the recoveries from each of the four matrices that were extracted
by all laboratories, calculated after outlier removal, are presented in Table 21. The mean recoveries
of TPHs from these four matrices ranged from 77.9 to 107 percent for analyses performed with the
PE-FTIR spectrometer, and from 75.9 to 101 percent for analyses performed with the BSci-IR
spectrometer (a discussion of the correlation between the data generated with die two IR spectrometers
is given later in this section). These recoveries are in agreement with data that we reported previously
for the single-laboratory evaluation of this method (1, 2); the percent differences between the mean
recoveries for the interlaboratory study and the single-laboratory study were ranging from 1.8 to 28.7
percent.
20
-------
TABLE 12. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN TPH-1 SOIL
EXTRACTS USING THE PE-FTIR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
xt
618
449
67
352
943
250
737
693
972
691
673
639
883
575
Concentration
X2
808
453
NDb
391
859
478
799
1,070
1,210
701
604
605
963
712
(mg/kg)
X3
1,000
487
77
420
854
44
933
789
1,100
656
620
580
885
832
Mean
809
463
51.3
388
885
257
823
851
1,090
683
632
608
910
706
Percent
RSD
23.6
4.5
70.5
8.8
5.7
84.5
12.2
23.0
10.9
23.6
5.7
4.9
5.0
18.2
The certified concentration of TPHs in this sample is 614 mg/kg.
ND - not detected; the estimated detection limit is 10 mg/kg.
21
-------
TABLE 13. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN TPH-1 SOIL
EXTRACTS USING THE BSci-IR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
x,
634
402
62.0
339
922
230
751
590
959
621
594
563
859
497
Concentration
X2
793
393
NDk
386
831
435
821
1,060
1,190
625
535
532
944
639
(mg/kg)
X3
987
433
80.5
404
825
52.5
940
690
1,080
593
544
510
856
747
Mean
805
409
50.8
376
859
239
837
780
1,080
613
558
535
886
628
Percent
RSD
22.0
5.1
72.0
8.9
6.3
80.0
11.4
31.7
10.7
2.8
5.7
5.0
5.6
20.0
a The certified concentration of TPHs in this sample is 614 mg/kg.
b ND - not detected; the estimated detection limit is 10 mg/kg.
22
-------
TABLE 14. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN TPH-2 SOIL
EXTRACTS USING THE PE-FTIR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
x,
1,840
1,940
257
1,350
1,490
1,850
1,480
1,880
2,140
2,320
1,900
2,130
1,850
1,780
Concentration
X2
2,160
1,990
410
1,160
2,000
1,730
1,100
2,240
2,200
2,440
2,010
1,900
1,900
1,920
(mg/kg)
X3
1,880
1,730
356
1,520
2,230
1,380
1,500
1,980
2,050
2,110
1,940
1,340
1,990
1,980
Mean
1,960
1,890
341
1,340
1,910
1,650
1,360
2,030
2,130
2,290
1,950
1,790
1,910
1,890
Percent
RSD
8.9
7.3
22.8
13.4
19.8
14.8
16.7
9.2
3.5
7.3
2.9
22.7
3.7
5.4
The certified concentration of TPHs in this sample is 2,050 mg/kg.
23
-------
TABLE 15. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN TPH-2 SOIL
EXTRACTS USING THE BSci-IR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
x,
1,620
1,760
250
1,290
1,370
1,670
1,350
1,660
1,960
2,110
1,670
1,900
1,700
1,530
Concentration
X2
1,950
1,790
374
1,120
1,740
1,550
1,050
2,010
2,040
2,210
1,760
1,680
1,730
1,680
(mg/kg)
X3
1,660
1,540
380
1,460
1,980
1,240
1,330
1,770
1,910
1,780
1,710
1,180
1,820
1,730
Mean
1,740
1,700
335
1,290
1,700
1,490
1,240
1,810
1,970
2,030
1,710
1,590
1,750
1,650
Percent
RSD
10.4
8.0
21.9
13.2
18.1
14.9
13.5
9.9
3.3
11.1
2.6
23.2
3.6
6.3
The certified concentration of TPHs in this sample is 2,050 mg/kg.
24
-------
TABLE 16. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN SRS103-100 SOIL
EXTRACTS USING THE PE-FTIR SPECTROMETER4
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
x,
32,600
28,900
4,130
23,200
25,200
46,400
25,500
19,500
28,900
33,200
30,900
26,500
31,100
21,600
Concentration
X2
35,600
29,600
11,100
20,800
29,500
45,100
24,400
14,800
27,800
32,700
30,900
29,200
32,200
24,300
(mg/kg)
X3
33,700
29,500
b
25,000
10,800
54,600
24,400
17,500
28,300
b
31,500
11,400
30,800
22,600
Mean
34,000
29,300
7,620
23,000
21,800
48,700
24,800
17,300
28,300
32,900
31,100
22,400
31,400
22,800
Percent
RSD
4.5
12.9
64.7
9.2
44.9
10.6
2.6
13.6
1.9
1.1
1.1
42.8
2.3
6.0
The concentration of TPHs measured in our laboratory for this sample is 32,600 mg/kg.
This laboratory did not submit an extract for this sample.
25
-------
TABLE 17. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN SRS103-100 SOIL
EXTRACTS USING THE BSci-IR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
X,
29,600
27,000
3,890
22,700
22,400
42,400
23,600
18,300
26,600
30,500
27,900
24,500
28,300
20,600
Concentration
x.
32,200
27,800
10,200
20,600
25,800
41,100
23,200
14,000
25,100
30,300
27,600
26,800
29,400
22,800
(mg/kg)
X3
30,500
27,900
b
24,400
9,790
49,100
23,800
16,200
25,700
b
28,600
11,300
28,200
21,300
Mean
30,800
27,600
7,050
22,600
19,300
44,200
23,500
16,200
25,800
30,400
28,000
20,900
28,600
21,600
Percent
RSD
4.3
1.8
63.3
8.4
43.7
9.7
1.3
13.3
2.9
0.5
1.8
40.0
2.3
5.2
a The concentration of TPHs measured in our laboratory for this sample is 32,600 mg/kg.
b This laboratory did not submit an extract for this sample.
26
-------
TABLE 18. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN SPIKED CLAY SOIL
EXTRACTS USING THE PE-FTIR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
x,
8,760
7,850
580
7,000
8,160
6,240
7,110
6,470
9,690
8,660
8,420
5,000
8,560
7,990
Concentration
X2
10,400
8,740
368
7,590
8,790
6,810
6,130
4,310
8,380
8,810
8,210
7,910
7,660
7,940
(mg/kg)
x,
8,680
9,740
963
7,270
9,290
4,100
5,780
3,200
9,400
9,270
8,830
8,970
9,020
8,590
Mean
9,280
8,780
637
7,290
8,750
5,720
6,340
4,660
9,160
8,910
8,490
7,290
8,410
8,170
Percent
RSD
10.5
10.8
47.3
4.1
6.5
25.0
10.9
35.7
7.5
3.6
3.7
28.2
8.2
4.4
The clay soil samples were spiked with motor oil at 10,000 mg/kg.
27
-------
TABLE 19. CONCENTRATIONS (mg/kg) OF TPHs DETERMINED IN SPIKED CLAY
SOIL EXTRACTS USING THE BSci-IR SPECTROMETER"
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
x,
9,220
8,020
586
7,180
8,440
6,040
7,460
7,310
9,880
9,110
8,720
5,400
8,770
7,970
Concentration
X2
10,800
9,380
411
7,670
9,120
6,610
6,650
5,230
9,000
9,210
8,670
8,360
8,120
8,120
(mg/kg)
X3
8,870
9,980
954
7,270
9,480
6,760
6,150
4,130
9,770
9,310
9,020
9,220
9,680
8,920
Mean
9,630
9,130
650
7,370
9,010
6,470
6,750
5,560
9,550
9,210
8,800
7,660
8,860
8,340
Percent
RSD
10.7
11.0
42.6
3.5
5.9
5.9
9.8
29.0
5.0
1.1
2.2
26.2
8.8
6.1
The clay soil samples were spiked with motor oil at 10,000 mg/kg.
28
-------
TABLE 20. SUMMARY OF RESULTS FROM ALL LABORATORIES (BEFORE OUTLIER REMOVAL)
Matrix
PE-FTIR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
BSci-lR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
Certified or
spike value
(mg/kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
Mean
concentration"
(mg/kg)
654
1,750
26,820
7,280
618
1,570
24,750
7,640
sr
(mg/kg)
111
206
4,320
968
113
188
3,740
88
SR
(mg/kg)
297
509
9,720
2,490
296
446
8,650
2,470
Percent
RSDr
17.0
11.8
16.1
13.3
18.2
12.0
15.1
11.5
Percent
RSDR
45.4
29.2
36.2
34.2
47.9
28.4
35.0
32.4
Percent
mean
recovery
107
85.4
82.3
72.8
101
76.7
75.9
76.4
Number of
laboratories
in the study
14
14
14
14
14
14
14
14
* The number of replicates per laboratory was three.
-------
TABLE 21. SUMMARY OF RESULTS FROM ALL LABORATORIES (AFTER OUTLIER REMOVAL)
Matrix
PE-FTIR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100soil
Clay soil spiked with
motor oil
BSci-IR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
Certified or
spike value
(mg/kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
Mean
concentration"
(mg/kg)
654
1,850
26,820
7,790
618
1,670
24,750
8,180
sr
(mg/kg)
111
213
4,320
1,000
113
194
3,740
910
SR
(mg/kg)
297
321
9,720
1,660
296
278
8,650
1,500
Percent
RSDr
17.0
11.5
16.1
12.9
18.2
11.7
15.1
11.1
Percent
RSDR
45.4
17.3
36.2
21.3
47.9
16.7
35.0
18.3
Percent
mean
recovery
107
90.2
82.3
77.9
101
81.5
75.9
81.8
Number of
laboratories
retained
14
13
14
13
14
13
14
13
' The number of replicates per laboratory was three.
-------
TABLE 22. SUMMARY OF RESULTS FROM LABORATORIES USING ISCO SYSTEMS (BEFORE OUTLIER
REMOVAL)
Matrix
PE-FTIR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
BSci-IR
ERA TPH-1 soil
ERA TPH-2 soil
SRS 103-100 soil
Clay soil spiked with
motor oil
Certified or
spike value
(mg/kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
Mean
concentration" sr
(mg/kg) (mg/kg)
748
1,890
25,900
8,220
693
1,680
23,700
8,530
68.3
256
6,170
1,030
67.7
222
5,350
1,020
SR
(mg/kg)
152
256
7,020
1,030
177
222
6,150
1,020
Percent
RSD,
9.1
13.5
23.8
12.5
9.8
13.3
22.6
12.0
Percent
RSDR
20.3
13.5
27.1
12.5
25.5
13.3
26.0
12.0
Percent Number of
mean laboratories
recovery retained
122
92.2
79.4
82.2
113
82.0
72.7
85.3
5
5
5
5
5
5
5
5
" The number of replicates per laboratory was three.
-------
TABLE 23. SUMMARY OF RESULTS FROM LABORATORIES USING ISCO SYSTEMS (AFTER OUTLIER REMOVAL)
N>
Matrix
PE-FTIR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
BSci-IR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
Certified or
spike value
(mg/kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
Mean
concentration" sr
(mg/kg) (mg/kg)
748
1,890
25,900
8,460
693
1,680
23,700
8,530
68.3
256
6,170
507
67.7
222
5,350
1,020
SR
(mg/kg)
152
256
7,020
507
177
222
6,150
1,020
Percent
RSDr
9.1
13.5
23.8
6.0
9.8
13.3
22.6
12.0
Percent
RSDR
20.3
13.5
27.1
6.0
25.5
13.3
26.0
12.0
Percent Number of
mean laboratories
recovery retained
122
92.2
79.4
84.6
113
82.0
72.7
85.3
5
5
5
4
5
5
5
5
' The number of replicates per laboratory was three.
-------
u>
TABLE 24. SUMMARY OF RESULTS FROM LABORATORIES USING EXTRACTION VESSELS 3.5 mL OR LESS IN
VOLUME (BEFORE OUTLIER REMOVAL)
Matrix
PE-FTIR
ERA TPH-1 soil
ERA TPH-2 soil
SRS 103-100 soil
Clay soil spiked with
motor oil
BSci-IR
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
Certified or
spike value
(mg/kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
Mean
concentration* sr
(mg/kg) (mg/kg)
603
1,710
30,650
7,510
568
1,530
28,300
7,860
127
224
4,580
1,070
117
203
3,960
927
SR
(mg/kg)
236
322
10,660
1,520
236
259
9,190
1,360
Percent
RSDr
21.1
13.1
14.9
14.2
,
20.6
13.3
14.0
11.8
Percent
RSDR
39.1
18.9
34.8
20.2
41.4
16.9
32.4
17.3
Percent Number of
mean laboratories
recovery retained
98.2
83.4
94.0
75.1
92.5
74.6
86.8
78.6
7
7
6
7
7
7
6
7
" The number of replicates per laboratory was three.
-------
TABLE 25. SUMMARY OF RESULTS FROM LABORATORIES USING EXTRACTION VESSELS 3.5 mL OR LESS IN
VOLUME (AFTER OUTLIER REMOVAL)
Matrix
PE-FTER
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
BSci-ER
ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100 soil
Clay soil spiked with
motor oil
Certified or
spike value
(mg/kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
Mean
concentration" sr
(mg/kg) (mg/kg)
603
1,710
32,300
7,510
568
1,530
29,820
7,860
127
224
2,600
1,070
117
203
2,200
927
SR
(mg/kg)
236
322
10,420
1,520
236
259
8,890
1,360
Percent
RSDr
21.1
13.1
8.1
14.2
20.6
13.3
7.4
11.8
Percent
RSDR
39.1
18.9
32.3
20.2
41.4
16.9
29.8
17.3
Percent Number of
mean laboratories
recovery retained
98.2
83.4
99.1
75.1
92.5
74.6
91.5
78.6
7
7
5
7
7
7
5
7
* The number of replicates per laboratory was three.
-------
TABLE 26. OUTLIER TESTING—THE COCHRAN STATISTIC, THE SINGLE
GRUBBS STATISTIC, AND THE DOUBLE GRUBBS STATISTIC"
Instrument1"
Number of
Matrix laboratories
Cochran
test
Single
Grubbs test
Double
Grubbs test
All laboratories
1
2
1
2
1
2
1
2
Laboratories
1
2
1
2
1
2
1
2
Laboratories
1
2
1
2
1
2
1
2
TPH-1 soil
TPH-1 soil
TPH-2 soil
TPH-2 soil
SRS103-100 soil
SRS103-100 soil
Clay soil spiked
with motor oil
Clay soil spiked
with motor oil
using Isco SFE system
TPH-1 soil
TPH-1 soil
TPH-2 soil
TPH-2 soil
STS 103-100 soil
SRS103-100 soil
Clay soil spiked
with motor oil
Clay soil spiked
with motor oil
using extraction vessels
TPH-1 soil
TPH-1 soil
TPH-2 soil
TPH-2 soil
SRS103-100 soil
SRS103-100 soil
Clay soil spiked
with motor oil
Clay soil spiked
with motor oil
14
14
14
13
14
13
14
14
14
13
14
13
5
5
5
5
5
5
5
4
5
3.5 mL or less
7
7
7
7
6
5
6
5
7
7
0.2725
0.3448
0.2781
0.2810
0.2739
0.2769
0.3768
0.3699
0.3219
0.3242
0.3703
0.3729
0.7097
0.6868
0.5046
0.5504
0.5036
0.4973
0.8041°
0.4649
0.7715
in volume
0.4151
0.3833
0.4690
0.4732
0.7314°
0.7850
0.7434°
0.7630
0.5280
0.6672
17.86
15.27
43.77°
14.06
45.42°
13.30
22.58
22.01
38.91°
20.78
45.13°
16.96
11.29
11.62
59.63
32.38
9.95
11.24
57.72
30.57
47.04
24.18
19.07
12.88
15.40
48.02
49.23
52.24
55.74
15.71
"
20.64
29.75
25.12
51.67°
38.22
52.68°
31.38
44.00
45.07
51.60°
36.35
54.43°
28.01
63.77
71.21
83.00
57.00
89.75
73.54
70.64
78.03
80.15
53.41
34.84
51.37
48.30
60.03
58.26
64.41
66.48
31.75
34.53
The critical values of the Cochran statistic, the single Grubbs statistic, and the double Grubbs statistic
are given in Reference 3.
Instrument 1 is the PE-FTIR spectrometer; instrument 2 is the BSci-IR spectrometer.
This value is an outlier.
35
-------
We pooled all data generated with the five Isco SFE systems used in this study. Tables 22
and 23 present the summary statistics for these data before and after outlier removal. The method
recoveries ranged from 79.4 to 122 percent for PE-FTIR analyses and from 72.7 to 113 percent for
BSci-IR analyses. When the results from all laboratories using extraction vessels with a volume of
3.5 mL or less were pooled (seven laboratories), the mean recoveries ranged from 75.1 to
99.1 percent for PE-FTIR analyses and from 74.6 to 92.5 percent for BSci-IR analyses (Tables 24 and
25).
We performed a 2-sample t-test to determine whether the data in Tables 21, 23, and 25
(specifically, percent recoveries for each of the four matrices) are different from each other at the 5-
percent significance level. When comparing the data in Tables 21 and 23, we find significant
differences only for the clay soil matrix; when comparing the data in Tables 21 and 25, we find
significant differences only for the SRS103-100 soil matrix; and finally, when comparing the data in
Tables 23 and 25, we find significant differences for all matrices.
To determine the overall method recovery, we performed a linear regression of the data
presented in Tables 12, 14, 16, and 18 (after removal of outliers). The measured concentrations for
the PE-FTIR analyses were plotted on the y-axis, and the true concentrations of TPHs in the four
matrices were plotted on the x-axis. The slope of the regression equation was 0.8291, and the
intercept was -18.87. The correlation coefficient was 0.9121 (160 data points). These data indicate
that the overall method recovery (for levels ranging from 614 mg/kg to 32,600 mg/kg) is
82.9 percent. Since the true concentrations for the three reference materials were obtained by
extracting them with Freon-113 and analyzing the extract by IR, and the fourth true value was the
spike concentration, we concluded mat the performance of Method 3560/8440 (SFE/IR) was
comparable to that of Method 9071A/8440 (Freon-113 extraction/IR analysis).
Method Precision
The interlaboratory standard deviation (SR, reproducibility) is the precision associated with
measurements generated by a group of laboratories; the single-analyst standard deviation (sr,
repeatability) is the precision associated with performance in an individual laboratory. The values for
sr and SR are given in Tables 20 through 25; they were used to calculate the method precision, given
as the repeatability relative standard deviation (RSDr) and the reproducibility relative standard
deviation (RSD,0.
When data from all laboratories were pooled and outliers removed, the RSDr ranged from
11.5 to 17.0 percent for PE-FTIR analyses and from 11.1 to 18.2 percent for BSci-IR analyses. The
RSDR ranged from 17.3 to 45.4 percent for PE-FTIR analyses and from 16.7 to 47.9 percent for BSci-
IR analyses.
When the data from the Isco SFE systems were pooled and outliers removed, the percent
RSDr ranged from 6.0 to 23.8 percent for the PE-FTIR analyses and from 9.8 to 22.6 percent for
BSci-IR analyses. The RSDR ranged from 6.0 to 27.1 percent for PE-FTIR analyses and from 12 to
26 percent for BSci-IR analyses. The interlaboratory method precision for the Isco SFE systems alone
(RSDR in Table 23) was better than that for all SFE systems (RSDR in Table 25), because the Isco
RSDR values were lower for each of the four matrices.
36
-------
The method precision for the seven laboratories using extraction vessels with a volume of
3.5 mL or less was almost similar to that achieved for all laboratories. For example, the RSDr in
^Table 25 ranged from 8.1 to 21.1 percent for PE-FTIR analyses and from 7.4 to 20.6 percent for
BSci-IR analyses, and the RSDR ranged from 18.9 to 39.1 percent for PE-FTIR analyses and from
16.9 to 41.4 percent for BSci-IR analyses. This seems to indicate that the vessel size within the range
used in this study was of little importance.
The precision estimates for the SFE/IR method are ±20 percent for the intralaboratory
performance and ±45 percent for the interlaboratory performance.
Analysis of Variance
Analysis of variance (ANOVA) was used to separate (mathematically) the total variation of
the experimental measurements into an intralaboratory portion and an interlaboratory portion, with a
corresponding split of the total number of degrees of freedom. The mean squares for the among
laboratories and within laboratories were calculated from the sum of squares and the number of
degrees of freedom. The ratio of the two mean squares (the mean of squares among laboratories and
the mean of.squares within laboratories) is distributed as F. Thus, we calculated the F and compared
it with the critical values of F for the upper 95 percent point of distribution at the corresponding
degrees of freedom. For example, in Table 27, we are performing an ANOVA for data submitted
by the 14 laboratories for the TPH-1 sample; in each case, we have three replicates per laboratory.
The among-laboratories degrees of freedom is the number of laboratories (14 minus 1). The within-
laboratories degrees of freedom is the number of laboratories multiplied by 2; the number 2 in this
case is the number of replicates minus 1. The ratio of the mean squares is 19.4. Since the critical
value of F is 2.09, we see that the variation from laboratory to laboratory was greater than that
attributed to the analytical error displayed within laboratories. The matrix and operational parameters
such as flow rate, extraction vessel design and orientation, mode of collection of the extracted
material, and temperature of the collection solvent/trap seemed to be important.
The ANOVA results for the other three matrices are presented in Tables 28 through 30. In
all cases, we find that the interlaboratory variance is greater than intralaboratory variance for the
reasons stated above.
To further substantiate that the interlaboratory variance can be attributed to matrix, we pooled
the data obtained with the Isco SFE systems (laboratories 05, 13, 14, 15, and 17). In this case, all
laboratories used an extraction vessel in vertical position, and the collection vial was kept in a beaker
with water at room temperature. The vessel dimensions, however, varied slightly, the restrictor
dimensions varied slightly, and the restrictor temperature was different for one laboratory (Table 3).
Nonetheless, the ANOVA results (Tables 31 through 34) indicate that the interlaboratory variance is
significantly less than that attributed to analytical error displayed within laboratories for three of the
matrices (TPH-2 soil, SRS103-100 soil, and spiked clay soil), but significantly greater than that
attributed to analytical error within laboratories for the TPH-1 soil.
One-way ANOVA was also performed for the laboratories using vessels of 3.5 mL or less
(Tables 35 through 38). In all four cases, we find that the interlaboratory variance is greater than that
attributed to analytical error displayed within laboratories.
37
-------
TABLE 27. ONE-WAY ANOVA FOR THE TPH-1 SOIL SAMPLES EXTRACTED BY ALL
LABORATORIES
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Source of
variation
Among laboratories
Within laboratories
Total
Concentration (mg/kg)
x,
618
449
67
352
943
250
737
693
972
691
673
639
883
575
Degrees of
freedom
13
28
41
x,
808
453
10
391
859
478
799
1,070
1,210
701
604
605
963
712
Sum of
squares
3,114,300
345.970
3,460,270
X3
1,000
487
77
420
854
44
933
789
1,100
656
620
580
885
832
Grand sum
Mean
square F
239,560 19.4
12,360
Sum
2,430
1,390
154
1,160
2,660
772
2,470
2,550
3,280
2,050
1,900
1,820
2,730
2.120
27,480
F '
1 cril
2.09
0 oj 13 28
38
-------
TABLE 28. ONE-WAY ANOVA FOR THE TPH-2 SOIL SAMPLES EXTRACTED BY ALL
LABORATORIES
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Source of
variation
Before outlier removal
Among laboratories
Within laboratories
Total
After outlier removal
Among laboratories
Within laboratories
Total
Concentration (mg/kg)
x,
1,840
1,940
257
1,350
1,490
1,850
1,480
1,880
2,140
2,320
1,900
2,130
1,850
1,780
Degrees of
freedom
13
28
41
12
26
38
X2
2,160
1,990
410
1,160
2,000
1,730
1,100
2,240
2,200
2,440
2,010
1,900
1,900
1,920
Sum of
squares
9,013,060
1.187.300
10,200,360
2,630,900
1.175.300
3,806,200
X3
1,880
1,730
356
1,520
2,230
1,380
1,500
1,980
2,050
2,110
1,940
1,340
1,990
1,980
Grand sum
Mean
square
(variance) F
693,310 16.4
42.400
248,790
219,240 4.6
45,200
Sum
5,880
5,660
1,020
4,030
5,720
4,960
4,080
6,100
6,390
6,870
5,850
5,370
5,740
5.680
73,350
Fcrtt
2.093
2.15b
Fcrit IS F0 05, 13t 28
Peril IS FQ.OS, 12, 26
39
-------
TABLE 29. ONE-WAY ANOVA FOR THE SRS103-100 SOIL SAMPLES EXTRACTED BY
ALL LABORATORIES
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Source of
variation
Among laboratories
Within laboratories
Total
Concentration (mg/kg)
x,
32,600
28,900
4,130
23,200
25,200
46,400
25,500
19,500
28,900
33,200
30,900
26,500
31,100
21,600
Degrees of
freedom
13
28
41
X2
35,600
29,600
11,100
20,800
29,500
45,100
24,400
14,800
27,800
32,700
30,900
29,200
32,200
24,300
Sum of
squares
3,052,860,000
484.740.000
3,537,600,000
x,
33,700
29,500
—
25,000
10,800
54,600
24,400
17,500
28,300
—
31,500
11,400
30,800
22,600
Grand sum
Mean
square F
234,835,400 13.6
17,312,140
Sum
101,900
88,000
15,230
69,000
69,500
146,100
74,300
51,800
85,000
65,900
93,300
67,100
94,100
68.500
1,085,730
Fa
ml
2.09
F is F
rcnt la l <
0.05. 13, 28
40
-------
TABLE 30. ONE-WAY ANOVA FOR THE SPIKED CLAY SOIL SAMPLES EXTRACTED
BY ALL LABORATORIES
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Source of
variation
Before outlier removal
Among laboratories
Within laboratories
Total
After outlier removal
Among laboratories
Within laboratories
Total
Fcrit IS F0 05 13> 28
be :c p
rcrit 1S rO.OS. 12. 26
Concentration (mg/kg)
x,
8,760
7,850
580
7,000
8,160
6,240
7,110
6,470
9,690
8,660
8,420
5,000
8,560
7,990
Degrees of
freedom
13
28
41
12
26
38
x,
10,400
8,740
368
7,590
8,790
6,810
6,130
4,310
8,380
8,810
8,210
7,910
7,660
7,940
Sum of
squares
217,320,000
26.250.000
243,570,000
74,870,000
26.070.000
100,940,000
X3
8,680
9,740
963
7,270
9,290
4,100
5,780
3,200
9,400
9,270
8,830
8,970
9,020
8,590
Grand sum
Mean
square F
12,783,500 13.6
937,500
6,239,200 6.2
1,002,700
Sum
27,840
26,330
1,910
21,860
26,240
17,150
19,020
13,980
27,470
26,470
25,460
21,880
25,240
24,520
305,640
F a
1 cril
2.09"
2.15a
41
-------
TABLE 31. ONE-WAY ANOVA FOR THE TPH-1 SOIL SAMPLES EXTRACTED BY
LABORATORIES 05, 13, 14, 15, AND 17
Laboratory
code
05
13
14
15
17
Concentration (mg/kg)
x,
943
673
639
883
575
x,
859
604
605
963
712
x,
854
620
580
885
832
Sum
2,660
1,900
1,820
2,730
2,120
Grand sum
11,230
Source of
variation
Degrees of
freedom
Sum of
squares
Mean
square
Among laboratories 4
Within laboratories K)
Total 14
239,780
46.600
286,380
59,950
4,660
12.9
3.48
is F
0.05, 4, 10
42
-------
TABLE 32. ONE-WAY ANOVA FOR THE TPH-2 SOIL SAMPLES EXTRACTED BY
LABORATORIES 05, 13, 14, 15, AND 17
Laboratory
code
05
13
14
15
17
Concentration (mg/kg)
x,
1,490
1,900
2,130
1,850
1,780
X2
2,000
2,010
1,900
1,900
1,920
X3
2,230
1,940
1,340
1,990
1,980
Sum
5,720
5,850
5,370
5,740
5,680
Grand sum
28,360
Source of
variation
Degrees of
' freedom
Sum of
squares
Mean
square
cril
Among laboratories 4
Within laboratories 10
Total 14
43,290
654.400
697,690
10,820
65,440
0.16
3.48
Peril 1S
O.OJ, 4, 10
43
-------
TABLE 33. ONE-WAY ANOVA FOR THE SRS103-100 SOIL SAMPLES EXTRACTED BY
LABORATORIES 05, 13, 14, 15, AND 17
Laboratory
code
05
13
14
15
17
Source of
variation
Among laboratories
Within laboratories
Concentration (mg/kg)
x,
25,200
30,900
26,500
31,100
21,600
Degrees of
freedom
4
10
X2
29,500
30,900
29,200
32,200
24,300
Sum of
squares
286,050,000 71
380.950.000 38
X3
10,800
31,500
11,400
30,800
22,600
Grand sum
Mean
square F
,512,500 1.9
,095,000
Sum
65,500
93,300
67,100
94,100
68,500
388,500
F a
1 crit
3.48
Total
14
667,000,000
Peril !S FO.OS, •». 10
44
-------
TABLE 34. ONE-WAY ANOVA FOR THE SPIKED CLAY SOIL SAMPLES EXTRACTED
BY LABORATORIES 05, 13, 14, 15, AND 17
Laboratory
code
05
13
14
15
17
Concentration (mg/kg)
x,
8,160
8,420
5,000
8,560
7,990
X2
8,790
8,210
7,910
7,660
7,940
X3
9,290
8,830
8,970
9,020
8,590
Sum
26,240
25,460
21,880
25,240
24,520
Grand sum
123,340
Source of
variation
Among laboratories
Within laboratories
Total
Degrees of
freedom
4
10
14
Sum of
squares
3,740,160
10.509.700
14,249,860
Mean
square
940,720
1,050,970
F F a
r rcrit
0.9 3.48
crit 1S
5, 4, 10
45
-------
TABLE 35. ONE-WAY ANOVA FOR THE TPH-1 SOIL SAMPLES EXTRACTED BY
LABORATORIES 01, 04, 06, 08, 13, 14, AND 17
Laboratory
code
01
04
06
08
13
14
17
Source of
variation
Among laboratories
Within laboratories
Total
Concentration (mg/kg)
xt
618
352
250
737
673
639
575
Degrees of
freedom
6
14
20
X2
808
391
• 478
799
604
605
712
Sum of
squares
804,350
227.060
1,031,410
x,
1,000
420
44
933
620
580
832
Grand sum
Mean
square F
134,060 8.3
16,220
Sum
2,430
1,160
112
2,470
1,900
1,820
2,120
12,670
Fa
ml
2.85
Peril 1S Fo.05, 6. 14
46
-------
TABLE 36. ONE-WAY ANOVA FOR THE TPH-2 SOIL SAMPLES EXTRACTED BY
LABORATORIES 01, 04, 06, 08, 13, 14, AND 17
Laboratory
code
01
04
06
08
13
14
17
Source of
variation
Among laboratories
Within laboratories
Concentration (mg/kg)
x,
1,840
1,350
1,850
1,480
1,900
2,130
1,780
Degrees of
freedom
6
14
x,
2,160
1,160
1,730
1,100
2,010
1,900
1,920
Sum of
squares
1,260,630
704.000
X3
1,880
1,520
1,380
1,500
1,940
1,340
1,980
Grand sum
Mean
square F
210,100 3.6
58,670
Sum
5,880
4,030
4,960
4,080
5,850
5,370
5,680
35,850
Fcrit"
2.85
Total
20
1,964,630
Fcrit IS FQ 03, 6, 14
47
-------
TABLE 37. ONE-WAY ANOVA FOR THE SRS103-100 SOIL SAMPLES EXTRACTED BY
LABORATORIES 01, 04, 06, 08, 13, AND 14*
Laboratory
code
01
04
06
08
13
14
Concentration (mg/kg)
xt
32,600
23,200
46,400
25,500
30,900
26,500
x,
35,600
20,800
45,100
24,400
30,900
29,200
X3
33,700
25,000
54,600
24,400
31,500
11,400
Sum
101,900
69,000
146,100
74,300
93,300
67,100
Grand sum
552,200
Source of
variation
Among laboratories
Within laboratories
Total
Degrees of
freedom
5
12
17
Sum of
squares
1,496,270,000
251.640.000
1,747,910,000
Mean
square F
299,254,000 14.3
20,970,000
F b
3.11
3 Laboratory 17 was not included because it used a 10-mL vessel for the extractions.
b p Jc P
rcrit lb r0.05, 5. 12
48
-------
TABLE 38. ONE-WAY ANOVA FOR THE SPIKED CLAY SOIL SAMPLES EXTRACTED
BY LABORATORIES 01, 04, 06, 08, 13, 14, AND 17
Laboratory
code
01
04
06
08
13
14
17
Source of
variation
Among laboratories
Within laboratories
Total
Concentration (mg/kg)
x,
8,760
7,000
6,240
7,110
8,420
5,000
7,990
Degrees of
freedom
6
14
20
X2
10,400
7,590
6,810
6,130
8,210
7,910
7,940
Sum of
squares
27,630,000
16.000.000
43,630,000
X3
8,680
7,270
4,100
5,780
8,830
8,970
8,590
Grand sum
Mean
square F
4,605,000 4.03
1,142,900
Sum
27,840
21,860
17,150
19,020
25,460
21,880
24,520
157,730
Fa
crit
2.85
crit 1S Fo.05, 6, 14
49
-------
Method Performance for the Clay Soil Samples Spiked with Corn Oil and Reference Oil
Table 39 shows the concentrations of oil/grease and TPHs determined in the extracts derived
from spiked clay soil samples by SFE with carbon dioxide. Oil/grease is defined as the material that
has been extracted from the soil sample with supercritical carbon dioxide and collected in PCE, but
has not been subjected to silica gel cleanup. The spiking level was 1,000 mg/kg for each the corn oil
and the reference oil. The oil/grease data indicate that six of the 14 laboratories achieved recoveries
ranging from 69 to 84 percent, five laboratories had recoveries ranging from 34 to 42 percent, two
had recoveries of 16 and 18 percent, and one laboratory did not submit an extract. These recoveries
may be biased low because the quantification of corn oil was done against the reference oil standard.
The TPH recoveries in these samples, after silica gel cleanup of the extract, ranged from 65 to
86 percent for seven laboratories, from 28 to 54 percent for four laboratories, 10 percent for two
laboratories, with one laboratory not submitting an extract.
These data are difficult to interpret. We would have expected (based on the results reported
earlier) much better and more consistent recoveries. It is possible that low recoveries of the reference
oil are partly due to the volatilization of isooctane or chlorobenzene during extraction. We were not
able to correlate these results with the flow rates of carbon dioxide in Table 9 for sample 8. For
example, the laboratories using HP systems (laboratories 10, 11, and 12), with a carbon dioxide flow
rate of 2 mL/min, recovered 53.6, 83, and 9.5 percent, respectively.
The data in Table 40 were obtained on the same extracts, but the IR analyses were performed
with a BSci-IR spectrometer. The recoveries were slightly lower than those given in Table 39, but
they followed the same trend.
Method Performance for the Wet Clay Soil Samples
Tables 41 and 42 show the method performance data for the wet clay soil samples spiked with
motor oil at 10,000 mg/kg. These samples were mixed with equal portions of anhydrous sodium
sulfate immediately prior to extraction. The recoveries were above 30 percent for only two
laboratories (30.7 and 37.7 percent), and eight laboratories had recoveries below 8 percent.
These results indicated that additional experimental work was needed to identify a better
drying method. Experiments were, therefore, performed at MRI-CO using anhydrous magnesium
sulfate and Hydromatrix (diatomaceous earth) as drying agents with clay soil samples spiked with
motor oil at 10,000 mg/kg. The water content of the samples was varied from 10.6 percent to
40 percent (Table 43); the extractions were performed at 340 atm/80°C/30 min (dynamic). We also
extracted spiked clay soil samples containing water at 10.6, 20, 30, and 40 percent to which no drying
agent was added, but the extractions were performed at 450 atm and 150°C for 30 min (dynamic).
The recovery data in Table 43 indicate that at 10.6 percent water, the recoveries obtained with the two
drying agents (e.g., 96 percent for anhydrous magnesium sulfate and 99 percent for Hydromatrix)
were identical with those achieved at higher pressure and temperature but with no drying agent (e.g.,
97 percent). As the water content increased, the recoveries decreased, and they became much lower
for the experiments performed without a drying agent. Since we noticed that the clay particles tended
to clump when we mixed them with anhydrous magnesium sulfate, we also used anhydrous magnesium
sulfate without mixing it with the clay soil sample, but adding it as a plug in the extraction vessel such
that the carbon dioxide would flow through the sample first and then through the bed of anhydrous
magnesium sulfate. The recoveries achieved in those experiments were 75 percent at 20 percent
50
-------
TABLE 39. RECOVERIES (PERCENT) OF OIL/GREASE AND TPHs DETERMINED IN
EXTRACTS OF CLAY SOIL SPIKED WITH CORN OIL AND REFERENCE
OIL (PE-FTIR SPECTROMETER)
Percent recovery
Laooratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Oil/grease
71.5
36.6
42.4
34.1
b
35.3
16.0
39.5
78.0
18.1
73.5
84.0
70.5
69.0
TPHs
71.8
28.5
76.5
31.0
b
32.2
10.1
53.6
83.0
9.5
78.3
85.8
65.2
68.4
The spiking level was 1,000 mg/kg for reference corn
oil and 1,000 mg/kg for reference oil. Single
determinations.
This laboratory did not submit an extract for this
sample.
51
-------
TABLE 40. RECOVERIES (PERCENT) OF OIL/GREASE AND TPHs DETERMINED IN
EXTRACTS OF CLAY SOIL SPIKED WITH CORN OIL AND REFERENCE
OIL (BSci-IR SPECTROMETER)'
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Percent
Oil/grease
55.0
30.6
38.4
25.8
b
29.0
12.5
30.3
62.0
14.3
53.5
68.0
53.5
55.5
recovery
TPHs
56.4
25.7
71.3
23.0
b
26.8
9.1
52.6
69.0
10.9
65.8
70.7
54.3
56.4
The spiking level was 1,000 mg/kg for reference corn
oil and 1,000 mg/kg for reference oil. Single
determinations.
This laboratory did not submit an extract for this
sample.
52
-------
TABLE 41. CONCENTRATIONS (mg/kg) AND PERCENT RECOVERIES OF TPHs
DETERMINED IN EXTRACTS OF WET CLAY SOIL SPIKED WITH MOTOR
OIL (PE-FTIR SPECTROMETER)'
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Concentration
(mg/kg)
3,770
1,920
135
2,210
489
3,070
76.0
69.0
1,780
533
289
608
2,330
775
Percent
recovery
37.7
19.2
1.4
22.1
4.9
30.7
0.8
0.7
17.8
5.3
2.9
6.1
23.3
7.8
a The clay soil was spiked with motor oil at
10,000 mg/kg. The water content was 30 percent.
Single determinations.
53
-------
TABLE 42. CONCENTRATIONS (mg/kg) AND PERCENT RECOVERIES OF TPHs
DETERMINED IN EXTRACTS OF WET CLAY SOIL SPIKED WITH MOTOR
OIL (BSci-IR SPECTROMETER)'
Laboratory
code
01
02
03
04
05
06
08
10
11
12
13
14
15
17
Concentration
(mg/kg)
3,730
1,850
125
2,150
489
2,930
86.0
71.8
1,740
561
316
612
2,320
754
Percent
recovery
37.3
18.5
1.3
21.5
4.9
29.3
0.9
0.7
17.4
5.6
3.2
6.1
23.2
7.5
The clay soil was spiked with motor oil at
10,000 mg/kg. The water content was 30 percent.
Single determinations.
54
-------
water, but they dropped to approximately 25 to 27 percent for samples with 30 and 40 percent water
(Table 43). It is possible that the low recoveries were due to restrictor plugging and subsequently to
reduced flow rate of carbon dioxide.
When spiked clay soil containing 40 percent water was mixed with anhydrous magnesium
sulfate (equal weights) and allowed to equilibrate overnight, or for 5 days, at room temperature, we
obtained much higher recoveries of TPHs when extracting at 340 atm/80°C for 30 min (dynamic).
The average recovery + one standard deviation of eight determinations (two sets of four samples
extracted in parallel) was 84.7 + 3.4 percent for overnight equilibration (14 hours) and
76.9 ± 6.1 percent for 5-day equilibration for clay soil spiked with motor oil (Table 44), and
74.3 percent for overnight equilibration and 77.1 + 4.6 percent for 5-day equilibration for clay soil
spiked with diesel oil (Table 44). We used in both experiments a plug of anhydrous magnesium
sulfate (1.5 g) and crushed the clumps that were formed upon adding magnesium sulfate. A
disposable glass pipet was used for this purpose, and the crushing was done directly in the extraction
vessel (to minimize losses of the more volatile petroleum hydrocarbons).
Correlation between the PE-FTIR Data and the BSci-IR Data
The development work for this method has been performed on an FTIR instrument, while
the method specified the use of a filter or fixed-wavelength instrument. Questions were raised
regarding how to compensate for the multiplex advantage of the FTIR and about the capability of the
FTIR for spectral subtraction, which cannot be done with a filter instrument. To address these
concerns, we analyzed all extracts for the interlaboratory study on two IR systems. The measurement
was done first on the PE-FTIR, then the IR cuvette was placed in the BSci-IR system and the
measurement was taken. The BSci-IR measurements and the PE-FTIR measurements (as TPH
concentrations in mg/kg for each matrix) were then plotted on the y-axis and the x-axis, respectively.
Table 45 shows the slope, intercepts, and the correlation coefficients of the FTIR and the IR data by
matrix. Excellent correlation was demonstrated, as shown by the values for the correlation
coefficients. The slopes indicate that the measurements obtained with the BSci-IR were always lower
than those obtained with the PE-FTIR system. We cannot explain the differences; however, because
these differences were 17 percent or less (the slopes of the linear regression equations ranged from
0.8267 to 0.9900 for four matrixes), we concluded that the two sets of data were comparable.
QUALITY ASSURANCE/QUALITY CONTROL
The quality control data generated in this study are presented in Tables 46 through 56. They
include all calibration data generated during the analysis of extracts (by instrument, by material used
in calibration, and by date) (Tables 46 through 52), the results from the analyses of the unspiked clay
soil samples (Table 53), the results of the analyses of the PCE blanks (Table 54), the results from the
analysis of all system blanks submitted by the participating laboratories (Table 55), and the results of
the sample storage study (Table 56).
55
-------
TABLE 43. PERCENT RECOVERIES OF TPHs DETERMINED IN EXTRACTS OF WET
CLAY SOIL SPIKED WITH MOTOR OIL (PE-FTIR SPECTROMETER)8
Percent recovery
Addition of anhydrous MgSO4b
Percent
water Plug only
10.6
20 74.9
30 24.8
40 26.9
Mixed with sample
95.9 ± 1.2
86.4 ± 9.2
55.8
45.3
Addition of
Hydromatrixb
98.8
72.2
53.1
47.1
No drying
agent0
96.8
74.8 ± 12.2
52.0
29.2
The value given is the average recovery ± one standard deviation (for triplicate determinations)
or the average recovery of duplicate determinations.
The extraction was performed at 340 atm/80°C/30 min (dynamic).
The extraction was performed at 450 atm/150°C/30 min (dynamic).
TABLE 44. PERCENT RECOVERIES OF TPHs DETERMINED IN EXTRACTS OF WET
CLAY SOIL (40 PERCENT WATER) SPIKED WITH MOTOR OIL OR DIESEL
OIL (PE-FTIR SPECTROMETER)'
Storage
Condition
14 hours at 22 °C
120 hours at 22°C
14 hours at 4°C
120 hours at 4°C
Motor oil
84.7 ± 3.4"
76.9 ± 6.1C
—
Diesel
74.3
77.1 ±
oil
d
4.6e
The extractions were performed at 340 atm/80°C/30 min (dynamic).
Each spiked sample (3 g) was mixed with 3 g of anhydrous
magnesium sulfate and stored as indicated. For extraction, a plug of
1.5 g of anhydrous sodium sulfate was put into each extraction vessel
at the outlet of the vessel.
The number of determinations was eight.
The number of determinations was seven.
Duplicate determination.
The number of determinations was three.
56
-------
TABLE 45. CORRELATION BETWEEN THE PE-FTIR DATA AND THE BSci-IR DATA"
Sample
Matrix identification
TPH-1 soil
TPH-2 soil
SRS 103-100
Spiked clay soil
Spiked clay soil
(oil/grease
analysis)
Spiked clay soil
(TPH analysis)
Wet clay soil
1
2
3
4,6,7
8
8
9
Slope
0.9900
0.8711
0.8880
0.9767
0.7689
0.8271
0.9727
Intercept
-29.5
50.3
945
534
21.8
13.9
12.4
Correlation
coefficient
0.9929
0.9953
0.9975
0.9831
0.9927
0.9891
0.9997
Number of
data points
42
42
40
42
13
13
14
The BSci-IR data were plotted on the y-axis and the PE-FTIR data were plotted on the x-axis.
57
-------
TABLE 46. CALIBRATION DATA FOR THE PE-FTIR SPECTROMETER USING
REFERENCE OIL IN PCE AND THE 10-mm PATH-LENGTH IR CELL"
Concentration
Date (Mg/mL)
04/03/92 10
25
50
100
250
500
04/06/92 10
25
50
100
250
500
04/07/92 10
25
50
100
250
500
04/10/92 10
25
50
100
250
500
04/16/92 10
25
50
100
250
500
04/17/92 10
25
50
100
250
500
Absorbance Slope
0.0560 0.0018
0.0406
0.0856
0.1797
0.4525
0.8672
0.0225 0.0017
0.0725
0.1220
0.2222
0.4736
0.8768
0.0361 0.0018
0.0748
0.1085
0.2109
0.4921
0.8992
-0.0058 0.0017
0.0785
0.1056
0.1899
0.4491
0.8722
0.0242 0.0017
0.0718
0.1082
0.2102
0.4749
0.8842
0.0239 0.0017
0.0715
0.1070
0.2092
0.4746
0.8828
Correlation
Intercept coefficient
-0.0016 0.9996
0.0314 0.9987
0.0288 0.9992
0.0121 0.9982
0.0247 0.9992
0.0242 0.9992
(continued)
Silica gel was added to the calibration standards.
58
-------
TABLE 46. (concluded)'
Date
04/21/92
04/23/92
04/24/92
04/28/92
05/01/92
05/05/92
Concentration
Cig/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance Slope
0.0240 0.0017
0.0713
0.1089
0.2115
0.4759
0.8869
0.0240 0.0017
0.0706
0.1070
0.2087
0.4738
0.8813
0.0236 0.0017
0.0715
0.1101
0.2097
0.4725
0.8848
0.0237 0.0017
0.0704
0.1074
0.2089
0.4731
0.8850
0.0230 0.0017
0.0704
0.1077
0.2086
0.4743
0.8903
0.0231 0.0017
0.0701
0.1066
0.2089
0.4722
0.8814
Correlation
Intercept coefficient
0.0247 0.9992
0.0242 0.9993
0.0245 0.9993
0.0234 0.9993
0.0226 0.9994
0.0234 0.9993
Silica gel was added to the calibration standards.
59
-------
TABLE 47. CALIBRATION DATA FOR THE BSci-IR SPECTROMETER USING
REFERENCE OIL IN PCE AND THE 10-mm PATH-LENGTH IR CELL8
Date
04/03/92
04/06/92
04/07/92
04/10/92
04/16/92
04/17/92
Concentration
0»g/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance Slope
0.006 0.0016
0.033
0.073
0.156
0.400
0.799
0.020 0.0016
0.072
0.109
0.198
0.422
0.813
0.031 0.0016
0.067
0.093
0.183
0.433
0.826
-0.003 0.0016
0.069
0.092
0.165
0.393
0.805
0.015 0.0016
0.058
0.089
0.178
0.412
0.806
0.024 0.0016
0.061
0.093
0.182
0.416
0.811
Correlation
Intercept coefficient
-0.0074 l.OOO
0.0260 0.9992
0.0197 0.9998
0.0055 0.9985
0.0112 0.9997
0.0162 0.9998
(continued)
Silica gel was added to the calibration standards.
60
-------
TABLE 47. (concluded)'
Date
04/21/92
04/23/92
04/24/92
04/28/92
05/01/92
05/05/92
Concentration
(jig/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance Slope
0.019 0.0016
0.060
0.093
0.182
0.417
0.813
0.019 0.0016
0.061
0.093
0.181
0.415
0.809
0.022 0.0016
0.064
0.098
0.185
0.419
0.813
0.012 0.0016
0.054
0.085
0.172
0.404
0.796
0.022 0.0016
0.063
0.096
0.187
0.420
0.815
0.022 0.0016
0.068
0.095
0.185
0.417
0.810
Correlation
Intercept coefficient
0.0143 0.9998
0.0147 0.9997
0.0184 0.9997
0.0075 0.9997
0.0179 0.9997
0.0190 0.9996
Silica gel was added to the calibration standards.
61
-------
TABLE 48. CALIBRATION DATA FOR THE PE-FTIR SPECTROMETER USING
MOTOR OIL IN PCE AND THE 10-mm PATH-LENGTH IR CELL8
Date
04/14/92
04/17/92
04/22/92
04/30/92
05/05/92
Concentration
Gtg/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance
0.0306
0.0701
0.1323
0.2576
0.6152
1.0992
0.0237
0.0629
0.1236
0.2499
0.6050
1.0987
0.0252
0.0640
0.1257
0.2515
0.6159
1.0310
0.0251
0.0640
0.1258
0.2514
0.6072
1.1016
0.0255
0.0643
0.1249
0.2508
0.6020
1.1023
Correlation
Slope Intercept coefficient
0.0022 0.0263 0.9984
0.0022 0.0175 0.9987
0.0021 0.0281 0.9963
0.0022 0.0189 0.9987
0.0022 0.0182 0.9989
Silica gel was added to the calibration standards.
62
-------
TABLE 49. CALIBRATION DATA FOR THE BSci-IR SPECTROMETER USING
MOTOR OIL IN PCE AND THE 10-mm PATH-LENGTH IR CELL"
Date
04/14/92
04/17/92
04/22/92
04/30/92
05/05/92
Concentration
Otg/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance Slope
0.030 ' 0.0020
0.057
0.109
0.215
0.516
1.002
0.020 0.0020
0.052
0.103
0.206
0.506
1.000
0.020 0.0020
0.050
0.101
0.205
0.502
1.001
0.021 0.0020
0.051
0.101
0.206
0.505
1.001
0.021 0.0020
0.053
0.103
0.207
0.506
1.002
Correlation
Intercept coefficient
0.0115 0.9999
0.0031 1.000
0.0014 1.000
0.0024 1 .000
0.0037 1.000
a Silica gel was added to the calibration standards.
63
-------
TABLE 50. CALIBRATION DATA FOR THE PE-FTIR SPECTROMETER USING
REFERENCE OIL IN PCE AND THE 10-mm PATH-LENGTH IR CELL (NO
SILICA GEL CLEANUP)'
Date
04/16/92
04/23/92
04/28/92
05/01/92
05/05/92
Concentration
Otg/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance
0.0305
0.0559
0.0977
0.1959
0.4784
0.8652
0.0225
0.0497
0.0960
0.1936
0.4597
0.8494
0.0202
0.0479
0.0933
0.1897
0.4569
0.8461
0.0214
0.0484
0.0941
6.1920
0.4576
0.8470
0.0208
0.0476
0.0944
b
0.4558
0.8468
Correlation
Slope Intercept coefficient
0.0017 0.0192 0.9989
0.0017 0.0146 0.9992
0.0017 0.0121 0.9992
0.0017 0.0132 0.9992
0.0017 0.0101 0.9993
" No silica gel was added to the calibration standards.
b Calibration was not performed with the 100-/ig/mL standard.
64
-------
TABLE 51. CALIBRATION DATA FOR THE BSci-IR SPECTROMETER USING
REFERENCE OIL IN PCE AND THE 10-mm PATH-LENGTH IR CELL
(NO SILICA GEL CLEANUP)'
Date
04/16/92
04/23/92
04/28/92
05/01/92
05/05/92
Concentration
Otg/mL)
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
10
25
50
100
250
500
Absorbance
0.027
0.046
0.084
0.171
0.429
0.820
0.019
0.042
0.083
0.169
0.411
0.804
0.018
0.043
0.083
0.169
0.410
0.801
0.019
0.043
0.083
0.171
0.412
0.806
0.019
0.041
0.083
b
0.411
0.804
Correlation
Slope Intercept coefficient
0.0016 0.0081 0.9998
0.0016 0.0046 0.9999
0.0016 0.0049 0.9999
0.0016 0.0051 0.9999
••
0.0016 0.0032 0.9999
a No silica gel was added to the calibration standards.
b Calibration was not performed with the 100-/ig/mL standard.
65
-------
TABLE 52. CONCENTRATIONS OF DAILY STANDARDS ANALYZED TO VERIFY
SYSTEM REPRODUCIBILITY
Date of analysis
04/06/92
04/06/92
04/06/92
04/09/92
04/09/92
04/10/92
04/10/92
04/10/92
04/10/92
04/14/92
04/14/92
04/14/92
04/14/92
04/16/92
04/16/92
04/16/92
04/16/92
04/17/92
04/21/92
04/21/92
04/23/92
04/23/92
04/23/92
04/24/92
04/28/92
04/30/92
04/30/92
05/01/92
05/01/92
05/01/92
05/05/92
Concentration
PE-FTIR
110
104
104
100
100
105
107
112
107
112
112
106
101
107
100
107
105
105
106
104
105
104
106
107
106
105
105
106
105
104
107
WmL).
BSci-IR
103
108
108
103
100
106
108
115
111
108
112
106
100
105
102
105
103
103
107
107
103
103
106
103
105
104
.103
102
108
109
104
a The true concentration of the calibration standard was 100 jig/mL.
66
-------
TABLE 53. CONCENTRATIONS OF TPHs IN THE EXTRACTS FROM THE UNSPIKED
CLAY SOIL SAMPLES
Laboratory
Concentration (rag/kg)
PE-FTIR
Extract not submitted for analysis.
BSci-IR
01
02
03
04
05
06
08
10
11
12
13
14
15
17
207
<10
492
43
306
103
a
423
204
58
325
398
328
251
204
<10
469
60
305
88
a
422
196
54
327
400
332
250
67
-------
TABLE 54. CONCENTRATIONS OF TPHs IN THE PCE BLANKS ANALYZED DURING
THIS STUDY
Number of
Date of blanks performed
analysis during day
Concentration Otg/mL PCE)
PE-FTIR
BSci-IR
04/03/92
04/06/92
04/09/92
04/10/92
04/14/92
04/16/92
04/17/92
04/21/92
04/23/92
04/24/92
04/28/92
04/30/92
05/01/92
05/05/92
3
5
3
4
5
2
1
2
2
1
1
2
3
3
68
-------
TABLE 55. CONCENTRATIONS OF TPHs IN THE EXTRACTS SUBMITTED AS
SYSTEM BLANKS
Laboratory
01
02
03
04
05
06
08
10
11
12
13
14
. 15
Number of
blanks submitted
0
1
2
1
2
0
1
0
5
0
1
4
4
Concentration (mg/mL)
PE-FTIR
a
<10
<10; <10
<10
<10; <10
a
<10
a
229; 18; 96;
< 10; 37
a
81
22; <10; < 10;
< 10
10; < 10; < 10;
BSci-IR
a
<10
< 10; < 10
<10
< 10; < 10
a
<10
a
238; 22; 98
27; 41
a
62
18; <10; <10;
11
14; < 10; < 10;
17
< 10; < 10; < 10;
< 10; < 10;
a Extract not submitted for analysis.
TABLE 56. PERCENT RECOVERIES OF TPHs FROM THE SPIKED CLAY SOIL
SAMPLES, STORED AT 4°C IN THE DARK, AS A FUNCTION OF TIME
Days of
storage
22
34
40
83.1
85.6
87.4
PE-FTIR
79.1
76.6
77.9
87.2
87.7
90.7
BSci-IR
82.2
81.2
81.7
The clay soil samples were spiked with motor oil at 10,000 mg/kg. The values given are for
duplicate determinations at each time.
69
-------
The following conclusions can be drawn from these data:
• The PE-FTIR and the BSci-IR calibration data generated during this study agreed to
within 10 percent. The slopes for the reference oil calibrations were 0.0017 for the
PE-FTIR system and 0.0016 for the BSci-IR system, and the correlation coefficients
for standards ranging from 10 to 500 /ig/mL were 0.999 or greater (Tables 46 and 47).
The slopes for the motor oil calibrations were 0.0022 for the PE-FTIR system
(Table 48) and 0.0020 for the BSci-IR system (Table 49).
• Both IR spectrometer systems gave reproducible results over the period of 1 month
during which IR analyses were performed. The multilevel calibration data (slopes,
intercepts, and correlation coefficients) from each instrument indicate deviations less
than 10 percent for the IR determinations performed over a period of 1 month. A
100-/ig/mL reference oil or motor oil standard was analyzed after every 10 analyses.
The largest deviation (+15 percent) was found on one day with the BSci-IR system.
The other deviations averaged approximately 5 percent over a period of 1 month
(Table 52).
• The extracts from the unspiked clay soil samples (sample 5) did not contain TPHs
above 10 mg/kg. We analyzed this material in our laboratory on several occasions and
found TPHs at levels ranging from 5 to 9 mg/kg. The sample-5 extracts submitted by
the participating laboratories, however, contained TPHs at concentrations as high as
492 mg/kg (Table 53). This implies that the levels reported for sample 5 are due to
cross-contamination from the previous extraction, since we instructed the laboratories
to analyze the clay soil blank immediately after the spiked clay soil (concentration
10,000 mg/kg). Thus, contamination of the SFE system, especially when dealing with
high-concentration samples, is likely, and the analyst must take the necessary
precautions to minimize it (e.g., clean extraction vessel and frits, replace restrictor,
perform system blanks).
• The Aldrich PCE (spectrophotometric grade) was acceptable for this study. The
solvent blanks analyzed during the study indicated TPH levels of less than 10 j*g/mL
(Table 54).
• The system blanks generated with the Aldrich PCE and the SFE-grade carbon dioxide
did not show contamination, with the exception of laboratories 11 and 13 (Table 55).
Since both laboratories used the Scott SFE-grade carbon dioxide, and since the PCE
was provided by MRI-CO, it appears that the source of contamination was in the SFE
system.
• The percent recoveries from the spiked clay soil samples, that had been stored at 4°C
in the dark and extracted after 22, 34, and 44 days, were independent of the storage
time.
70
-------
REFERENCES
1. Lopez-Avila, V., N. S. Dodhiwala, J. Benedicto, and R. Young, "SFE/IR Method for the
Determination of Petroleum Hydrocarbons in Soils and Sediments," EPA 600/X-92/046,
April 1992 (W.F. Beckert, Project Officer, EMSL-LV).
2. Lopez-Avila, V., J. Benedicto, N. S. Dodhiwala, and W. F. Beckert, "Development of an
Off-Line SFE/IR Method for Petroleum Hydrocarbons in Soils," J. Chromatogr. Sci. 30,
335-343, 1992.
3. Guidelines for Collaborative Study Procedure to Validate Characteristics of a Method of
Analysis, J. Assoc. Off. Anal. Chem. 72, 694-704, 1989.
4. "Lotus Spreadsheet Program for the Calculation of Performance Parameters from
Collaborative Study Data Including Outlier Analyses," revision 3/5/91, received from John
G. Phillips, Chairman, AOAC Statistics Committee.
5. EPA Method 418.1, "Total Recoverable Petroleum Hydrocarbons," in "Methods for
Chemical Analyses of Water and Wastewater," EPA 600/14-79/020, revised 1983.
71
-------
APPENDIX A
PROPOSED DRAFT PROTOCOL FOR SUPERCRITICAL FLUID EXTRACTION
OF PETROLEUM HYDROCARBONS (METHOD 3560)
A-l
-------
METHOD 3560
SUPERCRITICAL FLUID EXTRACTION OF
TOTAL RECOVERABLE PETROLEUM HYDROCARBONS
1.0 SCOPE AND APPLICATION
1.1 Method 3560 describes the extraction with supercritical fluids of total recoverable
petroleum hydrocarbons (TPHs) from soils, sediments, fly ash, and other solid materials that are
amenable to extraction with conventional solvents. The method is suitable for use with any
supercritical fluid extraction (SFE) system that allows extraction conditions (e.g., pressure,
temperature, flow rate) to be adjusted to achieve separation of the TPHs from the matrices of concern.
1.2 Method 3560 is not suitable for the extraction of low-boiling TPHs such as gasoline.
2.0 SUMMARY OF METHOD
2.1 A known amount of sample is transferred to the extraction vessel. The sample is then
extracted in the dynamic mode for up to 30 min with supercritical carbon dioxide at 340 atm and 80°C
and a gas flow rate of 500 to 1,000 mL/min. After depressurization of the carbon dioxide, the
extracted TPHs are collected in 3 mL of tetrachloroethylene or other appropriate solvent (see Section
5.3), or on a sorbent material, depending on the SFE system used. In the latter case, the analytes are
collected by rinsing the sorbent material with tetrachloroethylene or other suitable solvent. After
collection, the TPHs are analyzed by the appropriate determinative method.
3.0 INTERFERENCES
3.1 The analyst must demonstrate through the analysis of reagent blanks (collection solvent
treated as per Section 7.4) that the supercritical fluid extraction system is free of interferents. To do
this, perform a simulated extraction using an empty extraction vessel and a known amount of carbon
dioxide under the same conditions as those used for sample extraction, and determine the background
contamination by analyzing the extract by the appropriate determinative method (e.g., Method 8015
or 8440). If glass wool and a drying agent are used with the sample, they should be included when
performing a reagent blank check.
3.2 The extraction vessel(s), the frits, the restrictor(s), and the multiport valve may retain
solutes whenever high-concentration samples are extracted. It is, therefore, good practice to clean the
extraction system after each extraction. Replacement of the restrictor may be necessary when reagent
blanks indicate carryover. At least one reagent blank should be prepared and analyzed daily when the
instrument is in use. Furthermore, reagent blanks should be prepared and analyzed after each
extraction of a high-concentration sample (concentration in the high ppm range). If reagent blanks
continue to indicate contamination even after replacement of the extraction vessel and the restrictor,
the multi-port valve must be cleaned.
3560 - 1 Revised
December 1992
-------
4.0 APPARATUS AND MATERIALS
4.1 Supercritical fluid extractor and associated hardware.
WARNING - A safety feature to prevent overpressurization is required on the extractor. This feature
should be designed to protect the laboratory personnel and the instrument from possible injuries or
damage resulting from equipment failure under high pressure.
4.1.2 Extraction vessel -- Stainless-steel vessel with end fittings and 0.5- or 2-/*m frits.
Use the extraction vessel recommended by the manufacturer of the SFE system being used.
The volume of the extraction vessel should fit the sample size. PEEK (polyether ether ketone)
extraction vessels are acceptable only for use with specifically designed instruments.
4.1.2.1 Fittings used for the extraction vessel must be capable of withstanding
the required extraction pressures. The maximum operating pressure for most extractors
is 500 atm; however, extractors with higher pressure ratings are available. Check with
the manufacturer of the particular extraction system on the maximum operating pressure
and temperature for that system. Make sure that the extraction vessels are rated for such
pressures and temperatures.
4.1.3 Restrictor - 50-/xm ID x 150- or 375-/*m OD x 25- to 60-cm length piece of
uncoated fused-silica tubing (J&W Scientific or equivalent). Other restrictors may be used
including tapered restrictors, static pinhole restrictors, frit restrictors, and variable orifice
restrictors (manual and computer-controlled) or crimped metal tubing. Check with the
manufacturer of the SFE system on the advantages and disadvantages of the various restrictor
designs.
4.1.4 Collection device -- The extracted TPHs can be collected either in vials containing
solvent, or they can be trapped on a sorbent material (e.g., octadecyl-bonded silica, stainless-
steel beads).
4.1.4.1 When the analytes are collected in a solvent, install the restrictor
through a hole made through the cap and septum of the vial, and position the restrictor
end about 0.5 inch from the bottom of the vial. A syringe needle should also be inserted
through the septum of the vial (with the tip positioned just below the septum) to prevent
buildup of pressure in the vial. Use the type of vials appropriate for the SFE system
used.
4.1.4.2 When the analytes are trapped on a sorbent material, it is important
to ensure that breakthrough of the analytes from the trap does not occur. Desorption
from the trapping medium can be accomplished by increasing the temperature of the trap
and using a solvent to remove the analytes. To recover the analytes, use the conditions
suggested by the manufacturer of the particular system used..
4.2 Carbon dioxide cylinder balance (optional) - Balances from White Associates, catalog no.
30, Scott Specialty Gases Model 5588D, or equivalent, can be used to monitor the fluid usage. Such
a device is useful because carbon dioxide tanks used for SFE are not equipped with regulators, and
it is difficult to determine when the tank needs to be replaced.
3560 - 2 Revised
December 1992
-------
4.3 Tools required include: screwdriver (flat-blade), adjustable wrench, pliers, tubing cutter,
and various small open-end wrenches for small fittings.
4.4 Other materials - Magnesium sulfate monohydrate can be used as received. The
following materials require high-temperature treatment (muffling at 400°C for 2 to 4 hours) prior to
use since they may be contaminated with petroleum hydrocarbons. Alternatively, a blank can be
performed to determine whether these materials are sufficiently clean.
4.4.1 Silanized glass wool.
4.4.2 Drying agents such as anhydrous magnesium sulfate or diatomaceous earth.
5.0 REAGENTS
5.1 Carbon dioxide, C02 - Either supercritical fluid chromatography (SFC-grade) or
SFE-grade CO2 is acceptable for use in SFE. Aluminum cylinders are preferred over steel cylinders.
The cylinders are fitted with eductor tubes, and their contents are under 1500 psi of helium head
pressure.
5.2 Carbon dioxide (CO^ for cryogenic cooling — Certain parts of some models of extractors
(i.e., the high-pressure pump head and the analyte trap) must be cooled during use. The carbon
dioxide used for this purpose must be dry (< 50 ppm water content), and it must be supplied in tanks
with a full-length eductor tube.
5.3 Tetrachloroethylene, C2C14 (spectrophotometric grade) - Used for the collection of TPHs
for determination by IR. Analyze a blank to ensure no interferences are present at the TPH
wavelengths. Chlorofluorocarbons are not suitable for use with this method because of the risk to the
ozone layer.
5.4 Other appropriate pesticide-quality solvents may be used for the collection of TPHs for
determination by GC (i.e., methylene chloride). Chlorofluorocarbons are not suitable for use with
this method because of the risk to the ozone layer.
5.5 Copper filings — Copper filings added to remove elemental sulfur must have a shiny
bright appearance to be effective. To remove oxides from copper surfaces, treat with dilute nitric
acid, rinse with reagent water to remove all traces of acid, rinse with acetone (copper will darken if
acid is still present), and dry under a stream of nitrogen.
6.0 SAMPLE COLLECTION, PRESERVATION, AND HANDLING
6.1 Solid samples should be collected and stored in the same manner as any other solid
samples containing semivolatile organics.' See Chapter Four for guidance relating to semivolatile
organics (including holding times).
3560 - 3 Revised
December 1992
-------
7.0 PROCEDURE
7.1 Determination of sample percent dry weight - In certain cases, sample results are desired
based on dry-weight basis. When such data are desired, a portion of sample for this determination
should be weighed out at the same time as the portion used for analytical determination.
WARNING: The drying oven should be contained in a hood or vented.. Significant laboratory
contamination may result from a heavily contaminated hazardous waste sample.
7.1.1 Immediately after weighing die sample portion for extraction, weigh 5.00 to
10.00 g of the remaining sample into a tared crucible. Determine the percent dry weight of
the sample by drying it overnight at 105°C. Allow it to cool in a desiccator before weighing.
Calculate the percent dry weight as follows:
% drv weight = gof dry sample x 100
/o ary weigm g or sample
7.2 Safety considerations -- Read section 11.0 "Safety" before attempting this procedure.
7.3 Sample handling
7.3.1 Decant and discard any aqueous layer that has accumulated on a sediment sample.
Mix the sample thoroughly, especially composited samples. Discard any foreign objects such
as sticks, leaves, and rocks.
7.3.2 Weigh 3 g of sample into a precleaned aluminum dish. A drying agent (e.g.,
anhydrous magnesium sulfate or diatomaceous earth) may be added to samples that contain
water in excess of 20% to increase porosity or to bind water. Alternatively, magnesium sulfate
monohydrate is an excellent drying agent, and the amount of heat released (compared with
anhydrous magnesium sulfate) is small, thereby minimizing the loss of volatile petroleum
hydrocarbons. The amount of the drying agent will depend on the water content of the sample.
Typically, a ratio of 1:1 works well for wet soils and sediment materials. However, a certain
amount of water (up to 10%) in the sample has been shown to improve recoveries from certain
matrices; therefore, if the sample is dry, water may optionally be added to bring the water
content to approximately 20%.
7.3.2.1 If drying agent has been added to the sample, store the mixture of
sample and drying agent for several hours (preferably overnight) at 4°C, with a minimum
of headspace. This additional storage time is necessary to achieve acceptable analyte
recovery.
7.3.3 Transfer the weighed sample to a clean extraction vessel. The volume of the
extraction vessel should match the sample volume. Use two plugs of silanized glass wool to
hold the sample in place and fill the void volume (the use of drying agent or clean sand after
the second glass wool plug to fill the void volume is also recommended). Attach the end
fittings, and install the extraction vessel in the oven. Always use clean frits for each extraction
vessel.
3560 - 4 Revised
December 1992
-------
7.4 Sample extraction
7.4.1 Fill the collection vessel with 3 mL of tetrachloroethylene or other appropriate
collection solvent. Chlorofluorocarbons are not suitable for use with this method because of
the risk to the ozone layer.
7.4.2 Set the pressure at 340 atm and the temperature at 80°C. Follow manufacturer's
instructions in setting up the instrument. Extract for 30 minutes in the dynamic mode. Note
the safety precautions in Section 11 on venting the instrument into a chemical fume hood.
7.4.3 After the extraction time has elapsed, the system should automatically go to the
equilibrate mode. At this point, remove the collection vessel(s) containing the extract(s). Since
the depressurization of the carbon dioxide at the end of the restrictor outlet results in a gas flow
rate of about 500 to 1000 mL/min, part of the collection solvent will evaporate during the
extraction. However, cooling caused by the rapid expansion of the carbon dioxide limits the
loss of solvent, so that approximately 2 mL remains (when tetrachloroethylene is used) after
a 30-min extraction. To prevent the collection solvent from freezing during the extraction,
place the collection vial in a beaker with warm water (approximately 25°C). The extract is
then brought to the desired volume, or concentrated further. See Method 3510 for
concentration techniques by micro Kuderna-Danish or nitrogen biowdown. Concentration must
be performed in a chemical fume hood to prevent contamination of the laboratory environment.
7.4.4 Record the volume of liquid carbon dioxide used for extraction. Calculate the
average flow rate by dividing the volume of the carbon dioxide by the extraction time.
7.4.5 The extract is ready for analysis by Method 8015 - Nonhalogenated Volatile
Organics by Gas Chromatography, or by Method 8440 - Total Recoverable Petroleum
Hydrocarbons by Infrared Spectrophotometry.
7.5 SFE System Maintenance
7.5.1 Depressurize the system following manufacturer's instructions.
7.5.2 After extraction of an especially tarry sample, the frits may require replacement
to ensure adequate extraction fluid flow through the restrictor. In addition, very fine particles
contained in samples can clog the frits, necessitating replacement.
7.5.3 Clean the extraction vessel after each sample. The cleaning procedure depends
on the type of sample. After removing the bulk of the extracted sample from the extraction
vessel, the vessel should be scrubbed with an ionic detergent, water, and a bottle brush. After
extraction of tarry materials, use solvent rinses or an ultrasonic bath to clean the extraction
vessel.
7.5.4 For samples known to contain elemental sulfur, use copper filings to remove the
dissolved sulfur from the fluid. The copper filings (1 to 2 g per sample) can be packed in a
separate extraction vessel connected to the outlet end of the sample extraction vessel, or they
can be mixed with the sample, and a plug of copper filings can be loaded in the extraction
3560 - 5 Revised
December 1992
-------
vessel with the sample such that any sulfur extracted by the carbon dioxide can be removed
before the stream of carbon dioxide containing the analytes reaches the restrictor.
7.5.5 The procedure to be followed in emptying the syringe pump depends upon the
type of fluid being used. In the case of carbon dioxide, which is a gas at ambient temperature
and pressure, it is only necessary to vent the gas to a fume hood by allowing it to expand
across the purge valve. Follow manufacturer's instructions in emptying the syringe pump.
7.5.6 To change fluid supply cylinders, follow manufacturer's instructions.
7.5.7 Restrictor removal and installation - Follow manufacturer's instructions. When
using fused-silica restrictors, it may be necessary to replace the restrictor after each sample,
especially when extracting samples contaminated with heavy oils.
8.0 QUALITY CONTROL
8.1 Reagent blanks or matrix-spiked samples must be subjected to the same analytical
procedures (Section 7.4) as those used on actual-samples.
8.2 Refer to Chapter One for specific Quality Control procedures and to Method 3500 for
sample preparation quality control procedures.
8.3 All instrument operating conditions must be recorded.
9.0 METHOD PERFORMANCE
9.1 Refer to Method 8440 and 8015 for performance data.
9.2 Use standard reference materials to establish the performance of the method with
contaminated samples.
10.0 REFERENCES
1. Lopez-Avila, V., N. S. Dodhiwala, J. Benedicto, and R. Young, "SFE/IR Method for
the Determination of Petroleum Hydrocarbons in Soils and Sediments," EPA 600/X-92/046 (W.F.
Beckert, Project Officer), US EPA, Environmental Monitoring Systems Laboratory, Las Vegas, April
1992.
2. Pyle, S. M., and M. M. Setty, "Supercritical Fluid Extraction of High-Sulfur Soils with
Use of a Copper Scavenger," Talanta, 1991, 38 (10), 1125-1128.
11.0 SAFETY
11.1 When liquid carbon dioxide comes in contact with skin, it can cause burns because of its
low temperature. Burns are especially severe when the carbon dioxide is modified with organic
liquids.
3560 - 6 Revised
December 1992
-------
11.2 The extraction fluid, which may contain a modifier, usually exhausts through an exhaust
gas port on the rear of the panel of the extractor. This port must be connected to a chemical fume
hood to prevent contamination of the laboratory atmosphere.
11.3 When liquid carbon dioxide is used for cryogenic cooling, typical coolant consumption
is 5 L/min (as decompressed gas), which results in a carbon dioxide level of 900 ppm for a room of
4.5 m x 3.0 m x 2.5 m, assuming 10 air exchanges per hour. The NIOSH time-weighted average
concentration is 9,000 ppm (American Conference of Governmental Industrial Hygienists, 1991-1992).
3560 - 7 Revised
December 1992
-------
APPENDIX B
PROPOSED DRAFT PROTOCOL FOR DETERMINATION
OF TOTAL RECOVERABLE PETROLEUM HYDROCARBONS
BY INFRARED SPECTROPHOTOMETRY (METHOD 8440)
B-l
-------
METHOD 8440
TOTAL RECOVERABLE PETROLEUM HYDROCARBONS BY
INFRARED SPECTROPHOTOMETRY
1.0 SCOPE AND APPLICATION
1.1 Method 8440 (formerly Method 9073) is used for the measurement of total petroleum
hydrocarbons (TPHs) extracted with supercritical carbon dioxide from sediment, soil, and sludge
samples using Method 3560.
1.2 Method 8440 is not applicable to the measurement of gasoline.
1 .3 Method 8440 can detect TPHs at concentrations of 10 ^ig/mL in extracts. This translates
to 10 mg/Kg in soils when a 3-g sample is extracted by SFE (assuming 100 percent extraction
efficiency) and the final extract volume is 3 mL.
1.4 All organic solvents used in this method should be recovered and recycled.
2.0 SUMMARY OF METHOD
2.1 Soil samples are extracted with supercritical carbon dioxide using Method 3560.
Interferences are removed with silica gel, either by shaking the extract with loose silica gel, or by
passing it through a silica gel solid-phase extraction cartridge. After infrared (IR) analysis of the
extract, TPHs are quantified by direct comparison with standards.
3.0 INTERFERENCES
3.1 The parameter being measured (TPHs) is defined within the context of this method. The
measurement may be subject to interferences, and the results should be interpreted accordingly.
3.2 Determination of TPHs is a gross measure of mineral oils only, and does not include the
biodegradable animal greases and vegetable oils captured in oil-and-grease measurements. These non-
mineral-oil contaminants may cause positive interferences with the IR analysis if they are not
completely removed by the silica gel cleanup.
4.0 APPARATUS AND MATERIALS
4.1 Infrared spectrophotometer — Scanning or fixed wavelength, for measurement around
2950cnTl.
4.2 IR cells - 10-mm, 50-mm, and 100-mm path length, made of sodium chloride or IR-grade
glass.
4.3 Optional - A vacuum manifold consisting of glass vacuum basin, collection rack and
funnel, collection vials, replaceable stainless steel delivery tips, built-in vacuum bleed valve and gauge
8440 - 1 Revision WG 0
December 1992
-------
is recommended for use when silica gel cartridges are used. The system is connected to a vacuum
pump or water aspirator through a vacuum trap made from a 500-mL sidearm flask fitted with a one-
hole stopper and glass tubing.
5.0 REAGENTS
5.1 Reagent-grade chemicals shall be used in all tests. Unless otherwise indicated, all
reagents shall conform to the specifications of the Committee on Analytical Reagents of the American
Chemical Society, where such specifications are available. Other grades may be used, provided it is
first ascertained that the reagents are of sufficiently high purity to permit their use without adversely
affecting the accuracy of the determinations. For cleanup, use silica gel cartridges or loose silica gel.
5.2 Solvents — Spectrophotometric grade, or equivalent.
5.2.1 Tetrachloroethylene, C2C14
5.3 Materials for the preparation of the reference oil mixture -- Spectrophotometric grade, or
equivalent.
5.3.1 n-Hexadecane,
5.3.2 Isooctane, (CH3)3CCH2CH(CH3)2
5.3.3 Chlorobenzene, CfiHjCl
5.4 Silica gel
5.4. 1 Silica gel solid-phase extraction cartridges (40-/im particles, 60-A pores), 0.5 g,
Supelco, J.T. Baker, or equivalent.
5.4.2 Silica gel, 60 to 200 mesh, Davidson Grade 950, or equivalent (deactivated with
1 to 2 percent water).
5.5 Calibration mixtures
5.5.1 The material of interest, if available, or the same type of petroleum fraction, if
it is known and original sample material is unavailable, shall be used for the preparation of
calibration standards. Reference oil is to be used only for unknowns. Whenever possible, a
GC fingerprint should be run on unknowns to determine the petroleum fraction type.
5.5.2 Reference oil - Pipet 15.0 mL n-hexadecane, 15.0 mL isooctane, and 10.0 mL
chlorobenzene into a 50-mL glass-stoppered bottle. Maintain the integrity of the mixture by
keeping the bottle stoppered, except when withdrawing aliquots. Refrigerate at 4°C when not
in use.
8440 - 2 Revision WG 0
December 1992
-------
5.5.3 Stock standard - Pipet 0.5 mL calibration standard (Section 5.8.1 or 5.8.2) into
a tared 100-mL volumetric flask and stopper immediately. Weigh and dilute to volume with
tetrachloroethylene.
5.5.4 Working standards - Pipet appropriate volumes of stock standard (Section 5.5.3)
into 100-mL volumetric flasks according to the cell size to be used. Dilute to volume with
tetrachloroethylene. Calculate the concentrations of the standards from the stock standard
concentrations.
5.6 Calibration mixture for silica gel cleanup - Prepare a stock solution of corn oil by placing
about 1 mL (0.5 to 1 g) of corn oil into a tared 100-mL volumetric flask. Stopper the flask and weigh
to the nearest milligram. Dilute to the mark with tetrachloroethylene, and shake the contents to effect
dissolution. Prepare additional dilutions to cover the range of interest.
6.0 SAMPLE COLLECTION, PRESERVATION, AND HANDLING
6.1 Solid samples should be collected and stored as any other solid samples containing
semivolatile analytes.
6.2 Samples should be analyzed with minimum delay, upon receipt in the laboratory, and must
be kept refrigerated prior to analysis.
7.0 PROCEDURE
7.1 Prepare solid and sludge samples according to Method 3560.
7.2 Add 0.3 g of the loose silica gel to the 3-mL extract and shake the mixture for 5 minutes,
or pass the extract through a 0.5-g silica gel solid-phase extraction cartridge (conditioned with 5 mL
tetrachloroethylene). When working with loose silica gel, filter the extract through a plug of
precleaned silanized glass wool in a disposable glass pipet.
7.3 After the silica-gel cleanup, fill a clean IR cell with the solution and determine the
absorbance of the extract. If the absorbance exceeds the linear range of the IR spectrophotometer,
prepare an appropriate dilution and reanalyze. The possibility that the absorptive capacity of the silica
gel has been exceeded can be tested at this point by repeating the cleanup and determinative steps.
7.4 Select appropriate working standard concentrations and cell path lengths according to the
following ranges:
Path length (mm)
10
50
100
Concentration range
(/ig/mL of extract)
5 to 500
1 to 100
0.5 to 50
Volume (mL)
3
15
30
8440-3
Revision WG 0
December 1992
-------
Calibrate the instrument for the appropriate cells using a series of working standards. It is not
necessary to add silica gel to the standards. Determine absorbance directly for each solution at the
absorbance maximum at about 2950 cm"1. Prepare a calibration plot of absorbance versus
concentration of petroleum hydrocarbons in the working standards.
7.5 Determine the concentration of petroleum hydrocarbons in the extract by comparing the
response against the calibration plot.
7.6 Calculate the concentration of TPHs in the sample using the formula:
RxDx V
Concentration (mg/Kg) =
W
where:
R = mg/mL of TPHs as determined from the calibration plot
V = volume of extract in milliliters
D = extract dilution factor, if used
W = weight of solid sample in kilograms.
7.7 Recover the tetrachloroethylene used in this method by distillation or other appropriate
technique.
8.0 QUALITY CONTROL
8.1 Reagent blanks or matrix-spiked samples must be subjected to the same analytical
procedures as those used with actual samples.
8.2 Refer to Chapter One for specific Quality Control procedures and to Method 3500 for
sample preparation procedures.
9.0 PRECISION AND ACCURACY
9.1 Table 1 presents a comparison of certified values and the values obtained using Methods
3560 and 8440. Data are shown for both Freon-113 and tetrachloroethylene, since both solvents were
found to be acceptable as collection solvent. However, only tetrachloroethylene is recommended as
collection solvent for TPHs in Method 3560.
9.2 Tables 2 and 3 present accuracy and precision data from the single-laboratory evaluation
and the interlaboratory evaluation of Methods 3560 and 8440, respectively. These data were obtained
by extracting samples at 340 atm/80°C/60 min dynamic (Table 2) or at 340 atm/80°C/30 min dynamic
(Table 3).
8440 - 4 Revision WG 0
December 1992
-------
10. REFERENCES
1. Rohrbough, W. G.; et al. Reagent Chemicals. American Chemical Society Specifications. 7th
ed.; American Chemical Society, Washington, DC, 1986.
2. Methods for Chemical Analysis of Water and Wastes: U.S. Environmental Protection Agency.
Office of Research and Development, Environmental Monitoring and Support Laboratory.
ORD Publication Offices of Center for Environmental Research Information, Cincinnati, OH,
1983; EPA-600/4-79-020.
3. Lopez-Avila, V., N. S. Dodhiwala, J. Benedicto, and R. Young, "SFE/IR Method for the
Determination of Petroleum Hydrocarbons in Soils and Sediments," EPA 600/X-92/046,
U. S. EPA, Environmental Monitoring Systems Laboratory, Las Vegas, NV, April 1992.
4. Lopez-Avila, V., R. Young, and R. Kim, "Interlaboratory Evaluation of an Off-Line SFE/IR
Method for the Determination of Petroleum Hydrocarbons in Solid Matrices," EPA 600/X-
93/XXX, U.S. EPA, Environmental Monitoring Systems Laboratory, Las Vegas, NV, January
1993.
8440 - 5 Revision WG 0
December 1992
-------
TABLE 1. CERTIFIED AND SPIKE VALUES COMPARED WITH RESULTS
OBTAINED BY METHODS 3560/8440
Reference Material
Spike level or
certified level
(mg/Kg)
Methods
3560/8440
(mg/Kg)
Environmental Resource Assoc.
TPH-1 soil (Lot 91012) 1,830
Environmental Resource Assoc.
TPH-2 soil (Lot 91012) 2,230
Clay soil spiked with kerosene 100
Clay soil spiked with Ijght gas oil 100
Clay soil spiked with heavy gas oil 100
Clay soil spiked with medium neutral oil 100
Clay soil spiked with heavy neutral oil 100
Clay soil spiked with heavy lube oil 100
Environmental Resource Assoc.
TPH-1 soil (Lot 91017) 614
Environmental Resource Assoc.
TPH-2 soil (Lot 91017) 2,050
SRS103-100 soil 32,600
1,920 ± 126a
2,150 ± 380*
86.0; 93.0"
84.0; 98.0b
103; 108"
125 ± 19.4C
126 ± 15.8d
118; 155"
562; 447b
1,780; 1,780"
29,100 ± l,930e
Three 60-min extractions. The extracted materials were collected in Freon-113; the concentrations
were determined against the reference oil standard.
Duplicate 30-min extractions. The extracted materials were collected in tetrachloroethylene; the
concentrations were determined against standards made from the spiking materials, except TPH-1
and TPH-2 soils where reference oil was used to determine concentrations.
Six 30-min extractions. The extracted materials were collected in tetrachloroethylene; the
concentrations were determined against standards made from the spiking material.
Four 30-min extractions. The extracted materials were collected in tetrachloroethylene; the
concentrations were determined against standards made from the spiking material.
Three 30-min extractions. The extracted materials were collected in tetrachloroethylene; the
concentrations were determined against the reference oil standard.
8440 - 6
Revision WG 0
December 1992
-------
TABLE 2. SINGLE-LABORATORY METHOD ACCURACY AND PRECISION FOR
METHODS 3560/8440 FOR SELECTED MATRICES
Matrix
Clay soil3
ERA TPH-1"
(Lot 91016)
ERA TPH-2"
(Lot 91016)
SRS103-100b
Certified or
spike value
(mg/Kg)
2,500
2,350
1,450
32,600
Spike
Material
Motor oil
Vacuum oil
Vacuum oil
c
Method
accuracy
(% recovery)
104
80.3
88.6
94.2
Method
precision
(% RSD)
8.5
19.7
19.6
4.0
a Eight determinations were made using two different supercritical fluid extraction systems. The
extracted materials were collected in Freon-113.
b Ten determinations were made using three different supercritical fluid extraction systems. The
extracted materials were collected in Freon-113.
c This is a standard reference soil certified for polynuclear aromatic hydrocarbons. No spike was
added.
8440-7
Revision WG 0
December 1992
-------
TABLE 3. INTERLABORATORY METHOD ACCURACY AND PRECISION FOR METHOD 3560/8440 FOR SELECTED
MATRICES
Compound name
Perkin-Elmer FTIR
ERA TPH-1 soil
ERA TPH-2 soil
SRS 103-100 soil
Clay soil spiked with
motor oil
£ Buck-Scientific IR
00 ERA TPH-1 soil
ERA TPH-2 soil
SRS103-100
Clay soil spiked with
motor oil
True Mean
concentration concentration" srb
(mg/Kg) (mg/Kg) (mg/Kg)
614
2,050
32,600
10,000
614
2,050
32,600
10,000
654
1,850
26,820
7,790
618
1,670
24,750
8,180
111
213
4,320
1,000
113
194
3,740
910
SR°
(mg/Kg)
297
321
9,720
1,660
296
278
8,650
1,500
Percent
RSDrd
17.0
11.5
16.1
12.9
18.2
11.7
15.1
11.1
Percent
RSDRe
45.4
17.3
36.2
21.3
47.9
16.7
35.0
18.3
t
Percent
mean Number of
recovery laboratories
107
90.2
82.3
77.9
101
81.5
75.9
81.8
14
13
14
13
14
13
14
13
*—i (/3
I 2
The number of replicates per laboratory was three.
sr - repeatability standard deviation.
SR - reproducibility standard deviation.
RSDr - repeatability relative standard deviation.
RSDR - reproducibility relative standard deviation.
to o
-------
APPENDIX C
LIST OF INSTRUCTIONS FOR COLLABORATORS
C-l
-------
1- 2~ 3-
4 —
5~
6-
Dear 1 - 3 -:
Thank you for participating in the EPA's interlaboratory study on extracting total petroleum
hydrocarbons (TPHs) from environmental samples using supercritical fluid extraction (SFE). Its
purpose is twofold: first, to verify SFE as a viable alternative to the currently approved extraction
techniques for TPHs, and second, to establish a feasible non-CFC solvent for collecting the material
that is extracted by the supercritical carbon dioxide.
Your package should contain the following:
1. One set of instructions.
2. Forms to record SFE conditions and observations.
3. Nine vials containing soil samples labeled Vials No. 1 through 9. Vials No. 1 through 3
contain at least 10 g soil each. Vials No. 4 through 9 contain 3.00 g soil each. Refrigerate
all samples upon receipt.
4. Twenty empty vials with Teflon-lined screw caps for sending the extracts to MRI-California
Operations.
5. An additional vial is included (labeled MPJ# 000074) that contains anhydrous sodium sulfate
(Mallinckrodt AR); anhydrous sodium sulfate is recommended for use with the sample
contained in Vial No. 9.
If anything is missing or damaged, please contact me immediately for replacement. We made
arrangements with Aldrich to send you tetrachloroethylene (HPLC-grade, Lot No. 06626TX)
immediately. If the lot number differs from this number, do not use, but notify me so that the correct
tetrachloroethylene can be sent to you. SFE-grade carbon dioxide from Scott Specialty Gases will also
be sent to you. Please do not use other grades of carbon dioxide for this study.
After you have collected the 15 extracts, please send them to MRI-California Operations by March
27, 1992 for analysis by IR.
C-2
-------
Mr. Young . March 5, 1992
Page 2
For questions or comments, do not hesitate to write, call, or fax a message to:
Richard Young
MRI-California Operations
625-B Clyde Avenue
Mountain View, CA 94043
Phone: (415)694-7700
Fax: (415) 694-7983
For your information, there are 15 laboratories participating in the study. Our plans are to complete all
IR analyses and the statistical analyses by the end of April 1992, and to present the results of the study
at the 8th Annual Waste Testing and Quality Assurance Symposium in Washington, DC, in July 1992.
Also, when the study is completed, we will finalize the method protocols (Method 3560 for the SFE and
Method 8440 for the IR determination of TPHs) and send you a copy as soon as they are approved by
EPA.
Sincerely,
MIDWEST RESEARCH INSTITUTE -
CALIFORNIA OPERATIONS
Richard Young
Senior Chemist
Enclosures
C-3
-------
Mr. Young March 5, 1992
Page 3
INSTRUCTIONS FOR THE EPA INTERLABORATORY
STUDY ON SFE/IR METHOD FOR TPHs
Your Laboratory Code is: 15
REAGENTS
1. Tetrachloroethylene Aldrich Lot No. 06626TX. Please do not use any other grade of
tetrachloroethylene for this study.
2. Sodium sulfate (anhydrous) known to be free of interferences in the C-H stretch band range 3200
to 2700 cm'1
3. SFE-grade carbon dioxide from Scott Specialty Gases. Please do not use any other grade of
carbon dioxide for this study.
APPARATUS
1. Supercritical fluid extraction system. Follow manufacturer's instructions on how to operate the
system. Make sure that the extraction vessels and fittings are rated to at least 340 atm. Also
make sure that the temperature is maintained during the extraction at 80°C.
2. Since tetrachloroethylene freezes when carbon dioxide is expanded into it, the collection vessel
must be immersed in water of room temperature (20 to 25 °C).
SAFETY
1. A safety feature to prevent overpressurization is required on the extraction system. This feature
should be designed to protect the laboratory personnel and the instrument from possible injuries
or damage resulting from equipment failure under high pressure.
2. Liquid carbon dioxide can cause "burns11 because of its low temperature (-70°C).
3. The material safety data sheets (MSDS) for tetrachloroethylene and sodium sulfate should be
reviewed before their use.
INTERFERENCES
1. The analyst must demonstrate through the analysis of system blanks that the SFE system is free
from interferences. To do this, perform a simulated extraction, using an empty extraction vessel
and a known amount of carbon dioxide under the same conditions as those that normally will be
used for sample extraction, and analyze the extract by IR. The blank should have < 10 ppm
TPHs (based on using approximately 30 mL of carbon dioxide and collecting the extract in 3 mL
tetrachloroethylene). If you cannot perform the IR analysis of the system blank, send the extract
for the system blank, labeled with your laboratory code, the term "blank", and the form
recording SFE conditions and observations, to MRI—California Operations for analysis.
C-4
-------
Mr. Young March 5, 1992
Page 4
2. The extraction vessel(s), frits, restrictor(s), and selector valve may retain solutes. Therefore, it
is good practice to clean the extraction system after each extraction. Replacement of the
restrictor may be necessary when system blanks indicate carryover. At least one system blank
should be performed daily when the instrument is in use. Furthermore, system blanks should be
performed after each extraction of a high-concentration sample. If system blanks continue to
indicate contamination even after replacement of the extraction vessel and the restrictor, the
multiport valve must be cleaned.
3. Be aware of any organic solvents used to clean any parts of the extraction system or glassware
since these may introduce interferents.
SAMPLE PREPARATION
Remove the sample vial from the refrigerator and follow the specific directions for each vial.
1. Vials 1 and 2: perform triplicate extractions using 3-g portions of this material. Weigh 3.00 g
sample each time.
2. Vial 3: perform triplicate extractions using 3-g portions of this material. Due to the concentration
and nature of this sample (may plug restrictors), it is suggested that it be extracted last.
3. Vials 4 though 8: extract the entire contents of the vial weighing the vial before and after
emptying to verify that the vial contained 3.00 g sample. Record any differences from this
amount.
4. Vial 9: extract the entire contents, weighing the vial before and after emptying to verify that the
vial contained 3.00 g sample. Record any differences from this amount. Add 3 g sodium sulfate
as a drying agent.
For each extraction, transfer the sample to a clean extraction vessel and use two plugs of silanized glass
wool to hold the sample in place and to fill the void volume. Attach the end fittings and install the
extraction vessel in the oven. Always use clean frits for each extraction vessel.
EXTRACTION
1. Fill the collection vessel with 3 mL tetrachloroethylene Aldrich Lot No. 06626TX. Do not use
any other grade of tetrachlbroethylene. Since tetrachloroethylene freezes when carbon dioxide
is expanded into it, the collection vessel must be immersed in a beaker with water of room
temperature (20 to 25 °C).
2. Set the SFE pressure to 340 arm and the temperature to 80°C. Extract for 30 min at a flowrate
of 1 to 2 mL/min (as liquid carbon dioxide).
3. Record extraction conditions and observations on the forms provided. Identify each extraction by
vial number. When triplicate extractions are performed on the same sample, add the suffix A,
C-5
-------
Mr. Young March 5, 1992
Page 3
B, or C. That is, the extracts should be labeled 1A, IB, 1C, 2A, 2B, 2C, 3A, 3B, 3C for
matrices in vials 1, 2, and 3, respectively.
4. Transfer the tetrachloroethylene extract to one of the vials provided. Do not readjust the solvent
volume back to 3 mL. The vial should be sealed tightly with the Teflon-lined cap. Label the
vial in a legible and permanent manner hi the following way:
[15] - [extraction identification]
For example, if your lab code is 15 and you have just extracted the third sample from Vial 2,
the label should read "15-2C."
5. Store the extract hi a refrigerator at 4°C until it can be shipped to MRI-California Operations.
EXTRACT SHIPMENT
After all of the extractions have been performed, ship the extracts in the vials provided to:
Richard Young
MRI-California Operations
625-B Clyde Avenue
Mountain View, CA 94043
Ship the extracts packed hi dry ice via overnight carrier. Include the forms with the recorded SFE
conditions and observations. Please ship so that receipt occurs during weekdays (i.e., do not send out
on Friday). Fax a message indicating when you are sending the extracts. Please ship extracts back by
March 27, 1992.
Your participation is highly appreciated, and we look forward to receiving your extracts to make this a
successful study. We also welcome your comments and suggestions concerning this study.
C-6
-------
FORM I: INSTRUMENT OPERATING CONDITIONS FOR SUPREX, ISCO,
DIONEX-LEE SCIENTIFIC, AND CCS EXTRACTION SYSTEMS*
Instrument: Operator: Date:
Supercritical fluid: Supplier: Grade:
Modifier: Percent modifier:
Extraction vessel volume (mL): Number:
Extraction vessel dimensions: ID x length (mm)
Extraction vessel position: Horizontal Vertical
Restrictor: ID x OD x length (mm)
Collection system: Vial size:
Manufacturer: :
Catalog number: ^__
Solvent: Volume: Initial
Final
Operating conditions:
Dynamic or static .
Pressure (atm)
Temperature (°C)
Flowrate (mL/min)
Density (g/mL)
Time (min)
Restrictor temp. ("Q •
Volume of C02 (mL)
' If conditions vary for different samples, submit separate form.
C-7
-------
FORM H: INSTRUMENT OPERATING CONDITIONS FOR HP MODEL 7680A
EXTRACTOR
SAMPLE HISTORY
Original entry:
Last update:
SAMPLE INFORMATION
Sample name:
Sample ID:
Operator:
Sample amount:
Comments:
EXTRACTION STEP
FLUID DELIVERY
Density: g/mL
Pressure: bars (convert bars to atm)
Flowrate: mL/min
Extraction fluid: C02
EXTRACTION CHAMBER
Chamber temperature: °C
Equilibration time: min
Extraction time: min
Thimble size: mL
Thimble volumes swept:
ANALYTE TRAP
Analyte: Intermediate volatile
Trap material:
Nozzle temperature: °C
Trap temperature: °C
C-8
-------
FORM H: INSTRUMENT OPERATING CONDITIONS FOR HP MODEL 7680A
EXTRACTOR (continued)
FRACTION OUTPUT
Rinse Solvent Volume Rate Nozzle Trap Vial
substep name (mL) (mL/min) temperature temperature number
C-9
------- |