PROCEEDINGS
THIRD ANNUAL SYMPOSIUM
United States
Environmental Protection Agency
Symposium
on
^••MHHMHMH^HHMMHMHHM
SOLID WASTE TESTING
and
QUALITY ASSURANCE
HH|HHHH
Volume II
JULY 13-17, 1987
WASHINGTON, D.C.
WESTIN HOTEL
Symposium Managed by American Public Works Association
-------
PROCEEDINGS
THIRD ANNUAL SYMPOSIUM
*nfl^
United States
Environmental Protection Agency
Symposium
on
•HHMMH^MMHBMM^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^••••••BIMI
SOLID WASTE TESTING
and
QUALITY ASSURANCE
•••••••
Volume II
JULY 13-17, 1987
WASHINGTON, D.C.
WESUN HOTEL
Symposium Managed by American Public Works Association
-------
PROCEEDINGS INTRODUCTION
One of the major environmental problems facing the United States, as
well as other nations, is the need for safe handling and disposal of
hazardous waste. A fundamental component of all programs relating to
waste management is the need to perform measurements. These
measurements include waste composition and properties; effectiveness
of management processes; engineering properties of materials used in
constructing thanagement units; and, last but not least, long term
performance of such management units. Thus, the pivotal roles played
by the measurement methodology and, its attendent, quality assurance.
The analysis of complex waste matrices presents the environmental
community with demanding analytical problems for which solutions are
being developed at a rapid rate. This annual symposium series,
presented by the EPA’S Office of Solid Waste, is designed to focus on
recent developments in testing methods and quality assurance of
importance to both the RCRA and CERCLA programs.
The symposium highlights developing requirements for quality assurance
as well as new analytical procedures intended to be used in EPA’s
national RCRA and CERC A hazardous waste management programs. Our
purpose in holding these symposia is several fold. First, as a means
of communicating what EPA is doing regarding the activities EPA has
already initiated to upgrade the state—of—the—art as reflected in the
regulations and in SW—846. Second, to describe the direction EPA’s
program is taking with respect to testing and quality assurance
issues. Third, as a forum for discussion between Agency personnel and
representatives from public and private laboratories involved in waste
sampling and evaluation.
DhVID FRIEDMAN
CHIEF, METHODS SECTION
OFFICE OF SOLID WASTE
-------
PROGRAM COMMITI’EE
David Friedman
Chief Methods Section
Office of Solid Waste
U.S. EPA
Denise Zabinski
Chemist
Office of Solid Waste
U.S. EPA
David Bennett
Chief, Toxics Integration Branch
Hazardous Site Evaluation Division
(WH—548A)
U.S. EPA
Billy Fairless
Chief EMCM/ENSV
Region 7
U.S. EPA
Paul Friedman
Chemist
Office of Solid Waste
U.S. EPA
Duane Geuder
Chemist
Office of Emergency and
Remedial Response
U.S. EPA
Gary Ward
Chemist
Office of Remedial Response
U.S. EPA
Connie Glover
Manager
Lancy Environmental
Services Division
Liew Williams
Deputy Director
Quality Assurance and Methods
Research Division
U.S. EPA
Las Vegas, NV
Gail Hansen
Chemist
Office of Solid Waste
U.S. EPA
Kenneth Jennings
Environmental Scientist
Office of Waste Program
Enforcement
U.S. EPA
Tom Logan
Engineer
Environmental Monitoring
and Support Lab
U.S. EPA
Research Triangle Park,NC
William Loy
Chemist
Environmental Services
Division
Region 4
Athens, G
Theador Martin
Research Chemist
Environmental Monitoring
and Support Lab
U.S. EPA
Cincinnati, OH
Florence Richardson
Quality Assurance
Officer
Office of Solid Waste
U.S. EPA
Reva Rubenstein
Chief, Health Assessment
Section
Office of Solid Waste
U.S. EPA
Robert Stevens
Chief
California Department
of Health Services
-------
TABLE OF CONTENTS
Volume II
ORGANICS
Hierarchical Approach to the Analysis of Hazardous Organic
Compounds in Waste Waters
R. Kornfeld et al. 6—1
Review of Studies Concerning Effects of Well Casing Materials
on Trace Measurements of Organic Chemicals
R. Dowd 6-29
Application of Wide—Bore Capillary Column to the Analysis of
Volatile Organic Compounds by Method 8240
R. Slater, J. Longbottom 6—45
Determination of Formaldehyde in Samples of Environmental Origin
11. Bicking et al. 6—59
Heat Purge—Trap—Desorb Analysis of Volatile Water Soluble Compounds
S. Lucas et al. 6—81
Evaluation of Extraction Conditions for Appendix IX Compounds
T. Pressley, R. Einhaus 6—83
Preliminary Evaluation of Test Method for Volatile Organics
in Hazardous Waste: Batch Steam Stripping Distillation
A. Gholson 6—103
Preparation of Radioactive Mixed Waste Samples for Measurement
of RCRA Organic Compounds
B.A. Tomkins, J. Caton 6—105
Environmental Applications of Magic LC/T4S
A. Apffel 6—119
Comparison of Capillary Column and Packed Column Analysis
for Volatile Organics
R. Clark, J. Zalikowski 6—137
An Example of Interlaboratory Method Validation Studies in the U.S.
Environmental Protection Agency, Methods 3510 and 8270
J. Wesselman 6—151
Detecting Coeluting Compounds in CC/MS
N. Low 6-161
-------
The Identification of Selected Synthetic Surfacants From a
Cc lex Waste Matrix Using Thermospray Liquid Chromatography/Mass
Spectr m etry
P. Goodley 6—165
Evaluation of Method 3640 (GPC Cleanup) for l ppendix VIII Analytes
P. Marsden, J. Longbottom 6—167
Use of Wide—Bore Capillary Columns For the GC Analysis
of Environmental Samples
P. Marsden, 7. Fisk, D. Betowski 6—175
A Data Base for Establishment of Pre Analytical Holding Times
N. Maskarinec, J. Goodin, R. Moody 6—183
The Determination of Chlorophenoxyacid Herbicides by Liquid
thranatography Using Carbon—14 Tracers
R. Merriweather et al. 6—197
Novel Extraction Solvents for Environmental Samples
R. Zweidinger, L. Sheldon 6—199
An Expert System to M d in Using sw—846.
A. Bethke, A. Gaskill, R. ‘rruesdale 6—201
Liquid thra tography Mass Spectrometry: An Evolving Technique
D. Sauter 6—203
Mthi miin Detection Limts and Data Analysis
1. Warren 6—212
Development of Roboticized Analytical Methods
J. Cleland 6—213
Expert System for Interpretation of the Infrared Spectra
of Hazardous Waste Drum Samples
S. Levine, Y. Li—shi, S. Tcnnellini 6—215
Improving Sonication Techniques in CLP Organics Analysis
and Solid Waste Extraction
S. Berliner 6—229
Introducing the Third Edition of SW—846: Test Method
for Evaluating Solid Waste
N. Layne et al. 6—231
QUALITY ASSURANCE
U.S. Army Toxic and Hazardous Materials Agency Installation
Restoration Quality Assurance Program
K. Lang 7-1
Autcznated Maintenance and Reporting of Analytical Quality
-------
Assurance on a Personal Computer
R. Beaty, L. Richarson 7—7
Laboratory and Field Audits as Part of the EPA Hazardous Waste
Engineering Research Laboratory (HWERL) Quality Assurance Program
B. Blackburn, G. Simes 7—29
Review of Audits and Analyses in New Jersey’s Laboratory
Certification Program
D. Stainken, D. Bowyer, R. Fischer 7—39
Landfill Construction — Quality Assurance Beyond Testing
B. Woodward 7-49
Development of a Special Analytical Services (SAS) SOP for
Laboratory Performance on Volatile Method and Trip (Field)
Blanks Associated with Potable and Low—Level Monitoring Well
Samples
F. Genicola, Y. Lee, J. Rose 7—61
Development of Standards for EPA Hazardous Waste Methodologies
N. Mosesman, J. Crissman 7—85
Quality Assurance of Analytical Chemistry Through Auditing
E. Klesta, M. Marcus 7—95
Preparation of Natural Matrix Type Samples for Performance
Evaluation of Resource Conservation and Recovery Act (RCRA)
Contract Laboratories
H. Clements, R. Loebker, D. Klosterman 7—101
A Field Audit Program to Ensure the Quality of Environmental
Measurements
W. Lowry, S. Borgianini, C. Andreas 7—107
Performance Audits Recommended for Volatile and Semi—Volatile
Organic Measurements During Hazardous Waste Trial Burns
R. Jayanty et al. 7—117
Design and Use of Laboratory QC Programs to Satisfy DQOs for
Environmental Measurement Systems
E. Brantly et al. 7—126
RcRA Experience in Southeast Florida
C. Ouseph 7—127
Remedial Investigations Guidance Strategy
L. Grayson et al. 7—141
Developing an Environmental Laboratory Accreditation System
3. Locke 7—151
A Pre—Remedial Investigation Study as an Alternative Approach
in the Site Remediation Process
-------
in the Site Remediation Process
C, Andreas et al. 7—163
SAMPLING AND FIELD METHODS
Using Barcodes and Portable Computers for Sample Tracking
D. Hooton 81
Evaluation of a Prototype Field—Portable X—Ray Fluorescence System
for Hazardous Waste Screening
G. Raab et al. 8—5
Canister—Based Samplers for Volatile Organics
W. McClenny, J. Pheil 8—39
Catalog of Field Screening Methods
A. Szilagyi et al. 8—47
A Field Deployable Analytical Instrument for Analysis of
Semi Volatile Organic Compounds of Superfund Sites
E. Overton, S. Martin 8—55
Develo m nt of Sampling/Monitoring Guidance for the RCRA
Hazardous Waste Regulatory Program
G. Swanson, S. Schulberg, I. Show 8—71
Rapid Field Analysis of Volatile Organic Compounds in
Environmental Samples Utilizing a Microchip Gas Chromatograph
E. Overton 8—91
Sampling Techniques for Evaluation of Tarry Waste Impoundments
R. Dhonau, C. Ritzert 8—115
-------
ORGANIC
thai rpersons
Paul. Friedman William Loy
themist Chemist
Office of Solid Waste Environmental
U.S. EPA Services Division
401 M Street, S.W. Region 4
Washington, D.C. 20460 college Station Road
Athens, G 30613
-------
HIERARCHICAL APPROACH TO THE ANALYSIS
OF HAZARDOUS ORGANIC COMPOUNDS IN WASTE WATERS
Richard A. Kornfeld, Projects Manager, Judith E. Gebhart, Section
Manager, J. Scott Warner, Research Leader, and Samuel V. Lucas,
Principal Research Scientist, Battelle Columbus Division, Columbus,
Ohio; James E. Longbottom, Chief, Organic Analyses Section, U.S.
Environmental Protection Agency, Cincinnati, Ohio
INTRODUCTION
The Resource Conservation and Recovery Act (RCRA)( 1 ) specifies over 300
toxic organic compounds in its Appendix VIII listing which may be used
to identify hazardous wastes. In response to a petition by the state of
Michigan, the U.S. Environmental Protection Agency (EPA) has proposed
the amendment of RCRA Appendix VIII by the addition of over 100 ott,e
organic compounds to give a total of over 400 organic constituents. ’ 2
Gas chromatographic methods for determining organic compounds in wastes
are given in SW—846, Test Methods for Evaluating Solid In
many cases these methods are modifications of procedures for the
determination of some, but not all, of the Appendix VIII and Michigan
List compounds in wastewater. EPA is currently attempting to validate
analytical methods for as many of these 400 plus compounds as possible.
A hierarchical approach to these validation efforts is being pursued.
An example of a hierarchical approach to the development and validation
of analytical methods for the determination of organic compounds in
wastes is presented in Figure 1. This figure is presented to show the
context in which the work reported here leads to subsequent method
development activities. For example, compounds which are not amenable
to determination by GC-MS are to be evaluated in future programs for the
feasibility of alternate approaches including special GC conditions,
derivatization prior to GC analysis, high performance liquid
chromatography (HPLC), and/or non—chromatographic methods. Compounds
which are amenable to GC—MS are to be subjected to further method
performance evaluation.
Volatile compounds will be evaluated for amenability to the purge-trap-
desorb (PTD) techniques described in Method 5030. For compounds which
are not successfully recovered from an aqueous matrix using this
procedure, alternate procedures such as heated PTD, direct liquid
injection (DLI), or the Method 5020 headspace sample introduction
technique can then be evaluated. Compounds which are successfully
recovered using the Method 5030 room temperature PTD procedure can be
included in a validation study of Methods 5030/8240 using procedures
described in the Single Laboratory Method Validation Protocol which was
developed at Battelle.
The semivolatile compounds shown to be amenable to GC-MS determination
will be evaluated for amenability to the extraction procedures described
6-1
-------
I. Add to list
2. Test In
no .,-MS
t Piods
1. Add to list
2. Test in
-P ’S
u thods
Figure 1. Hier3rcPlical Approach for Analytical ethod Development
for Organic RCRA Analytes
1. HPtC
2. Derlvatizat1oi
3. Non-chromitograptuic
Volatile
.1
(vdluate Other OçtiOns:
1. Heated P lO
2. Direct injectloø
3. Micro eitraCtiofl
4. Oistt 1ation
S. Non -CRC i tPiods
I. Nodifled extrac-
tion conditions
( 5ff. other sos-
vents, etc.)
2. Develop/improve
c1e nup pro-
cedures
6-2
-------
in i . etY:od 8270. For compounds which are not successfully re vE red from
an aqueous matrix using this procedure, alternate extraction sche: es nd
novel techn ques, such as so Id phase extraction (SPE , car Le
considered. Compounds successfully recovered sing the Method 820
extraction procedure can be included in evaluations of several sample
cleanup procedures. These investigations will inv lve fortifying a wide
variety of waste matrices with the compounds to permit selection of
appropriate cleanup procedures for compounds/sample matrix combinations.
Ultimately, these compounds will be included in a validation of Method
8270 using the Single Laboratory Method Validation Protocol.
Implementation of the hierarchical method development approach is
expected to result in the development of a suite of analytical methods
consisting of a limited number of procedures for the determination of a
large proportion of the more than 400 organic compounds. This limited
number of analytical procedures for extraction steps, cleanup steps, and
determinative steps will form the core of a generic approach to the
selection of appropriate analytical methods addressing hazardous wastes.
The system will be generic in the sense that the specifications of type
of analyte, type of matrix, and type of sensitivity and required
specificity will generate, from the limited suite of component
analytical procedures, the most appropriate set of analysis conditions.
The reduction in the number and variety of methods required to
characterize wastes will provide cost benefits both to the government
and to the regulated coninunity. This generic approach.will facilitate
periodic updates as new information becomes available about specific
analytes in specific matrices. Another advantage of this generic
approach is that the areas requiring method modification or method
development will be clearly identified. Consequently, this collection
of research requirements can then be prioritized for resource
allocation.
Results from the first phases of this approach are reported in this
manuscript and presentation and in two other presentations by Battelle
scientists. The first phase defines the scope of work for the research
program and involves the ideitification of compounds which are amenable
to GC separation and MS detection. These evaluations involved the
analysis of standard solutions using the GC-MS conditions described in
the Contractor Laboratory Protocol (CLP) for the application of SW-846
Methods 8240 and 8270 for volatile and sernivolatile organic compounds,
respectively. Compounds suitable for GC separation and MS detection
were classified as candidates for further Method 8240 or Method 8270
testing.
Two subsequent phases were also explored. Volatile compounds determined
to be amenable to gas chromatographic separation were evaluated for
purging efficiency using three non—MS detection methods, SW-846 Methods
8010, 8015, or 8020. Using results from room temperature PTD and DLI
experiments, purging efficiencies can be calculated. The purging
efficiencies from room temperature PTD experiments, coupled with GC-MS
6-3
-------
data for these compounds, can also be used to select those compounds to
be validated by Method 8240.
Semivolatile organic compounds found suitable for GC separation and MS
detection were further evaluated for Method 8270 suitability by
determining their recovery from water and their seven day aqueous
stability using a modified SW-846 Method 3510 extraction procedure.
Compounds exhibiting acceptable recovery and stability are candidate
compounds to be validated by Method 8270.
EXPERI MENTAL
GC-MS Suitability
The initial set of ar alytes consisted of organic compounds included I
RCRA Appendix viii(’i plus those included in the Michigan petition( 2
minus the EPA priority pollutants. After eliminating redundancies in
the two lists, remaining compounds were classified as to their predicted
suitability for SW-846 Method 8240 (volatiles), Method 8270 (semi—
volatiles), or for their predicted inability to be determined by either
method.
Sources for the selected analytes were identified in the following order
of priority: 1) the EPA repositories of reference compounds and
pesticides (EMSL — Las Vegas and RIP), 2) the EPA repository of
certified solutions (EMSL - Cincinnati), and 3) commercial suppliers.
GC-MS suitability studies were performed using analyte mixtures prepared
after consideration of chemical reactivity.
Individual analyte concentrations in the volatile mixtures were 200
ug/mL for most of the analytes or 400 ug/mL for a few analytes predicted
to exhibit lower response factors. Injections of volatile analytes
provided a minimum of 300 ng of analyte on column. The concentrations
of individual semivolatile analytes in the injection standards were 40
ug/rnL. For analytes not detected on the first attempt, higher
concentrations were employed, ranging from 50-400 ug/mL. Injections of
semivolatile analytes provided a minimum of 80 ug of analyte to the
splitless injection evaporator cavity. The usual packed GC column, 1
percent SP1000/Carbopack B (Supelco), was used for volatile compounds,
and a 30 meter x 0.25 mm ID fused silica coated with 0.25 micron
immobilized methyl phenyl silicone (J&W DB-5) was used for semivolatile
compounds.
Internal standards specified in the CLP for both volatile and
sernivolatile analyses were used to provide measures for both GC relative
retention indices and MS detection response factors. Surrogate
standards specified in the CLP were included in volatile analyte
mixtures but not for the semivolatile mixtures. The significantly
greater number of these in the latter case would have made data
interpretation more difficult without increased usefulness of the
results obtained. The CLP GC and MS analysis conditions and MS quality
control checks on ion source tuning were used in all cases.
6-4
-------
Volatiles Purging Efficiency
The compounds included in this study were chosen from two sources, 1)
volatiles listed for each of the three SW—846 Methods and 2) volatiles
deemed suitable for GC separation from the GC-MS suitability studies.
Methods 8010, 8015, and 8020 provide packed-column gas chromatographic
conditions for the determination of certain volatile organic compounds.
Samples were analyzed using these Methods in conjunction with purge-
trap-desorb, Method 5030 or direct liquid injection (DLI). Detection is
achieved by a halogen specific detector for Method 8010, a flame
ionization detector for Method 8015, and a photoionization detector for
Method 8020.
Individual standard solutions were combined to form spiking mixtures
that would avoid co-elutions. PTD analyses were performed after adding
3 uL of the spiking solution to 5 mL of water. Compound concentrations
ranged from 40-800 ug/L and were estimated from previous experience to
be high enough to give sufficient response for a reliable calculation of
purging efficiency.
Sernivolatiles Recovery and Stability
The compounds included in the original scope of Method 3510 performance
testing are derived from Appendix VIII( 3 ), the Michigan list’ 2 J and the
proposed Appendix IX and borderline chemicals iists.( ) After
eliminating redundancies and compounds predicted to be unsuitable for
this study, sources were sought for analytes from first, the EPA
repositories of reference compounds and pesticides and second,
commercial suppliers.
Individual stock and internal standard solutions were prepared.
Spiking mixtures were designed to eliminate coelutions since flame
ionization detection (FID) was used after GC separation. In addition,
acidic and basic compounds were segregated to avoid chemical
interactions; different GCs were used to analyze mixtures containing
acids or bases.
Spiking mixtures were added to one liter aliquots of water and prepared
for GC-FID quarititation using SW—846 Method 3510 modified for the base
and acid extraction steps to use one 300-mL aliquot of methylene
chloride for each step instead of three 60—mL aliquots in each
extraction step. Peak identification was accomplished using an internal
standard while quantitation resulted from external calibration.
Compound mixtures were injected into a splitless injection evaporator
cavity and separated with a 30 m x 0.25 mm ID fused silica capillary
column coated with a 0.25 urn film of immobilized methyl phenyl silicone
(Supelco SPB—5).
Stability studies used the same analytical procedure for spiked water
samples that had been stored at approximately 4 C in the dark for 7
days.
6-5
-------
RESULTS AND DISCUSSION
GC-MS Suitability
After eliminating EPA priority pollutants from further consideration,
328 compounds were considered for suitability testing. After omitting
compounds that were not obtainable and those which were believed to not
be within the scope of work for this project, the remaining compounds
were classified either as volatiles or semivolatiles.
The 54 volatile analytes tested with Method 8240 GC conditions are
listed in Table 1 in which the status is indicated as satisfactorily
detected CS), detected with a response factor versus benzene—D6 below
0.02 (LR), or not detected (ND). Thirty—three analytes were
satisfactorily detected and 6 were detected with low response factors.
The very low response factors are predicted to result in unacceptably
high method detection limit (MDL) values for Method 8240.
Table 1 also lists 15 volatile analytes that were not detected using the
Method 8240 conditions. All of these non-detected analytes were
analyzed at least twice, with the repeat analysis usually at 2 to 5-fold
higher levels than the original 300 ng level. Three of those 15
analytes, hexachioropropene, tetranitromethane, and thiophenol, were
thought to have failed to elute due to boiling points and/or polarities
that were too high for the SP1000/Carbopack B column, and these
compounds were retested using the Method 8270 (semivolatile analyte)
conditions.
Non—detection of the hydrazines and aziridines (6 analytes) was probably
due to extreme GC peak tailing on the SPI000/Carbopack B column. Five
of these nitrogen bases were also tested with the semivolatile analytes.
The sixth, N(2-hydroxyethyl)ethyleneimine, was not tested due to its
extreme polarity.
2-Butanone peroxide apparently quantitatively decomposed to 2-butanone
in the injector. Methyl mercaptan apparently coelutes with methanol on
the SP1000/Carbopack B column and, in this case, would be substantially
lost at the jet separator due to the presence of the methanol vapor
displacement of the helium carrier. The remaining 4 undetected volatile
compounds, 2 haloethers, methyl isocyante, and 2-methyllactonitrile,
were not repeated in the semivolatile set since they were both too
volatile to be recovered in a Kuderna-Danish (KD) distillation of
extraction solvent and were also known to be chemically and/or
hydrolytically labile.
The 185 semivolatile analytes plus the 8 volatile analytes to be
retested with Method 8270 conditions are listed in Table 2. The status
is indicated as satisfactorily detected (S), expected to be satisfactory
for GC—MS determination based on other information (ES), detected with
a response factor less than 0.02 versus phenanthrene—D 10 (LR), or not
detected (ND). There were 128 analytes detected with satisfactory
response factors and 9 analytes detected with low response factors. All
6—6
-------
9 of these analytes are highly polar and reasonably expected to be
sensitive to thermal decomposition in the injection port.
Table 2 contains 11 analytes that are indicated with status “ES , which
designates that non-detection of the analyte is considered anomalous.
All 11 of these analytes were found to be suitably analyzed by gas
chromatography in other projects. Thus, although MS data was not
previously obtained, it seems reasonable to classify their non-detection
in the present work as anomalous. Except for the two organophosphates,
these anomalously non-detected Table 2 analytes are strongly basic
molecules, so that a possible explanation for their non—detection would
be that the GC column used was somewhat acidic precluding satisfactory
elution.
The 45 analytes for which non-detection in the GC—MS data cannot be
classified as anomalous are also listed in Table 2. Generally, these
analytes are highly polar or are labile to decomposition before or
during chromatography. Four of the 45 analytes are aromatic diamines,
1,2- and 1,3—phenylenediamine, 2,4-diaminoanisole, and 1,5-
naphthalendiamine. These 4 analytes can probably be analyzed by fused
silica GC if special precautions are taken to ensure good performance
for basic materials. Ethylene thiourea (ETIJ) has been shown in previous
work in our laboratory to be amenable to GC analysis using special
conditions. For another six analytes (acrylamide, cycloheximide, 2-
fluoroacetamide, niclosamide, oxydemeton-methyl, and thioacetamide)
polarity, volatility and lability considerations apparently do not
account for the non—detection, and, therefore, a more thorough attempt
to develop GC—based methods might be successful. For the remaining 34
analytes, the causes of non-detection can be classified as one or more
of the following: exceptionally high polarity, thermal or chemical
lability, or insufficient volatility. Recommendations for further
method development for these analytes focus on HPLC techniques,
especially ion chromatography or post column derivitization methods.
Volatiles Purging Efficiency
Fifty—seven of the 84 volatile organic compounds included in these
experiments produced a measurable recovery for room temperature PTD.
The largest group of compounds, those classified as being suitable for
Method 8010 evaluation, produced measurable purging efficiencies for 40
of the 53 compounds tested. Ten of the 21 Method 8015-classified
compounds and 7 of 10 Method 8020-classified compounds could be detected
at room temperature PTD conditions. Tables 3, 4, and 5 show the results
for Methods 8010, 8015, and 8020 respectively.
There were 7 compounds (1,2-dibromo—3-chloropropane in Method 8010
experiments and 1,4—dioxane, isobutanol, 2-butanone, 4-methyl-2-
pentanone, beta-propiolactone, and propionitrile in Method 8015) for
which PTD recoveries were low enough (20% or less) to question their
satisfactory validation using room temperature PTD methodology. Four of
these compounds have undergone preliminary evaluation using heated PTD;
the results will be reported at this symposium.
6—7
-------
Precision, as measured by RSD, was 15% or more for only 11 of the 84
tested compounds. Chlorobenzene, chloroform, chloromethane, 1,1-
dichioroethane, 1 ,2-dichloropropane, tetrachioroethylene,
trichioroethylene, and vinyl chloride in Method 8010 all have recoveries
in the 50—90% range while 1,4—dioxane, 4-methyl-2-pentanone and beta-
propiolactone in Method 8015 all had low recoveries.
Semivolatiles Recovery and Stability
An average recovery or stability of less than 70% was considered an
indicator of potential incompatibility for future validation testing
since a compound must be adequately stored and extracted before being
analyzed. Table 6 lists 41 compounds that did not fulfill both 70%
criteria. Two of these compounds, dioxathion and dibenz(a,e)pyrene,
were not detected in either sample extracts or calibration standards.
Dioxathion is perhaps unstable to heat while apparently
dibenz(a,e)pyrene did not elute from the GC column during data
collection. Possible reasons for poor recovery and/or stability are
given in Table 6 and include unfavorable distribution coefficients,
hydrolysis during extraction or storage, and oxidation during extraction
or storage.
The 115 compounds exhibiting suitable recovery and stability are listed
in Table 7. While all 115 compounds are considered candidates for
subsequent validation testing, 13 had poor reproducibility. Five
compounds, 1,2:7,8—dibenzacridine, 3,3’-dimethoxybenzidine, p-
dimethylaminoazobenzene, methyl parathion, and 4,4’-oxydianiline, had
relative standard deviation (RSD) values for extraction efficiency
greater than 15% on Day 0 and/or Day 7. Ten compounds, carbophenothion,
1,2:7,8—dibenzacridine, dichiorovos, 3,3’-dimethoxybenzidine, 3,3’ -
dimethylbenzidine, fenthion, 4,4’-methylenebis(N,N-dimethylaniline), N-
nitrosopyrrolidine, 1,4-phenylenediamine, and sulfallate, had RSD values
greater than 15% for GC-FID response factors from the highest level
calibration standard (100 ug/mL). Two compounds, 1,2;7,8-dibenzacridene
and 3,3’-dimethoxybenzidine exhibited poor reproducibility for both
recovery and response factor measures. These 13 compounds may prove not
to be suitable for semivolatile analysis when validation studies are
performed.
CONCLUSIONS AND RECOMMENDATIONS
The principal conclusion of the studies reported in this manuscript is
that the hierarchical approach is a logical means for developing and
validating analytical methods for the determination of organic compounds
in wastes. Not only will this scheme allow for the logical evaluation
of alternative analytical approaches, but the hierarchical system will
allow new candidate compounds to be rapidly screened with a limited
suite of component analytical procedures.
Recommendations for specific parts of the hierarchical approach
discussed in this manuscript are:
6-8
-------
o Of the volatile organic compounds tested for GC-MS
suitability and purging efficiency, the 57 listed in
Table 8 are recommended for inclusion in SW-846 Method
8240 validation studies.
o Based on their suitable GC-MS analysis, extraction, and
stability, the 115 semivolatile organic compounds listed
in Table 7 should be included in SW—846 Method 8270
validation studies. Preliminary experiments should
include 1) extraction/stability studies at
concentrations lower than the 500 ug/L values used in
this investigation and 2) extraction of compounds under
conditions milder than the strongly acidic and strongly
basic conditions employed in Method 3510. Testing of
the compounds failing extraction! stability criteria may
allow some of them to be included in subsequent
validation experiments which use milder extraction
conditions.
o Based on results for semivolatile, non—priority
pollutant organic compounds, the following changes to
proposed Appendix IX are recommended. These compounds
affected are contained in Table 9.
- the 12 compounds listed in Table 9A should
be removed from Appendix IX. -
— 3 compounds on the borderline list should
be added to Appendix IX (see Table 9B).
- 10 compounds from Appendix VIII that did
not appear on proposed Appendix IX should
be included on the Appendix IX list (see
Table 9B).
REFERENCES
1. 40 Code of Federal Regulation, Part 261.
2. Federal Register, 49, No. 247, December 21, 1984, pp. 49784—49793.
3. Federal Register, 49, No 191, October 1, 1984, pp. 38786-38809.
4. Federal Register, 51, No. 142, July 24, 1986, pp. 26632—26642.
6-9
-------
DISC LA I MER
Although the research described in this article has been funded by the
U.S. Environmental Protection Agency, it has not been subjected to
agency review and therefore does not necessarily reflect the views of
the agency and no official endorsement should be inferred.
6-10
-------
TABLE 1. GC-MS SUITABILITY TESTING RESULTS FOR VOLATILE ANALYTES
-i Acetonltrile
2 Allyl alcohol
3 Allyl chloride
4 Benzyl chloride
5 Sls—(2-chloroethyl) sulfide
6 Bls(chloromethyl) ethir
7 Brosoacetone
8 2-Butanon. p.roxid*
9 2—Butanone
10 Carbon disulfide
11 Chioral hydrate
12 2-Chloroethanol
13 Chloro.eth methyl ether
14 Chloroprene
15 3—Chloropropionitrile
16 1.2-Dlbroso-3—chloropropane
17 Dlbro.o.ethane
18 1,4—01chloro-2-but ne
19 D I chlorodi fl uorom.thane
20 1 .3—Dlchloro—2-propanol
21 1,2,3,4—Dlepoxybutane
22 1 1 1-Dlmethylhydrazine
23 1,2—Dlmethylhydrazlne
24 1,4—Dloxane
25 Epichlorohydrin
26 Ethylene dlbromide
27 Ethylene ox1d
28 Ethylenimine
29 Ethyl methacrylate
30 Hexachloropropene
31 N—(2—Hydroxyethyl )ethyleneimlne
32 2—Hydroxyproplonl tn le
33 Isobutyl arEohol
34 Malononitrile
35 Methacrylonltnlle
36 2—Methylaziridfne
37 Methyl hydrazine
38 Methyl Iodide
39 Methyl Isocyanate
40 2—Methyllactonltrile
41 Methyl mercaptan
42 Methyl methacrylate
43 Pentachloroethane
44 2-Plcoline
45 Propargyl alcohol
46 —Proplolactone
47 Proplonltrile
48 N—Propyla.lne
49 PynidIne
50 Styrene
51 1,1,1.2—Tetrachloroethane
52 Tetranitromethane
53 Thiophenol
54 1,2,3-Trlcnloropropane
(a) 8 = Appendix Ylli; N • Michigan list
(b) LR; low response factor
S: suitable for GC-MS analysis
ND: not detected In GC-MS data
8 75-05-8
8 107-18-6
8 14 107-05-1
8 100-44—7
8 M 505-60-2
8 542-88—1
8 598—31-2
8 1338-23-4
8 78-93—3
8 75—15—0
8 75-87-6
14 107-07-3
8 107-30—2
8 14 126—99—8
8 542-76—7
8 96—12-8
8 74—95—3
8 764—41—0
8 75—71—8
8 96—23—1
8 1464—53—5
8 57-14—7
8 540-73-8
8 123—91—1
8 106-89—8
8 106-93-4
8 75—21—8
8 151-56—4
8 97-63—2
8 1888-71-7
N 1072-52—2
M 78—97-7
8 78-83-1
8 109—77-3
8 126—98—7
8 75—55—8
8 60-34-4
8 74-88-4
8 624-83-9
8 75—86-5
8 74—93—1
8 80-62-6
8 76—01—7
8 109-06-8
8 107-19—7
14 57—57—8
8 107- 12-0
8 107-10-8
8 110-86-1
N 100-42-5
8 630—20-6
8 509-14-8
8 108-98-5
8 96—18-4
RCRA
Status
No. Substance Llst(a) CAS No. Number
Code(b)
U00 3
P005
U317
P028
P158
Poll
U 160
U034
P133
U046
U276
P027
U074
U085
U098
U099
U 108
U115
P054
Ui 18
U243
U289
Ui 49
Ui 52
P067
P068
U 138
P064
P069
U162
U 191
P102
U302
P101
U 194
U 196
U323
P112
P104
S
S
S
LR
ND
S
ND
S
S
LR
LR
ND
S
S
S
S
S
S
S
S
ND
ND
S
S
S
S
ND
S
ND
ND
LR
S
S
S
ND
ND
S
ND
ND
ND
S
S
S
LR
S
S
IR
S
S
S
ND
ND
S
6-11
-------
TABLE 2. GC-MS SUITABILITY TESTING RESULTS FOR SEMIVOLATILE ANALYTES
98-86-2 S
8 53—96—3 S
8 591-08—2 LR
8 79—06-1 ND
8 116—06—3 ND
N 117—79—3 S
N 60-09-3 S
8 N 92-67-1 S
N 132—32-1 ES
8 61-82—5 ND
N 101-05—3 S
8 62-53-3 ES
N 90-04-0 S
8 N 140—57-8 S
8 492-80-8 ND
N 86-50-0 S
N 101-27-9 IR
N 17804—35-2 NI)
8 106—51—4 S
N 1689-84-5 S
8 357-57-3 ND
N 2425-06-1 S
N 133-06-2 S
N 63-25-2 S
N 1563-66-2 S
N 786-19-6 S
N 470-90-6 S
8 106—47-8 S
8 510—15—6 S
N 95-79-4 S
N 6959—48-4 S
N 5131—60—2 ES
N 95-83—0 ES
H 56-72-4 S
N 120-71-8 S
N 7700—17—6 5
N 135-20—6 NO
H 66—81-9 ND
8 131—89—5 LR
8 50—18-0 MD
N 8065-48-3 S
8 2303-16-4 S
H 39156—41-7 ND
8 N 95-80-7 S
H 333—41—5 ES
8 224-42-0 S
8 192—65—4 S
N 117-80-6 S
8 87-65-0 S
N 62-73—7
N 141-66—2 S
8 56-53—1 S
H 64—67—5 LR
8 56312-13—1
8 60—51—5 S
8 119-90—4
8 60—11—7 S
8 57—97—6 5
8 119-93—7 S
8 57—14—7 MD
8 540—73—8 ND
8 122-09—8 S
8 99—65—0 S
8 528-29—0 S
RCRA
Status
No. Substance Llst(a) CAS No. ISus ber
Code(b)
1 Acetophenone
2 2-Acetyl asinofluorene
3 1-Acetyl—2-thiourea
4 Acrylaaide
S Aldicerb
6 2-ui noanthraquinons
7 A.irtoazob.nzens
8 4—Aainobipheeyl
9 3-i .ino—9—ethylcarbozole
10 Aaitrol.
11 Anliezine
12 Aniline
13 o—Anisid1
14 Arepite
15 Auraalne
16 AZlflDhOS - thy1
17 Barban
18 knayl
19 p-8.nzoqulnone
20 8rosoxyi 1l
21 Brucin.
22 Ciptifol
23 Captan
24 Cirbaryl
25 Carbofuran
26 Carbop enothion
27 Chlorfenvlnphos
2$ 4—Chloroaniline
29 Chlorobenzilate
30 5—Chloro—2-eethylanhllrle
31 3—(Chloro.ethyl )pyridlne hydrochloride
32 4 —Chioro—1,3 -phenylefledlaullfle
33 4—Chloro—1,2-pheflylefledlaelfle
34 CousapPuos
35 p-Cresldlne
36 Crotoxyphos
37 Cupferron
38 CyclohexiMide
39 2—CycloPtexyl-4 ,6-dlnltrophenol
40 Cyc 1 ophosphaml d i
41 Demeton
42 Diallate
43 2,4—D1.mlnoanisole sulfate
44 2,4-Olaminotoluene
65 Dlazlnon
46 1,2:7,8—Dlbenzacrldlfli
47 1 .2:4 • 5—01 benzopyrerle
48 Dlchione
£9 2,6 -OicPulorODhenOl
50 Dlchlorovos
51 Dlcrotooftos
52 Dlethylstllbestrol
53 Diethyl sulfate
54 1hydrOSafr01e
55 Dimetnoate
56 3,3’ —Dirnetnoxybenzldlfle
57 1,4 -Dimethylam lnoazObenzefle
58 7 ,12_ Dlmethylbenz(a)aflthracefle
59 3,3• -Dlmethylbeflzldlne
60 1,1—Dlnethylhydrazlfle
61 1,2 -Dlnethylhydrazlfle
62 _Olnethylphenethylanlfle
63 1,2-Dlnltrobel zene
64 1,3-Dlnltrob.nzenl
0004
0005
P002
0264
0257
0274
U253
0011
0333
0260
0326
0014
P 1 51
0280
0271
0197
0272
P018
0285
0266
U279
0121
0148
P143
0038
0329
0319
0305
U306
P130
0262
0238
U290
P134
P034
0058
P155
0062
0307
U327
0313
U299
P144
P146
U086
U325
U090
1J09 I
U09 3
0094
0095
U098
0099
P046
6-12
-------
TABLE 2. (Continued)
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
1 78
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
100- 5—4
N 39300—45—3
88-85-7
N 78-34-2
$ 57-41-0
122-66-7
298-04-4
N 2104—64—5
N 563- 12—2
51-79-6
151—56—4
96-45—7
62-50—0
52—85—7
N 115—90-2
N 55-38—9
N 33245—39—5
640—19—7
70—30—4
1888—71—7
N 680—31-9
N 123—31-9
465—73—6
N 54—85—3
120—58—1
143—50—0
N 21609—90—5
N 121—75—5
108— 31—6
123—33—1
N 72—33—3
91-80-5
16752—77—5
7 2-43—5
7 5—55—8
56—49—5
101—14—4
N 101—61—1
60-34—4
Ui 64
P131
P128
U297
P147
U 309
U298
1 )166
1)167
1)168
P072
1)321
P075
U 250
1)263
1)275
U288
P132
S
S
S
S
S
ND
S
S
S
S
NO
ND
S
S
S
S
S
MO
S
S
S
S
S
ND
S
S
S
S
S
M D
S
S
ND
S
ND
S
ES
MD
S
MD
S
S
S
S
MD
S
S
S
S
S
MO
S
S
ES
MO
MO
S
S
S
S
S
S
I0
No. Substance
Lfst(a) CAS
No.
RCP.A
Number
Status
Coce(b)
--
U284
P153
U 109
P141
P154
U238
P054
UI 19
P097
P156
U330
P057
1)132
1)243
U312
P060
1)141
P14.0
1)324
U 147
1)301
1)155
P066
P067
Ui 57
U158
1)255
P068
U163
1,4— Oinitrobenzene
Dinocap
Dinoseb
Dioxithion
5, 5—01 phenyl hydantol n
1,2-Dipheny lhydraz lne
Di sul foton
EPN
Ethf on
Ethyl carbiaate
Ethylen l.in.
Ethylene t b1ourea
Ethyl .ethan.sul fonate
Fa hur
Fensulfothion
Fenthion
Fluchioralin
2- Fl uoroacetaai d .
Hexachlorophene
Ntxachl oroprcpene
Hexasethyl phosphoraside
Hydroqul none
I sod ri n
Isonicotinic acid hydrazloe
Isosafro le
Kepone
Leptophos
Malathion
Maleic anhydride
Maleic hydrazide
Mestranol
Met! pyr1 le n.
Metho.yl
p,p’-Methoxychlor
2-Methyl azirid lne
3-Methylcholanthre n.
4,4k -Methyleneb ls(2—chloroanhllne)
4,4’-Methylenebis(M,iI—d lmethylanil lne)
Methyl hydrazine
Methyl nethanesul fonate
N—Methyl—N—ni trv—N—nltrosoguanidine
Methyl parathion
2-Me thyl phenol
3-Methy lphenol
4-Me thy lpheno I
Methyl thiouracil
Mevinphos
Mexacarbate
Ml rex
Monocrotophos
Na led
I ,5—Naphthalenedlatine
1, 4-Napntroguthone
1-Naphtnyta.lne
2—Na phthylaml re
1—Naphthyl-2-th lourea
Niclosamiøe
N’ cotine
5-Ni troacenappthene
4—Nltroanlllne
5—NI tro—o— an ls 1dm.
4-NI trobiphenyl
Nitrofen
Nitrogen e sUrd
70— 25—7
298—00—0
95-48-7
108— 39—4
106—44—5
56-04—2
$ 7786—34—7
N 315—18-4
$ 2385—85—5
N 6923—22—4
N 300—76-5
N 2243—62—1
8 130-15-4
8 134—32-7
8 91—59-8
8 86—88-4
N 50—65-7
8 54—11—5
N 602-87—9
8 100—01-6
$ 99—59—2
N 92—93-3
$ 1836—75—5
8 N 51—75-2
6—13
-------
TABLE 2. (Continued)
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
14$
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
171
178
179
180
181
182
183
184
185
186
181
NI troqiycerin.
5—NI trp—o-toluidln.
4-4 5ltroqulnol me—i-oxide
N—NI trosodi butyl ulne
N-N ltr osod lethanola lne
N-N itr osodlethylaaine
p-Nltrosodl phenyl aalne
N-Nltroso-N-.thylurea
i .Nltroscsethy ithyl ailile
N-il tro*o-N-uethyl urea
i-NItroso-N-..t$yluretI ane
N-li troscuorphoi lea
N—MitrosoØperid lne
N-Nltrosopyrrol I dine
Octaisthyl pyrophosphorsalde
Oxydemiton-aethyl
4,4’—OxydianI lIne
Parathion ethyl
Pentaclilorobenzene
Pentacbloronl trobonzene
Phenadetin
Phenazopyridine hydrochloride
Phenobarb ltal
1.2—P lienyleead la.Ine
1 ,3—Phenyl enedl amine
, .4—Phenylenedl i.ine
N—Phenyl thiourea
Phorate
Phosi lone
Phosaet
Phospflaa ldon
Phthallc anhydride
Plperenyl sulfoxide
Pronaa lde
1,3-Propane sultone
Propyl thiouracl 1
Resorc lnol
Rotenone
Saccharin
Safro le
Strychnine
Sulfa I late
Terbufos
1.2,4 ,5—Tetrachlorobenzene
2,3,4 ,6—Tetrachlorophenol
Tetrachlorvinphos
Tetrdethyl dl thiopyrophosphate
Tetraethyl pyrophosphate
Tetrani tromethan,
Thi oacetami de
Thiofanox
Th lonaz In c
Thiophenol
Toluene dilsocyanate
o-ToluIdine
Trichiorfon
2,4,5—Trich lorophenol
0,0,O—Trlethyl phosphorothioate
Trifluralin
55-63—0
99—55-8
56—51—5
924—16—3
1116—54—7
55-18-5
N 156-10-S
159—73—9
10595—954
684—93—5
615—53-2
59—89-2
100-75-4
930-55—2
152-16—9
N 301-12—2
N 101-80—4
8 56-38—2
8 608-93-5
8 82-68—8
8 62-46-2
N 136-40-3
II 50-06-6
8 9 5-5 4 -5
8 108-45-2
8 106-50-3
8 103—85—5
8 298-02-2
8 2310-17—0
N 732—11—6
N 13171—21—6
8 85-44-9
N 120-62—7
8 23950—58—5
8 1120-71—4
8 N 51—52—S
8 108-44-3
N 83-79-4
8 81—07—2
8 94-59—7
8 57-24-9
N 95-06-7
N 1307 1—79—9
8 95-94—3
8 58-90—2
N 961-11-5
8 3689—24—5
8 107-49—3
8 509-14—8
8 62—55—5
8 39196—18—4
8 297-97-2
8 108-98—5
8 584-84—9
8 N 95-53—4
N 52-68-6
8 95—95—4
8 126—68—1
N 1582-09-8
ND
S
S
S
ND
S
ES
ND
S
ND
ND
ES
S
S
LR
NO
S
S
S
S
S
NO
S
ND
NO
S
NO
S
S
S
S
S
S
S
ND
LR
S
NO
ND
S
S
$
S
S
S
S
ES
S
NO
S
‘4D
S
S
S
S
S
ES
S
RCRA
Status
No. Substance Llst(a) CAS No. Number
Code(b)
8
8
8
8
8
8
8
8
8
8
8
8
8
8
Foal
0181
0173
0287
0176
0177
0178
U116
P085
P157
0303
0187
0320
0268
P093
P145
U193
U334
0273
0202
0203
0277
P149
0308
P109
P112
U128
P045
P040
P104
U223
U328
P139
U332
6-14
-------
TABLE 2. (Con :nL eci’
a-”-
--
©.
S4bstanca
LlSt(i)
i
N jpber
tat .s
Coce )
188
2 4,5—Tr1 et ylan 1tne
N
: ;-17—7
2 9
189
T metnyl P ospt ate
N
E 1 2-56—1
u O
I
i90
1.3,5-Tr1n tro enzene
8
99354
L2 4
I
191
192
Tris(2,3— lbro o ropyfl
Trl-p—tolyl pnosphate(c)
os ata
8
N
126—72—7
78—32-3
L235
.R
S
193
Warfarln
8
81-81—2
POOl
ND
(a) 8 • Appendix VIII; N M1c 1gar 1 st
(b) S — apparantly suitable for GC—MS analysis
LR - low sponse; response factor, versus phenanthr ne—D1O, less than 0.02
ND - not detected
ES - expected to be suitable for 5C-MS analysis but not detected in this study
(c) Substituted for the non—specific rnixture, tricresy phosphate
6—15
-------
TABLE 3. ROOM TEMPERATURE PTD RECOVERIES OF COMPOUNDS BY USE OF SW-846 METHOD 8010
= fi == = Sannflaa flfla.aflnfl aSS anna saaana eaannaaaaxa —raaaanaasn—a=z =
No. Substance List(a) CAS No.
Conc.
(ug/L)
40
800
(c)
400
(c)
400
100
200
200
100
100
(c)
400
100
100
400
160
100
80
100
400
400
400
80
80
160
80
160
80
160
160
80
100
200
100
200
100
400
400
400
40
100
80
(f)
200
Recovery
(percent) (b)
Mean RSD
ND (d)
ND
81
107
65
77
81
ND
51
85
ND
ND
88
76
73
ND
90
ND
83
109
14
71
83
82
80
30
109
86
103
78
107
90
ND
100
ND
78
86
87
102
4
1
12
12
6
20
13
18
3
18
2
4
4
5
8
6
1
2
2
7
1
16
1
3
3
18
1
4
14
3
2
88 2
25 6
1 Allyl chloride
2 Benzyl chloride
3 Bis(2-chloroethoxy)methane
4 Bis(2-chloroethyl)sulfide
5 Bis(2-chloroisopropyl)ether
6 Bromoacetone
7 Bromobenzene
8 Bromodichioromethane
9 Bromoform
10 Bromomethane
11 Carbon tetrachloride
12 Chioroacetaldehyde
13 Chioral hydrate
14 Chlorobenzene
15 Chloroethane
16 2-Chioroethanol
17 2-Chioroethyl vinyl ether
18 Chloroform
19 1-Chiorohexane
20 Chioromethane
21 Chloromethyl methyl ether
22 Chioroprene
23 3-Chioropropionitrile
24 Chiorotoluene
25 Dibromochioromethane
26 1 ,2-Dibromo-3-chloropropane
27 1,2-Dibromoethane
28 1,2-Dichlorobenzene
29 1,3-Dichlorobenzene
30 1 ,4-Dichlorobenzene
31 1,4-Dichloro-2-butene
32 Dichiorodifluoromethane
33 1,1-Dichloroethane
34 1,2-Oichloroethane
35 1,1-Dichloroethylene
36 trans-1,2-Dichloroethylene
37 1,2-Dichloropropane
38 1..3-Dichloro-2-propanol
39 1,3-Dichioropropene
40 Epichiorohydrin
41 Methylene bromide
42 Methylene chloride
43 Methyl iodide
44 Pentachioroethane
45 1,1,2,2-Tetrachloroethane
9 B 107-05-1
8 100-44-7
9 111-91-1
M 505-60-2
9 108-60-1
8 598-31-2
(e) 108-86-1
9 75-27-4
9 75-25-2
9 74-83-9
9 56-23-5
8 107-20-0
8 75-87-6
9 108-90-7
9 75-00-3
M 107-07-3
9 110-75-8
9 67-66-3
(e) 544-10-5
9 74-87-3
8 107-30-2
9 B 126-99-8
9 B 542-?6-7
(e) 95-49-8
9 124-48-1
9 96-12-8
9 106-93-4
9 95-50-1
9 541-73-1
9 106-46-1
9 110-57-6
9 B 75-71-8
9 75-34-3
9 107-06-2
9 75-35-4
9 156-60-5
9 78-87-5
B 96-23-1
9 10061-01-5
8 106-89-8
9 75-95-3
9 75-09-2
9 B 74-88-4
9 76-01-7
9 79-34-5
6-16
-------
TABLE 3. (Continued)
== =====fl—_=fl st_st s==sttt—_=s=flzss ss=fl=t—__=_=s_s = = 55=5. =s== s===n========
Recovery
(percent) (b)
Conc.
No. Substance List(a) CAS No. (ugh) Mean RSD
46 1,1,1,2-Tetrachioroethane 9 630-20-6 160 85 7
47 Tetrachioroethylene 9 127-18-4 100 51 17
48 1,1,1-Trichioroethane 9 71-55-6 200 97 4
49 1,1,2-Trichloroethane 9 79-00-5 40 83 5
50 Trichloroethylene 9 79-01-6 100 85 15
51 Trichlorofluoromethane 9 75-69-4 100 82 13
52 1,2,3-Trichioropropane 9 96-18-4 40 50 7
53 Vinyl chloride 9 75-01-4 100 81 16
nnz.._fl_sstfls_SSfltfl__s__fl__s S..... SSssSSSflaSSSSSSSfl ..s.sns.......s.ss.n.n. . S
(a) 9 — Proposed Appendix IX to Part 264 as published in the Federal Register,
51, No. 142, July 24, 1986 pp 26639-26642.
B — Borderline chemicals considered for additions to proposed Appendix IX to part
264 and published in the Federal Register, 51, No. 142, July 24, 1986, p 26637.
M Michigan list of chemicals proposed to be added to Appendix VIII to Part 261 and
and published in the Federal Register, 49, No. 247, December 21, 1984, p 49793.
8 = Appendix VIII to Part 261 as revised and published in the Federal Register, 51,
No. 247, August 6, 1986, pp. 28305-28310. Compounds so designated are not present
on proposed Appendix IX , borderline chemicals, or Michigan Lists.
(b) Calculated from three replicates by comparing PTD vs DLI peak areas.
(c) Apparent decomposition in solution and/or on-column.
(d) ND — not detected.
(e) Included on SW-846 Method 8010 list but not present on proposed Appendix IX,
borderline chemicals, Michigan, or Appendix VIII lists.
(f) Poor chromatographic behavior on the system used.
6—17
-------
TABLE 4. ROOM TEMPERATURE PlO RECOVERIES OF COMPOUNDS BY USE OF SW-846 METHOD 8015
nsnnnan ssa_flsannflassfl flmflflSt.tflSfl
Recovery
(percent) (b)
Conc.
No. Substance List(a) CAS No. (ug/L) Mean RSD
1 Acetonitrile B 75-05-8 800 (c)
2 Acrylamide 8 79-06-1 800 ND (d)
3 Allyl alcohol 9 B 107-18-6 800 ND
4 Carbon disulfide 9 75-15-0 200 ND
5 1,2,3,4-Diepoxybutane 8 1464-53-5 800 MD
6 Diethyl ether (e) 60-29-7 200 90 11
7 1,4-Dioxane 9 8 123-91-1 800 1 100
8 Ethylene oxide 9 B 75-21-8 800 (f)
9 Ethyl methacrylate 9 97-63-2 200 55 14
10 2-Hydroxypropionitrile P 1 78-97-7 (g)
11 Isobutanol 9 78-83-1 800 2 (b)
12 Malononitrile 9 B 109-77-3 800 ND
13 Methacrylonitrile 9 B 126-98-7 800 37 (b)
14 Methyl ethyl ketone 9 78-93-3 200 14 12
15 Methyl mercaptan (h) 74-93-1 200 (f)
16 Methyl methacrylate 9 80-62-6 200 55 11
17 4-Methyl -2-pentanone 9 108-10-1 200 20 35
18 Paraldehyde 8 123-63-7 800 ND
19 Propargyl alcohol 9 B 107-19-7 800 ND
20 beta-Propiolactone K 57-57-8 800 1 100
21 Propionitrile 9 8 107-12-0 800 7 14
= .wS.fl—_= r nnnsnnnnsflflfl flasasanssn SSSSsflnfl a
(a) 9 Proposed Appendix IX to Part 264 as published in the Federal Register,
51, No. 142, July 24, 1986 pp 26639-26642.
B — Borderline chemicals considered for additions to proposed Appendix IX to part
264 and published in the Federal Registers 51, No. 142, July 24, 1986, p 26637.
M Michigan list of chemicals proposed to be added to Appendix VIII to Part 261 and
and published in the Federal Register, 49, No. 247, December 21, 1984, p 49793.
8 = Appendix VIII to Part 261 as revised and published in the Federal Register, 51,
No. 247, August 6, 1986, pp. 28305-28310. Compounds so designated are not present
on proposed Appendix IX, borderline chemicals, or Michigan Lists.
(b) Calculated from three replicates except for isobutanol and methacrylonitrile
which had two replicates. Recovery calculated for PTD vs DLI peak areas.
(c) Interference from methanol, the solvent used to prepare the spiking solutions.
(d) ND not detected.
(e) Included on SW-846 Method 8015 list but not present on proposed Appendix IX,
borderline chemicals, Michigan, or Appendix VIII lists.
(f) Not retained on the trap used for SW-846 Method 8015.
(g) Poor chromatographic behavior under conditions specified by SW-846 Method 8015.
(h) Included on an earlier version of Appendix VIII, not on the August 6, 1986 revised
list.
6-18
-------
TABLE 5. ROOM TEMPERATURE PTD RECOVERIES OF COMPOUNDS BY USE OF SW-846 METHOD 8020
================s===========n=============__===
Recovery
(percent) (b)
Conc.
No. Substance List(a) CAS No. (ug/L) Mean RSD
1 Benzene 9 71-43-2 200 77 8
2 Benzenethiol 9 B 108-98-5 (c)
3 Ethylbenzene 9 100-41-4 200 94 1
4 2-Picoline 9 B 109-06-8 (c)
5 Pyridine 9 110-86-1 (c)
6 Styrene 9 100-42-5 200 86 7
7 Toluene 9 108-88-3 200 99 2
8 o-Xylene 9 (d) 200 92 8
9 rn-Xylene 9 (d) 200 99 1
10 p-Xylene 9 (di 200 98 1
(a) 9 = Proposed Appendix IX to Part 264 as published in the Federal Register,
51, No. 142, July 24, 1986 pp 26639-26642.
B = Borderline chemicals considered for additions to proposed Appendix IX to part
264 and published in the Federal Register, 51, No. 142, July 24, 1986, p 26637.
M Michigan list of chemicals proposed to be added to Appendix VUl to Part 261 and
and published in the Federal Register, 49, No. 247, December 21, 1984, p 49793.
8 = Appendix VIII to Part 261 as revised and published in the Federal Register, 51,
No. 247, August 6, 1986, pp. 28305-28310. Compounds so designated are not present
on proposed Appendix IX, borderline chemicals, or Michigan Lists.
(b) Calculated for three replicates by comparing PTD vs DLI peak areas.
(c) Poor chromatographic behavior under conditions specified by SW-846 Method 8020
(d) Listed in proposed Appendix IX as, Xylene (total) with CAS No. 1330-20-7.
6-19
-------
TABLE 6. COMPOUNDS WITH AVERAGE RECOVERIES AND/OR AQUEOUS STABILITIES LESS THAN 70
PERC ENT
Percent Percent
No. Substance List(a) Recovery Stabiflty(b) Coninents(c)
I Aramite(Isomers 1 and 2) 9 54,52 KS
2 Azinphos-methyl M 62 KS
3 Benzenethiol 9 B 33 (d) CP,OE,OS
4 p-Benzoquinone 9 B 0 OE
S Captafol M 55 KS
6 Captan N 40 KS
7 Demeton-O N 68 KS
8 2,4-Diaminoanisole sulfate N 60 DC,OE
9 2,4-Diaminotoluene N 42 DC,OE
10 Dibenzo(a,e)pyrene 9 B (d) (d) LE
11 Dichlone N 0 OE
12 Diethylstilbestrol 8 67 AW,OS
13 Dihydrosafrole 8 10 KS
14 Dimethoate B 31 HE,HS
15 7,12-Dimethylbenz(a)anthracene 9 45 cP
16 1,4-Dinitrobenzene 9 14 HE
17 Dinocap N 28 CP,HS
18 Dioxathion N (d) (d) CP
19 Ethyl carbamate 8 28 DC
20 Hexachiorophene 9 62 AW,CP
21 Isosafrole 9 B 46 DC
22 Malathion N 5 KS
23 Maleic anhydride 8 0 HE
24 Malononitrile 9 B 9 CP,DC
25 4,4’-Methylenebis(2-chloroaniline) 9 33 0 OE,OS
26 Mexacarbate M 68 HE,HS
27 Monocrotophos N 26 HE
28 1-Naphthylamine 9 44 OS
29 Nicotine 8 67 OE
30 Pentachloroethane 9 64 4 HE,HS
31 1,2-Phenylenediarnine 8 32 DC,OE
32 1,3-Phenylenediamine 8 19 DC,OE
33 Phosalone (e) 65 MS
34 Phosmet M 15 KS
35 Phospharnidon N 63 HE
36 Phthalic anhydride 8 1 67 CP,DC,HE,HS
37 Resorcinol 9 B 10 DC,OE
38 Strychnine 8 55 AW,OS
39 Thioacetamide 8 1 DC
40 Toluene diisocyanate 8 6 HE
41 Trimethyl phosphate M 60 HE
*flSZSZ*flt*S
(a) 9 — Proposed Appendix IX to Part 264 as published in the Federal Register,
51, No. 142. July 24, 1986, pp. 26639 26642.
6- 20
-------
TABLE 6. (Continued)
B Borderline chemicals considered for additions to proposed Appendix IX to part
264 and published in the Federal Register, 51, No. 142, July 24, 1986, p 26637.
M Michigan list of chemicals proposed to be added to Appendix VIII to Part 261 and
and published in the Federal Register, 49, No. 247, December 21, 1984, p. 49793
8 Appendix VIII to Part 261 as revised and published in the Federal Register, 51,
No. 247, August 6, 1986, pp. 28305-28310. Compounds so designated are not present
on proposed Appendix IX, borderline chemicals, or Michigan lists.
(b) Percent Stability = Ave Recovery (Day 7) x 100/Ave Recovery (Day 0).
(c) Comments:
AW Adsorption to walls of glassware during extraction and storage.
CP Nonreproducible chromatographic performance.
DC = Unfavorable distribution coefficient.
HE Hydrolysis during extraction accelerated by acidic or basic
conditions.
MS — Hydrolysis during storage.
LE Late eluting compound.
OE — Oxidation during extraction accelerated by basic conditions.
OS — Oxidation during storage.
(d) Compound not detected in either sample extracts or calibration standards.
(e) Included in an earlier version of Appendix VIII but not on the August 6, 1986 revised
list.
6-21
-------
TABLE 7. COMPOUNDS TENTATIVELY RECOPVIENDED FOR INCLUSION IN SW-846 METHOD 3510(a)
— ss.anaaa..nasnasnsanaannwnnanannsssssassan flaflflflflaflflflflsfls
RCRA
No. Substance CAS No. No. list(b)
I Acetophenone 98-86-2 U004 9
2 2-Acetyla inofluorene 53-96-3 U005 9 B
3 2-Aminoanthraquinone 117-79-3 U264 N
4 Aainoazobenzene 60-09-3 U257 N
5 4-Aalnobiphenyl 92-61-1 U274 9
6 3-Amino-9-ethylcarbazole 132-32-1 U253 N
7 Anilazine 101-05-3 U333 N
8 Aniline 62-53-3 9 8
9 o-Anisidine 90-04-0 U260 N
10 Benzyl alcohol 100-51-6 9
11 Broinoxynil 1689-84-5 U272 I I
12 2-sec-Butyl-4,6-dinitrophenol 88-85-7 9 B
13 Carbaryl 63-25-2 U279 N
14 Carbofuran 1563-66-2 U127 N
15 Carbophenothion(d) 786-19-6 U148 N
16 Chlorfenvinphos 470-90-6 P143 N
17 4-Chloroaniline 106-47-8 9
18 Chlorobenzilate 501-15-6 9
19 5-Chloro-2-methylaniline 95-79-4 U329 N
20 3-(Chloromethyl)pyridine hydrochloride 6959-48-4 U319 K
21 4-Chloro-1,2-phenylenediamine 95-83-0 U306 N
22 4-Chloro-1,3-phenylenediamine 5131-60-2 U305 N
23 Coumaphos 56-72-4 P130 N
24 p-Cresidine 120-71-8 U262 .14
25 Crotoxyphos 7700-17-6 U238 N
26 2-Cyclohexyl-4,6-dinitrophenol 131-89-5 P034 8
27 Demeton-S 126-75-0 14
28 Diallate 2303-16-4 U062 B
29 Diazinon 333-41-5 U313 N
30 1,2:7,8-Dibenzacridine(c,d) 224-42-0 8
31 Dibenzofuran 132-64-9 9
32 1,2-Dibromo-3-chloropropane 96-12-8 9
33 2,6-Dichlorophenol 87-65-0 9
34 Dichlorovos(d) 62-73-7 P144 14
35 Dicrotophos 141-66-2 P146 14
36 3,3’-Dimethoxybenzidine(c,d) 119-90-4 U091 9
37 p-Dimethylaininoazobenzene(c) 60-11-7 U093 9
38 3,3-Dimethylbenzidine(d) 119-93-7 U095 9
39 1,2-Dinitrobenzene 528-29-0 8
40 1,3-Dinitrobenzene 99-65-0 8
41 Diphenylamine 122-39-4 9 8
42 5,5-Diphenyihydantoin 57-41-0 N
43 Disulfoton 298-04-4 9
44 EPN 2104-64-5 P141 14
45 Ethion 563-12-2 P154 14
46 Ethyl methanesulfonate 62-50-0 U119 8
47 Ethyl parathion 56-382 9
48 Famphur 52-85-7 P097 9
6-22
-------
TABLE 7. (Continued)
S s_Sans
RCRA
No. Substance GAS No. No. List(b)
49 Fensulfothion 115-90-2 P156 N
50 Fenthion(d) 55-38-9 N
51 Fluchioralln 33245-39-5 U330 N
52 Hexachioropropene 1888-71-7 U243 9 B
53 Hexamethyl phosphoramide 680.31-9 U312 N
54 Isodrin 465-73-6 P060 9 B
55 Kepone 143-50-0 9 B
56 Leptophos 21609-90-5 P140 M
57 Mestranol 72-33-3 U301 N
58 Methapyrilene 91-80-5 U155 9
59 Methoxychior 72-43-5 9
60 3-Methyicholanthrene 56-49-5 U157 9
61 4,4’-Methylenebis(N,N-dimethylaniline)(d) 101-61-1 U255 N
62 Methyl methanesulfonate 66-27-3 9
63 2-Methylnaphthalene 91-57-6 9
64 2-Methyl-5-nitroaniline 99-55-8 U181 9
65 Methyl parathion(c) 298-00-0 9
66 2-Methyiphenol 95-48-7 9
67 3-Nethyiphenol 108-39-4 B
68 4-Methyiphenol 106-44-5 - 9
69 2-Methylpyridine 109-06-8 U191 9 B
70 Mevinphos 7786-34-7 P131 M
71 Mirex 2385-85-5 U297 N
72 Haled 300-76-5 U309 N
73 1,4-Naphthoquinone 130-15-4 U166 9
74 2-Naphthylamine 91-59-8 U168 9
75 5-Nitroacenaphthene 602-87-9 U250 H
76 2-Nltroaniline 88-74-4 9
77 3-Nitroaniline 99-09-2 9
78 4-Nitroaniline 100-01-6 9
79 5-Nitro-o-Anisidine 99-59-2 U263 N
80 4-Nltrobiphenyl 92-93-3 U275 N
81 Nitrofen 1836-75-5 U288 N
82 4-Nitroquinoline-1-oxide 56-57-5 8
83 N-Nitrosodi-n-butylarnine 924-16-3 9
84 N-Nitrosodlethylamine 55-18-5 9
85 p-Nitrosodiphenylamine 156-10-5 U287 N
86 N-Nitrosomethylethylamine 10595-95-6 9 B
87 N-Nitrosomorpholine 59-89-2 9
88 N-Nitrosopiperidine 100-75-4 1.1176 9
89 N-Nitrosopyrrolidine(d) 930-55-2 9
90 4,4’-Qxydianiline(c) 101-80-4 U303 H
91 Pentachlorobenzene 608-93-5 9
92 Pentachloronitrobenzene 82-68-8 9
93 Phenacetin 62-44-2 U187 9
94 Phenobarbital 50-06-6 U268 N
95 1,4-Phenylenediamine(d) 106-50-3 8
96 Phorate 298-02-2 9
6-23
-------
TABLE 7. (Continued)
a. ......a. ....a.aa ...a...a. asses.
RCRA
No. Substance CAS No. No. LIst(b)
97 Plperonyl sulfoxlde 120-62-7 U270 N
98 Pronamlde 23950-58-5 9
99 Safrole 94-59-7 U203 9
100 Sulfallate(d) 95-06-7 U277 M
101 Terbufos 13071-19-9 P149 N
102 1,2,4,5-Tetrachlorobenzene 95-94-3 9
103 2,3,4,6-Tetrachlorophenol 58-90-2 9
104 Tetrachlorvinphos 961-11-5 U308 N
105 TetraethyI dithiopyrophosphate 3689-24-5 P109 9 B
106 Tetraethyl pyrophosphate 107-49-3 8
107 Thionazine 297-97-2 P040 9
108 o-Toluldine 95-53-4 U328 B N
109 2,4,5-Trlchiorophenol 95-95-4 9
110 0,0,0-Triethyl phosphorothioate 126-68-1 8
111 Trlfluralin 1582-09-8 1J332 N
112 2,4,5-Trimethylanlline 137-17-7 U259 N
113 1,3,S-Trinitrobenzene 99-35-4 U234 8
114 Tri-p-tolyl phosphate 1330-78-5 N
115 Tris-(2,3-dibromopropyl) phosphate 126-72-7 U235 9 B
fltmssssflsaflssflsss.ssflnn w.ssszsass.a.ssssasflsas.ssssssflssssssssssna
(a) Based on a compound having demonstrated both extractability and aqueous
stability values equal to or greater than 70%.
(b) 9 . Proposed Appendix IX to Part 264 as published in the Federal Register,
51, No. 142, July 24, 1986, pp. 26639-26642.
B — Borderline chemicals considered for additions to proposed Appendix IX to part
264 and published in the Federal Register, 51, No. 142, July 24, 1986, p 26637.
N — Michigan list of chemicals proposed to be added to Appendix VIII to Part 261 and
and published in the Federal Register, 49, No. 247, December 21, 1984, p. 49793
8 Appendix VIII to Part 261 as revised and published in the Federal Register, 51,
No. 247, August 6, 1986, pp. 28305-28310. Compounds so designated are not present
on proposed Appendix IX, borderline chemicals, or Michigan lists.
(c) Compounds exhibiting either day 0 or day 7 recoveries with relative
standard deviations greater than 15 percent.
(d) Compounds with calibration response factors having relative standard deviations
greater than 15 percent.
6-24
-------
TABLE 8. COMPOUNDS RECOMMENDED FOR INCLUSION IN SW-846 METHOD 8240
VALIDATION STUDY
.sss=asss .aas a. assasas ansflanssnasssss sasaassssssassss ass.sss assess... ass
No. Substance List(a) CAS No.
1 Allyl chloride 9 B 107-05-1
2 Benzene 9 71-43-2
3 Benzyl chloride 8 100-44-7
4 Bromobenzene (d) 108-86-1
5 Bromodichioromethane 9 75-27-4
6 Bromoform 9 75-25-2
7 Bromomethane 9 74-83-9
8 Carbon tetrachioride 9 56-23-5
9 Chlorobenzene(b) 9 108-90-7
10 Chioroethane 9 75-00-3
11 Chloroform(b) 9 67-66-3
12 1-Chlorohexane (d) 544-10-5
13 Chloromethane(b) 9 74-87-3
14 Chioroprene 9 B 126-99-8
15 Chiorotoluene (d) 95-49-8
16 Dibromochloromethane 9 124-48-1
17 1,2-Dibromo-3-chloropropane(c) 9 96-12-8
18 1,2-Dibromoethane 9 106-93-4
19 1,2-Dichlorobenzene 9 95-50-1
20 1,3-Dichlorobenzene 9 541-73-1
21 1,4-Dichlorobenzene 9 106-46-7
22 1,4-Dichloro-2-butene 9 110-57-6
23 Dichiorodifluoromethane 9 B 75-71-8
24 1,1-Dichloroethane(b) 9 75-34-3
25 1,2-Oichloroethane 9 107-06-2
26 1,1-Dichioroethylene 9 75-35-4
27 trans-1,2-Dichloroethylene 9 156-60-5
28 1,2-Dichloropropane(b) 9 78-87-5
29 1,3-Dichioropropene 9 10061-01-5
30 Diethyl ether (e) 60-29-7
31 1,4-Dioxane(b,c) 9 B 123-91-1
32 Ethylbenzene 9 100-41-4
33 Ethyl methacrylate 9 97-63-2
34 Isobutanol(c) 9 78-83-1
35 Methacrylonitrile 9 B 126-98-7
36 Methylene bromide 9 75-95-3
37 Methylene chloride 9 75-09-2
38 Methyl ethyl ketone(c) 9 78-93-3
39 Methyl iodide 9 B 74-88-4
40 Methyl methacrylate 9 80-62-6
41 4-Methyl-2-pentanone(b,c) 9 108-10-1
42 beta-Propiolactone(b,c) M 57-57-8
43 Propionitrile(c) 9 B 107-12-0
44 Styrene 9 100-42-5
6—25
-------
TABLE 8. (ContInued)
No. Substance list(a) CAS No.
45 1,1,1,2-Tetriebloroethane 9 630-20-6
46 1,1,2,2-Tetrachioroethane 9 79-34-5
47 Tetrachloroethylene(b) 9 127-18-4
48 Toluene 9 108-88-3
49 1,1,1-Trlchioroethane 9 71-55-6
50 1,1,2-Trichioroethane 9 79-00-5
51 TrichIoroethylene(b) 9 79-01-6
52 Trlchlorofluoro.ethane 9 75-69-4
53 1,2,3-Trlchioropropane 9 96-18-4
54 VInyl chlorlde(b) 9 75-01-4
55 a-Xylene 9 (f)
56 0-Xylem. 9 (f)
57 p-Xylene 9 (f)
ssa.eSSaasmSaflsssastassssssasas.aaa.as.san.flnn.ena.sssasSna.Sns
(a) 9 — Proposed Appendix IX to Part 264 as published In the Federal Register,
51, No. 142, Jury 24, 1986 pp 26639-26642.
8 — Borderline chemicals considered for additions_to proposed Appendix IX
to part 264 and published in the Federal Register, 51, No. 142, July 24,
1986, p 26637.
N — Michigan list of chemicals proposed to be added to Appendix VIII to Part
261 and published in the Federal Register, 49, No. 247, December 21,
1984, p 49793.
8 Appendix VIII to Part 261 as revised and published in the Federal
Register, 51, No. 247, August 6, 1986, pp. 28305-28310. Compounds
so designated are not present on proposed Appendix IX, borderline
chemicals, or Michigan lists.
(b) Compounds with relative standard deviations of 15 percent or more.
(c) Compounds with mean recoveries of 20 percent or less.
(d) Included on SW-846 Method 8010 list but not present on proposed Appendix IX,
borderline chemicals, Michigan, or Appendix VIII lists.
(e) Included on SW-846 Method 8015 list but not present on proposed Appendix IX,
borderline chemicals, Michigan, or Appendix VIII lists.
(f) Listed In proposed Appendix IX as Xylene (total) with CAS No. 1330-20-7.
6.- 26
-------
TABLE 9. RECOMMENDED CHANGES FOR SEMIVOLATILE ORGANIC COMPOUNDS ON
PROPOSED APPENDIX IX
snssaasa S*nSaSssaas*nssasssaaessse..aa.a.aasaaas.sa.saS*.
No. Substance List(a) CAS No.
9A. Reconinended Deletions
1 Aramite(Isomers I and 2) 9 140-57-8
2 Benzenethiol 9 B 108-98-5
3 p-Benzoquinone 9 B 106-51-4
4 Dibenz(a,e)pyrene 9 B 192-65-4
5 7,12-Oirnethylbenz(a)anthracene 9 57-97-6
6 1,4-Dinitrobenzene 9 100-25-4
7 Hexachiorophene 9 70-30-4
8 Isosafrole 9 B 120-58-1
9 Malononitrile 9 B 109-77-3
10 4,4’-Methylenebis(2-chloroaniline) 9 101-14-4
11 1-Naphthylamine 9 134-32-7
12 Pentachioroethane 9 76-01-7
13 Resorcinol 9 B 108-46-3
9B. Reconvnended Additions
1 2-Cyclohexyl-4,6-dinitrophenol 8 131-89-5
2 Diallate B 2303-16-4
3 1,2:7,8-Dibenzacridine(b,c) 8 224-42-0
4 1,2-Dinitrobenzene 8 528-29-0
5 1,3-Dinitrobenzene 8 99-65-0
6 Ethyl methanesulfonate 8 62-50-0
7 3-Methyiphenol B 108-39-4
8 4-Nitroquinoline-1-oxide 8 56-57-5
9 1,4-Phenylenediamine(c) 8 106-50-3
10 Tetraethyl pyrophosphate 8 107-49-3
11 o-Toluidine B 95-53-4
12 0,0,0-Triethyl phosphorothioate 8 126-68-1
13 1,3,5-Trinitrobenzene 8 99-35-4
(a) 9 Proposed Appendix IX to Part 264 as published in the Federal Register,
51, No. 142, July 24, 1986, pp. 26639-26642.
B Borderline chemicals considered for additions to proposed Appendix IX to part
264 and published in the Federal Register, 51, No. 142, July 24, 1986, p 26637.
8 — Appendix VIII to Part 261 as revised and published in the Federal Register, 51,
No. 247, August 6, 1986, pp. 28305-28310. Compounds so designated are not present
on proposed Appendix IX, or borderline chemicals lists.
(b) 1,2:7,8-Dibenzacridine exhibits a day 0 recovery with relative standard
deviation greater than 15%.
(c) Compounds with calibration response factors having relative
standard deviations greater than 15%.
6-27
-------
REVIEW OF STUDIES CONCERNING EFFECTS OF WELL CASING MATERIALS
ON TRACE MEASUREMENTS OF ORGANIC COMPOUNDS
Richard M. Dowd, President, R. M. Dowd & Company, 1317 F Street,
N.W., Washington, D.C.
ABSTRACT
This report analyzes the results of laboratory and field studies
that allow a direct experimental comparison among commonly used
monitoring well casing materials (stainless steel, Teflon, rigid
PVC) in terms of their potential effects on measurements of trace
organic compounds. Each of the studies analyzed attempts to
determine experimentally how much —— if any —— sorption occurs, or
what differences result among measured concentrations of a series of
organic compounds.
Because the compounds tested were not consistent among all the
studies —— although some of the same compounds were represented in
several of them —— the analysis compares effects of the casing
materials on sorption of different chemicals. The laboratory
studies analyzed in this report all relate the measurements taken to
a control, and the field Investigation to measurements of the same
trace compounds in adjacent wells constructed of different casing
materials.
In comparing the measured trace concentrations to determine whether
the well casing materials cause significant differences in results,
a ratio was formulated to reflect the relative sorption effects of
each of the materials; sensibly constant ratios over a reasonable
range of trace concentrations would indicate few, or relatively
minor, differences between the various materials, while varying
ratios would Indicate larger differences.
The report first reviews the methodology and result8 of each
individual Investigation analyzed; these results are then compared
across studies through the averaged ratios; and conclusions are
drawn about similarities and differences in sorption behavior.
Additional observations about sample variation, effects of well
purging, and limited measurements of non—volatile compounds are
noted.
INTRODUCTION
This review compares the results of four studies that allow a direct
experimental comparison of the potential effects of commonly used
well casing materials (stainless steel, Teflon, and rigid PVC) on
measurements of trace levels of organic compounds. 1 ’ 2 ’ 3 ’ 4 ’ 5
6—29
-------
Each of the four studies was designed to determine experimentally,
for a water solution in contact with well casing materials (or
coupons made from them), how much sorption occurs, or what
differences results among measured concentrations, using several
organic compounds. The tested organics were not identical in each
of the studies, although several compounds are represented more than
once. Three of these studies are laboratory experiments that relate
the measurements to a control, while the fourth is a field study
that compares measurements of the same trace concentrations on
different casing materials in wells located close together.
If well materials affect trace level measurements significantly,
then a ration can be formulated to reflect the relative sorption
effects of the materials. Such a value should be sensibly constant
over a reasonable range of trace level concentrations.
Although the exact procedures and the organic compounds measured
differ among the studies, nevertheless any superiority of one well
casing material over another ought to be observable if the studies
are sufficiently sensitive. It is possible, of course, that at the
trace levels measured —— 100 ppb to 20 ppb for the laboratory
studies and subparts per billion for the field study —— other
sources of variation so overwhelm the results that no meaningful
differences can be observed, and this information is in itself
useful.
LABORATORY STUDIES
Some preliminary observations about the varying lengths of the test
periods in the laboratory studies are in order. The Reynolds and
Giliham study tested only the effects on virgin materials over times
up to 7 days. This approach is important in addressing the initial
effect of a material On trace level measurements and on the
mechanism of sorption. However, this approach does not mimic actual
field protocol, such as followed by Barcelona. Barcelona et al.
show that purging is essential to the correct operation of
monitoring wells.
Both the ChemWaste and the Radian laboratory studies also
Incorporate a 7—day exposure period; however, beyond that they each
add 1—hour and 24—hour re—exposures to represent results from both
initial and much more closely calibrated samplings. These
differences should be kept in mind throughout this review.
1. The Reynolds & Giliham Study . This laboratory study compared
effects of six organic polymer materials —— PVC, Teflon, nylon,
polypropylene, polyethylene, and latex rubber —— on a series of
different trace level organics, ranging from 20 ppb to 45 ppb of
1,1,1 trichiorethane, 1,1,2,2 tetrachioroethane, hexachioroethane,
6—30
-------
bromoform, and tetrachioroethylene. Samples were withdrawn from
exposure at times varying from 10 minutes to 7 days and analyzed.
From the published results, the ratios of the concentrations of the
chemicals in contact with Teflon and PVC coupons respectively can be
ca1c ulated and compared for the five different compounds measured.
These ratios, representing the relative effects of sorption on
virgin materials, are presented in Table 1. This table indicates
that, for four of the compounds measured, the Teflon/PVC ration is
very close to 1, implying little difference among the materials’
effects. The exception is tetrachloroethylene, which shows much
greater sorption, and hence less sensitivity, for Teflon.
Table 1
RATIO OF CONCENTRATION MEASUREMENTS
OVER TIME FOR DIFFERENT COMPOUNDS
Reynolds Study: TEF/PVC
Average
Compound/Time 10 mm. 100 miii. — 7 days 7 day
1,1,1 Trichioroethane 1.06 0.93 0.68
1,1,2,2 Tetrachioroethane 1.05 1.01 0.94
Hexachioroethane 1.08 1.07 1.30
Bromoform 1.06 1.17 1.68
Tetrachioroethylene 0.9 0.46 0.10
Average 1.03 0.93 0.94
In general, this shows typical differences at the 7—day time of ±
30L Some of the chemicals are detected more easily with PVC (1,1,1
trichloroethane, 1,1,2,2 tetrachloroether, tetrachioroethylene) and
some more easily with Teflon (hexachiorethane, bromoform). Based on
this study alone, it would be difficult to establish clear
superiority of either Teflon or PVC.
2. The ChemWaste Management Study . This study measured effects of
coupons of stainless steel, Teflon, and rigid PVC on six organic
compounds: methylene chloride, 1, 2 dichioroethane, trans—1,2
dichioroethylene, trichioroethylene, chlorobenzene, and toluene.
The test solutions were prepared In the same way as for the Reynolds
study; the organics were dissolved in a concentrated methane
solution and then exposed to the coupons at two diluted
concentrations of 50 ppb and 100 ppb.
Each casing material was first exposed for an initial 7 days
(similar to Reynolds) after which the solutions from each coupon and
control were sampled. The coupon materials were then re—exposed for
6—31
-------
Compound/Time
Meth Chi
1,2—DCE
1,2—DCY
Trichioroethylene
Tol uene
Chiben
BLE 2
CWM Report
RATIO OF CONCENTRATION MEASUREMENTS
OVER TIME FOR DIFFERENT COMPOUNDS
(Soppb nominal)
ss/PVC
Reexposed
1 hour 24 hour
1.02 1.07
1.05 1.08
0.98 1.12
1.00 1.10
1.05 1.13
1.18 1.14
Initial
ldays
0.92
0.92
1.00
1.02
0.88
0.95
avg
Compound/Time
Meth Chl
1, 2—DCE
1,2-DCY
Trichioroethylene
Toluene
Chiben
avg
1.11
Compound/Time
Meth Chi
1,2-DCE
1,2-DCY
Tr ichi.oroethylene
Toluene
Chiben
avg
TABLE 3
CWM Report
RATIO OF CONCENTRATION MEASUREMENTS
OVER TIME FOR DIFFERENT COMPOUNDS
(lOOppb nominal)
SS/PVC
Reexposed
1 hour 24 hour
1.03 0.93
1.03 1.03
0.96 1.04
1.03 1.03
1.05 1.03
1.06 1.04
1.01
Compound/Time
Meth Chi
l,2-DCE
1 ,2-DCY
Trichioroethylene
Tol uene
Chiben
avg
TEF/PVC
Reexposed
1 hour 24 hour
0.79 0.90
0.81 0.92
0.78 0.88
0.80 0.87
0.79 0.90
0.97 0.90
0.83
0.90
6—32
Initial
7days
1.12
1.14
0.91
0.89
1.01
1.04
1.02
1.05
TEF/PVC
Reexposed
1 hour 24 hour
1.02 0.94
1.03 0.94
0.98 0.93
1.02 0.92
1.03 1.06
1.00 1.19
1 ..01 1.00
0.95
Initial
7days
0.89
0.90
0.76
0.73
0.79
0.84
0.82
Initial
7days
1.06
1.03
1.13
Lii
1.07
1.13
1.03
1.09
-------
one hour and resampled. A third sampling was performed following a
final 24—hour re—exposure.
Results from the initial 7—day conditioning period — — which may have
very little significance since groud ’ater sampling protocols require
purging wells prior to drawing samples —— show that PVC is a better
measuring materials 12 times, while the Teflon or stainless steel Is
better 12 tImes.
To compare the remainder of the results, the concentrations of the
other experimental exposure was used to calculate ratios. Table 2
shows the SS/PVC and TEF/PVC ratios for the 50 ppb nominal
concentration, and Table 3 shows the same calculations for the 100
ppb nominal concentration, for all six of the tested chemicals.
There are not enough control samples or replicates in this study to
estimate the standard deviation. The best that can be done Is to
compare the ratios. Inspection of Tables 2 and 3 shows that PVC and
Teflon exhibit very similar behavior; the ratios are close to 1, a
result similar to that from the Reynolds study (although the latter
shows wider variation). 1here the coupons were re—exposed for one
hour, Teflon is slightly better (2—3%) at the 50 ppb nominal case
for four chemicals and PVC slightly better (2—3%) for one chemical;
at the 100 ppb nominal, PVC is better (207.) for all six chemicals.
After 24 hours re—exposure, at the 50 ppb nominal PVC is better
6—8%) for four chemicals and Teflon better (6 to 20%) for two, while
at the 100 ppb nominal. PVC appears about 10% better for all six
chemicals.
The re—exposure samples show stainless steel performing better once
better (2—10%) than PVC ten times and PVC performing better once for
each of the two nominal concentrations. For the 1—hour
re—exposures, which are likely to most nearly represent a well
monitoring protocol after purging, only one of the ratios
(cblorobenzene for SS/PVC) is more than 5% greater than 1.
Without any estimate of uncertainties, It is impossible to know if
the differences shown in this experiment (on the order of 10%) are
significant.
3. The Radian Study . This study followed the same general protocol
as the ChemWaste study, with several significant differences. The
same six chemical compounds were used In the experiment but were
dissolved to a nominal 100 ppb concentration in a water solution as
a carrier, instead of methanol as for Reynolds and ChemWaste. The
exposure periods were similar: an initial 7 days exposure of the
well casing coupons, followed by 1—hour and 24—hour re—exposures.
The 7—day and 24—hour re—exposures were held at 5°C, while the
1—hour re—exposure was at room temperature. Table 4 shows the
6—33
-------
results of these analyses, with trichioroethylene deleted since it
was not stable.
Table 4 also presents the results of the nine control samples, with
averages and standard deviations. An analysis of variations shoved
that the controls were not drawn from different populations and thus
can be averaged. Therefore the standard deviation gives an estimate
of the variation likely to be seen in any set of measurements.
Inspection of Table 4 shows that, for the re—exposure experiments
for the 20 paired differences possible between concentrations (of
PVC and Teflon, and PVC and stainless steel) with the various
coupons, no difference exceeds two standard deviations, and five
exceed one standard deviation. For the initial 7—day exposure, none
of the differences between PVC and Teflon are greater than two
standard deviations.
However, the stainless steel results are quite unusual. The
stainless steel concentrations are more than two standard Ieviations
greater than the controls, as well as more than two standard
deviations larger than both PVC and Teflon for all of the
chemicals. It seems likely that some contamination has entered the
system. The stainless steel coupons were used just as received from
the manufacturer and may have contained cutting fluids or other
organics which affected the spiking solutions; there was no control
using a coupon without a Rpike to check for this. In any event, it
appears that the 7—day stainless steel concentrations are highly
suspect and should be redone.
As before, the ratios (SS/PVC and TEF/PVC) can be calculated from
these data to establish relative sorption effects between the
various compounds. The results, presented in Table 5, are similar
to those from the hemWaste Management study, yielding ratios very
close to 1, except for the 7—day SS/PVC ones.
Table 5 also presents an eBtimated standard deviation for the
ratios, based on the standard deviation of the control samples,
rather than on the email number (3) of the replicates. While this
is not a satisfactory statistical analysis, it gives an order of
magnitude to the variations in the ratios which may be present due
to uncertainties in the experiment. (A more complete statistical
analysis will be performed.)
The results of the re—exposure show that, after one hour, PVC
appears somewhat better than both Teflon (2—10%) and stainless steel
(10—30%); however, the difference does not appear significant, never
exceeding two standard deviations. After 24 hours, PVC appears
better than Teflon (0 to 6%) but not as good as stainless steel
(8—12%), although again neither set of differences appears
significant.
6—34
-------
B E
Radian Study
CONCENTRATIONS REMAINING IN
ppb
SS/PVC
24 hour
1.12
1.08
1.01
1.12
1.11
1.09
TEF/PVC
24 hour
1.01
0.96
0 . 94
1.00
0.96
0.98
1.20
std dev
0. 29
0.45
0.17
0.45
0.43
std dev
0 . 21
0.31
0.13
0.23
0. 29
ONE HOUR
AVG
PVC
TEFLON
S.S.
CONTRLS
STD
Meth Chi
129
127
119
127
16
1,2—DCE
68
64
48
57
10
1,2—DCY
109
100
96
103
9
Toluene
42
38
32
37
5
Chiben
67
60
55
58
8
Meth Chi
PVC
117
24 HOUR
TEFLON
118
S.S.
131
AVG
127
STD
16
1,2—DCE
53
51
57
57
10
] .,2—DCY
103
97
104
103
9
Toluene
34
34
38
37
5
Chiben
56
54
62
58
8
Meth Chi
PVC
114
7 DAY
TEFLON
136
S.S.
189
AVG
127
STD
16
1,2—DCE
47
56
82
57
10
1,2—DCY
95
102
136
103
9
Toluene
31
35
69
37
5
Chiben
40
56
84
58
8
TABLE 5
Radian Study
RATIO OF CONCENTRATION MEASUREMENTS
OVER TIME FOR DIFFERENT COMPOUNDS
Reexposed
1 hour
std dev
0.16
0.18
0.11
0 . 15
0.17
Meth Chi
0.92
1,2—DCE
0.73.
1,2—DCY
0.88
Toluene
0.76
Chiben
0.82
avg
0.82
1
hour
Meth Chi
0.98
1,2—DCE
0.94
1,2—DCY
0.92
Toluene
0.90
Chiben
0.90
avg
0.93
Reexposed
std dev
0.19
0.28
0.12
0.22
0.23
std dev
0.18
0.25
0. 11
0.20
0.20
Initial
7days
1.66
1.74
1.43
2.23
2.10
1.83
Initial
7days
1. 19
1. 19
1.07
1.13
1.40
std dev
0 . 17
0.24
0.11
0.18
0.18
6—35
-------
It is difficult to conclude from this experiment that there is a
consistent difference between stainless steel and PVC or Teflon and
PVC. Based on the Radian study alone, if a well were purged and
subsequently sampled within one hour, PVC would seem to be somewhat
more sensitive than either stainless steel or Teflon, but not
significantly so. Twenty—four hours after purging, the PVC is still
somewhat more sensitive than Teflon but less sensitive than
stainless steel —— again, not significantly so. While the initial
7—day stainless steel experiment appears flawed, Teflon is more
sensitive after 7 days —— again, not significantly so.
C. Barcelona’s Field Experiment
The experiment by Michael Barcelona and John Heif rich was designed
to provide a comprehensive field study which tests the differences
between three different well casing materials: PVC, stainless
steel, and Tef ion. 6 The experiment has potential advantages over
the lab studies in that It investigates the detection of chemicals
actually in the groundwater at two different contaminated sites.
At each site there were six veils, with one each of the three
different casing materials In a cluster upgradlent and one each of
the three materials clustered dowagradlent of the site. Each of the
wells at a given cluster was installed within two meters of the
other two, thereby attempting to assure that each cluster was
sampling the same groundwater.
Each site was sampled monthly, six times starting in May and
extending through October. At each of the sites, samples were taken
prior to purging the stagnant water. 7 The wells were purged until
parameters such as pH stabilized, and then samples were taken for an
extensive list of groundwater parameters that included p11,
conductivity, temperature, alkalinity, and total iron, and for a
series of organic compounds that included total non—volatile organic
compounds (NVOC), inethylene chloride, 1,1—dichioroethane (l,l—DCE),
cls—l,2 dlchloroethylene (c—i, 2—DCY), trichloroethylene, 1,1,1 TCE,
and chlorobenzene.
The largest concentrations of total volatile halocarbons were
detected at Site 2 down—gradient at a few parts per billion.
However, the concentrations are so low that, In many cases, clear
differences between concentrations at the different wells cannot be
seen.
As Barcelona points out in the paper, there were problems at Site 1
with apparent grout contamination at both up— and down—gradient
wells. This was a factor in the abnormally large pH levels seen in
five out of the six wells at this site. The only well which
apparently did not have high pH levels was the down—gradient PVC
well. Such grout contamination obviously is of concern if it could
6—36
-------
affect the measurement of organic constituents, since those are
crucial to the determination of sorption effects. A fairly simple
correlation of total NVOC with the pH values for the stainless steel
and Teflon wells at Site 1 suggests that there may be a direct
relationship between pH and the non—volatile organic compounds. The
correlation coefficient was 0.5, with a possibility of 6%.
Therefore, in reviewing the Site 1 data, the problems with pH must
be kept in mind, since it is apparent that grout contamination
occurred and the NVOC values may be affected as well. The wells at
Site 2 apparently were constructed in such a way that there was no
grout contamination, and the purged pH was as expected.
With respect to well casing material, no definitive conclusions can
be drawn because the well casing effect is confounded with spatial
variability and, at Site 1, with grout contamination. Each type of
well casing is used only once for each experimental sampling. As a
result, the differences seen could be a result of either well casing
or spatial differences. A so—called mixed model analysis may be
used to help disentangle the effects.
Prior to that, however, the data can be inspected to observe if any
consistent differences across sites are apparent.
Table 6 shows the results of ratios calculated, as before, for
ssIPvc and TEF/PVC for each chemical measured and a group of NVOC
chemicals and Total Volatile Halocarbons (TVOC) for Sites 1 and 2
up— and down—gradient.
As can be seen, there are no consistent results indicating that PVC
is inferior to the other two materials. For example, for NVOC at
Site 1 down—gradient (with grout contamination) stainless steel and
Teflon have superior detection capability, but at Site 1
up—gradient, where all wells have grout contamination, PVC is
superior. Since, presumably, the differences should be the same, it
is likely either spatial variability or grout contamination cause
such a large variation.
If, therefore, attention is focused on Site 2 where volatile
organics were detected, it can be seen that PVC is consistently
superior to both Teflon and stainless steel, except for the
non—volatiles. These values, however, have a significant variation
and only the measurements for 1,1 DCE suggest that PVC is
significantly (more than two standard deviations) superior to
stainless steel and Teflon.
Indeed, the combination of time series data, the fixed spatial
distances, and the material differences may allow a determination of
errors due to spatial variability, which seem likely to be larger
than those due to the analytical variability.
6—37
-------
TABLE 6
Ratio of
Barcelona et. al.
Groundwater Measurements
for Two Wells
sS/Pvc
NVOC
Meth.Chl.
NVOC
Meth.Ch]..
TVOC
12DCE+111TCE
11DCE
C12DCY
Ratio
Site 1 down
1.83
1.98
Site 2 down
1.11
0.90
0.61
0.70
0.43
0.63
STD Ratio
Site 1 up
0.80
4.25
0.35
0.44
0.24
0.39
0.16
o . 22
STD
0.40
5.65
6.36
0.39
0.73 0.30
TEF/PVC
NVOC
Meth.Ch l.
Ratio
Site 1 down
1.52
1.35
STD Ratio
Site 1 up
0.69 0.92
0.92 1.46
STD
0.47
1.49
NVOC
Meth.Chl.
TVOC
12DCE+111TCE
11DCE
C12DCY
average
Site 2 down
0.89
0.80
0.43
0.58
0.62
0.72
0.67
0.31
0.31
0.16
0.12
0.13
0.17
0.20
10.35
0.59
* 1,2 DCE & 1,1,1 TCE
TEF/PVC
0.66
0.95
Site 2 up
3.85
0.91
average
Site 2 up
12.99
0.98
Table 7
COMPARISON OF SPECIFIC VOLATILE CHEMICALS
BETWEEN LABORATORY AND FIELD STUDIES
RATIOS OF CONCENTRATION MEASUREMENTS
sS/PVC
Barcelona
1,2 DCE
C 1,2 DCY
1,2 DCE
C 1,2 DCY
CWM
1.03
0.96
Radian
0.71
0.88
0.94
0.92
C M
1.05
0.98
1.03
0.98
0.81
0.78
0.70 + 39*
0.63 .22
0.58 + .12*
0.72 .17
6—38
-------
D. Comparison of Results
In comparing the different studies, this review has analyzed the
ratios of SSIPvc and TEF/PVC derived from the experiments to assess
whether one material is consistently better than another in
measuring compounds common to the experiments. Table 7 compares 1,2
DCE and c1 —l,2 DCY, two compounds measured (at 1—hour re—exposures)
in the ChemWaste, Radian and Barcelona experiments (two laboratory
and one field experiment). The results are generally consistent,
although the Barcelona experiment suggests that PVC is somewhat
better than is suggested by the Radian or ChemWaste studies. This
implies that PVC would detect concentrations of the chemicals more
efficiently than would the other two materials.
A second way to compare the study results is to average all of the
chemicals in each of the experiments. Obviously, it is necessary to
be very careful about averaging different chemical compounds with
different sorption behaviors. But, since it is generally not known
which compounds are likely to be in groundwater, averagIng the
ratios indicates what effects might occur with an unknown compound
or suite of compounds in a groundwater situation. Table 8 shows
average ratios for ssIPvc and TEF/PVC for two categories for the
four studies reviewed. It assumes that the laboratory 1—hour
re—exposures are roughly comparable to the Barcelona field
experiment when the well is purged and that the 7—day exposures for
Reynolds, ChemWaste and Radian are roughly comparable to the
Barcelona field experiment when samples are taken under stagnant
conditions. 8 If the average for the 1—hour re—exposure is
considered for both ssIPvc and TEF/PVC, the ratios are determined by
all four experiments are very similar.
There is somewhat more variation when looking at the 7—day stagnant
situation for stainless steel and PVC. One reason for this may be
the possibility of contamination in the Radian 7—day stainless steel
experiment.
E. Conclusion
Based on this review of the existing studies, several conclusions
are possible relating to the effects of well casing materials on the
measurement of organic compounds, to apparent sample variations that
occur in an actual measurement situatIon, and to judging the
efficacy of various well materials.
o These four experiments suggest that, when groundwater is purged
from a well, there are no consistent differences between the
effects of stainless steel and PVC on volatile organic compound
measurements. The laboratory experiments also show no
significant differences between Teflon and PVC. In the field
experiment, there is a s nall difference that may be significant,
6—39
-------
Table 8
RATIO OF WELL CASING MATERIAL MEASUREMENTS
(average of all volatile organic compounds, laboratory & field tests)
Radian CWM CWM Reynolds Barcelona
1 hr SS/PVC
re—exposure 0.82 1.03 1.05 0.93* (1,2)
(purged)
1 hr. TEF/PVC
re—exposure 0.93 0.83 1.01 0.81* (1,2)
(purged)
7 day SS/PVC 1.64k 1.09 0.95 0.46** (2)
(stagnant)
7 day TEF/PVC 1.16 1.02 0.82 0.94 0.63** (2)
(stagnant)
* Purged; numbers indicate Sites
** Stagnant waters; number (2) indicates site 2
+ Possible contamination of the SS values
6—40
-------
showing that PVC may be more sensitive; if further work confirms
this, PVC would detect volatile organics better than Teflon. If
stagnant water is sampled, the comparison between studies is not
so clear. Non—volatile chemicals, tested only In the Barcelona
study, also do not show a -significant difference across sites for
any of the three casing materials.
o The effects of well materials on measurements of trace organic
compounds need to be disentangled from sample variability. This
variability may contribute an error larger than any error from
analytical variability. Further investigation could shed light
on whether this is an artifact of this particular field study ——
which seems unlikely —— or is consistent at other waste sites.
o It seems clear that, in order to judge the efficacy of various
well materials, comparisons must be made in the context of normal
experimental variation. If it is to be judged that well material
A has a different sorption than well material B, the difference
in concentrations (reflecting different absorption behaviors)
between them must be greater than the normal variation in the
samples themselves. Further experiments could determine whether
effects associated with the well easing material are larger than
the 15% to 25% sample variation that seems likely.
Finally, a conservative position at the present may be to allow a
choice of any of these three well casing materials: Teflon,
stainless steel and PVC, at this time, excluding PVC could result in
less detection of organic compounds.
FOOTNOTES
i-Reynolds and Giliham, “Absorption of Halogenated Organic
Compounds by Polymer Materials Commonly Used in Groundwater
Monitors.” Proceedings, Second [ 1985] Canadian—American Conference
on Hydrogeology , Banf, Alberta, 1986, pp. 125—132.
2 ChemWaate Management, Inc., “Absorption of Organics by
Monitoring Well Construction Materials,” unpublished technical note.
3 Barcelona, Michael J. and John A. Heifrich. “Well Construction
and Purging Effects on Ground—Water Samples,” Environmental Science
& Technology , Vol. 20, No. 11, 1179—84.
4 Sykes, McAllister and Homolya. “Sorption of Organics by
Monitoring Well Construction Materials.” Radian Corporation (to be
published).
5 The review does not include a separate paper by Barcelona et al.
that reported an investigation of the relationship between Teflon
tubing and PVC tubing because, as Barcelona indicated, fiexible PVC
6—41
-------
tubing differs from rigid PVC pipe, and the sorption behavior is
likely to differ greatly.
6 The Teflon well at Site 2 was actually a Teflon/aluminum oxide
dedicated sampler.
7 Wlth regard to well purging, Barcelona concluded that the
experiment shows that purging is essential in order to eliminate
spurious results from 8tagnant water. The investigation provides a
useful data base for studying the necessity of purging.
8 The stagnant conditions are not exactly the same, since the
Barcelona experiment allowed 30 days exposure to the well casing
materials, while the laboratory experiments had only 7 days; in
addition, the Barcelona casings were not virgin materials.
6—42
-------
APPLICATION OF WIDE—BORE CAPILLARY COLUNN TO THE ANALYSIS OF
VOLATILE ORGANIC COMPOUNDS BY METHOD 8240
Robert W. Slater, Jr. James E. Longbottom, Environmental Monitoring
and Support Laboratory, U.S Environmental Protection Agency,
Cincinnati, Ohio
ABS TRACT
Under the Resource Conservation and Recovery Act (RCRA), the U.S.
Environmental Protection Agency (USEPA) has revised the regulations
concerning the ground water monitoring for suspected contamination
from hazardous waste treatment, storage and disposal facilities.
Over 240 compounds are specified for monitoring in Appendix IX, with
approximately 60 compounds having volatilities which would make them
amenable to purge and trap techniques.
Method 8240 with the packed column suffers from limitations of
resolving power and limited analyte range. In several cases,
structurally similar compounds are not resolved, precluding positive
identification or quantitation of the individual isomers. Also, the
boiling range of compounds to be chromatographed is effectively
limited to the dichlorobenzenes. Late eluting peaks become very
broad and diffuse, and quantitation limits become higher.
Recently, large diameter capillary columns C 0.5 mm ID) have become
available with several phases specifically designed for retention
and separation of volatile compounds. These columns afford the high
resolving power of capillary columns with geometric and positional
isomers well separated; at the same time, they accept high flow
rates which make them compatible with the purge and trap apparatus.
Additional benefits of the capillary column are a broadening of the
analyte range, and a decrease in the analysis time. Good
sensitivity is achieved for tetrachlorobenzene with an analysis time
is achieved for tetrachlorobenzene with an analysis time of 32 mins,
compared to an analysis time of 45 mins for dichlorobenzene on a
packed column.
Accuracy and precision data are presented on the analysis of reagent
water and publicly—owned treatment works (POTW) effluent using
Method 8240 with a capillary column.
INTRODUCTION
Packed columns have been used for the separation of volatile organic
compounds in the standard U.S. Environmental Protection Agency
(USEPA) approved methods, such as Method 524 for drinking water
analyses, Method 624 for wastewater contaminants and later Method
8240 for Resource Conservation and Recovery Act (RCRA) compounds.
These methods have been effective when the number of measured
6—45
-------
compounds was small; but with increasing monitoring regulations it
has become apparent that packed columns can not provide the
necessary column efficiency to resolve the many additional compounds
of interest. Capillary columns offer sufficient efficiency to
complete the separations.
Some attempts to use narrow—bore capillaries (<0.32 mm ID) for these
analyses resulted In some success, but the methods lacked ruggedness
due primarily to the problem of interfacing the purge and trap
apparatus to the analytical column. The purge and trap requires
relatively high flows for efficient desorption, 15—30 mL/min,
whereas the column requires only a 2—3 mL flow for efficient
operation. In addition, these columns have a thin film thickness
which limits the analytical range to low concentrations.
More recently, wide—bore capillaries (Ø .53 mm ID) having specialty
phases designed for the retention and separation of volatiles have
been introduced. When operated in the traditional capillary mode
with a flow rate of 5—7 mL/min, the column offers the excellent
separations and a greatly increased analyte range. Also, the
columns have relatively thick liquid phases, 1.5 to w3m, to give a
greater sample capacity. But a more practical mode of operation
when a mass spectrometer is used is the “high—flow” mode, where
column flows of 15 to 20 mL/min are utilized. At these flows only a
small loss of column efficiency is noted, but the time of analysis
is greatly shortened. In this mode, the column can be a direct
substitution for the packed column for purgeable analyses.
The announced drinking water regulations of November 1985 provided
the initial impetus toward development of a capillary column gas
chromatography/mass spectrometer (GC/NS) method. The announcement
cited 60 organic compounds for monitoring, with volatilities ranging
from chioromethane (BP, —24°C) to 1, 2, 3—trichlorobenzene (BP,
219°C). These compounds were included in the scope of two packed
column methods. The Agency proposed a packed column GC/MS method
f or the monitoring requirement. In addition, the monitoring
regulations required an analytical range from l/j g/L for vinyl
chloride to 200 pg/ L for l,l,l—trichloroethane. In September,
1986, Environmental Monitoring and Support Laboratory—Cincinnati
(EMSL—cincinnati) proposed Method 524.2, providing single laboratory
validation data on 58 of the volatile compounds. Method detection
limits (MDL) for most compounds were less than 0.5 .tgIL, indicating
the necessary method sensitivity when a 25 mL sample is used.
The success with drinking water analytes led to investigate
application of capillary column GC/MS to Appendix IX compounds. In
our laboratory, and also through a contract with Battelle Columbus
Laboratories, we have identified approximately 60 compounds which
are amenable to purge and trap isolation and GC/MS detection. While
many of these compounds are common with the drinking water method,
6—46
-------
many new compounds are listed, several of which are beyond the
volatility range previously described.
EXPERIMENTAL
Five mL aqueous samples were purged on a Tekmar LSC—2 concentrator
under the conditions given in Table 1. Following concentration, the
analytes were thermally desorbed directly onto a 60 m x 0.75 mm ID
VOCOL capiUary column maintained at 10°C in a Hewlett—Packard 5895
GC/MS. The column effluent was interfaced with the ion source
through a jet separation. Column conditions and MS parameters are
presented in Table 2.
Concentrated standards of the volatile analytes were obtained from
the USEPA Repository or prepared from neat material in methanol at
approximately 5000 j..&gImL. From these concentrates spiking mixtures
were produced containing each analyte at 50 pg/mL in methanol. A
list of the method analytes from Appendix IX along with their
retention times is given in Table 3.
Calibration for each analyte was accomplished by purging aqueous
standards of the compounds spiked at 5, 10, 50 and 100 .g/L In
Milli—Q water. The response factors at each level were calculated
from the peak area of the ion chromatogram for that compound. For
most compounds that ion was the base peak for the compound (Table
3), but in the few cases where the ion was common to co—eluting
peaks, a secondary ion was selected. A linear fit was calculated
for the detector responses. In all cases, the coefficient was 0.994
or greater, so that use of the average response factor was used over
the entire range.
RESULTS AND DISCUSSIONS
Wide and narrow bore capillary columns shown to be effective for 58
compounds and included In the June final regulations.
Initial experiments for the Appendix IX list were to determine the
purge characteristics of a number of potential analytes for which we
had no prior data. From this list of potential analytes (Table 4)
percent recovery and retention time data were gathered. Compounds
having a purge recovery of 60Z or greater were selected for further
study.
Method accuracy and precision were determined by analyzing eight
replicates of spiked reagent water at 10 Ag/L of each component.
The average recoveries relative to the purged standards for most
components were in the range of 90—110% with relative standard
deviations of 5 to 9%. Slightly lower recoveries were found for the
vary gaseous halocarbons at 80—85%. A summary of these data Is
presented in Table 5. Only the meta— and p—xylene isomers were not
6—47
-------
separated and quantitated; and are reported as single value of 20
p g/L for the isomer pair.
We have began to demonstrate the applicability of the method to the
analysis of complex sample matrices. For example, a wastewater
sample from a publicly—owned treatment works (POTW), a secondary
effluent, was obtained from the Cincinnati Municipal Wastewater
Plant. Since the sample was known to be biologically active, the
sample was acidified with 1:1 IICL to pH 2 prior to spiking. As
before, the target level for each analyte was 10 p.gIL. This level
was verified for all compounds except methylene chloride which
appeared as a large contaminant peak at approximately 150 gJL. The
eight aliquots of spiked POTW were analyzed and the average
recoveries calculated. Figure 2 presents chromatograms of the POTW
samples. The upper trace depicts the unspiked POTW sample. Peaks
marked with a “s are Internal standards or surrogate compounds.
The lower trace Is the chromatogram of the spiked sample. For most
compounds the recoveries were slightly lower than those found from
reagent water, ranging from 80 to 95Z; again RSDs were in the range
of 4 to 9% (Table 6). Notable exceptions to these recovery
percentages are the dichlorobenzenes and trlchlorobenzenes which
appear as background contaminants. When corrections for these
background levels are made the recoveries fall more into line with
the other data.
SUMMARY
The applicability of wide—bore capillary columns to Method 8240 for
some Appendix IX compounds has been demonstrated. The capillary
column provides superior resolving power and a greatly increased
range of analytes. Method accuracy and precision for most compounds
are improved from those achieved with the packed column. We believe
that the wide—bore capillary column is a means to provide better
analytical data across all media. We are at present sponsoring
formal single laboratory validation studies at Battelle with
Appendix VIII and IX compounds En groundwater and leachate. Future
plans Inhouse call for us to demonstrate the method applicability to
other wastewaters and municipal sludges.
6—48
-------
Table 1. Purge and Trap conditions
Sample Volume — 5 mLs
Sample Temperature — M bient C 25°C)
Purge Time — 11 minutes
Purge Flow — 40 niL/mm.
Desorption — 4 mins at 180°C
Desorption Flow — 15 niL/mm.
Table 2. Chroinatographic and Mass Spectrometric Conditions
Column — 60m x 0.75 mm ID (glass) VOCOL
CC Column Flow - 15 niL/mm
CC Column Conditions — hold 5 nuns at 10°C, then program
to 180 at 6°C/mm.
MS Scan Range — 45 to 300 amu
MS Scan Rate — 0.7 sec/scan
6—49
-------
Table 3. CHROIVtATOGRAPHIC RETENTION TIMES FOR VOLATILE ORGANIC
COMPOUNDS ON APPENDIX IX LIST
RETENTION ANALYTICAL
ANALYTE TIME IONS
(mins) PRIMARY SECONDARY
Dichiorodifluoromethane 1.55 85 87
Chioromethane 1.63 50 52
Vinyl chloride 1.71 62 64
Bromomethane 2.01 94 96
Chioroethane 2.09 64 66
Trichiorofluoromethane 2.27 101 103
1,1—Dichioroethene 2.89 61 96,63
Methylene Chloride 3.60 84 49,86
trans—1,2—Dichloroethene 3.98 96 61,98
1,1—Dichioroethane 4.85 63 65,83
Chloroform 6.40 83 85,47
1,1,1—Trichioroethane 7.27 97 61,99,117
Carbon Tetrachioride 7.61 117 119,44
Benzene 8.23 78 ——
Trichioroethene 9.59 95 130,131
1,2—Dichioropropane 10.09 63 112
Bromodichioromethane 10.59 83 85,127
Dibromomethane 10.65 93 95,174
trans—1,3 —Dichloropropene 11.98 75 49,110
Toluene 12.43 92 91,65
cis—1,3—Dichloropropene 13.22 75 49,110
1,1,2—Trichloroethane 13.41 97 83,61
Tetrachioroethene 13.74 166 129,94
Dibromochioromethane 14.39 129 127,131
1,2—Dibromoethane 14.73 107 109,188
Chlorobenzene 15.76 112 77,51
1,1,1,2—Tetrachioroethane 15.94 133 131,117
Ethylbenzene 15.99 91 106,51
p—Xy lene 16.12 106 91
m—Xylene 16.17 106 91
o—Xylene 17.11 106 91
Styrene 17.31 104 78
Bromoform 17.93 173 171,254
1,1,2,2—Tetrachioroethane 18.72 83 131,85
6—50
-------
Table 3. (Continued)
RETENTION ANALYTICAL
ANALYTE TIME ION
(mins) PRIMARY SECONDARY
1,2,3—Trichioropropane 19.02 75 77
trans—1,4--Dichlorobutene --2 19.48 75 77,53
1,3—Dichlorobenzene 21.22 146 148,111
1,4—Dichlorobenzene 21.55 146 148,111
1,2—Dichlorobenzene 22.52 146 148,111
Hexachioroethane 23.22 117 201,166
1,2—Dibromo—3--Chloropropane 24.53 75 155,157
1,2,4—Trichlorobenzene 26.55 180 182,145
Hexachiorobutadiene 26.99 225 223
Naphthalene 27.17 128 — —
Hexachioropropane 27.19 213 211,141
1,2,4,5—Tetrachlorobenzene 30.81 216 214,179
2—Chloronaphthalene 32.42 162 127
6—51
-------
Table 4. ABSOLUTE RECOVERIES OF SOME APPENDIX IX
COMPOUNDS BY PURGE AND TRAP GAS CHROMATOGRAPHY!
MASS SPECTROMETRY (GC/MS)
Retention
Time
Compound (mins) % Recovery
trans—i ,3—dichloropropene 11.98 86
cis-1,3--dichloropropene 13.22 100
N—Nitrosodmetrylanino 13.64 0
trans—i, 4-dichlorobutene—2 19.48 57
bis—(2—chloroethyl) ether 21.39 3
Hexachioroethane 23.22 102
Isophorene 25.39 0.9
Hexachioropropene 27.19 95
Hexachiorocyclopentadrene 30.57 43
1 ,2,4,5-Tetrachlorobenzene 30.81 78
2- Ch1oronaphtha1ene 32.44 51
Acenaphthylene 34.51 19
Acenaphthene 35.25 25
4—Chiorophenyiphenyl ether 37.32 7
4—Bromophenyiphenyl ether 40.17 0.3
6—52
-------
Table 5. ACCURACY AND PRECISION DATA FOR VOLATILE
ORGANIC COMPOUNDS IN REAGENT WATER DETERMINED
WITH A WIDE BORE CAPILLARY COLUMN
Standard Percent
Conc. Recoverya Deviation Rel. Std.
Analyte pg/L of Recoveryb Dev.
Benzene 10 105 3.8 3.6
Bromodichioromethane 10 102 4.7 46
Bromoform 10 110 8.4 7.6
Bromomethane 10 100 9.1 9.1
Carbon tetrachioride 10 84 5.5 6.5
Chlorobenzene 10 101 5.0 4.9
Chioroethane 10 79 7.8 9.9
Chloroform 10 95 5.0 5.3
Chioromethane 10 98 4.8 4.9
1,2—Dibromo—3—chloropropane 10 94 5.7 6.1
Dibromochioromethane 10 102 5.6 5.5
1,2—Dibromoethane 10 108 6.0 5.6
Dibromomethane 10 105 5.6 5.3
1,2—Dichlorobenzene 10 101 5.1 5.1
1,3—Dichlorobenzene 10 99 8.1 8.2
1,4—Dichlorobenzene 20 103 5.3 5.2
Dichiorodifluoromethane 10 85 5.9 7.0
1,2—Dichioroethane 10 96 4.1 4.3
1,1—Dichioroethene 10 95 7.2 7.5
trans—1,2—Dichloroethene 10 104 4.9 4.7
1,2—Dichioropropane 10 101 5.0 4.9
Ethylbenzene 10 102 3.9 3.8
Flexachlorobutadiene 10 102 9.2 9.0
Methylene chloride 10 99 4.6 4.6
Naphthalene 10 114 12.0 10.5
Styrene 10 109 6.5 6.0
1,1,1,2—Tetrachioroethane 10 96 4.6 4.7
1,1,2,2—Tetrachloroethane 10 98 6.0 6.1
Tetrachioroethene 10 90 3.8 4.2
6—53
-------
Table 5. (Continued)
Standard Percent
Conc. Recoverya Deviation Re]. Std.
Analyte Mg/L of Recoveryb Dev.
Toluene 10 109 11.6 10.8
1,2,4—Trichlorobenzene 10 108 9.0 8.3
1,1,1-Trichioroethane 10 100 8.0 8.0
1,1,2—Trichioroethane 10 113 10.9 9.6
Trichioroethene 10 92 5.2 5.6
Trichiorofluoromethane 10 84 7.1 8.4
1,2,3—Trichioropropane 10 78 6.7 8.6
Vinyl chloride 10 93 8.3 8.9
o—Xylene 10 106 10.8 10.2
m—Xylene 10 96 4.6 4.8
p —Xylene 20 104 8.3 7.1
a. Recoveries were calculated using internal standard method. Internal
standard was fluorobenzene.
b. Based on eight analyses.
6—54
-------
Table 6. ACCURACY AND PRECISION DATA FOR VOLATILE
ORGANIC COMPOUNDS IN POTW WATER DETERMINED
WITH A WIDE BORE CAPILLARY COLUMN
Average Percent Recovery
Conc. Recoverya Rel. Std. Corrected for
Analyte ug/L % Dev. Background
Benzene 10 86 4.8 b
Bromodichioromethane 10 92 3.8
Bromoform 10 83 4.2
Bromomethane 10 77 11.8
Carbon tetrachloride 10 87 6.9
Chlorobenzene 10 89 3.5 88
Chioroethane 10 78 18.3
Chloroform 10 101 4.8 85
Chioromethane 10 79 4.7 74
2—Chloronaphthalene 10 120 6.4 102
1,2—Dibromo--3-chloropropane 10 82 3.9
Dibromochioromethane 10 86 3.7 85
1,2—Dibromoethane 10 82 4.0
Dibrornomethane 10 87 3.9
1,2—Dichlorobenzene 10 131 4.1 78
1,3—Dichlorobenzene 10 106 9.8 82
1,4—Dichlorobenzene 20 116 4.0 81
1,2—Dichioroethane 10 88 4.4
1,1—Dichloroethene 10 84 4.8 84
trans—1,2- -Dichloroethene 10 92 5.2
1,2—Dichioropropane 10 90 3.6
cis—1,3—Dichloropropene 10 91 3.1
trans—1,3—Dichloropropene 10 94 3.6
Ethylbenzene 10 91 3.9 86
I-Iexachlorobutadiene 10 84 4.5 83
1-lexachioroethane 10 92 4.0
Methylene chloride 150 91 6.9 89
Naphthalene 10 91 4.2 87
Styrene 10 87 3.8
1,2,4,5 —Tetrachlorobenzene 10 101 4.5 88
1,1,1,2—Tetrachioroethane 10 85 4.4
1,1,2,2—Tetrachioroethane 10 87 4.2
Tetrachioroethene 10 87 5.2 86
6—55
-------
Table 6. (Continued)
Average Percent Recovery
Conc. Recoverya Rel. Std. Corrected for
Analyte ug/L Dev. Background
Toluene 10 90 4.0 80
1,2,4—Trichlorobenzene 10 263 3.7 51
1,1,1—Trichioroethane 10 88 6.1 88
1,1,2—Trichioroethane 10 86 3.7
Trichioroethene 10 91 3.3
1,2,3-Trichioropropane 10 87 5.1
Vinyl chloride 10 87 5.4 87
o—Xylene 10 95 3.7 88
m—Xyl ene
p-Xylene 20 94 3.6 85
a. Recoveries were calculated using internal standard method. Internal
standard was fluorobenzene.
b. A blank value indicates the compound was not found in the blank and no
correction was made.
6—56
-------
B LAN K
Ln
SPIKED SAMPLE
4 6 B 10 12 14 16 18 20 22 24 PA PR .RC) 32
FIG, 1 CHROMATOGRJ\19 OF VOLATILE COMPOUNDS IN POTW EFFLUENT
-------
DETERMINATION OF FORMALDEHYDE IN SAMPLES
OF ENVIRONMENTAL ORIGIN
Merlin K. L. Bicking, W. Marcus Cooke, Battelle Columbus Division,
Columbus, Ohio; Fred K. Kawahara, James E. Longbottom, Environmental
Monitoring and Support Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Cincinnati, Ohio
ABSTRACT
An analytical method was developed for the determination of
formaldehyde in liquid samples and leachates of solid samples.
After a review of the current literature, five candidate methods
involving chemical derivatization were chosen for evaluation.
Methods involving both liquid and gas chromatographic procedures was
considered. The derivatization method which satisfied all
performance criteria was reaction with 2,4—dinltrophyenylhydrazine
(DNPH) at pH 5, followed by quantification using liquid
chromatography with absorbance detection. Mildly acidic
derivatization conditions were employed to avoid unwanted generation
of formaldehyde from ubiquitous precursors. Optimized experimental
procedures include a 30 minute reaction time, followed by extraction
using methylene chloride or a solid sorbent method. The
derivatization of formaldehyde proceeded in high yield with
excellent reproducibility. Laboratory blank levels were in the 10 —
15 ag/L range. The method satisfied performance criteria over the
range of 15 — 1400 g/L. Several authentic sample matrices were
used to evaluate the method. A study of the kinetics of formation
of derivative indicated that the method was indeed measuring free
formaldehyde in solution and was not generating formaldehyde from
precursors. The method was subjected to a single laboratory
validation protocol.
INTRODUCTION
The goal of this program was to develop and validate an analytical
method for the determination of formaldehyde in samples of
environmental origin. The potential application of the method to
other carbonyl compounds was also of interest. Formaldehyde is of
primary concern because of its potential environmental hazard and
the fact that there are numerous sources for this chemical in the
environment. These sources may be classified as direct, consisting
of free formaldehyde in solution, or indirect, resulting from
decomposition of various formaldehyde precursors.
The problem of formaldehyde generation from precursors is a well—
documented phenomenon C i .). Numerous chemical entities, both
naturally occurring and synthetic, will release formaldehyde upon
application of an appropriate pH change and/or heat. The desired
6—59
-------
analytical method should be capable of measuring free formaldehyde
as well as any formaldehyde generated under representative
environmental conditions. These levels must be distinguished from
formaldehyde which is generated only by unusual conditions used in
the anplysis. Although suitable analytical methods exist for the
determination of formaldehyde in air, methods available for
determination of formaldehyde in other environmental matrices are
not suitable in terms of sensitivity, selectivity, pH and
temperature. Analysis of real samples is complicated by the fact
that existing literature procedures require high acid concentrations
for demigration, which may result in unwanted generation of
formaldehyde from precursors. The method development activities
described here have been designed to avoid this problem.
PROCEDURE
Liquid samples are initially filtered using centrifugation and a
glass fiber filter. A 100 mL aliquot of sample is adjusted to pH
with 4 mL of 5 M acetate buffer. A 6 niL aliquot of a 1 mg/mL
solution of 2, 4-dinitrophenylhydrazine (DNPH) in ethanol is added
and the solution is placed on a wrist action shaker for 30 minutes
at room temperature. After 30 minutes, a 10 niL aliquot of saturated
Naci is added and the formaldehyde—DNPH derivative is extracted from
the reaction solution using either methylene chloride or a reverse
phase solid sorbent. If the methylene chloride Is used, the solvent
Is concentrated and a solvent exchange to methanol is performed. If
a solid sorbent extraction method is used, the sorbent is eluted
with ethanol. The final volume is 10 niL in each case. The ethanol
or methanol solution is injected directly into a liquid
chromatography system employing an absorbance detector. The
principal chromatographic operating parameters are — Injection
volume: 20i4 ; column:4.6 x 250 mm Zorbax ODS; mobile phase:
methanol/water (75/25); flow rate: 1.0 rnL/mln.; detector: 360 rim.
Quantification is achieved from injection of independently
synthesized formaldehyde—DNPH standards.
RESULTS AND DISCUSSION
Literature Search
A search of the existing literature was conducted using the Chemical
Abstracts data base. Three main categories were searched:
analytical method, reagent, and analyte. The purpose of this search
was to assess the current status of methodologies for determination
of formaldehyde and similar compounds and to identify candidate
analytical methods for further study. The original search yielded a
core list of references which was supplemented with individual
references obtained by project personnel.
6—60
-------
The analytical methods considered involved either liquid or gas
chromatography. An initial survey of the citations resulted in
identification of five derivatization reagents which appeared to
satisfy method requirements in terms of sensitivity, selectivity,
rate of reaction, and limited severity of reaction conditions. The
five reagents were: 2,4—dinitrophenylhydrazine (DNPH),
2, 4—pentanedione (acetylacetone), 3—methyl—2—benzothiazolinone
hydrazone (MBTH), pentafluorophenylhydrazine (PFPH), and
pentafluorobenzyloxyamine (PFBOA). A subsequent search of the data
base was concerned with citations involving these reagents and
carbonyl compounds. More than 60 citations were obtained. The
citations are provided in Table 1 which is organized by reagent.
THE PH DEPENDENCE OF THE DNPH REACTION
Experiments designed to optimize the DNPH reaction at a pH above 3
were conducted. A method employing mild acid conditions was desired
to minimize formation of formaldehyde from samples during the
analysis itself. Additional method development work was necessary
since existing literature procedures used high acid conditions
(i.e., pH
-------
TABLE 1. LITERATURE SEARCH RESULTS
a. an flaa aafl ae at t*flts*fl*Sa aS ttfl
REPHO AUTHOR REFERENCE YEAR REAGENT(s)
1 reaal.y, Ronald R Ho(fmann, Cathertn. S.: Ruepp.1, MelvIn L., Worley, ..llmAnal. them.. 32(1), 111014 1980 DMPHISILICA
2 Cant, P. A. E.s Wa11.r, J • J. Chrom.togr., 130, 26773 1971 0MPH
3 Chtavart, Gluceppes 8er ssnlnt, CecilIa .3, Chromatogr., 318(2). 42732 1985 0MPH
4 Deki, Mitcuos YoahLmus -a, Minoru Chess. Pharm. SuIt., 23(6), 1174.-S 1975 0MPH
S Fund, Rochy, Groaje.n, Daniel Anal Chess., 33(2), 168-71 1981 0MPH
6 Guenler, J. P.: SImon, P. Delcourt, is 01d1.rJ.an, N. F. Lef.vre, C.; MuChromato raphla, 18(3), 13744 1984 0MPH/SILICA GEL
I Haivaraon, Hans .3. Chromato r. , 57(5), 406-9 1971 0MPH
8 Honda, Suaumu ; Rakeht, Razuakt .3. Chromato t., 132(2), 403-11 1978 0MPH
9 Honda, Suausnu: Rakehl, katuakL s Taklura, kL,osht Anal. ChIm. Acta, 11, 25-31 1913 0MPH
10 Heihika, Yaiuyukls Takata, Yoshinori .3, Chrossatogr., 120(2), 57989 1976 0MPH
11 Jacob ,, W. A.; ELastn et, P. T. .3. LLq. Chromatogr., 3(4), 669-16 1982 0MPH
12 Ralilo, H.ikkIs Linko, Relno R. 3. Chromato;r., 166(1), 229-32 1973 0MPH
13 RsLtIo, HeLkkL ; Linko, Retno R. ; Rattaranta, Jukka .3. Chrossatogr., 65(2), 335-60 1972 0MPH
14 RaLo, Paavo .3. Chromato r., 205(1), 39-47 1981 0MPH
“13 Rosnarek, Rarel; Novakova, JanoL iss Ventura, Earet; Churacek, Jaroatav Collect. Czech, Chem. Conueun., 47(8), 2121-7 1982 0MPH, MPH
16 Rorot, A.N.i Doubush, TI. Chess. Anal. (Warsaw), 27 (3-6), 449.-Si 1982 0MPH
17 Ruwata, Raguhlro; Uebort, Mlchlko; Yainasaki, Yoshtakl .1. Chromato r. Sd, 11(5), 264-8 1979 0MPH
18 Lawrence J. P.; 1yen ar, .3. R. mt. J. Environ. Anal. the.., 13(1). 41-32 1983 0MPH
19 Lavln, Jan Otof : Andersaon, Kurt: Lindahi, Roger; N11,jon. Carl Aa.l Anal. Chess. • 57(6), 1032—S 1985 DMPR(HSPO4(MECN
20 Llebezelt, C. HRC CC, 5(4), 215-16 1982 0MPH
21 Linko, P. P.; P.1110, H.; Rainlo, P. 3, Chromatogr., 155(1), 191-4 1978 0MPH
22 Lynch, Catherine: Llm, C. K.: Thomas, Mervyn; Peters, Timothy 3. Gun. Chtm. Acts, 130(1), 117-22 1983 0MPH
23 MansfIeld, C. 1.; Hod e, Srenda 1.; Kage, Robett 8., Jr.; Rasnlln, H. C. J. Chromato r. Sd., 15(8), 301-2 1977 0MPH
24 Haskatinec, N. P.; Mannini, 0. 1,.: Oldham, P. 3. L,iq. Chromato8r,, 4(1), 31-9 1981 0MPH
23 Neng, 2.: Tanner, R. L. Report, BNL-51725 ; AvaiL. MTIS 1984, 9(15), Ab.tr. No. 29191 1983 0MPH
26 Nichols, l’roy ; Svarnas, Geor;e: Thomas, Robert S. Report. MERADCO4-2216; Avail. NTIS Indea (U.S.) 1919, 79(24), 138 1979 0MPH
21 Pap., L. 3.: Turner, L. P. .3. Chromato r. Sd., 10(12), 14 .4-7 1912 0MPH
28 Papa, L. .3.: Turner, L. P. .3. Chromato r. Sd, 10(12), 741-30 1972 0MPH
29 Pias, .3. 8.: Gasco, L. Chromatographla, 8(6), 270-3 1973 0MPH
30 Retndl, Berirams Stan, Hans Juer en .3. Agrlc. Food Chess., 30(5). 849-54 1982 0MPH
31 Reindi, B.; Stan, H. .1. J. Chrom.togr.. 235(2), 481-8 1982 DMPH
32 R,tnecctus, C. A.: Anderson, H. C.: Felska, S. .3. .3. Food Sci., 43 (5), 1494-6 1978 0MPH
33 Ronkalnen, P.ntti: Bruniner, Sasra .3. Chromato r. - 28(2), 253-8 1961 0MPH
34 Scum, Sami 3. Chromato r. - 136(7). 211-1 1971 0MPH
3!’ SmIth, P. A.: Dr,um ,ond, I. Analyst (London), 106(1242). 875-7 1979 I3NPH
-------
TABLE 1. (Continued)
REFNO AUTHOR REFERENCE YEAR REAGENT(s)
36 Steinberg, Spencer: Kaplan, I , R. tnt. J. Environ. Anal Chem. 18(4), 253-66 1984 0MPH
31 Swarm, Si.: Liparl, F .7. Liq. Chromatogr. 6(3), 425-44 1983 0MPH
38 Takami, Katsushlge ; Kuwata, Katuhiro; Suglmae, Aklyoshi ; Nakainoto, Masao Anal. Chem. , 57(1), 2435 1985 0MPH
39 Tuss, H: Ni .itzert, V.: Seller, W. ; Neeb, 8. Fresenius’ 2. Anal. Chem. , 312(7), 61341 1982 0MPH
40 Uraiets, V. P.: RLjks, J. A.: L.clercq, P. A. .7. Chromatogr., 194(2) 135-44 1980 0MPH
41 Van Hoof, P Wificox, A.; Van 8uggenbout,0. ; Jonasens, .7. Anal. Chim. Ada, 169. 419-24 1985 0MPH
42 Van Langenhove, Herman 8.; Van Acker, Marc; Schsmp, Niceas N. Analyst (London), 108(1284), 329-34 1983 0MPH
43 Van SchaIm, K Neth. Milk Dairy .7., 31(1-2), 3964 1983 0MPH
44 Vigh, C.: Varga-Puchony, Z. ; Hlavap, .7.; Pptro-Turc a, M.: SEarfoldl-Szalmaj. Chromatogr., 193(3). 432-6 1980 0MPH
45 Sawickl, 8.; Hauser, T. 8.: Stanley, T.W. Elbert, W. Anal. Che. , 33(1) 93-96 1961 MRTH
46 Kakehi, Kazuaki: Konishi, Tad.o; Sugitnoto, Ikuo, Honda, Susumu .7. Chromatogr. , 318(2), 367-72 1985 P0
47 Okamoto, N. J. Chromatogr. , 202(1), 55-61 1980 PD
48 Kobayaihi, Keiko: Tanaka, HichLru : Kawai, Satoshi; Ohno, Takeo .7. Chromatogr. , 176(1), 118-22 1919 PFPH
49 Hoshika, Yasupukl; Muto, CtLchL .7. Chromatogr. , 152(1). 224-7 1918 PFPH(DNPH-AB)
50 Pobayashi, Kalko; Tanaka, MIchiru; Eaval, Satoahl .7. Chromatogr., 181(2), 413-17 1980 PFBOA, PFPH
51 Stahovec, W.L.; Mopper, K. 3. Chromatogr. , 298(3). 399 406 1986 CHD
‘52 Fret, 8W.: Lawrence, .7. ?. .7. Chrcmatogr., 83, 321-30 1913 DAMSYL-CL+
51 Parsons, James S.: Mitzner, Startiey Environ. Sd. Technol., 9(12), 1053-8 1915 DESORB
54 Chlan. E.S.K.: Kuo, P.P.E.; Cooper, V.3. Environ. Sd. Technol., 11(3), 282-5 1977 DISTILLATION
55 Carey. N. A.; Perslnger, H. 8. 3. Chromatogt. Set.. 10(9), 531 43 1912 DPNH
56 Soyce, Scott 0.: Hornig, James F’. Water Res. , 11(6), 68597 1983 DLI.
57 K.nuth. Michael L. ; Hoglund, Narilynn 0. J. Chromatogr. , 285(1), 15360 1984 DLI.
58 Zlatkls, A.: Wang, P.S.; Shanfleld, H. Anal. Chem., 55(12), 1848-52 1983 DLI.
J. Chromatogr. , 285(2), 383-8 1984 HCN
60 Metvally, Mohamed N. 8.; knundson, Clyde H.: Richardson, Thomas .7. Amer. 011 Chem. Soc., 48(4), 149-54 1911 HYDRAZ
61 Ci1-Av. S.; Schurlg, V. Anal. Chem., 43(14), 2030-3 1971 METAL COMPLEX
62 Guebltz, C.; W1nter telget, 8.: Pr L, 8W. .7. Liq. Chromatogr., 7 (4), 839-54 1984 NBDH
63 Savlcki, 8.: Sawlckl,C.R. Academic Press, London 1978 REF.
64 Sawi ki E. : Sawlcki, CR. Academic Press Longon 1975 REF.
65 ‘ ‘ikalnen, Penttl: Brunmer, Saara .7. Chromatogr. , 28(2), 259-62 1967 HYDRAZ
66 Euo. P.P.K.; Chlan. 8.5K.; DeWalle, P.8. Water Res. ll(11),100511 1977 DISTILLATION
67 Cap, Lubomir: Tayebe, Marghiche Acts Univ. Palacki. Oloniuc. , Fac. Rerum Nat., l9ICbem. 25), 95-tOO 1984 CCIZINC
68 Jatighorbani, Morteza: Eliinger, Max; Starke, Kurt N8S Spec. Pubi. (U. 5.), 464, 131-6 1977 METAL COMP.
49 (Jden, Peter C.: BIgl y, Imogene 8.: Walters, Frederick H. Anal. Chim. Aeta, 100, 555-61 1978 METAL GOMP.
(a) See text for Identification of pertinent acronyms.
-------
100.0 —
90.0 —
80.0 —
70.0 —
60.0 —
a) -
>-
50.0 —
a)
‘3
a,
°- 40.0 —
30.0 —
20.0 —
10.0 —
00.0 — -- _____ ____
0.00 3.00 4.00 8.00
Reaction pH
Figure 1. Effect of reaction pH on percent yield for derivatization of
formaldehyde with DNPH. Circles: phosphate buffer; square:
acetate buffer
I I i I . I . I
1.00 2.00
I I I i I
5.00
6.00
7.00
6—64
-------
separation, sensitivity, reproducibility, and chromatographi
interferences.
The room temperature DNPH reaction was found to provide almost
quantitative derivatization of formaldehyde in less than one hour.
Recovery of the DNPH derivative was consistent at approximately 90
percent over a spike range from 50 to 1000 g/L. Consistent
laboratory blank levels were observed in the 10 to 15 iigIL range for
formaldehyde and 100 to 120 gIL for acetaldehyde. The source of
the analytes appearing in the laboratory blanks was not determined
but appeared to arise from multiple sources. Instrumental detection
limits were estimated to be in the low gIL range for this
derivative. No chromatographic interferences were observed; excess
reagent, formaldehyde derivative, and acetaldehyde derivative were
readily separated using a simple isocratic mobile phase with
absorbance detection at 360 um.
The MBTH reaction provided high yield (i.e., 90 percent) of the
formaldehyde derivative in less than one hour over a spike range
from 100 to 2000 i- .gIL. Chromatographic separation of the
derivatives was easily accomplished using the same system employed
f or the DNPH derivatives, except that the absorbance detector was
operated at 310 urn. Estimated method detection limits were in the
low tg/L range. However, the reagent eluted as a broad band in the
same time window as the formaldehyde derivative, which made
quantification more difficult at low analyte levels. These results
indicate that this derivative may be useful for other analyses.
Additional work is needed since liquid chromatographic separations
with this derivative have not been reported previously in the
literature.
The product of the reaction of acetylacetone, ammonia, and
formaldehyde was monitored by liquid chromatography with
fluorescence detection. The rate of derivatization was slow with
derivative still being formed after six hours. Although
fluorescence detection offered excellent detector sensitivity with
no interferences, the method was considered inappropriate for the
present analytical problem due to the slow kinetics of
derivatization.
The PFPH derivative of formaldehyde could not be readily synthesized
in large quantities, so quantification was not possible for this
reaction. The chromatograms obtained using gas chromatography with
electron capture detection showed a readily identified
formaldehyde—PFPH derivative which could be confirmed by mass
spectrometry. However, numerous small peaks were also present
throughout the chromatogram, indicating impurities in the reagent or
reaction side products. Chromatographic performance deteriorated
after multiple injections which suggested the presence of
nonvolatile material.
6—65
-------
Similar problems were encountered with the PFBOA derivative. Fewer
impurities were observed in the chromatogram but chromatographic
performance deteriorated after multiple injections.
The DNPH reaction was chosen for additional study as a result of
these experiments. The excellent recovery, reproducibility,
sensitivity and freedom from interferences were primary factors in
choosing this method over the other four candidates.
MATRIX STUDIES
The initial DNPH analytical method was evaluated further with five
matrices to assess the need for any additional modifications. Five
authentic environmental matrices, three liquid and two solid, were
chosen in consultation with EPA technical personnel. One liquid
matrix consisted of a groundwater sample removed from a well below a
landfill site. Another liquid sample consisted of a final
effluent. The third liquid sample was a landfill leachate. One
solid sample was a dry solid sludge taken from a facility using
phenol—formaldehyde glue. The second solid sample was wood dust
collected from a wood product using urea—formaldehyde glues.
The solid samples were extracted according to the Toxicity
tharacteristics Leaching Procedures (TCLP) which used a pH 5 acetate
buffer. This method produced leachate which was amenable to
derivatization with the existing procedure. The liquid samples were
derivatized as received, after centrifugation.
In an initial storage experiment the samples were stored at 4°C.
Aliquots were removed at regular intervals and analyzed in
triplicate with and without formaldehyde spikes. These experiments
provided data on analytical reproducibiliity, spike recovery, and
stability upon storage. The groundwater sample showed formaldehyde
levels which were not statistically different from laboratory
blanks. The analyzed levels remained constant over a two week
storage period and spike recoveries were approximately 90 percent.
The final effluent showed low levels of formaldehyde and was stable
upon storage, but spike recoveries were erratic due to emulsion
formation during extraction. Formaldehyde levels in the landfill
leachate decreased steadily upon storage, indicating that the sample
was still biologically active. Reproducibility and spike recoveries
were also poor due to emulsion formatIon. The sludge’s leachate was
str ble up n storage, p ovi ing good reproduci I1Ity a d eo e ’
spikes. Formaldehyde levels were approximately 80 gIg. Similar
results were obtained for the wood dust extract, except that the
formaldehyde levels were in excess of 700 g/g and increased with
time. Reproducibility and recovery for the two solids leachates
were superior to the results obtained for the liquid samples because
the solid extracts did not form emulsions upon extraction.
6—66
-------
METHOD REFINEMENT
The DNP}l derivatization procedure was further refined to ensure
complete reaction of formaldehyde in solution and subsequent
extraction of the derivative. The minor changes involved an
increase in buffer and DNPH concentrations, and the addition of NaC1
before extraction.
While the recovery obtained using the new procedure was not
statistically different from that obtained using the original
method, the new reaction conditions allowed the substitution of the
methylene chloride extraction step with a reverse phase solid
sorbent extraction procedure. Solid sorbents available from three
different manufacturers were evaluated in this experiment. The
results indicated that equivalent recovery of derivative, compared
to solvent extraction, was obtained using a total of l.5g of a C18
sorbent obtained form J.T. Baker Chemical Company. This alternative
extraction procedure reduced the analysis time for each sample and
avoided reproducibility problems caused by emulsion formation during
methylene chloride extraction.
The modified reaction conditions and solid sorbent extraction
procedure were used in all subsequent experiments. Representative
chromatograms for standards and each matrix type are provided in
Figures 2 through 6.
GENERAL APPLICATION OF THE DERIVATIZATION METHOD
Efficacy of the procedure for determination of other analytes of
Interest to EPA was assessed also. Potential Interferences from
paraformaldehyde, a polymer of formaldehyde, were examined by
derivatizing a constant level of formaldehyde in the presence of
varying levels of paraformaldehyde. These experiments demonstrated
that a significant Increase in analyzed formaldehyde levels only
occurred when paraformaldehyde was present at higher levels than
free formaldehyde.
Four other carbonyl compounds were also derivatized. Acrolein,
chioroacetaldehyde, beuzaldehyde, and cyclohexanone were derivatized
using the procedure optimized for formaldehyde. Derivatives were
tentatively identified in the chromatogram for each compound. The
highest yields were obtained for acrolein and cyclohezanone. Since
the reaction occurred to a measurable extent for each compound,
acceptable yields could probably be obtained by adjusting
appropriate experimental variables such as DNPH concentration,
reaction time, and reaction temperature.
KINETICS OF FORMATION OF THE FORMALDEHYDE—DNPH DERIVATIVE
This study was undertaken to measure the rate at which the
formaldehyde—DNPH derivative was formed under the reaction
6-67
-------
4.0
t ., , - i— t ‘ r
‘ •I i
i—r i- i
t -t r -t -t’—v i.r r.l
r— 1
2.0 6.0
10.0
M c x i
Mi
125. 000 .
MINUTES
X Siux, 125.000
0.000
Figure
2. Representative chromatogram from
optimized method. Abbreviations
derivatizatlon of a 50
-- FOR-D: formaldehyde
ppb solution
derivative;
of formaldehyde using
ACET-D: acetaldehyde
the
uerlva Jve.
I
—a
-a
-a
S.
I
S.
ACET-D
9.28
FOR—u
7.09
co
- 4
S.
-J
1 -,
8.0
12.0
14.0
16.0
10.0 20.0
-------
a-
a,
.4
.4
.4
.4
-4
-a
-I
I
-.
.4
•0
•1
I
.4
00 02
FOR-O
7.13
003
10.41
13 .1214.13
6.0 8.0 12.0
MINUTES
14.0 18.0 28.0 20.0
x Sui,xi
ZSMnI
200.000
0. 000
Max,
4. 0
20
200.000 iu
0. 000
r-i r , v st i --i-j-’r r--r-v-——r-r i—--’r-r’——j
Figure 3.
Representative chromatogram from derivatization of final effluent sample.
-------
005
004 13.80
12.80
1 • ’ • F
Mawi
Mini
2.0
150.000 .v
0.000 i v
8.0 10.0 12.0 14.0 16.0
18.0 20.0
150.000
0. aoo
ACET-O
9.22
0
FOR—O
7.09
003
10.28
4.0
I I• •••I - P • - I P 1 r P •1 t P P I I• I I I- ? I -1 P• I I t
MINUTES
X Sima
2 Spun.
Figure 4.
Representative chromatogram from derivatization of landfill leachate sample.
-------
005 14.12
13.03
‘•T”’’ I “-‘r?”r—I-j- - -,-t.?. .r- , —J-r-1--1”r--I-—t t—Ir-r tir—r 1
2.0 4.0 6.0
isaoo
0.000 my
8.0 12.0 14.0 16.0 18.0 20.0
MINUTES
2 Sm,
X Sum,.
150,000
0. VJOO
ACET-D
8.32
.1
.4
.4
.4
-4
.4
*
-J
.4
.1
-4
-1
-1
-4
-4
-4
-j
-.4
.4
.1
—4
.4
.4
I
FOR-U
7.15
004
10.42
4 T h. ?.
Max,
Mine
Figure 5. R ’rresentative chromatogram from derivatization of sludge extract.
-------
-J
-4
-a
. 8
.4
-i
-.4
.4
.4
—I
Mi ,.
150.000 NV
0.000 NV
FUR -U
7.01
ACET—D
Q.15
003
18.21
004
12.78
-5
13.04
‘I’ ‘‘•i’ ,• iiji’,-i t it’ ,1ttt*1 -t—-ri ‘r ’ -r ? I
48 8.0 80 10.0 12.0 14.0 18.0 18.0 20.0
— Z Siww, 150.008
MINUSES xsIuul, 8.000
Figure 6.
Representative chromatogram from derivatizatlon of wood dust extract.
-------
conditions employed. The reaction conditions assured an excess on
all reagents, providing pseudo—first—order conditions for the
reaction for formaldehyde, and first—order formation of the
derivative. The purpose of these experiments was to determine the
apparent rate constant, k, for the formation of the derivative using
spiked reagent water, and compare this rate to the value obtained
from derivatization of formaldehyde in sample matrices. The data
would also allow a comparison of the form of the kinetic equation
for each sample, as evidenced by the shape of the kinetic plots.
Linear first—order plots were obtained for formation of the
forrnaldehyde—DNPH derivative matrices, indicating no complex kinetic
relationships. In addition, the observed rate constants (Table 2)
for the samples were not statistically different from the value
obtained from standards. In other words, the derivatization
proceeded at the same rate in all cases, and the analysis did not
generate additional formaldehyde from precursors. These data
suggest that the analytical results obtained using this
derivatization procedure represent levels of free formaldehye in the
samples.
SINGLE—LABORATORY VALIDATION OF THE METHOD FOR FORMALDEHYDE
A single—laboratory validation protocol was used to evaluate the
final, optimized method for the determination of formaldehyde. The
protocol consisted of six steps —— instrumentation range,
preliminary method evaluation, ruggedness testing, method range and
detection limit, referee validation, and matrix validation. All
steps except the referee validation were performed.
The liquid chromatography system with absorbance detection was
determined to be linear from 1.43 to 1430 wg/L formaldehyde. The
preliminary method evaluation consisted of multiple replicates at a
spike level of 75 wgIL. A mean percent recovery of 88 percent with
a relative standard deviation of under four percent was obtained in
these experiments (Table 3). Ruggedness testing was performed to
demonstrate that reasonable variation in method parameters did not
affect the analytical results. Four method variables (reaction
time, reaction pH, volume of DNPH reagent, and number of sorbent
cartridges) were chosen because they were considered most likely to
be subject to variation in the course of an analysis. Relatively
broad limits were selected for each variable. However, none of the
experimental variables was found to significantly affect the
analytical results, suggesting that the method was sufficiently
rugged to be useful in other laboratories (Table 4). The method
detection limit was calculated to be 7.2 wg/L (Table 5). Finally,
the method was evaluated with two matrices (final effluent and
sludge). Although some significant differences in recovery were
observed between standards and spike matrices, the precision and
recovery data for the matrix experiments easily satisfied the
established method performance criteria (Table 6). Additional
6—73
-------
TABLE 2. SUMMARY OF FIRST-ORDER RATE CONSTANTS FOR
APPEARANCE OF FORMALDEHYDE DERIVATIVE
Average k
Matrix
.
Standard
Deviation
Standard (250 ppb)
Landfill Leachate
1.83 x i0
1.70 x i0
0.12 x i0
0.50 x i0
Final Effluent
Wood Dust
1.98 x i0
1.90 x i0
0.36 x i0
0.34 x i0
Sludge
1.97 x i0
0.47 x iO
6—74
-------
TABLE 3. RESULTS OF PRELIMINARY METHOD EVALUATION
Spike Level (ppb) 75.0
Mean Response For Eight Replicates 65.8
Mean Percent Recovery 87.7
Lower Confidence Bound 85.2
Upper Confidence Bound 90.2
Percent Relative Standard Deviation 3.42
6—75
-------
TABLE 4. RESULTS OF RUGGEDNESS TEST EXPERIMENTS
Parameter Low Condition High Condition
Parametçr
Effect . a)
Significance(b)
Reaction Time
25 mm
40 mm
-0.447
Not
significant
Reaction pH
4.8
5.2
2.27
Not
significant
DNPH Reagent
5 ml
7 ml
4.13
Not
significant
Number of
Cartridges
2
4
3.49
Not
significant
(a) A measure of the effect of the parameter on the analytical results.
(b) Critical Effect Level = 6.92 for 7 degrees of freedom.
6—76
-------
TABLE 5. RESULTS FOR METHOD RANGE AND METHOD DETECTION LIMIT EXPERIMENTS
C’
—4
—4
Spike Level (a)
(ppb)
Percent Recovery(a)
Confidence Bound
for Recovery
(Lower/Upper)
Percent Relative
Standard Deviation
Confidence Bound
For Percent Relative
S.D. (Lower/Upper)
Method Detection
Limit (ppb)
Number of
Replicates
15.0
70.6
57.3/84.0
22.6
15.0/46.1
7.19
8
46.8
94.6
86.8/102
9.77
6.46/19.9
8
146
88.9
85.5/92.3
4.57
3.02/9.31
8
457
86.4
82.9/89.9
4.37
2.82/9.63
7
1430
91.7
87.9/95.5
4.91
3.25/10.0
8
(a) All values are corrected for a laboratory blank value of 12 • 5 tig/L.
-------
TABLE 6. SUMMARY OF MATRIX VALIDATION EXPERIMENTS
Data from Method Range experiments.
Difference is not significant if confidence bound range includes zero.
Difference is not significant if confidence bound range includes
unity (1).
Spike Level
ppb
Final
Effluent
Sludge
Extract
46.8
1430
457
1430
99.1
94.6
80.5
91.7
106
86.4
82.2
91.7
6.84
9.77
4.57
4.91
3.24
4.37
1.98
4.91
4.5
-11
20
-9.5
-6.3/15
-16/-6
15/25
-14/-5
Percent Recovery
Test Matrix
Standard Matrix(a)
Relative Standard Deviation
Test Matrix
Standard Matrix
Percent Recovery Difference
Confidence Bounds for
Percent Recovery Difference
(Lower/Upper)
Significant Difference(b)
Test/Standard
Standard Deviation Ratio
Confidence Bounds for
Standard Deviation Ratio
(Lower/Upper)
Significant Difference(c)
No Yes Yes
Yes
0.86
0.82
1.0
0.38
0.39/1.9
0.37/1.8
0.45/2.5
0.17/.85
(a)
(b)
(c)
No No No Yes
6—78
-------
Table 7 provides a summary of single oeprator accuracy and
precision.
TABLE 7. SINGLE OPERATOR ACCURACY AND PRECISION
Parameter
Matrix
Type
Average
Percent
Recovery
Standard
Deviation
(Percent)
Spike
Range
(jig/L)
Number
of
Analyses
Formaldehyde
Reagent
Water
86
9.4
15.0-1430
39
Final
Effluent
90
11
46.8-1430
16
Sludge
93
12
457-1430
15
6—79
-------
matrix experiments may be necessary, but it can be concluded that
the analytical results provided by this method accurately reflect
the amount of formaldehyde in the samples.
CONCLUSIONS AND RECOMMENDATIONS
Quantification of formaldehyde at low wg/L levels is possible by
employing DNPH derivatization at pH 5. The reaction solutions can
be readily extracted using solid sorbent cartridges rather than
solvent extraction. Under the conditions employed, the formation of
the formaldehyde—DNPH derivative appears to follow first—order
kinetics In all matrices studies. The results suggest that the
procedure is indeed measuring free formaldehyde in solution and
significant amounts of formaldehyde are not being formed from
precursors dUring the derivatization. The single laboratory
validation data indicate that the method is rugged and should
perform well in other laboratories.
Additional work with this general derivatization approach should
focus on application of the method to other matrix types.
Derivatization of other carbonyl compounds may also be possible with
minor modifications of the reaction conditions.
REFERENCES
“Aldehydes — iotometric Analysis,” E. Sawicki and C. R. Sawicki,
Volume 5; Academic Press, London: 1975.
Toxicity characteristics Leaching Procedure, 40 CFR, Volume 51,
No. 114, Friday, June 13, 1986, Page 21685.
6—80
-------
HEATED PIJRGE-TRAP-DESORB ANALYSIS OF VOLATILE, WATER SOLUBLE COMPOUNDS
S. V. Lucas, Principal Research Scientist, H. M. Burkholder,
Researcher, 3. S. Warner, Research Leader, R. A. Kornfeld, Projects
Manager, Battelle Columbus Division, Columbus, Ohio, and 3. E.
Longbottom, Chief, Organic Analyses Section, U. S. Environmental
Protection Agency, Cincinnatti, Ohio
ABST1 ACT
In an on—going research program sponsored by the U. S. EPA, heated
purge—trap--desorb (H—PTD) methodology has been investigated for its
applicability to the analysis of high to moderate volatility organic
pollutants which are too water soluble for conventional PTD analysis.
The 33 analytes selected for this study are Appendix VIII and Michigan
List compounds. Some were selected due to an a priori expectation
that they would i equire an H—PTD approach. Some were compounds which
had failed prior SW—846 PTD method testing. Other volatile but highly
polar compounds which had failed to elute in GC—MS suitability testing
using the SP1000/Carbopack B packed GC column were also included. The
analytes included 9 nitriles, 8 nitrogen bases, 5 alcohols, 6 carbonyl
compounds, 3 thiols, and two others. Although most of the analytes
displayed good GC behavior on the chosen GC column (30mm x 0.53mm ID
fused silica; 1.0 micron Supelcowax 10), many were found to be
hydrolytically unstable to the heated purge step, thermally unstable
toward the trap desorption step, or non—recoverable by H—PTD for
unknown reasons. H—PTD testing for method development was performed
principally with acrolein, acetonitrile, propionitrile, acrylonitrile,
methacrylonitrile, methyl ethyl ketone, isobutanol and dioxane.
Most of the recovery and reproducibility problems encountered in H—PTD
analysis are associated with the large amount of water evaporated from
the heated purge vessel. Three approaches were investigated in an
attempt to control this water and prevent its condensation in the
plumbing and trap of the H—PTD device: 1) A small condenser was
used at the purge vessel outlet, 2) purge stream dilution with
trapping at a temperature above the resultant dew point, and 3)
chemically selective water removal. A fourth approach using a Naf ion
tube had already been shown to be impractical with highly polar
analytes and, therefore, was not investigated. The first approach was
the only water control method found practical. This approach was
explored with regard to the breakthrough volumes of critical analytes,
the use of more retentive traps, the temperature of the condenser, and
the “salting out” effect of chlorides and sulfates of sodium and
magnesium.
Results to date indicate that the developed method, which employs a
standard 5—niL purge vessel is effective for hydrolytically stable
analytes with water solubilities, vapor pressures and polarities
similar to, or more favorable than, acetonitrile and isobutanol.
Absolute H—PTD recoveries ranging from 75 to 100 percent with RSD’s
<10% (3 replicates) have been obtained at the 50 ug/L spiking level
6—81
-------
using 5—mL samples and FID detection.
Future work expected to be completed before presentation to the
symposium includes precision and accuracy studies with multiple
replicates for all analytes of the original 33 for which the method
provides sou recovery. Results will be reported in terms of absolute
H—PTD recovery (versus septum injection of the H—PTD sample), recovery
(versus H—PTD calibration standards), precision (based on 6 to 8
replicates at high and low spiking levels), and estimated method
detection limits for the tested analytes in reagent water.
6—82
-------
EVALUATION OF EXTRACTION
CONDITIONS FOR APPENDIX IX COMPOUNDS
Thomas A. Pressley, Environmental Monitoring and Support Laboratory,
U.S. Environmental Protection Agency, Cincinnati, Ohio; and Robert
L. Einhaus, Technology Applications, Inc., Cincinnati, Ohio.
ABSTRACT
An analytical method utilizing gas chromatography/mass spectrometry
(GC/NS) was selected as the mode of analysis for the extractable
semi—volatile organic compounds comprising the Appendix IX list
(Fed. Reg. Vol. 51, No. 142, Thursday, July 24, 1986). The
chroinatographic conditions specified in the Contract Laboratory
Program (CLP) have been selected for analysis of the Appendix IX
compounds, but the extraction conditions specified in the CLP, for
the analysis of Appendix IX compounds are the subject of much
criticism and variance. This is especially true when using the
optional continuous extractors which supposedly remove the art form
the extraction process, and enable the extraction of
emulsion—forming water samples. The main criticism of the CLP
extraction scheme seems to arise from the low recoveries of the more
polar neutral and the acidic compounds by the base neutral
extraction at elevated pH followed by acidification and extraction
of the acidic compounds. These losses may be attributed to
occlusion. Initial extraction under acidic conditions, however,
eliminates many of these problems, but produces acid/neutral
extracts that sometimes contain more interferences and other organic
constituents. Subsequent pH elevation for the extraction of the
basic constituents gives less precipitation and flocculation than
when extracted initlaily at elevated pH.
In these studies, a set of experiments was set up to compare the
recoveries of Appendix IX analytes from reagent water under various
extraction conditions, all utilizing continuous extraction with
methylene chloride. The following schemes were examined:
L. Acid/neutral extraction followed by basic extraction
2. Base/neutral extraction followed by acidic extraction
3. Neutral extraction followed by acidic extraction
4. Extraction at pH 4 only
Percent recovery data for the analytes under the various extraction
schemes was presented with statistical evaluation.
INTRODUCTION
As mandated by Sections 304(h) and 501(a) of the Clean Water Act of
1977, The U.S. Environmental Protection Agency proposed fifteen test
procedures for measuring the concentration of priority pollutants in
6—83
-------
industrial wastewaters (1) which, after extensive public comment,
laboratory testing and subsequent editorial revision, were
promulgated as approved methods (2). Among the most ubiquitously
applied of these procedures has been EPA Method 625, an omnibus
method for measurement of trace, semivolatile, organic pollutants by
gas chromatography/mass spectrometry (CC/MS following extraction of
a water sample with metbylene chloride using separatory funnel or
continuous extractor and concentration of the extract by
Kuderna—Danish (K—D) evaporation. The method was designed to
recover basic, neutral and acidic compounds. Thus, water samples
are first extracted under basic (pH 11) conditions, acidified to a
p11 2 and reextracted. The two extracts are then analyzed
independently under different chromatographic conditions. The basic
and neutral compounds contained in the first extract are separated
by a packed, 1.8 meter glass column with a stationary phase of 3%
SP—2250 (or equivalent) while separation of acidic compounds is
accomplished with a similar column packed with SP—1240DA (or
equivalent). With the advancement in recent years of fused silica
capillary column (FSCC) technology, however, separations adequate
for mass selective detection of basic, neutral and acidic organic
compounds have been achieved with a single, bonded—phase, silicone
P 5CC. Accordingly, the Superfund (CERCLA) Contract Laboratory
Protocol (CLP) and the SW—846 (RcRA) program have adopted capillary
column CC/MS procedures which analyze combined, base/neutral and
acid extracts.
Since the inception of Method 625, it has been understood that while
the base/neutral—acid extraction sequence provides acceptable
analytical recoveries for a preponderance of priority pollutants,
chlorinated hydrocarbon pesticides, such as a—BHC, —BHC (lindane),
endosulfans I and II and endrin, displayed impaired recoveries due
to decomposition under alkaline extraction conditions (3).
Subsequently, a neutral (p11=7) — acid (pH 2) extraction procedure
was shown to outperform Method 625 In the recovery of these
compounds (4). Similar losses have also been demonstrated for
certain phthalate esters which were attributed to hydrolysis under
basic conditions (5).
The most strident criticism of the B/N—A extraction procedure
regards the low recoveries of polar neutral and acidic compounds
form groundwater and wastewater samples. These losses have been
attributed to occlusion. The acid/neutral—base extraction sequence
has found favor with many analysts because of its ability to
mitigate the flocculation and precipitation which engenders
occlusion of trace organic compounds in the sample. There is some
concern, however, that an Initial acidic condition may produce
Interferences or modification of constituents.
In an effort to arrive at some consensus on extraction methodology,
EMSL—Cincinnati has designed an experiment to evaluate 4 candidate
6—84
-------
extraction procedures when applied to approximately 200 represen-
tative compounds from the priority pollutants, RCRA Appendix VIII
and IX and Hazardous Substance lists. The 4 procedures included
base/neutral extraction at a pH 11 followed by acidic extraction at
a p13 2(B/N—A), acid/neutral extraction at a ph 2 followed by basic
extraction at a ph ll(A/N—B), neutral extraction at pH7 followed by
acidic extraction at pH 2(N—A) and, finally, a single extraction at
a pH of 4 (pH4). This last procedure has been proposed as an option
for the continuous extraction technique on the premise that a
prolonged extraction period can compensate for less favorable
partition coefficients. All extractions were accomplished using
continuous, liquid—liquid extractors (method 3520). Continuous
extraction was employed in favor of the separatory funnel technique
(Method 3510) to simulate the extended extraction conditions which
would be most conducive to analyte decomposition and for formation
of interferences and to minimize random variance that would
accompany numerous repetitions of the labor intensive manual
procedure.
The initial phase of this study consisted of an investigation of
analyte recoveries form spiked reagent water samples using the 4
extraction procedures. The 200 compounds were divided into 8
analysis groups. Extracts of five of these groups were measured by
GC/MS analytical procedures for semivolatile, organic pollutants
detailed in the CLP. The remaining three analysis groups comprised
R RA Appendix VIII compounds which for reason of involatility or
thermal lability were deemed unsuitable for GC/MS analysis (6).
Following the same extraction protocol, these compounds were
analyzed by high performance liquid chromatography (HPLC) using
SW—846 Methods which are currently in the late developmental phase
here at EMSL—Cincinnati.
At the completion of the above experiment, spiked POTW samples will
be analyzed by the same protocol In order to evaluate the matrix
dependent characteristics of these extraction procedures.
DESIGN
Approximately two hundred compounds from the priority pollutant,
R RA Appendix VIII and Hazardous Substance Lists were grouped for
analytical convenience into eight analysis sets. These compounds
were obtained as standard concentrates from both commercial and EPA
sources. Each analysis set was spiked into reagent water and
processed by continuous extraction according to each of the
following treatment procedures.
*A/N_B: Extraction at pH 2 followed by adjustment to
pH 11 and reextraction.
*B/N..A: Extraction at ph 11 followed by acidification to
ph 2 and reextraction.
6—85
-------
*N_A: Extraction of neutral sample (p11’ 7 ) followed by
reextraction at pH 2.
*pH4: Single extraction at pH4.
Four replicate extractions were performed for spikes of each
analysis set under each of the four extraction conditions itemized
above. The resultant extracts were then concentrated by K—D
techniques and analyzed by CC/MS or HPLC procedures. Percent
recoveries achieved for the analytes under each extraction condition
were then compared. Recovery differences were identified at p 0.01
significance level using ANOVA statistical processing.
(Statistician’s MACE, Matrix Calculating Engine, Inc., Madison,
Wisconsin).
EXP tIMEWTAL DETAIL
Apparatus and Qiemicals :
The continuous, liquid—liquid extractor used in this study is
described in Figure 1 • The remaining glassware, the solvents and
reagents, obtained from a variety of commercial suppliers, all
conformed to specifications presented in Sections 2.3 and 2.4 of the
CLP.
Analytes :
CC/MS analytea were acquired from commercial and EPA sources.
Supeipreme — HC Standards (Supelco, Inc., BeUefonte, PA) served a
source of priority pollutants and HSL compounds. The Quality
Assurance Branch (QAB) of (SL — Cincinnati provided spiking
concentrates of RCRA Appendix IX compounds which were prepared by
QAB for the Interlaboratory validation study of SW—846 Method 8270.
HPLC analytes were obtained for the Aldrich and Sigma aiemical
Companies as neat materials and from the EPA Repository for
Hazardous Materials as reference standards.
Spiking Solutions :
The 200 compounds were divided into 8 analysis sets. A spiking
solution constituting each analysis set was prepared by dilution
with either methanol, for the CC/MS spikes, or acetonitrile, for the
HPLC spikes.
Set A: Priority Poilutants and HSL Compounds — This spiking
solution was composed of the Supelpreme — BC mixes listed below.
All mix constituents were at concentrations of 2000 ug/rnL.
Base/neutral Mix 1 (4—8900)
Base/neutral Mix 2 (4—8901)
Pesticides (4—8903)
6—86
-------
Phenols (4—8904)
Polynuclear Aromatic Hydrocarbons (4—8905)
Beuzjdines (4—8906)
One—half mL of phenol mix and 0.25 mL of each of the remaining mixes
were diluted with methanol to 25.0 mL in a volumetric flask. To
constitute the sample, 1.0 niL of the spiking solutions was pipetted
into 2.0 L of reagent water.
Set B: Method 8270 MV Nix 1 — Six niL of Mix 1, Ampule 3 spiking
concentrate was placed in a 50.0 niL volumetric flask and diluted to
the mark with methanol. Two liters of reagent water were spiked
with 2.0 niL of this solution. Parameters in Mix 1, Ampule 3
concentrate were at concentrations of either 200, 160, and 120 ug/niL.
Set C: Method 8270 MV Mix 2 — Compounds in Mix 2, Ampul 2
concentrate were uniformly at a concentration of 400 ug/niL. Three
niL of this concentrate was diluted to 50.0 mL with methanol and
spiked at a level of 2.0 niL per 2.0 L of reagent water.
Set D: Method 8270 MV Mix 3 — The Ampul 2 concentrate of this mix
was diluted and spiked in a manner identical to Mix 2, above.
Sets F, G and H: Appendix VIII HPLC Compounds — Spiking solutions
were prepared by dilution of neat materials and repository standards
with acetonitrile. Spiking volume was 2.0 niL for 2 L volume of
reagent water. The spike concentration of each analyte was such
that a 100% recovery would yield a detector response equivalent to
20 x the EDL. Analyte set constituents are listed below:
SetF SetG
Ethylene thiourea Mitomycin C
1,2 — Diphenyihydrazine Methoiny l
Rotenone 1, 2—Phenyldiamine
N—Nitroso—N--ethylurea 1, 5—Napthalenediamine
Benomyl Actinomycin D
Set H
Thiourea
l—Acety l—2—thiourea
1, 3—Phenyldiamine
l—(o—chloropbenyl ) —2—thiourea
Diethyls tilbestrol
Surrogate Standards
Each water sample was spiked with CLP, acid and base/neutral
surrogate compounds. Spiking solutions consisted of iuethanolic
dilutions of surrogate standard mixes supplied by Supelco, Inc.
6—87
-------
Calibration Standards
Supelpreme—HC Standard Mixes and CLP Surrogate Mixes were also used
to prepare CC/MS calibration standards for priority pollutants
(Analysis Set A). Calibration standard concentrates for the R RA
Appendix IX compounds were provided by Quality Assurance Branch of
F 1SL—Cincinnati. Calibration curves were generated using three
standards prepared by serial dilution of the standard mix
concentrates with methylene chloride. HPLC calibration standards
were made by dilution of neat materials and repository standards
with acetonitrile.
Internal Standards
As per CLP a 1OWL aliquot of internal standard solution was added to
all calibration standards and sample extracts just prior to GC/MS
analysis. The internal standard solution was supplied by Supelco,
Inc.
Instrumentation
The instrument used for CC/MS analysis was a Hewlett Packard 5890 GC
coupled with an HP 5970B Mass Selective Detector (MSD). Instrument
output was processed using a }IP—200 data system.
With HPLC, a Waters Associates modular system was used for all
analyses. The system included a WISP autosainpler (Model 71DB), two
solvent pumps (Model 510), a fixed wavelength UV detector (Model
440) and system controller (Model 721). An IBM PC/AT with Nelson
Analytical software (Rev. 3.6) and a Nelson Analytical Interface
OOodel 762B) collected and processed the chromatographic data.
Continuous Extraction
The continuous extractor was a 2—L capacity, all glass system
fabricated by Paxton Woods Glass Ship, cincinnati, Ohio. A diagram
of the extractor is presented in Figure 1. The extractor was first
loaded with 150 mL of methylene chloride after which 300 mL of
methylene chloride was placed In the distilling flask. Two liters
of spiked water sample were then poured gently into the extractor
with the aid of a stirring rod. The stirrer—condenser assembly was
attached and actuated, after which the methylene chloride reservoir
(distilling flask) was warmed to 40°C with a heating mantle. The
solvent volume cycled at a rate of approximately 0.6 hour -.
Samples were extracted overnight for a period of 20 ± two hours.
Adjustment of the sample pH was accomplished just prior to
Initiating the extraction with the sample in the extractor. Sample
pH was measured by a pH meter to the nearest tenth of a pH unit. To
d t’ t aci lity to pH 2 d asiTc ty t p N H 2 SO ; nd 6
6—88
-------
NaOH(aq) solutions were employed. Adjustments to pH7 and p11=4 were
performed using phosphate buffers.
Following the initial extraction, samples were readjusted to the
second p 1 1. Methylene chloride in the flask was collected for K—D
concentration and replaced by fresh solve it. The sample was then
reextracted in the manner described above. A second extraction was
not required for the sample at pH4.
Extracts were dried with acidified sodium sulfate and concentrated
to a volume of 1.0 mL by K—D techniques as prescribed in Method
8270. Concentrate pairs of the same sample were combined in 0.50 mL
portions in 1.5 niL injector vial. Following the above procedures,
extracts prepared for HPLC analyses required a solvent exchange
step. One niL of combined extract was dosed with 2.0 niL of
acetonitrile and reconcentrated to 1 niL.
GC/MS Analysis
Analyses of Sets A through E were accomplished using procedures for
GC/MS quantitation of seinivolatile extractables detailed in the
CLP. Recommended chromatographic conditions were followed using a
30 m x 0.32 mm (I.D.), 1 urn film thickness, DB—5 fused silica
capillary column (J. & W. Scientific). Application was made of
SW—846 Method 8270 (7) in selection of ion masses (mlz) for
quantitation of RI RA IX compounds not included in the CLP.
HPLC Analysis
Sets F, C and H were measured by HPLC. These methods employed
reverse phase separation with a 4.6 mm x 25 inn, 10 urn ODS analytical
column (Resolver C18, Fisher Scientific) in a mobile phase
consisting of acetonitrile (Mec N) and water. chromatographic
separations were obtained using the following linear solvent
gradients:
(1) 20% Me N:80Z H 2 0 to 100% Me N (20 minutes)
(2) 100% 1120 (5 minutes), 100% 1120 to 100% NecN (20 minutes)
Analytes were detected by tJV at a wavelength of 254 nanometers.
RESULTS
At the time of this writing, we are able to report on recoveries of
the priority pollutants and RCRA Appendix VIII, HPLC compounds.
GC/MS analysis and data reduction for the RCRA Appendix IX compounds
are still in progress. These results will be incorporated into this
report for presentation at the Symposium For Solid Waste Testing and
Quality Assurance , 13 — 17 July 87.
6—89
-------
FIgure 1. LIquid/liquid Continuous Extractor
tor
driven
AUIhn
type
condenser
r Joint
45/50
50 cm
r Joint
24/40
Capacity
500
m l
6—90
-------
Percent recovery data for priority pollutants and RCRA Appendix
VIII, HPLC compounds are shown In Tables 1 and 2, respectively.
Table 1 also includes results obtained for surrogate standards,
which, for each of the four extraction methods tested, fell within
the performance based limits set forth in Exhibit E, Section IV of
the CLP. Percent recoveries were calculated as follows:
PR = Cexp x 100%
Ct
where, Cexp is the experimentally determined analyte concentration
in the water sample and Ct is the nominal, true concentration.
Analyte recoveries were not corrected (normalized) to responses
obtained from a synthetic, reference solution of the spiking
concentrate. Reference solutions were employed, however, for
quality assurance purposes. Results were statistically processed
using the ANOVA t—test to Identify method dependent recovery
differences with a null hypothesis significance of p(t) 0.01.
While all extraction method combinations were compared, this report
focuses on comparisons of the CLP method (base/neutral—acid) with
each of the other tested procedures. Table 3 lists priority
pollutants which displayed significant recovery differences and
Table 4 provides a similar listing of RCRA Appendix VIII, HPLC
compounds.
Priority Pollutants
In comparing the B/N—A and the A/N—B extraction schemes, 18 out of
57 compounds presented Improved recoveries under the A/N—B procedure
while only 2 had significantly diminished recoveries. Thirty—three
differences were found for B/N—A versus N—A with 15 cases of greater
recovery and 18 cases of lesser under the N—A procedure. The pH4
extraction method fared worst in comparison with B/N—A where the
latter outperformed the former for 27 out of 32 compounds presenting
significant recovery differences.
The comparative performance of the pH4 extractions was a
puzzlement. It was expected that this method would mimLck A/N—B and
N—A procedures in recovering neutral and weakly acidic compounds.
However, the pH4 extraction, with a few exceptions, yielded
recoveries uniformly lower than A/N—B and N—A. The most obvious
explanation for this discrepancy is the pH4 procedure was performed
as a single 20—hour continuous extraction whereas the other
processes consisted of 2 such extractions conducted sequentially on
the same sample. Given the high partitioning efficiency of the
continuous extractors, it seems unlikely that any molecular
compounds would remain in the sample after a 20—hour extraction,
thus rendering the above explanation as specious. It will be
tested, nevertheless, by repeating the pH4 method employing a second
extraction.
6—91
-------
The most consistent and pronounced differences in extraction
performance occurred with the alpha, delta and gamma isomers of BHC
(1,2,3,4,5, 6—hexachiorocyclohexane) and the phthalate esters
(Table 3).
Decomposition of d..—BHC and ‘—BHC(lindane) under alkaline extraction
conditions has been documented for EPA Method 625. Slayton and
Trovato have noted diminished recoveries of lower molecular weight
phthalate esters under the B/N—k regime. The latter findings were
attributed to basic hydrolysis at the initial extraction pH which
could be mitigated by sterical effects (5). Phthalate ester results
reported in Table 1 suggest this hypothesis, as well. A compilation
of pertinent phthalate ester data is given on the next page:
((A/N—B)—(B/N--A))
methyl — 932
ethyl — 82.3
butyl—beazyl — 60.0
phthalates — di—N—butyl — 32.2
bis(2—chloroethoxy) — 6.0
bis(2—chloroethyl) — 3.0
bis(2—chloroisopropyl) — 2.9
di—N—octyl — —5.0
bis(2—ethy lhexyl) — —12.0
A. majority of phenols demonstrated improved recoveries when not
exposed to initial, basic extraction conditions as evidenced by the
results of the B/N—A and N—A extraction schemes detailed in Table 5.
As the table shows the significant increases were fairly modest (9%
— 16%) and the B/N—A phenolic recoveries by themselves were quite
acceptable.
TABLE 5
PERCENT RECOVERIES OF PHENOLS BY B/N—A, A/N—B
AND N—A EXTRACTION SCHEMES
Compound B/N-A A/N-B N-A
D—5 Phenol 79.5 93.3 87.8*
phenol 82.0 92.5 94.8
4—Qi loro—3-iiethylpheno l 92.8 101.0* 106.0
2—Qilorophenol 70.8 80.2* 68.8*
2, 4, 6—Tribromophenol 81.3 89.5 92.8
2,4—dich.loropheno l 88.0 97.8 103.0
2—Nitrophenol 89.8 95•3* 104.0
2,4,6—Trichiorophenol 80.3 89.0 96.0
2—Fluorophenol 73.0 86.3 86.5
Pentachiorophenol 114.0* 116.0* 136.0*
*Not significantly different from B/N—A recovery at p 0.01.
6—92
-------
R.CRA APPENDIX VIII, HPLC COMPOUNDS
The Appendix VIII, HPLC parameters are an assortment of
approximately 50 compounds sharing the properties of low volatility
and/or thermal lability, characteristics which make them
inappropriate for inclusion in CC/MS protocols. This list comprises
a fairly high percentage of therapeutic agents with characteristic
suseptibility to decomposition or rearrangement at extremes of pH.
The 19 compounds reported in Table 2 are those for which acceptable
liquid chromatography has been developed thus far. Using B/N—A
extraction, acceptable recoveries were achieved for only 8 of 19
compounds. The remaining methods each provided significant recovery
increases for 8 or more compounds when compared to B/N—A (Table 4).
In terms of the number of acceptable recoveries the methods rank as
follows:
ph4(16) N—A(14) ) A/N—B(13) > B/N—A(8).
As expected, a number of compounds demonstrated interesting
extraction performances under the 4 procedures. Initial basic
extraction conditions apparently caused N—nitroso—Nethylurea,
l—acetyl--2—thiourea, methomyl and actinomycin D to decompose, while
even mildly acidic conditions elicited the same response from
Mitomycin C. Three compounds, Rotenone, Benomyl and N—nitroso—
N—ethylurea were found to achieve better recoveries under mildly
acidic conditions than at the pH extremes.
CONCLUS IONS
Thus far in this experimentation, the acid/neutral—base extraction
scheme has demonstrated significantly higher recoveries than the
base/neutral—acid procedure for 18 priority pollutants and 10 R RA
Appendix VIII, HPLC compounds, most noteworthy of which are the BHC
isomers and phthalate esters, supporting the previously reported
findings of Slayton and Travato (5). Like A/N—B, the neutral acid
and pH4 methods also achieved recovery Improvements over B/N—A for
many of the same compounds, but also presented a number of cases of
diminished recovery. A/N—B, on the otherhand, was significantly out
performed by B/N—A in only 5 cases, and one of these,
l,2,4—trichlorobenzene, may have been artifactual. The N—A and pH4
extractions yielded better overall recoveries for the RCRA VIII HPLC
compounds alone. However, A/N—B recovery performance for these
compounds was more than ample to make it feasible as a cost
effective approach for an omnibus, GC/MS — LC/MS protocol where
extract splitting would be employed.
EMSL — Cincinnati will continue its research on standard matrix
recoveries of R RA Appendix IX compounds employing continuous
extraction and analysis methods described herein. Upon completion
of the list of 200 compounds in the experimental design, the second
phase of this experiment will Investigate recoveries of the same
compounds from municipal and industrial wastewaters.
6—93
-------
Tabie 1. Average Percent Recoveries of Priority Pollutants Using Four,
Nethylene Chloride, Continuous Extraction Procedures
rue
Value A/N - B B/N - A N - A p 84
Coapound (ug/t) PR SD PR SD PR SD PR SD
Anthracene 10 81.5 5.2 83.3 4.4 72.3 7.0 63.3 5.7
8— BBC 10 97.0 5.9 96.3 5.7 98.5 4.0 81.3 6.3
Benao(a)anthracene 10 86.3 3.8 84.0 1.6 62.0 9.1 60.5 8.7
Benzo(a)pyrene 10 101 7.9 106 7,6 78.5 11.1 61.8 11.5
Benzo(b)fluoranthene 10 101 1.8 100 10.0 12.5 14.5 63.3 18.4
Bis(2 -ethylhexyl)phthalate 10 104 11.4 116 16.7 86.7 13.1 60.2 8.8
Bis(2-chloroethoxy)phtha late 10 94.0 4.8 88.0 3.4 95.5 5.8 83.3 3.8
BisU-chloroethy l)phtha late 10 109 6.7 106 2.9 114 2.2 96.5 4.8
Bis(2-ch loroisopropy l)phtha late 10 87.2 8.5 84.3 5.0 92.5 3.7 76.3 6.9
Butylbentylphthalate 10 112 14.8 52.0 34.7 96.3 8.7 63.0 5.6
Chrysene 10 86.3 3.8 84.0 1.6 62.0 9.1 60.5 8.7
0-5 Nitrobenzene(S) 25 91.0 5 ,0 85.3 1.9 95.3 2.2 74.3 2.7
D-5 Phenol(S) 50 93.3 7.1 79.5 5.8 87.8 3.7 71.5 3.3
D— BSC 10 90.5 3.5 21.0 2.4 92.3 3.4 75.2 3.8
Di—N—buty lphtha late 10 99.5 8.5 67.3 3.9 98.5 8.3 75.3 6.9
Di—N-octy lphtha late 10 120 13.5 125 13.4 83.5 15.1 64.2 9.3
Die ldrin 10 98.2 10.2 98.5 3.9 72.0 10.6 69.3 8.1
Diethylphthalate 10 89.8 5.0 1.5 8.7 92.2 5,4 78.2 5.1
Diaethylphtha late 10 93.2 4.9 0 - 95.3 6.0 76.5 5.8
Endosulfansulfate 10 98.5 8.9 74.3 5.6 84.0 7.8 11.5 8.5
Fluoranthrene 10 96.3 7.4 94.0 4.2 73.? 9.1 62.3 6.8
Fluotene 10 84.8 4.4 77.8 6.3 71.0 4.7 64.3 3.9
ileptachior 10 109 12.6 106 13.1 76.0 16.2 44.0 29.4
Beptachior Spoxide 10 104 12.8 105 5.1 82.0 10.6 64.8 6.6
Rexach lorobenaene 10 82.8 2.1 81.7 5.3 60.3 9.5 49.5 8.9
Hexachiorobutadiene 10 31.5 3.? 27.8 4.6 26.5 3.0 25.3 1.0
Bexacbloroethane 10 38.5 4.9 33.5 1.7 34.5 3.0 40.0 4.5
-BBC (Lindane) 10 93.5 5.7 0 — 92.3 3.8 76.5 5.2
N-nitroso—di-N—propy laaine 10 117 8.3 111 7.7 116 5.7 95.0 7.1
U-nitrosodipheny laaine 10 64.5 11.4 88.5 4.4 95.5 1.7 72.3 2.2
$apbtha lene 10 74.3 2.5 68.0 3.8 74.3 8.2 64.3 1.0
Nitrobenaene 10 99.0 5.4 94.3 5.0 101 1.3 92.3 3.9
Pentachlorophenol 20 116 7.7 114 15.2 136 9.5 84.8 6.9
Phenanthrene 10 84.5 5.8 81.0 6,5 71.5 3.9 60.5 5.3
Pyrene 10 95.0 6.3 92.0 5.0 70.5 8.2 60.3 8.5
Phenol 20 92.5 3.3 82.0 8.1 94.8 2.1 81,5 3.1
Terphenyl(S) 25 88.0 5.8 87.0 2.9 93.8 4.2 70.5 3.0
1,2,4-Trich lorobenzene 10 53.5 1.3 128 11.1 53.8 4.3 47.0 2.4
1,2-Dichlorobenaene 10 59.5 1.9 53.5 2.1 59.8 4.3 55.8 1.7
6—94
-------
Table I
Continued
True
Value A/N - B BIN - A N - A p114
Co pou PS SB PS SB PP. SB
i, —Dicb1oroben ene 10 55.5 3.9 48.5 2.4 53.3 4.3 52.5 3.3
2,4,6- ribroiophen i(S) 50 89.5 7.0 81.3 7.5 92.8 4.9 71.8 5.2
2,4,6-Trichloro heno1 20 89.0 2.2 80.3 11.6 96.0 5.2 70.0 2.6
2,4—Dich1oropher l 20 97.8 4.6 88.0 11.6 103 3.4 83.3 2.6
2,4-Dinitroto lue e 10 89.5 11.8 95.0 108 108 8.0 92.0 2.4
2-Chloronaphthyler .e 10 74.5 5.0 64.0 8.3 71.8 3.9 63.0 3.2
2-Chiorophenol 20 80.2 2.8 70.8 10.2 68.8 4.6 55.5 5.7
2—Cbioropheny l—pberiy lether 10 87.3 2.9 78.3 7.? 91.8 5.1 77.5 4.8
Z-F luorobiphenyi(S1 25 78.3 2.1 12.3 5.5 79.0 5.6 66.5 1.0
—F1uorophenoi S 50 86.3 8.8 73.0 1.4 86.5 5.? 75.8 4.3
M
2—Nitrophenel 20 95.3 5.6 89.8 10.0 104 5.8 86.8 5.1
4,4—ODD 10 105 8.8 106 5.4 79.3 13.8 65.5 8.3
4,4—BBS 10 92.5 5.4 93.5 4.8 70.0 9.1 59.0 6.1
4,4-DDT 10 99.0 11.1 94.3 9.2 70.5 14.3 59.3 7.11
4-Broiopheny1-p eny1ether 10 86.8 3.1 82.3 7.1 71,0 4.2 54.0 5.8
4-Ch1 oro-3- ethyIpheno1 20 101 7.4 92.8 6.8 106 7.4 88.5 1.0
A-B11C 10 89.3 6.8 15.3 10.2 86.3 3.3 70.5 2.6
Acenaphthene 10 81.2 4.6 71.3 8.7 75.3 3.5 63.3 3.5
(S) - Surrogate Standard
6—95
-------
Table 2. sean Percent Recoveries of ECRA Appendix VIII, HPLC Coapounds Using
Four. Nethylene Chloride, Continuous Extraction Procedures
TRUE
VALUE A/N - B B/N - A N - A Ph:4
C0 P0U1 0 tug/h) PR SO PR SD PR SD PR SO
Ethylene Thiourea 40 31.2 87.1 46.2 9.5 20.6 50.5 13.8 72.5
1,2—Diphenylhydrazine 40 55.4 16.2 95.8 8.6 85.6 3.3 86.6 7.4
Rotenone 130 53.8 4.8 59.2 8.1 86.6 5.1 90.2 7.8
N—nitroso-N-ethylurea 30 48.8 13.9 0 — 26.8 14.2 61.4 5.2
Beno.yl 190 67.6 6.5 51.2 18.0 80.2 2.5 85.4 9.6
5-Nitro—o-tu luidine 7 92.6 7.3 102.2 5.1 81.6 4.2 83.2 7.5
N-nitroso—N—.ethylurea 30 72.4 10.2 0 - 21.6 11.1 82.0 12.9
Crotanaldehyde 130 93.8 14.9 44.0 17,1 69.8 4.0 72.0 7.2
ito.ycin C 60 0 - 75.6 5.1 0 - 0
Metho .y l 100 97.8 5.8 0 — 103.9 8.7 99.3 3.2
1,2-Pheny ldia.ine 80 103.1 7.6 102.4 2.1 111.7 7.1 103.5 2.0
1,5-Ilapthalenediaiine 80 51.0 6.2 43.8 9.1 45.2 14.1 21.5 23.1
Actinoiycin B 100 81.7 4.3 0 - 98,9 3.3 88.8 10.1
1—Acety l—2—thiourea 87 86.7 1.6 0 — 91,3 6.0 81.5 4.0
1,3-Phenyldiaiine 103 85.0 11.3 92.9 190 80.2 4,5 79.0 3.2
1-Phenyithiourea 113 96.5 1.9 87.1 9.8 92.5 4.5 84.3 2.4
1—to-chioropheny l)—2-thiourea 110 99.4 3.2 86.2 10.2 93.7 4.9 87.1 3.1
1-Naphthyi—2—thiourea 44 95.6 3.3 82.5 13.5 84.6 3.2 81.1 2.6
Diethyistilbestrol 127 75.4 1.4 93.9 8.5 102.6 4.9 95.3 3.5
6—96
-------
Table 3. Priority pollutants presenting significant recovery differences;
a eoaparison of CLP extraction iethod with acid/neutral-base,
neutral-acid and p8:4 procedures.
Average Percent Eecovery
B/N - A A/N - B
8-Nitrosodiphenylaise 88.5 64.5 -24.0
Phenol 82.0 92.5 10.5
1,2,4-Trichiorobeozene 128.0 53.5 —14.5
1,2-Dich lorobenzene 53.5 59.5 6,0
1,34 lichlorobenzene 48.5 55.5 7.0
Z,6,6—Tribroiopheno l(s l 81.3 89.5 8,2
2 ,4,6—Trichlorophenol 80.3 89.0 8.?
Z,4—Dichioropbeno l 88.0 97.8 9.8
Z-Chloronapbthalene 64.0 74.8 10.8
2—F luoropheno l(sl 73.0 86.3 13.3
Buty lbenzylphthalate 52.0 112.0 60.0
0-5 Phenoitsl 79.5 93.2 13.8
d-BHC 21,0 90.5 69.5
Di-N-butyipbtbalate 67,3 99.5 32.2
Dietbylphthaiate 7.5 89.8 82.3
Diiethylphtha late 0.0 93.2 93.2
Endosulfansu lfate 74.3 98.5 24.2
X-BIIC(Lindane) 0.0 93.5 93.5
15.3 89.2 74.0
Acenaphthene 71.3 81.2 9.9
Average Percent Recovery
COMPOUND 81W - A N - A
Anthracene 83.3 72.3 -11.0
Benzo(a)anthracene 84.0 62.0 -22.0
8en o(a)pyrene 106.0 78.5 -27.5
Ben!o(b)f luoranthene 100.0 72.5 -27.5
Bis(2-ethylhexyl)phthalate 116.0 86.? -29.3
Chrysene 84.0 62.0 -12.0
0—5 Nitrobenzene(s) 85.3 95.3 10.0
d-BHC 21.0 92,3 71.3
Di—W—butylphthalate 67.3 98.5 31.2
Di-W-octylphtha late 125.0 83.5 -41.5
Die ldrin 98.5 72.0 —26.5
Diethyiphthalate 7.5 92.2 84.7
Diiethylphtha late 0 .0 95.3 95.3
Endosulfansuifate 74.3 84.0 9.7
Fluoranthrene 94.0 ‘ 13.7 -20.3
Reptachlor 106.0 76.0 -20.0
Reptachior epoxide 105.0 82.0 -23.0
llexachiorobenzene 81.1 60.3 -21.4
%-BHC Liodare) 0.0 92.3 92.3
Pyrene 92.0 70.5 —21.5
Phenol 82.0 94.8 32.8
6—97
-------
Table 2 Co tir ued
Page 2
Average Percent Recovery
COMPOUND B/N - A N - A
1,1-Trich1oroben eoe 128.0 53.8 -74.2
Z,4,6-Trichloropheno l 80.3 96.0 15.7
2,4-Dichiorophenol 88.0 103.0 15.0
Z-Ch lcropheny l-phenyletber 78.3 91.8 13.5
2-P uoropheno1 73.0 86.5 13.5
2-Nitrophenol 89.8 104.0 14.2
4,4—DOD 106.0 79.3 —26.7
4,4—DDE 93.5 70.0 —23.5
4,4-DDT 94.3 70.5 -23.8
4-Broiophenyl-pheny lether 82.3 71.0 -11.3
4-ch loro-3— .ethylpheao l 92.8 106.0 13.2
a-EEC 15.3 86.3 71.0
Average Percent Recovery
COMPOUNL 8/N - A p8 : 4
Anthracene 83.3 63.3 -20.0
b-DEC 96.3 81.3 —15.0
Ben o(a)anthracene 84.0 60.5 -13.5
Be n o(a pyrene 106.0 61.8 -44.2
Benzo(b)fluoranthrene 100.0 63.3 -36.7
Bis 2—ethy1hexy1}phtha1ate 116.0 60.2 -55.8
Chrysene 84.0 60.5 -23.5
0-5 $itrobenzene(s} 85.3 74.3 -21,0
d-BHC 21.0 75.2 54.2
Di-N-octylphtba late 125.0 64.2 -60.8
Dieldrin 98.5 69.3 -29.2
Diethylphtha late 7.5 78.2 70.7
Diaethy lphtha late 0.0 76.5 76.5
Pluorene 77.8 64.3 -13.5
Deptachior 106.0 44.0 —62.0
Heptachior epozide 105.0 64.8 -40.2
Herach lorobenzene 81.7 49.5 -32.3
1-EllClLindane) 0.0 76.5 76.5
N-Nitrosodipropyiaiine 111.0 95.0 -16.0
N-Nitrosodiphenylaiine 88.5 72.3 -16.2
Pentachiorop eno1 114.0 84.8 -29.2
Phenanthrene 81.0 60.5 —20.5
Pyrene 92.0 60.3 -31,7
Terphenyl(s) 87.0 70,5 16.5
1,Z,4-Trich lorobenzene 128.0 47.0 87.0
2—Ch lorophenol 70.8 55.5 15.3
2-Methy1-4,6-dinitrophenol 109.0 81.0 28.0
4,4-DOD 106.0 65.5 40.5
4,4—DDE 93.5 59.0 34.5
4,4- 00 ? 94.3 59.3 35.0
4-Bro.ophenyl-phenyiether 82.3 54.1
15.3 70.5
6—98
-------
Table 4. RCRA Appendix VIII, BPLC compounds presenting significant recovery
differences; a comparisor of CLP extraction method with acid/neutral —
base, neutral-acid and p114 procedures.
Average Percent Recovery
Compound Bi - A A/k - B
1,2-Diphenyihydrazine 95.8 55,4 -40.4
N-kitroso-}4-etby lurea 0,0 48.8 48.8
N - kitroso- k-methy lurea 0.0 72.4 72.4
Benomyl 51.2 67.6 16.4
Crotonaidehyde 44.0 93.8 49.8
1-Acety l-2-thiourea 0.0 86.7 86.7
1-Phenyithiourea 87.1 96.5 9.4
1(o —Cblorophenyll-Z-thiourea 86.2 99.4 13.2
1- apbthy1-Z-thiourea 82.5 95.6 13.1
Diethylstibestrol 93.9 75.4 —18.5
Mitomycin C 75.6 0.0 -75.6
Nethomyl 0.0 97.8 97.9
Actinomycin D 0.0 81.7 17.7
Aft?age Perctnt geeo,etj
Compound - A N • A
8otenone 53.2 86.6 27.4
N-kitroso-k-ethylurea 0.0 26.8 26.8
k-Mitrogo-k-methy lurea 0.0 21.6 21.6
Benomyl 51.2 80.2 29.0
5-kitro—o-toluidine 102.0 81.6 -20.4
Crotonaldehyde 44.0 69.8 25.8
1-Acetyi-Z-thjourea 0.0 91.3 91.3
Mitoiycin C 75.6 0.0 -75.6
ethoiyi 0.0 104.0 204.0
Actinumycin 0 0.0 98.9 98.9
Average Percent kecovery
Compound 81k - A pH 4
Ethylene tbiourea 46.2 13.8 -32.4
Ro tenone 59.2 90.2 31.0
k-kitroso-N—etby lurea 0.0 61.4 61.4
k—kitroso-N-aethy lurea 0.0 82.0 82.0
Benoayl 51.2 85.4 34.2
t1—Nitro-o—tojujd ne 102.0 83.2 -18.8
Crotonaidebyde 44.0 12.0 28.0
1 —Acetyl-2 -tbiourea 0.0 81,5 81.5
Mito yci C 75.6 0.0 -75.6
ethomyl 0.0 99,3 99,3
1 ,5-Waphtha lenediamine 43.8 21.5 -22.3
Actinomycin 0 0.0 86.8 88.8
6—99
-------
Table 5. Percent recoveries of phenols by B/K - A, 1111 - B
and N - A extraction schenes.
ONPOJJND B/N-A A/N-B N-A
D—5 Phenol 79.5 93.3 87.8*
phenol 82.0 92.5 94.8
4-Ch loro—3-.ethglphenol 92.8 101.0* 106.0
2-Chlorophenol 70.! 80.2* 68.8*
2 ,4,6 -Tribroaopheno l 81.3 89.5 92.8
2,1—dichiorophenol $8.0 91.8 103.0
2-Nitrophenol 89.8 95.3* 104.0
2,4,6-Trich lorophenol 80.3 89.0 96.0
2—Fluorophenol 73.0 86.3 86,5
Pentachiorophenol 114.0* 116.0* 136.0*
* Not significantly different frou B/N — A recovery at p 0.01.
6—100
-------
REFERENCES
“Guidelines Establishing Test Procedures For The Analysis Of
Pollutants; Proposed Regulations”, 40 CFR, Part 136, Federal
Register , December 3, 1979, Vol. 44, 69464.
“Guidelines Establishing Test Procedures For The Analysis of
Pollutants Under The Clean Water Act; Final Rule and Interim
Final Rule and Proposed Rule”, 40 CFR, Part 136, Federal
Register , October 26, 1984, Vol. 49, No. 209, Part VIII.
Longbottoin, J.E. and Lichtenberg, J.J., Eds., Methods of Organic
Chemical Analysis of Municipal and Industrial Wastewater ,
EPA — 600/4—82—057, July 1982.
Elchelberger, J.W., Kerns, E.H., and Budde, W.L., Analytical
Chemistry , Vol. 55, No. 9, August 1983, pp 1471—1479.
Slayton, J.L. and Trovato, E.R., “Review of Acid/Neutral Continuous
Liquid/Liquid Extraction of Priority Pollutants and Hazardous
Substance List Compounds,” Proceedings , Rocky Mountain
Conference, Denver, Colorado, August 6, 1986.
Lucas, S.V. and Kornfeld, R.A., “GC—MS Suitability Testing of RcR.A
Appendix VIII and Michigan List Analytes,” Final Report ,
Battelle, Columbus, Ohio, February 20, 1987.
Method 8270 — Gas Chromatography/Mass Spectrometry For Semivolatile
Organics: Capillary Column Technique, Test Methods For
Evaluation Solid Waste, Vol. 1B, U.S. Environmental Protection
Agency, Office of Solid Waste and Emergency Response, Washington,
D.C., November 1986.
6—101
-------
PRELIMINARY EVALUATION OF TEST METHOD FOR VOLATILE ORG NICS
IN HAZARDOUS WASTE: BATCH STEAM STHIPPING DISTILLATION
A.R. Gholson, S.B. Balik, R.K.M. Jayanty, Research Triangle Institute,
Research Triangle Park, NC; G.D. McAlister, R.T. Harrison, U.S.
Environmental Protection Agency, Research Triangle Park, NC
ABSTRACT
A test method for determining the volatile emission potential of
hazardous waste using batch steam stripping/distillation has been
evaluated for predicting whether pretreatment of a waste is needed to
remove volatile organics compounds prior to disposal. The procedure
provides an estimate of the minimum amount of volatile organics that
can be recovered in a full—scale steam stripping system and some
indication of the potential emissions from the original waste and its
residue.
The test method was evaluated by preparing synthetic wastes which
approximated the chemical and physical properties of six different
waste categories. The six waste types included mixed phase aqueous,
aqueous waste plus sludge, dilute aqueous waste, organic waste plus
sludge, dry solvent waste, and organic solvent waste. Each type of
waste contained known amounts of methylene chloride, 2—butanone,
1—butanol, pyridine, toluene, phenol, and naphthalene. These seven
compounds were chosen to represent a wide range of volatilities,
polarities, and reactivities of compounds that may be contained in
actual wastes. In addition, three real wastes were tested; two
contained chlorinated organics in water and the third contained
hydrocarbons in a refinery sludge matrix. All distillations were
performed in basic and acidic conditions to see if pH effected
recovery. The steam distillations of the wastes were performed by
collecting 40 percent of the total volume of the waste in several
liquid fractions. Volatile compounds that did not condense were
collected on a cold trap or in a bag. The various fractions were
analyzed by gas chromatography (GC) and the amount of each component
was determined. The total atmDunt of a compound in all the fractions
was related to its respective total percent recovered.
The recoveries from the waste ranged from 70 to 140 percent for most
compounds. This range of recovery compared favorably with recoveries
by other methods. There was, however, some variation of recovery in
the different waste matrices. The recovery of semi—volatile compounds
such as phenol was improved by extending the distillation. The test
method was duplicated for each waste and the results were found to be
reproducible. GC injections were also duplicated and variance was
found to be less than 20 percent. The manner in which the synthetic
waste were prepared, screening of the wastes, apparatus involved,
sample analysis, and data reductions will be presented.
6—103
-------
PREPARATION OF RADIOACTIVE “MIXED” WASTE SAMPLES FOR MEASUREMENT
OF RCRA ORGANIC COMPOUNDS
Bruce Tomkins, John Caton, Organic Chemistry Section, Analytical
Chemistry Division, Oak Ridge National Laboratory, Oak Ridge,
Tennessee
ABSTRACT
A radioactive “mixed” waste typically contains alpha, beta—, and/or
gamma—emitting radionuclides and varying quantities of semivolatile
and/or volatile organics, some or all of which may be named
specifically by the Resource Conservation and Recovery Act (RCRA).
Because there are no acceptable procedures currently available for the
disposal of “mixed” wastes, they are presently stored above—ground at
great cost to the user. For this reason, analytical procedures which
can identify the presence, or at least confirm the absence, of RCRA
organics in radioactive waste are necessary for deciding the proper
approaches for their disposal. An important aspect of this is the
development of methods for preparing mixed waste samples which allow
the R RA organics in radioactive waste are necessary for deciding the
proper approaches for their disposal. An important aspect of this is
the development of methods for preparing mixed waste samples which
allow the RCRA organics to be measured in conventional organic
analytical laboratories.
Our general approach to characterizing organic compounds in “mixed”
wastes is two—fold: First, organic species are removed from the
sample in such a way that transfer of the active radionuclides is
minimized. Analytical equipment and procedures listed in EPA—approved
methods are used wherever possible. However, methods such as
solid—phase extraction of semivolatiles, which minimize solvent
volumes and operator exposure, have been investigated as reasonable
alternatives. Secondly, the final determinations are performed on the
nonradioactive concentrated sample in conventional laboratories using
EPA analytical methods.
Decontamination factors (the ratio of the initial to the final 1
radioactivities of the samples) ranging between 1,000 and 10,000 are
achieved for water samples contaminated with Co—60, Cs—137, or Sr—90,
using either simple continuous liquid—liquid extraction with pentane
or solid—phase extraction employing octadecyl reversed—phase columns.
At the same time, recoveries of semivolatile organic compounds, such
as polycyclic aromatic hydrocarbons and neutral pesticides, is
quantitative at ppm or ppb concentrations. Soxhlet extraction of
radioactive sludges (mixed with anhydrous sodium sulfate to remove
excess water) with pentane or methylene chloride achieves similar
‘The submitted manuscript has been authored by a contractor of the
U.S. Government under contract No. DE—ACO5—840R21400. Accordingly,
the U.S. Government retains a nonexciusive, royalty—free license to
publish or reproduce the published form of this contribution, or allow
others to do so, for the U.S. Government purposes.
6—105
-------
decontamination factors while extracting organics in an EPA—approved
fashion.
Our initial experiences with volatile organic species indicate that a
nExlified, inexpensive purging vessel coupled to a Tenax—GC sorbent
trap is capable of achieving both substantial decontamination factors
and the quantitative recovery of target compounds. This methodology
has already been tested on supernatants of “mixed” waste sludges.
Both recovery data and decontamination factors will be presented.
INT XXJCTI(
A radioactive “mixed” waste typically contains alpha—, beta—, or
gaimna—emitting radionuclides and varying quantities of semivolatile or
volatile organic species, some or all of which may be named
specifically by the Resource Conservation and Recovery Act (RCRPt).
Because there are no acceptable means available currently for
disposing of these mixed wastes, they are presently stored
above—ground in sealed drums. For this reason, analytical procedures
which can determine RCRA organics in radioactive waste are necessary
for deciding the proper approach for disposal. n important goal of
this work is the develorm nt of methods for preparing mixed waste
samples in a manner which allows the RCRF organics to be measured in
conventional organic analysis laboratories without special
precautions.
nalytica.l procedures developed for handling mixed waste samples must
satisfy not only the usual constraints present in any trace—level
organic chemical determination, but also those needed to insure the
protection of the operator from radioactive contamination.
Consequently, procedures should be designed to use the least amount of
radioactive sample comuensurate with achieving acceptable sensitivity
with the RCRA analytical methods. Furthermore, the usual laboratory
glassware which would normally be used should be replaced with
disposable materials wherever possible, in order to reduce the
“clean—up” time required, and thereby reduce the operator’ s exposure
to radioactivity. Actual sample handling should be reduced to the
absolute minimum. Finally, the final isolate must exhibit a
surfficiently low level of alpha, beta, or galmna activity to permit
detailed characterization in a conventional organic analysis
laboratory.
Clearly, not all traditional or EPA—approved sample preparation
procedures will prove satisfactory for mixed waste samples, given
these additional restrictions. This paper describes our experiences
in analyzing mixed waste aqueous and solid samples using a variety of
conventional and supplemental procedures. In general, it is entirely
feasible to prepare isolates from radioactive samples which are
enriched in either semivolatile or volatile species, yet which are
essentially decontaminated, and are therefore suitable for analysis in
conventional organic analysis laboratories.
6—106
-------
EXPERIMENTAL
Samples
The radioactive sludge materials tested were provided by the
Operations Division of the Oak Ridge National Laboratory, Oak Ridge,
N.
Reagents, Solvents , and Radioactive Tracers
All solvents used in this work were purchased in “Distilled in Glass”
purity from American Burdick & Jackson (Muskegon, MI), and were used
as received.
Standard mixtures of the EPA priority polycyclic aromatic hydrocarbons
(PAH), phenols, and organochlorine pesticides were purchased from
Supelco, Inc. (Bellefonte, PA).
Acidic solutions with a known activity of the radionuclides Co—60 and
Cs—137 were supplied by the Radiochamical and Activation Analysis
Group, Analytical Chemistry Division, Oak Ridge National Laboratory,
Oak Ridge, I lL
The pH 4.9 leaching solution used is that described in EPA Method
1310, Extraction Procedure (EP) Toxicity Test Method and Structural
Integrity Test (1). 64.3 mL of 1 N sodium hydroxide and 4.7 niL of
glacial acetic acid were made to exactly 1 L with distilled water.
Equipment
Liquid—liquid extraction of aqueous samples with solvents of lesser
specific gravity than water (e.g., peritane) were performed using a
glass apparatus similar, but not identical, to Kontes Part No.
K—584000 Continuous Extraction Apparatus (Vineland, NJ).
Liquid—liquid extraction of aqueous samples with solvents of greater
specific gravity than water (e.g., methylene chloride, were performed
using Kontes Part No. K—584100, Continuous Extraction Apparatus.
All high—pressure liquid chromatographic separation and measurements
were performed with a Beckman Model 334 Gradient Liquid Chromatograph
(Berkeley, GA). This instrument was equipped with UV detector (254
nni), a Spectra/Gb Filter Fluorometer (excitation filter, 280 nm;
emission filter, “blue from fluorescaxnine”) purchased from Gibson
Medical Electronics (Middleton, WI), an HP 3390 integrator, and a
stripchart recorder. The analytical column was protected by a 0.5 u
porosity high—pressure inline filter (Scientific Systems, Inc., State
College, PA, part nos. 05—0149 and 05—1055) and a guard column packed
with Perisorb RP—18 (Upchurch Scientific, Inc., Oak Harbor, WA)
connected in series. A Vydac No. 201 TP 5415 octadecyl column (The
Sep/ /Ra/rions Group, Hesperia, CA), 15 cm x 4.6 mm o.d., 5 u silica,
was used for both PAH and phenol separations. The column was
maintained at 30 degrees C using a Model 7931 column heater purchased
from Jones Chromatography (Littleton, CO). The methanol and water
reservoirs, as well as the detectors, operated at room temperature. A
6—107
-------
10 uL sample volume was introduced onto the column during each
determination.
The solid phase extraction columns, manifold, pump, and pressure
controllers were purchased from 3. T. Baker, Inc. (Phillipsburg, NJ).
Two manifolds and pressure controllers were used, such that the same
one was always used for collecting organic analytes from radioactive
aqueous samples, while the other was always used for eluting these
analytes from loaded columns which had been essentially washed free of
radioactivity.
The organochiorine pesticides were separated and quantitated according
to EPA Method 608 using a Perkin Elmer Sigma 300 Capillary Gas
Chromatograph (Norwalk, CT) equipped with a Model AS—300 autosaxnpler
(Perkin Eln r) and an SE—54 fused silica capillary column (0.25 u film
thickness, 0.25 mu I.D., 30 m) purchased from Supelco, Inc.
(Bellefonte, PA). The 0 injector and electron capture detector
temperat ,res were 0 both 30(9 C. The column teu erature ias programmed
from 140 C to 180 C at 4 C/mm, and from 180 C to 250 C at 2 C/mm.
The oven temperature was held at 250°C for 5 mm before returning to
the starting temperature. A 3 uL injection was employed. The carrier
gas was 90/10 (v/v) argon,’ uethane, flowing at 5 uL,4nin. Data were
collected, displayed, and analyzed using a Model 3000 Chromatography
Data System (Nelson Analytical, Inc., Cupertino, CPk) and an IBM PC/XT
personal computer.
The collection of volatiles was performed using equipment custom—made
at the Oak Ridge National Laboratory. A special Teflon sampling head
equipped with a Teflon—faced silicon rubber septum screwed snugly onto
a 40 mL EPA VOPk sampling vial. The head provided a 10/32 screw port
for a reusable “Fingertight” fitting, manufactured by Upchurch
Scientific (Oak Harbor, W! ). A length of capillary Teflon (1/16” o.d.
x 0.3 ian id.) passed through the fitting into the VOL vial. The
other end of the Teflon tubing was attached to a nitrogen cylinder
with a flow controller. The head provided an additional port for a
l/8”—to--1/4” Swagelok reducing union; the 1/8” side was screwed into
the collection head. The other end of the Swagelok union was
connected to a 25 cm x 4.6 mm o.d. stainless steel column dry—packed
with Tenax CC 35/60 mesh, which was purchased from Ailteck Associates
(Deerfield, IL). A simple calibrated rotameter was connected to the
free end of the Tenax trap with a piece of rubber tubing.
The volatiles were analyzed using a procedure based on EPA Method
8240; however, modifications were introduced to allow the volatile
analytes to be desorbed directly from a stainless steel trap packed
with Tenax CC. A Model LSC—2 Tekmar Liquid Sample Concentrator
(Tekmar Co., Cincinnati, OH) was used to desorb the organics of
interest. The existing 1/8” o.d. trapping column and corresponding
column heater in the oven were both removed and replaced with the 1/4”
o.d. sample trap and corresponding heater. Exactly 250 ng each of
brcmochloromethane, 1—chloro—2-bromo— propane, and 1,4-dichiorobutane
from a Purgeables Internal Standard Mix—624 (Supelco, Inc.,
Bellefonte, PA) was sparged onto the Tenax sample trap using an 11 mm
wet purge and a 4 mm dry purge. The internal standards and analytes
6—108
-------
present on the trap were desorbed at 180°C for 3 mm using helium
carrier gas. The components were swept onto the head of a 6.6’ x 1/4!!
o.d. glass column packed with Carbopack B coated with a 1% SP—l000
(Supelco) located in a Hewlett—Packard 5995 GC/11S. The oven
temperature was programmed rom 45°C (hold for 3 mm) to 180°C (hold
for 45 mm) at a rate of 8 C/mm. The analytes were detected by the
mass spectrometer, which operated at 70 eV ionization potential, mass
range of 35—260 amu, and a scan time of 0.24 scan/sec.
All measurements of ganirna activity were performed using a proportional
scintillation counter. Two sodium iodide crystal well counters
heavily shielded with lead were coupled to an EG & G Ortec (Oak Ridge,
TN) system consisting of two Model 776 Counter/Timer boards, two Model
478 0—2 key bias supply boards, and a Model 779 Interface/Controller
board. The counter exhibits a typical efficiency of Ca. 35% for
Cs—137.
All measurements of beta activity were performed using a beta RIDL
Proportional Counter consisting of a Model 40—9B voltage module
(operating at 2300 V) 1 a Model 30—19 sensitivity module, and a Model
49—25 timing and readout module. The lead sample chamber was
continuously sparged with 10% methane/90% argon during each
measurement. The counter typically exhibits 8% counting efficiency
for Sr—90.
Procedures
Soxhlet Extraction of Sludges . Approximately 10 g sludge were mixed
with 20 g anhydThus sodium sulfate, then transferred to a
pre—extracted cellulose Soxhlet thimble. The sludge was then
extracted overnight with methylene chloride or pentane. The extract
was concentrated to exactly 1 mL; portions were taken for gross beta
and gross gamma counting (see below).
Leaching of Sludges . Approximately 10 g of sludge and 200 mL of the
pH 4.9 iE tic acid/sodium acetate solution were stirred briskly
overnight using an overhead electric stirrer.
Extraction of p 4.9 Leachate . Exactly 100 mL of the pH 4.9 leachate
described aB ve were extracted continuously with pentane or methylene
chloride overnight. The pentane extract was then concentrated using a
Kuderna—Danish concentrator to exactly 1 mL. Aliquots were taken for
gross beta and gross gamma counting (see below).
Solid Phase Extraction of EPA Priority Pollutant PAIl from 4.9 Buffer .
Three Baker—lO SPE 3 mL OCT DECYL columns were conditioned with
methanol followed by 15% (v/v) isopropanol in water. An 18 mL portion
of isopropanol was added to 100 mL water which had been spiked at
40—400 ppb with each PAIl and 47,000 cpm (Ca. 2200 Bq) Co—60. The
resulting solution was passed through the OCT1 DECYL column using the
“radioactive” manifold. The column was rinsed with a few mL of 15%
(v/v) isopropanol in water. The PAH were eluted with exactly 1 mL of
methylene chloride using the “clean” manifold (2). The final organic
fraction was screened for gross gamma activity before performing the
6—109
-------
final quantitation by HPLC.
Solid Phase Extraction of EPA Organochlorine Pesticides from Water .
Three Baker—lO SPE 3 mL or inL OCTN)ECYL columns were conditioned
with methanol and water adjusted to pH 2 with HC1. A 100 niL volume of
water was spiked with 120 ug of each priority pollutant phenol and 5.5
E+6 cpa (ca. 2.6 E+5 Bq) Cs—137. The solution was adjusted to pH 2
with HC1, and the phenols were “salted out” by adding 25 g of NaC1.
The test solution was then passed through the column, which was later
rinsed with 0.01 N HC1 and permitted to air—dry briefly using the
“radioactive” manifold. The phenols were eluted with 5 mL of methanol
using the “clean t ’ manifold (4). The final organic fraction was
screened for gross gamma activity before performing the final
quantification by HPLC.
HPLC Separation and Quantitation of EPA Priority Pollutant P2\H. PI H
were eluted fro ithe HPLC coluiiiii ii ing a gradient whic1T hanged
linearly from 70% methanol/ ’ater (hold for 10 mm) to 99%
methanol/wrater (hold for 10 mm) over 15 mm. Quantitation was
performed using the integrator, which monitored the flurometer signal.
HPLC Separation and Quantitation of EPA Priority Pollutant Phenols .
The heno1s were eluted using gradient program which changed
linearly from 30% methanol/1% (v/V) acetic acid in water to 95%
methanol/l% (v/v) acetic acid in water over 20 mm. Quantitation was
performed using the integrator, which monitored the UV detector signal
(254 run).
Collection of EPA Priority Volatiles from Water . Three 5 niL aqueous
solutions con€ Thing 50 ppb each of the EPA volatiles and the
purgeables surrogate standard were prepared in 40 niL EPA VO1 vials.
The Teflon sampling head was screwed onto the top of each vial, and a
Tenax GC stainless trap was connected to the reducing union. The
solutions were sparged with nitrogen at 90 mL/min for 15 mm. The
trap was then sealed and analyzed for volatiles, as described above.
In separate experiments, 5 niL of unspiked water served as the blank,
and 5 niL of water spiked with E+5 Eq Co—60 served as the test for
transfer of radiation. In the latter, a Tenax column was unpacked,
and the front end of the column was removed and tested for gross gamma
emission.
Determination of Gross Gamma Radioactivity . A known volume (usually 1
mL) of organiãisolate was placed into a 10 x 75 mm glass test tube,
which was then stoppered and placed into the well counter. Gross
gamma activity was typically counted to Bq, but only assuming that the
gamma emitter had the same efficiency as Cs—137 (standard available).
Determination of Gross Beta Radioactivity . Exactly 200 uL of organic
isolate was placed onto a 25 nun o.d. watch glass, and the solvent was
allowed to evaporate. The watch glass was then placed in a cardboard
mount featuring a mylar film window. The mount was placed into the
lead sample chamber, and gross beta activity was measured for 10 mm.
The activity is reported in Eq.
6—110
-------
RESULTS 1 ND DISCUSSION
The barrel sludge and filter cake samples studied, which were obtained
from a process water treatment clarifier, exhibited a nominal beta
activity of 20,000 Bq,/g (mostly as Sr—90). Historically, similar
samples exhibit not more than twice that value. The remaining solid
is an impoundment pond sediment in which Sr—90 is also the principal
radionuclide, but at a nominal activity of 2000 Bq/g.
The usual conditions employed for leaching a solid as described in EPA
Method 1310 (1) were deemed unsuitable for these radioactive solids.
The practice of tumbling 100 g of solid and 2 L of leaching solution
for at least 18 hours in a glass jar presented a significant risk of
seepage or radioactive spill. Thus, a substitution was made for
Method 1310 in which (a) both the mass of sample and volume of liquid
were scaled down by a factor of ten, and (b) brisk stirring of the
sample and leach solution using an overhead electric laboratory
stirrer was substituted for the tumbler. In this manner, the spirit
of Method 1310 — intimate contact of the solid and leach solution —
was maintained, while minimizing the likelihood of a radioactive
spill.
The subsequent liquid—liquid extraction of the radioactive leachate
also was modified. EPA Method 3510 (1), Separatory Funnel
Liquid—Liquid Extraction, similarly was deemed unsuitable for
radioactive liquids. EPA Method 3520 (1), Continuous Liquid—Liquid
Extraction, was substituted. In this procedure, the radioactive
aqueous sample remains essentially undisturbed in a glass vessel,
while the organic extraction solvent moves slowly through it.
Continuous liquid—liquid extractors are available for solvents which
have densities greater than water (e.g., methylene chloride) and less
than water (e.g., pentane). In practice, the use of pentane yielded
recoveries of test organic compounds comparable to those obtained with
methylene chloride. Hence, a subsequent filtration step could be
avoided. While many potential RCRA organic compounds are indeed less
soluble in pentane than in methylene chloride, this deficiency may be
readily overcome by simply extracting for a longer period of time than
that used for methylene chloride. Furthermore, the RCRA compounds are
frequently present at trace or ultratrace levels — concentrations
where the analyte would be reasonably soluble in pentane.
Table 1 follows the reduction in beta activity (due primarily to
Sr—90) during the leaching of three test radioactive solids and the
subsequent acid/base continuous extractions with pentane. Leaching
with pH 4.9 sodium acetate/acetic acid solution reduced the activity
by an order of magnitude. The acid/base extractions of the leachate
using either methylene chloride or pentane yielded a further reduction
to near—background levels, thereby suitable for analysis in a
conventional organic laboratory. Thus, traditional extraction
procedures are capable of providing a concentrated organic extract
containing the extractable organics without carrying over a
significant quantity of beta activity from Sr—90, an alkaline—earth
radionuclide. A cursory HPLC examination of these concentrates
revealed no UV—absorbing organic compounds present in these samples at
6—111
-------
part—per—billion levels.
Sludge samples similar to those described above are also readily
decontaminated when Soxhiet extracted with organic solvents such as
methylene chloride. In a typical experiment, in which the final
organic extract is concentrated to 1 mL, the total radioactivity
present is <0.5 liqAnL gross beta and <1 Bqj nL gross ganm a (referenced
to Cs—137) radiation over background — an activity level entirely
compatible with conventional organic analysis laboratories.
The potential advantages of solid phase extraction (SPE) warranted
careful evaluation of the technique applied to simulated aqueous
samples near—quantitative recovery (>95%) of the EPA priority
polycyclic aromatic hydrocarbons and the EPA organochiorine pesticides
(both representing neutral compounds) at simulated environmental
levels (40—400 ppb and 5 ppb, respectively) while simultaneously
reducing the gama activity present from a starting level of 20,000 or
40,000 cpu (Ca. 1,000 or 2,000 Bq) to background or near—background
levels (within instrumental uncertainty). The decontamination factors
were Ca. 2,000 for these samples. Note that in SPE, the sample
preparation is complete in less than a half hour, more than one sample
(up to ten) may be processed simultaneously, there is minimal cleanup
required (columns are disposable), and the operator’s exposure to the
radioactive sample is minimized. It is assumed (but not confirmed)
that most water—soluble radionuclides would behave similarly to Cs—137
and Co—60, and can be removed almost completely from the final organic
extract.
Solid phase extraction did not prove as successful in recovering the
EPA priority pollutant phenols from a test mixture, even though the
g rm activity ‘was reduced from a starting value of 5.5 E-s-6 cpu (ca.
2.6 E÷5 Bq) to background levels (within instrumental uncertainty). A
variety of columns, including OCTADECYL (two sizes), OCTYL, and
PHE YL, were tested; however, the OCTADECYL columns yielded the best
recoveries, as shown in Table 4. Clearly, the columns tested did not
recover either phenol or 4—nitrophenol quantitatively at the test
level of Ca. i ppn, although they did recover most of the other
phenols at or somewhat below the 95% level. These data indicate that
while SPE certainly does have advantages in the analysis of phenols in
mixed waste aqueous samples, the OCTADECYL columns should not be
considered optimal. Columns packed with resins or porous polymers
might give better recoveries of phenols; however, it must be
demonstrated that these columns will also yield a final extract with a
near—background level of radiation.
The determination of volatile constituents in radioactive aqueous
samples was simplified greatly using the custom—made sampling head and
V vial containers. The recovery of many EPA priority volatile
materials using the sampling head and a dry—packed Tenax GC trap was
quite reasonable at the 50 ppb level, as shown in Table 5, while
maintaining a very low level of activity in the Tenax trap itself.
During one of the trials using only Co—60 tracer as the “volatile”
analyte, the Tenax from the front 3—5 cm of the trap was removed after
sparging and subjecting to gross ganuna counting. No gama activity
6—112
-------
was detected. Reducing the sparge gas flow rate from 90 mL/min to
25—40 mL/rnin and the sparge time from 15 mm to 11 mm would have
probably led to an improved recovery of volatiles without adding
additional activity to the trap. Furthermore, if the liquid sample is
sparged gently , the sampling head itself should remain noncontaminated
despite repeated use. The only part of the sampler which would become
contaminated, therefore, would be the capillary Teflon tubing which is
in contact with the sample. Even here, contamination is minimized
because the Teflon tubing is not wetted with water, and should absorb
very little radioactive solution. Hence, it is quite feasible to
sparge volatiles from a radioactive aqueous sample onto a Tenax trap,
and take the trap into a conventional organic analysis laboratory for
detailed characterization because the trap is itself noncontaminated.
In general, then, our preliminary experiences in preparing low—level
radioactive samples for analysis suggest that concentrates not
contaminated by the ganhrna— or beta—emitting alkaline or alkaline earth
metals may be prepared readily without expensive and exotic equipment.
In addition, several of the major pollutant compound classes, such as
the PAR, organochlorine pesticides, and phenols, may be readily
concentrated using simple solid phase sorbent technology.
Our experiences have not addressed the effect of these sample
preparation procedures on several important classes of radionuclides.
These include common low—energy beta emitters such as tritium,
volatile beta/gamma emitters such as iodine—l31, and transuranic
radionuclides such as americium—24l. Furthermore, samples
contaminated with alpha emitters such as Pu—239 require special
handling techniques, such as glove boxes and inert atmospheres,
regardless of the activity level present. Nevertheless, the
procedures described here should be considered a reasonable starting
point for preparing suitable isolates of RCRA organic compounds.
These procedures may also be entirely proper for preparing such
isolates from alpha—contaminated samples provided that all
manipulations are performed in a contained environment.
ACKN LEDGMENTS
The authors gratefully acknowledge the assistance of the Inorganic and
Physical Analysis Group, Analytical Chemistry Division, Oak Ridge
National Laboratory for many helpful discussions concerning the
handling of radioactive materials, and Mr. William F. Fox for his many
comments concerning radiation safety. Mr. Norman A. Teasley supplied
radioactive tracer solutions of known activity. Finally, Ms. Cheryl
A. Treese is acknowledged for performing the detailed characterization
of ultratracelevel organochlorine pesticides and organic volatiles
described in this work.
REFERENCES
1. Test Methods for Evaluating Solid Waste, Physical/Chemical
Methods, Sw—846, Second Edition, Revised , Washington, D.C., U.S.
Environmental Protection Agency, April, 1984.
6—113
-------
2. “Rapid Extraction of P.P H’s from Water with ‘BAKER’—lO SPE Column”,
in ‘ Baker’—lO SPA Applications Guide, Vol . II, Phillipsburg, NJ, J.T.
Baker Chemical Co., 1984, P. 126.
3. “Rapid Extraction of Organochiorine Pesticides from Potable Water
with ‘BAKER’—lO SPE Disposable Column OCTYL (C8), 6mL”, in ‘ Baker’—lO
SPE Applications Guide, Vol . I, Phillipsburg, NJ, J.T. Baker Chamical
• -:, 1982, p. 12. —
4. “Rapid Extraction of Phenols from Water for HPLJC with ‘BAKER’—lO
SPE Disposable Cohunn OCTYL (C8), 6 mL, in ‘ Baker’—lO SPE Applications
Guide, Vol . I, Phillipsburg, NJ, J.T. Baker Chemical Co., 1982, p. 26.
6—114
-------
Table 1
SUMMARY OF
EXTRACTION
AND
LEACHING
ACTIVITY
DATA
FOR THREE SOLIDS
Treatment
Pond
Sludge
g
Barrel
Sludge,
11
20
g
Filter
Cake
20 g
3513-A6, 30
Initial beta
activity, Bq 62,400 316,000 426,000
Activity present
after pH 5
leaching, Bq
(1) 1,200 34,000 39,200
(2) 3,200 31,200 37,200
Activity present
after acid/base
extraction of the
leachate with pentane,
Bq
(1) <22 <4 0
(2) <26 <38 34
Activity present
after acid/base
extraction of the
leachate with methylene
chloride, Bq
(1) <1 120 53
(2) 100 103 73
6—115
-------
Table 2
Trial I Trial 2 Trial 3
83 83 73
90 83 80
100 95 90
93 70 68
89 85 81
100 100 100
98 98 95
98 95 98
90 94 90
93 95 93
93 95 93
96 99 95
96 91 90
91 93 90
Samples spiked with 47,000 cpni (Ca. 2200 Bq) Co-60. Final activity
<10 cpm (Ca. <0.5 Bq), with measurement limited by instrumental
uncertainty. Decontamination factor ca. 2000.
Samples prepared using solid phase extraction, as described in text.
Recovery of Priority PAIl from 100 mL pH 4.9 Buffer
Spike Level 40-400 ppb
Compound
Naphtha lene
Acenaphthene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
BaA
Chrysene
B (b) fluoran
B(k) fluoran
BaP
Dibenz(a,h)anth
Benzo(ghi)pery l
Indeno [ l,2, 3-cd]pyrene
Mean % rec.
80
84
95
77
85
100
97
97
91
94
94
97
92
91
6-116
-------
Table 3
Recovery of EPA Organochiorine Pesticides at Test Level of 5 ppb
Recovery, % at 5 ppb
Lindane Endrin
Hexane -1 89 88
Hexane -2 100 100
Hexane -3 98 100
Average Recovery, % 96 96
All samples spiked with 18,000 cpm (ca. 900 Bq) Cs-137. Final activity
< 10 cpm (ca. <0.5 Bq). Estimated decontamination factor >2000.
Samples prepared by solid phase extraction, as described in text.
6—117
-------
Table 5
Test of Sparger for Purgeables in Radioactive Aqueous Samples at
50 ppb
Recovery in ppb
Compound Trial 1 Trial 2 Trial 3 Avg % Rec.
Methylene chloride 28 240 23 194
Acetone 87 140 73 200
Carbon disulfide 16 10 10 23
1,1-Dichloroethene 20 16 15 34
l,l-Dichloroethane 27 24 22 49
1,2-Dichioroethene 25 22 20 45
Chloroform 14 34 30 52
1,2-Dichioroethane 32 28 26 57
2-Butanone 97 105 92 196
1,l,l-Trich1oroethan 23 24 22 46
Carbon tetrachioride 25 25 25 50
Vinyl acetate 32 30 27 59
Bromodichioromethane 26 27 27 53
1,2-Dichioropropane 28 28 27 55
cis-l,3-Dichloropropene 21 30 27 52
Trichioroethene 29 29 27 57
Dibromochioromethane 18 32 30 53
1,1,2-Trichioroethane 24 33 31 59
Benzene 30 44 26 67
trans-1,3-Dichloropropene 20 23 21 43
Bromoforin 6 31 28 43
4-Methy l-2 -pentanone 37 51 46 89
2-Hexanone 22 55 51 85
Tetrachioroethene 9 33 32 49
l,1,2,2-Tetrachloroethane 21 21 22 43
Toluene 21 35 22 52
Ch lorobenzene 10 22 21 35
Ethylbenzene 9 21 21 34
Styrene 5 17 16 25
Xylenes 7 26 18 34
Radioactive spike added: 5.0 E+6 Bq Cs-l37.
Final activity: background (within instrumental uncertainty)
Estimated decontamination factor >1 E+6
6—118
-------
ENVIRONMENTAL APPLICATIONS OF MAGIC LC/MS
Alex Apffel, Hewlett—Packard, Scientific Instruments Division, Palo
Alto, California
ABSTRACT
MAGIC LC/MS (Monodispersed Aerosol Generation Interface Combining
Liquid Qiromatography and Mass Spectroscopy) is a technique for
generating electron Impact spectra from analytes separated by High
Performance Liquid Qironiatography. The technique, developed by
Browner et al.(1) consists of an aerosol generator, a desolvation
chamber and a two stage momentum separator. The HPLC effluent is
first nebulized by the aerosol generator. The resulting droplets
are then desolvated at atmospheric pressure yielding a mixture of
analyte particles and solvent vapor. In the momentum separator,
the solvent vapor is pumped away and the analyte particles are
allowed to enter the MS source where they are flashed vaporized and
ionized by electron impact.
The technique requires very little modification of either the HPLC
methodology or the MS operation. Typical LC flow rates range from
0.2 to 1.0 mi/mm. A wide range of solvents can be used; the
primary restriction being the mandatory use of volatile buffers.
The Mass spectrometer can be used in either electron impact or
chemical ionization modes.
The current work examines the application of MAGIC LC/MS to the
analysis of pollutants in environmental samples. Specifically, the
analysis of a group of compounds cited in the EPA Appendix
VIII(2,3) which are not easily amenable to analysis by GC/MS,
including thiourea, ethylene thiourea, naphthyl thiourea, maleic
hydrazide, warfarin, benzindine, thiram, reserpine and others.
Although these compounds have been analyzed by Thermospray
LCIMS(4), this technique, while versatile is limited in its
information content. MAGIC LC/NS generates standard El spectra
which can be computer—searched in standard libraries.
Additionally, data concerning basic performance specifications and
optimization of the operational parameters will be discussed.
INTRODUCTION
Monodispersed Aerosol Generation Interface Combining Liquid
chromatography and Mass Spectrometry (MAGIC LC/MS), introduced by
Browner et al. is a technique which allows Electron Impact (El)
spectra to be obtained from compounds which are not amenable to gas
chromatographic (GC) separation due to low volatility or thermal
lability. As such, MAGIC LC/MS offers an attractive complement,
rich in structural information, to Thermospray LC/MS which
6—119
-------
typically produces simple spectra consisting primarily of
pseudo-molecular ions.
This approach offers some distinct advantages in applications in
which the samples cannot be separated by CC on in which HPLC is the
separation method of choice and In which structural data of
unknowns or confirmatory data for target compounds is needed.
EXP I1IENTAL
Liquid (2iromatography
The HPLC consisted of a Hewlett—Packard Model 1090 HPLC equipped
with a DR—5 ternary pumping system, autosampler and a filter
photometric detector. All HPLC columns were lOOx2.lmm i.d.. The
following column systems were used; Browalee 5cm RP—l8 and RP—8
(Brownlee Labs, Santa Clara, Ca) or Phase Separation 3 urn RP—18
(Phase Separations, Norwalk, CT.).
Mass Spectronetry
All mass spectral measurements were performed on a standard,
unmodified HP 5988 differentially pumped quadrapole mass
spectrometer with high nasa (2000 aiiu) option. In order to accept
the MAGIC interface 0.5 inch probe, a standard DIP vacuum port was
mounted on the left side of the source manifold. The standard
CC/MS interface occupies the port on the right side.
Typical operating conditions are as follows: tuning using autotune
routine; source temperature @ 330°C; source manifold pressure @
lxlO 6 torr; Electron multiplier @ 2500—3000 V. For data
acquisition In the scan node: 50—500 amu aec -.
MAGIC IC/MS
The experimental MACIC Interface hardware is shown schematically in
Figure 1. The interface consists of four main sections; aerosol
generator, desolvation chamber, nomentum separator and 0.5 inch
transfer probe.
MAGIC IC/MS operates in the following manner: initially an aerosol
Is generated which consists of droplets within a narrow range of
diameters. As the droplets travel through a desolvation chamber
which is maintained at near ambient pressure and temperature, the
volatile solvent is evaporated leaving a particle of non—volatile
material (Including analyte) behind. This mixture of solvent vapor
and analyte particles enters a two stage momentum separator In
which the vapor Is pumped away while the relatively massive (and
consequently high momentum) particles pass into the source of the
mass spectrometer in a narrow particle beam. The particles Impact
6—120
-------
the heated source wall and are flash vaporized. Analyte molecules
are then ionized by electron impact and mass analyzed in the
standard manner.
The aerosol generator is shown schematically in Figure 2. The HPLC
effluent enters vertically through a short piece of 25 cm fused
silica capillary, forming a liquid jet which breaks up into a
stream of droplets due to instabilities in the jet . These
droplets are dispersed by a jet of helium gas which enters at right
angles to the liquid jet through a short length of 24 gauge
hypodermic needle. The helium flow rate is approximately 1 L
tain 1 -. The hypodermic needle tubing through which the helium
enters is mounted on an x—y positioner to allow precise control of
the intersection point with the droplet stream.
The desolvation chamber consists of a glass tube with formed
details to accommodate the aerosol generator and pressure and
temperature sensors. The chamber is wrapped with a heating tape,
the power to which is generated by a feed back control system.
The momentum separator is constructed from stainless steel and
consists of a cylindrical nozzle and two conical skimmers. The
First stage is pumbed by a Baizers (Hudson, NH) UnoOl6B mechanical
pump while the second stage is evacuated by a Baizers DuoOl6B
mechanical pump. Since these pumps run relatively hot, it has not
been found necessary to use a cold trap to isolate the HPLC mobile
phase vapor.
Typical pressures in the system are as follows: desolvation
chamber 200 torr; momentum separator (first stage) 2—5 torr;
momentum separator (second stage) 0.1—0.5 torr; source manifold
7xl0 6 —3xlO 5 torr.
Chemicals
Solvents were HPLC grade purchased from EM Science (Cherry Hill,
NJ.).
Environmental standards were obtained courtesy of the EPA
depository (Las Vegas, NV).
RESULTS AND DISCUSSION
Optimization
During the investigation and development of a viable experimental
MAGIC system, it was necessary to evaluate those control factors
which play a major role in performance. The most critical
operational parameters were found to be: helium gas flow rate;
desolvation chamber temperature and pressure; and the mass
6—121
-------
spectrometer source temperature. These factors were evaluated
under a range of conditions to determine optimal settings.
The helium gas flow rate and the position of the helium gas jet
relative to the liquid jet depend on the characteristics of the
HPLC mobile phase composition and flow rate. Referring to figure
2, at a certain point above the 25 cm fusio silica the liquid
column breaks up into a stream of droplets due to instability
inherent to the system. 3 The helium dispersion gas must enter the
system slightly above this point with sufficient velocity to
disperse the droplets at 900 relative to the initial liquid jet.
As long as the dispersion point is above the droplet formation
point, maximum signal is obtained. For a given flow rate, in a
mobile phase consisting of mixture of either acetonitrile or
methanol and water, the position of the droplet formation point
varies over a 2mm range. The gas dispersion point, then is set
optimally slightly above the highest position. For a typical
mobile phase flow rate of 0.4mLIrain, the helium gas flow rate is
approximately 1 L/sin. It is possible, however to operate the
system with mobile phase flow rates between 0.1 and 1.2 mL/min by
adjusting the gas flow rate accordingly.
The main purpose of therinostatting the desolvation chamber is to
replace the heat which is absorbed by the desolvation process. As
such, the temperature is optimally held a slightly above ambient
temperature C 35—40°c). This temperature does vary towards
slightly higher temperatures as the aqueous content of the mobile
phase increases, reflecting the relatively larger amount of heat
required to vaporize water. The effects of mobile phase and
desolvation chamber temperature on the signal produced by 1 L
injections of 10 ng of caffeine are shown in figure 3. Note that
at approximately 35°c, the optimum response can be obtained for all
mobile phase conditions.
The desolvation chamber pressure also plays an important role. In
principle, the higher the pressure in the desolvation chamber, the
more efficient the heat transfer and desolvation process. However,
this does not take into account the pressure characteristics of the
supersonic nozzle—skimmer system in the momentum separator. At
sufficiently high pressures, the divergence of the nozzle jet
becomes too large and transfer efficiency decreases through the
system 6 . Due to the interplay of these two factors, there is an
optimum desolvation chamber pressure around 200 torr.
The third important factor in the operation of the system has been
found to be the mass spectrometer source temperature. Somewhat
surprisingly, best performance was obtained with a relatively hot
source (330°c). Our initial expectations were that thermal
decomposition and decreased signal would result from high source
temperatures. As shown in figure 4, however, this proved not to be
6—122
-------
the case. For a range of compounds examined, signal intensity
increased to a maximum at 330°c, while the normalized spectrum is
essentially constant over this range. The compounds examined do
show variations in their temperature dependence. Comparing for
example, caffeine and reserpine (see figure 4), caffeine produces a
nearly constant signal, while reserpine shows a significant
Increase in signal Intensity (but not in spectral character).
There are exceptional cases, however, where temperature shows a
major effect on the spectrum. In particularly, cholesterol
(MW386) undergoes an increased loss of water, producing the
fragment at m/z=368, as a function of temperature. This is shown
graphically In figure 5 where the ration of the intensities of
fragments 368 and 386 are plotted against source temperature. In
our experience, it must be noted, this kind of temperature
dependency is an exception rather than the rule.
PERFORNAN
The quantitative performance was evaluated using caffeine as a
probe and is shown in figures 6 and 7. In scan mode (50—300
amu/sec) the minimum detectable quantity (S/N5) (fig. 6) was found
to be 5 ng. In SIN mode (fig. 7.), monitoring m/z ’l94, MDQ was
found to be 5Opg. In both cases the r 2 value for the regression
was 0.998 (0.5—500ng SIN: 5—l000ng Scan)(fig. 8). Note that for
the SIN experiment, the signal obtained for the 1 ug injection
saturates the detector and shows a loss in linearity. It should be
noted that our experience has been that caffeine is neither the
most sensitive nor the least sensitive compound examined and MDQ’s
may vary by a factor of 10 either up or down.
Qualitative performance was evaluated by examining a wide range of
compounds (listed in table 1) in a flow injection mode. These
compounds were injected at a lOOng level to generate reference
spectra. Of these compounds, the majority can be found in the
NBS/EPA/NIH spectral library. The spectra represent classical El
spectra. The advantage of this is that the data can generally be
interpreted by computer library searches, and in those cases where
the spectra are not found in the library, the interpretation can be
performed according to well known rules. It Is interesting to note
that a large number of spectra that are in the library are not
easily analyzed by GC.
APPLICATIONS
Several applications have been evaluated to demonstrate the
potential of the technique. The applications examined are
environmentally oriented, but the sample handling capabilities of
HPLC due to precolumn technology suggest a wide variety of
applicable matrices.
6—123
-------
figure 9 shows the separation of 12 triazine pesticides in full
scan mode (50—400amu) at the lOOng level. Due to the excellent
signal to noise demonstrated by this trace, average MDQ for these
compounds is approximately 1 ng. The use of multiple ion
monitoring generates useful detection and identification data at
the 5Opg levels. Although triazine pesticides can, in fact, be
separated by GC/MS, this is an excellent example of a situation in
which HPLC is the separation method of choice due to the ability to
handle direct injection of relatively large volumes of aqueous
samples. Mass spectrometry shows one of its particularly
attractive features as a chroinatographic detector in its ability to
separate nearly coeluting compounds due to spectral differences.
Figure 10 shows the separation of 12 phenylurea herbicides in full
scan mode (50—400ainu) at lOOng level. Although there is an evident
variation in the response factors of these compounds, It can be
clearly seen that the detection limits are similar to those stated
above. In selected ion monitoring, the detection limits in this
application can be reduced slightly to an average of 20 pg, due to
the fact that each of the compounds generates one of 4 fragments
(61, 72, 91 or 114).
Finally, figure 11 shows the separation of 11 compounds listed on
the EPA appendix IX list. 5 These compounds cannot easily be run
using CC/MS. As above the date shown is for full scan mode (50—400
ama) at the lOOng levels, indicating similar detection limits.
This application provides a rapid straight forward technique to
separate, identify and quantitative these compounds.
Conclusions
Recent advances in MAGIC LC/MS have lead to significant
improvements in sensitivity and ease of use which make the
technique a viable approach to LC/MS coupling In the analytical
laboratory. The technique generates RI spectra from compounds
separated by HPLC, which are difficult or impossible to separate by
CC.
In Its current stage of development, MAGIC LC/MS offers a useful
complement of Thermospray LC/MS. Whereas Thermoapray LC/MS has
been found to be particularly useful for molecular weight
information, MAGIC LC/MS generates El spectra which are familiar,
library searchable and easily Interpreted. Where Thermospray LC/MS
has been found very useful for relatively large molecular weight
peptides and proteins, we have found MAGIC LC/MS to be most useful
(but not limited to) for relatively low ((1000 ama) molecular
weight compounds in which electron impact leads to a small number
of identifiable fragments.
Our investigations Indicate, however, that MAGIC LC/MS offers a
6—124
-------
MAGIC
LCiMS
SYSTEM
MOMENTUM
SCHEMATIC
P-2 0 0TORR
NEBULIZER DESOLVATION CHAMBER SEPARATOR TRANSFER HP5988A GC/MS SYSTEM
P-5X1c1 6 TORR
N)
U,
1 [
P-
10 TORR
P
- 0.1 TORR
PUMP B
PUMP A PUMP B
BALZERS IJNOOI6B BALZERS 0U00168
DIFFUSION PUMP
-------
MONODISPERSED AEROSOL GENERATOR
0
/DROPLEDISPERSION\
He fAS IN I
HeIN- ’
N
DROPLET
FORMATION POINT)
25 micron I
FUSED SILICA,/
/
/
IN/’
X—Y POSITIONER
/
I
\\
\\
LIQUID IN
-------
DESOLVATION CHAMBER
TEMPERATURE (°C)
long CAFFEINE (SIM)
N.)
w
0
z
a
z
% METHANOL
-------
S CAFFEINE
‘ •O”• ALDICARB
O*_ CORTISOL
—0--- CHOLESTEROL
‘—O—— RESERPINE
I
I
C
03
SOURCE TEMPERATURE
-------
1.
I
I
THERMAL DEGRADATION OF CHOLESTEROL
I
H 3 C
I
R
A
T
I
0
1 .
HO
H
H
0..
3
6
8
/
3
8
6
LOSS OF WATER
M —18 — 368
0.
0
0.
0.
—0-— 360/386
220
230
240
250
260
270
200
290
300
310
320
330
340
350
SOURCE TEMPERATURE
-------
1 . E6-
S a 0E5
8. E5
a) 7.0E5
C i
6aOE5
-o
C 5. E5-
4. E5 -
3 a 0E5
2aøE5:
1 .0E5-
0-
ljtg
20 22
SCAN (50-250) (CAFFEINE: TIC)
6.0E4
a)
0 5a0E4
C
It,
4.0E4
C
3a0E4
c i:
2.0E4
5ng
8a0
500ng
lOOng
k_ L I ___ ___ __
Time (mm.)
15.5
10
12 •1
Time (mm.)
18
-------
SELECTED ION MONITORING (CAFFEINE: 194 )
3. E5
lOOng
ing
500pg
o 5O O
C
2.0E5 ‘Ti
0
C 4 Ø •
J
C a:
loopg 5Ong
3 Ø -
a 1.0E5- _____________________
( LLJ .
Time (mm.)
Ong
o - ‘ •
0 2 4 B 8 10 12 14 16. 18
Time (mm.)
-------
MAGIC LINEARITY
a.
I
S
G
N
A
L
SIN
LINEAR REGRESSION (SIN)
LINEAR flELflE:T . AN),
I I I I I I I——i
200 300 ** 500600100800 900 00O
0
SCAN
ECAFFEJNE (NS)J
6—132
-------
AMITROt.E ATRAZINE DIPROPETRYN (MW - 255)
C I
10000 (MW.84) (MW2I S)
•8000
o X 200 N ’N 255
N •8000
o 68 H,C))HN N”NHCH(CH ,
6000 N—N 6000 ) ,!1 H C 6000 / 152
0
4000 4000 ‘ H N H (C , 4000 104
C
1?3
.0 .0 / .0
CE 2000 109 20? 2000 138 I
0 i r 1I Li i I
— v.-. .
100 150 200 250 100 150 200 100 150 200 250
Ma #’Charge 11ass Charge Mass .’Charga
z
>-
w >.
2 W m
5.0E5- Z W z
w
0 Z )-0
0
Ri W WO -J 01-
Ion range \ 1 .eo to 256.00 amu. from D TR:TRISCRN5.D /
ZW > 0.
Zz RãZ - ICa. GRADIENT3O-70% B/IOn*i
W 3C HELIUM FLOW 12 (it*i
C w
-J
DESOIVATION CHAMBER
‘H
0
0 jj MS SCAN: 60-300 .c
EM: 3000 V
2.0E5-
I-
0.. SOURCE TB F: 33 0 0
1. OES Li L .L
I • I • I • I
2 4 B 8 10
Time (mm.)
TRIAZ1NE PESTiCiDES
-------
10000
9000
6000
2
V
C)
C
0
4000
. 0
a:
DIURON
(MW 232)
c os
2000
/
93
S
U
C
0
232 C
/ 3
.0
16? L a:
C II.
124
/
L
10000
9000
6000
4000
2000
SIDURON
AW 232)
QM O
I
200
hi.
250
1
UNURON
(MW — 248)
0CM,
119
/
£
S
U
C
0
. 5
C
3
.0
161
/
1
6E5
232
/
200
1
• 4E5
250
z
124
/
1
• 2E5
0
C
‘U
C
-o
1
200
•0E5
250
8
0E4
6
0E4
4
HPC.C
2.
z
w
z
2
COLUM t 100 2.1 Buns HYPERSIL 00$
NOBLE PHASE A: H 2 0
B: ACE1’ONtTI*.E
ORADENT 00.100% BllOn4n
FL RATE 0.4 n4At*i
MGC HEL )UUF1.OWO 12 IA *i
DESOLVATION CHAMBER 0 35C
MS SCAN: 50-25Msc
EM 2500 V
SC TE : 00 C
SA s U
4
T I me
6
(mm.)
PHENYLUREA HERBICIDES
-------
2E4
103
/
U
C
0
.5
C
.0
a:
81
41
21
WARFARIN
131 (MW.308)
—
cC
CM CM,
187
1
Ii
Th (RAM
(MW • 240)
8 S $
I II
5. 0E4
0E4
U
3.oE:4
.5
2. 0E4
.0
a: 10000
5
Iii.
120
.1
V
(3
C
0
C
.0
a:
D CHLORO 8EN2IDIN
(MW - 252)
II - II-
8.
6. 0E4
4. 0E4
2.0 4
tL
178
/
240
I I
DATA: RPPIX7. D
4.0E5
I
L )
L i i
3.0E5
II )
(1
C
‘U
C
cc
200
250
2.0E5
w
z
c
Er
0
w
z
w
>
w
1. E5
cOND
0
I
HPI.C COLUMN: 100 2.1 Swim SPHERISORB 00S2
MOBILE PHASE A: .OIM NH 4 OAC
B: ACETONIThILE
O ADIEWr 0 90% aSn
SLOW RATE 0.4 n miIn
MACC HELIUM 12 n
DESOLVATION CHAMBERO 40C
MS SCAN: 5O500 ec
EM: 3000
SOURCE TEMP: 330C
SAMPLE 100nWco ,one ,fl
2
Time (mm.)
EPA APPENDIX IX COMPOUNDS
-------
number of further areas of development which could, in principle,
significantly expand its areas of applicability. In particular,
the use of alternative ionization modes show promise for ionizing
compound which can be introduced to the mass spectrometer by MAGIC,
but which do not undergo electron impact ionization. Initial
experiments have generated chemical ionization (CI) spectra, and
further investigation is underway.
FO0TN0T
1 wiiloughby, R..C. and Browner, R.F.., Anal. them. 56 (1984) 2626
2 Federai. Register 45, 99, May 19, 1980, p. 33132—3
3 Federal Register 49, 191, October 1, 1984, p. 38797—8
4 Thermoepray LC/HS of EPA Appendix VIII Compounds. Goodley, P.C.
and Thorp, 3., Hewlett—Packard LC/MS Applications Note AN176—41
REF ENC
WiUoghby, R.C. and Browner, R., Anal. them 1984 (56) 2626.
Browner, R.F., Winkler, P.C., Perkins, D.D. and Abbey, L.E.,
Submitted for publication, Microchemical Journal.
Rayleigh, Lord, Proc.Roy. Soc. 1879 (29) 71.
Israel, G.W. and Friedlander, J., J.Col.Int.Sci 1967 (24) 330.
Federal Register 45, 99, Nay 19, 1980, 33132—3.
6—136
-------
COMPARISON OF CAPILLARY COLUMN AND
PACKED COLUMN ANALYSIS FOR VOLATILE ORGANICS
R. R. Clark, J. A. Zalikowski, Montgomery Laboratories, Pasadena,
California
INTRODUCTION
The EPA GC/MS methods for volatile organic analysis in water,
wastewater, and hazardous waste matrices (EPA methods 524, 624 and
8240) have always stipulated the use of packed columns. Only
recently has EPA proposed the use of capillary column GC/MS analysis
by printing Method 524.2 which covers volatile analysis in raw and
finished drinking water for SDWA analyses. EPA is expected to issue
formal capillary GC/MS methods for wastewater and hazardous waste
matrices in the near future.
Capillary columns are superior to packed columns in peak resolution
and sensitivity, thereby leading to better identification and
quantification of compounds present in the samples. This is
especially true for samples containing complex matrices. This also
results in shorter runtimes, with Increased productivity in the
laboratory and reduced analytical costs. Capillary column use was
not originally permitted because they were never part of the
original ruggedness testing and method validation studies performed
by EPA. Consequently, few environmental laboratories have had much
experience with the use of capillary column GC/MS on a routine basis
for analysis of volatiles.
Montgomery Laboratories has routinely used capillary columns for
volatile organic analysis of water, wastewater, and solid waste
matrices for over eight years. We were one of the few laboratories
nationwide to be granted a formal EPA variance for using capillary
columns in volatile organic analyses by GCMS. We will present our
experience in working with the QA requirements described in the
published volatile methods (524.1, 524.2, 624, and 8240). We have
undertaken a comparison of the capillary column method versus the
packed column method using both the Finnigan 5100 and Hewlett
Packard MSD.
EQUIPMENT
The packed column analyses were performed on a Hewlett Packard
59980C Mass selective detector (MSD). The MSD was installed with a
Tekmar LSC—2 Purge and trap device, a packed column injection port,
a capillary column injection port, and a jet separator. Our
original intention was to compare the packed and capillary columns
on the MSD alone. However due to problems with the design of the
MSD we were unable to consistently use it with a capillary column
for volatile analysis. The packed column used was a 2 meter x 2 mm
6—137
-------
ID glass column packed with Carbopack B (60—80) mesh coated with 12
SP—1000 The capillary column used was a 30 meter x 0.25 mm ID DB—5
FSCC with 0.25 urn film thickness. Cryogenic focusing was performed
by immersing a loop of the capillary column in a cup of liquid
nitrogen. The capillary column was inserted directly into the
source.
METHODS
The methods used were the EPA published methods 624, 524.2 and
8240. Method 524.2 was modif led to use a loop of the capillary
column for cryogenic focusing rather than the unit specified in the
method.
COMPARISON OF THE PUBLISHED METHODS
EPA methods 524.1, 524.2, 624 and 8240 were all developed for the
analysis of volatile organics in water. However, each method
specifies different compounds, procedures and quality assurance
procedures. Table 1 lists many of these differences. Method 524.1
Is not included because it is the same as method 542.2 with the
exception that 524.1 requires the use of packed columns.
Methods 624 and 8240 both specify only packed columns for analysis.
Method 524.2 specifies either narrow or wide bore capillary
columns. Method 524.1 is an equivalent packed column method. Each
method also contains a different list of compounds which can be
analyzed by the method.
The quality control requirements of the three methods differ
substantially. Methods 524.2 and 8240 require that a five point
initial calibration be performed, while method 624 requires only
three. Method 524.2 requires that the percent RSD (relative
standard deviation) of the response factors determined for Initial
calibration be within 10 percent in order to be called linear. If
the percent RSD is higher, then a curve must be used. Method 624
allows a 35 percent RSD to be called linear, no curves allowed.
Both methods 524.2 and 624 require BFB tuning and continuing
calibration check to be performed “daily” while method 8240
specifies every 12 hours. The methods also have differing
requirements for the frequency of analysis of duplicates, spikes and
external standards.
Method 624 has a large list of compounds which can be used as
internal and surrogate standards. Method 8240 has a smaller list,
while the list for method 524.2 is even smaller. One of the
8urrogate standards to use for methods 624 and 8240 is a compound
specified as an analyte in method 524.2.
6—138
-------
TABLE 1
Comparison QC requirements for Several EPA
Methods for GCMS Analysis of Volatiles
524.2 624 8240
Column type Capillary Packed Packed
Wide bore Cap
Number of Compounds 58 30 34
Sample size 25 ml 5 ml 5 ml
Initial Calibration 5 point 3 point 5 point
% RSD for Standards 10% linear 35% linear 30% for certain
Curve othervise Curve OK compounds - linear
Cont. Cal. Frequency “Daily” “Daily” 12 hours
Cont. Cal. Criteria 20% difference Varies with 25% for certain
compound compounds
Tuning Frequency “Daily” “Daily” 12 hours
Duplicate Frequency Not specified Not specified Matrix spike dups
Spike Frequency Not specified 5% 5%
External Std 10% frequency 5% frequency “daily”
All compounds All compounds 6 Compounds
60-140% recov. Table of criteria Mm RF=0.300
Surrogate and Fluorobenzene Large table D8—Toluene
Internal Standards D4-1,2-DCB of acceptable BFB
BFB compounds D4-1,2-DCA
BrChMe thane
1,4—DFB
D5-ChBenzene
6—139
-------
DETECTION LIMITS
The detection limits published by the EPA In their methods differ
greatly. Table 2 lists the published Method Detection Limits (MDLs)
for a selected group of compounds in water. MDLs are defined as the
minimum concentration of an analyte which must be present in a
sample to be 99 percent confident that the measured signal comes
from the analyte and not background noise. Note that the limits
published for method 8240 (SW846 3rd edition) are not MDLs but
rather PQLs (practical limits of quantitation). Both methods 624
and 524.1 are packed column methods, however the MDLs for 524.1 are
much lower than tho8e for 624. One reason for this is the larger
volume of sample purged using method 524.1 (25 ml vs. 5 ml). Method
8240 is also a packed column method, however the PQL listed Is
approximately 5 ugh. PQLa are defined as an estimate of “the
lowest level that can be reliably achieved within specified limits
of precision and accuracy during routine laboratory operating
conditions.” Therefore the PQLS listed for 8240 are more practical
when considering environmental samples.
The method detection limits listed for the capillary column method
(524.2) are all in the 0.02 to 0.2 ugh range. Some of the numbers
published originally were actually below the Instrument detection
limit (IDL) for the particular compound. These are now undergoing
revision.
Table 3 shows the results of MDLs determined in our laboratory using
method 624 for the packed column and method 524.2 for the capillary
column. Note that the MDL on the capillary column was determined at
two different concentrations. These MDLs were determined under
routine working conditions, rather than under optimum conditions.
The !‘IDLs determined on the packed column at 5 ug/l are similar to or
slightly better than those presented for method 624. However, they
are higher than those presented for method 524.1. The MDLs
determined by the EPA for method 524.1 were spiked at approximately
1.5 ug/l. Since the MDL is determined at approximately three times
the standard deviation of a set of seven replicates, a lower MDL can
be obtained by spiking the replicates at a lower value. However at
some low concentration the precision becomes poor and the MDL values
do not become smaller. In addition the lower limit of the MDL Is a
function of the absolute instrument sensitivity. The sensitivity
(IDL) of the MSD used to determine the MDLs was such that we could
not spike at a lower value and obtain a smaller MDL for all
compounds and spiking was therefore restricted to 5 ugh.
The MDL for the capillary column was determined at a spiked value of
5 and 1 ugh. At 5 ug/l the MDL determined was about the same for
both the packed and capillary columns, which indicates that the
precision of both methods is approximately the same. However, due
6—140
-------
TABLE 2
Detection Limits for Published EPA Methods.
(ugh in Water)
EPA METHOD NUMBER
Compound 524.1 524.2 624 8240
Benzene 0.10 0.03 4.4 5
Bromoform 0.66 0.20 4.7 5
Bromomethane -- 0.06 -— 10
Carbon Tetrachioride 0.28 0.02 2.8 5
Chlorobenzene 0.14 0.03 6.0 5
Chioroethane —— 0.21 -- 10
2-Chioroethylvinylether -- -- -- 10
Chloroform 0.24 0.04 1.6 5
Chioromethane -- 0.05 —- 10
Dibromochioromethane 0.30 0.07 3.1 5
Dichlorobromomethane 0.28 0.03 2.2 5
1,1-Dichioroethane 0.17 0.03 4.7 5
1,2—Dichioroethane 0.22 -— 2.8 5
1,1—Dichioroethene 0.J.9 0.05 2.8 5
t-1,2—Dichloroethene 0.19 0.03 1.6 5
1,2—Dichioropropane —- 0.02 6.0 5
c—1,3—Dichloropropene —— —— 5.0 5
t—1,3-Dichloropropene -- -- -- 5
Ethylbenzene -- 0.03 7.2 5
Methylene chloride 0.13 0.09 2.8 5
1,1,2,2—Tetrachioroethane 0.41 0.20 6.9 5
Tetrachioroethene 0.29 0.05 4.1 5
Toluene 0.12 0.08 6.0 5
1,1,1—Trichioroethane 0.26 0.04 3.8 5
1,1,2—Trichioroethane —— 0.08 5.0 5
Trichioroethene 0.36 0.02 1.9 5
Trichiorofluoromethane 0.21 0.07 -- 5
Vinyl chloride 0.31 0.04 10
Note: EPA methods 524.1, 624 and 8240 are packed column methods
while method 524.2 is a capillary column method.
Detection limits listed for method 8240 are really PQLs.
6—14 1
-------
Table 3
Determined MDL values
(ugh in Water)
PA(XED CAPILLARY
Compound 5 ugh S ugh 1 ugh 1
Benzene 0.74 1.1 0.15
Bromoforin 0.94 0.68 0.31
Bromomethane 1.6 1.4 0.12
Carbon Tetrachioride 0.60 0.68 0.17
Chlorobenzene 0.54 0.91 0.14
Chloroethane 2.6 6.3 0.27
2-Chloroethylvinylether —- 0.63 0.15
Chloroform 0.60 0.90 0.15
Chioromethane —- 5.14 0.15
Dibromochioromethane 1.0 0.65 0.10
Dichiorobromoinethane 0.64 6.3 0.48
1,1-Dichioroethane 0.50 0.80 0.26
1,2-Dichloroethane 0.5 1.2 0.23
1,1-Dichloroethene 1.2 1.1 0.20
t—1,2--Dich loroethene 1.1 0.93 0.33
1,2—Dichioropropane 0.81 1.8 0.33
c—1,3-Dichloropropene —— 0.62 0.21
t-1,3-Dichloropropene -- 0.80 0.25
Ethylbenzene 2.1 0.78 0.14
Methylene chloride 1.2 1.06 0.09
1,1,2,2—Tetrachioroethane 0.60 0.68 0.21
Tetrachloroethene 0.77 1.1 0.16
Toluene 0.40 0.64 0.11
1,1,1—Trichioroethane 0.57 0.69 0.17
1,1,2 -Trichloroethane 1.1 0.69 0.08
Trichloroethene 0.91 1.1 0.36
Trichlorofluoromethane 1.2 0.99 0.23
Vinyl chloride 2.2 3.6 0.33
Note: The packed column data was obtained using laboratory water
spiked a 5.0 ug/l. The capillary column data vas obtained
using laboratory water spiked at both 5.0 and 1.0 ugh.
6—142
-------
to the sensitivity of the capillary column, we were able to
determine an MDL at a spike level of 1 ugh. Spiking at the lower
concentration of 1 ugh gave lower MDL values. However, our
determined values are much higher than those presented for method
524.2. The MDLs published for method 524.2 were spiked at 0.1 to
0.5 ugh. If we were to determine our MDLs at a lower spike
concentration we might expect to approach the published values
(Table 2). However, as the EPA has noted, some of the values in
Table 2 are below the IDL and are therefore unrealistic.
Table 4 lists the MDLs obtained in our laboratory using both packed
and capillary columns compared to the value published in EPA method
8240 for soils. The replicates for both the packed and capillary
column MDL determinations were spiked at 0.25mg/Kg and used 1 gram
of soil diluted to lOmis (a 10 fold dilution) compared to method
8240 which uses 5 grams of soil in 5 ml of water (no dilution). MDL
values in Table 4 assume no dilution factor. We feel that the MDL
values for the capillary column could be pushed lower by spiking at
0.024 mg/Kg or lover.
CHROMATOGRAMS
The superiority of capillary co1umn over packed columns becomes
most apparent when chromatograms are examined. Figure 1 is an
example chroinatogram of a 5 ugh standard run by packed column on
the MSD. Figure 2 is a chromatogram of a 2 ug/l standard run also
run on the MSD using a capillary column. The packed column run
shows that the 5 ugh standard has a low signal to noise ratio. The
capillary run, which was run at a lower concentration, has a much
better signal to noise ratio. This is due to the capillary column
being directly inserted into the MS source allowing all of the
sample to be detected, rather than being split in the jet separator
as with packed columns. The capillary peaks are also much sharper,
because of the higher signal to noise ratio.
Figure 3 is a chromatogram of a 1 ug/l standard run on the 5100.
Note that this chromatogram shows a lot less noise than the
capillary run on the MSD. The 5100 is a more sensitive instrument,
in part due to the 3KV conversion dynodes.
Figures 4 and 5 are chromatograms of jet fuel (JP—4) on both the
packed and capillary columns. Both samples were run at 1.0 mg/i.
JP—4 on the packed column does not show many of the lighter
hydrocarbon which are present at the beginning of the
chromatographic run using the capillary column. This may be impart
due to discrimination of the jet separator. The capillary column
chromatogram also shows many more heavy hydrocarbons than does the
packed column chromatogram.
6—143
-------
Table 4
Deteruined MDL values
(aglKg in soil)
____ EPA
PAt ED WI Li ART METHOD
Cospound 8240
Benzene 0.0097 0.0066 0.005
Broaofor. 0.0054 0.0072 0.005
Broao.ethane 0.011 0.0081 0.01
Carbon Tetrach].orjde 0.014 0.0063 0.005
Chlorobenzene 0.012 0.0071 0.005
Chioroethane -— 0.0220 0.01
2-Chloroethylviny lether -- 0.0063 0.01
Chiorofor. 0.011 0.0068 0.005
Chloro.ethane - - 0.0169 0.01
DibromochloroMethane 0.012 0.0068 0.005
Dich lorobromomethane 0.0086 0.0067 0.005
1,1-Dlchioroethane 0.0095 0.0067 0.005
1,2-Dichioroethane 0.0070 0.0063 0.005
1,1—Dichioroethene 0.011 0.0066 0.005
t-1,2—Dichloroethene -- 0.0068 0.005
1,2-Dichioropropane 0.011 0.0068 0.005
c-1,3-Dichloropropene 0.0092 0.0070 0.005
t—1,3--Dichloropropene —— 0.0070 0.005
Ethylbenzene 0.0094 0.0075 0.005
Methylene chloride 0.0071 0.0072 0.005
1,1,2,2—Tetrachioroethane 0.011 0.0079 0.005
Tetrachioroethene 0.011 0.0074 0.005
Toluene 0.011 0.0081 0.005
1,1,1-Trichloroethane 0.012 0.0064 0.005
1,1,2—Trichioroethane 0.0082 0.0052 0.005
Trichloroethene 0.011 0.0071 0.005
Trichlorofluoromethane 0.011 --
Vinyl chloride 0.014 0.0256 0.01
Note: The packed and capillary column data was obtained by
spiking a soil sample at 0.25 mg/Kg and was corrected for
dilution as explained in the text.
Detection limits listed for method 8240 are really POLs.
6—144
-------
lime (rr in. )
FIGURE 1
5 ugh STANDARD PACKED COLUMN
(MSD)
20
6—145
-------
12
Tstr e (mm.
FIGURE 2
2 ugh STANDARD CAPILLARY COLUMN
(MSD)
6—146
I -
6 8
-F-
14
-------
L L JLJ, J LL JL L _
49e 600 800 1000 1200
5:20 8:00 10:40 16:00
FIGURE 3
1 ugh STANDARD CAPILLARY COLUMN
(5100)
-------
10
30
20
Time (rn’in )
FIGURE 4
JP—4 PACKED COLUMN
6—148
-------
:48 13:29 29:09
509 1009 1500
FIGURE 5
2000
26:40
JP-4 CAPILLARY COLUNN
-------
PROBLEMS WITH MSD
We could not use the NSD for capillary column purge and trap
analyses on a consistent basis. When a capillary column is used
with a GQ1S the column is inserted directly into the source. When
used with a purge and trap device, water vapor can be introduced
with the sample. This water vapor triggered an error detection
circuit In the MSD which shut down the source due to “excessive
source pressure.” The HP MSD was originally specified to allow 0.8
mi/mm total flow into the source. Attempts to remove the water by
employing the Tekmar 4000’s “dry purge” option again failed to work
consistently. Since the “excessive source pressure” error was
triggered on an inconsistent basis we were unable to produce any MDL
data from the MSD using capillary column. Since our capillary
problem and those of other laboratories have been brought to their
attention, HP has reinvestigated the maximum pressure limit and
determined that the 0.8 mi/mm was overly conservative. HP has
since come up with a solution to the problem (by—passing the error
detection circuit) and is offering it an option.
LIMITATIONS OF CAPILLARY TECHNIQUES
Although capillary columns provide much greater resolution and
sensitivity then packed columns, this higher sensitivity also limits
the linear range. An attempt to meet CLP protocols for
standardization required by the CLP methods (20, 50, 100, 150 and
200 ugh). Using capillary columns we can routinely see compounds
below 0.1 ugh, however above 20 ugh the detector becomes
saturated. This problem can be overcome by diluting samples 80 that
the concentrations fall within the linear range of the standards.
SUMMARY
The EPA’s publication of capillary column methods for volatile
organic analysis will increase the sensitivity of the analysis and
will allow more complex matrixes to be examined. However it is
Important to note the limitations of capillary columns. In our
laboratory we feel that for low level samples containing complex
mixtures capillary column analysis is preferred. For samples with
only a few components at high levels we prefer the use of packed
columns. For high level complex mixtures we prefer to dilute the
sample and to analyze It using capillary columns.
6—150
-------
AN EXAMPLE OF INTERLABORATORY METHOD VALIDATION STUDIES
IN THE U.S. ENVIRONMENTAL PROTECTION AGENCY, METHODS 3510 AND 8270
Raymond J. Wesselman, Staff Scientist/Chemist, Project Management
Section, Quality Assurance Branch, Environmental Monitoring and
Support Laboratory - Cincinnati, U.S. Environmental Protection
Agency, Cincinnati, Ohio
ABSTRACT
One of the activities of the Quality Assurance (QA) Branch of the
Environmental Monitoring and Support Laboratory in Cincinnati
(EMSL-Cincinnati) is to perform interlaboratory method validation
studies. Through the use of a contractor, the QA Branch developed
three mixes containing 59 compounds which were spiked into three
matrices (reagent water, ground water from a dump site, and leachate
from a dump site), at six different concentration levels. These 59
compounds represent the Appendix 1X 1 list minus those compounds which
already have performance data. The six concentration levels form
three Youden pairs which are used to determine accuracy, single
laboratory precision and overall precision. The mixes were tested in
a control laboratory for applicability to Method 3510 (liquid/liquid
extraction) and Method 8270 (gas chromatograph/mass spectrometer
(GS/MS)) for extraction efficiency, sensitivity, resolution, and
stability. Fixed fee contracts were awarded to ten laboratories.
Each laboratory was required to analyze three matrices for all 59
compounds at six concentrations. A quality control sample and a
blank were also required for analyses each day. The resulting data
were evaluated statistically by the interlaboratory method validation
study (IMVS) system of computer programs which is consistent with
ASTM procedure D2777, “Standard Practice for Determination of
Precision and Bias of Methods of Committee D-l9 on Water”. IMVS
tests for the rejection of outliers (both whole laboratory per matrix
and individual data points), estimation of mean recovery (accuracy),
estimation of single-analyst and overall precision and tests for the
effects of matrices on accuracy and precision. A final report,
containing a description of the study, statistical treatment of data,
results, discussion and conclusions, completed the interlaboratory
method validation process.
INTRODUCTION
The Environmental Monitoring and Support Laboratory - Cincinnati
(EMSL-Cincinnati) develops analytical methods and provides quality
assurance (QA) support for the various offices of U.S. Environmental
Protection Agency (USEPA) for maximum reliability and legal
defensibility of environmental data collected under the water and
waste regulations. In EMSL-Cincinnati, QA responsibilities are
6—15 1
-------
assigned to the Quality Assurance Branch. One of its activities is
to conduct interlaboratory method validation studies to generate
accuracy and precision statements for the analytes specified in each
method. This paper describes one such study now being conducted for
the Office of Solid Waste (OSW), on two of its methods, 3510 and
8270.
METHOD SUMMARIES
Method 3510 is an extraction and concentration procedure for
water-insoluble and slightly water soluble organic analytes from
aqueous samples. In Method 3510, one liter of water or wastewater is
extracted successively at pH >11 with three 60 mL volumes of
methylene chloride. The pH of the sample is then readjusted to <2
and again extracted successively, with three 60 mL volumes of
methylene chloride. The extracts are combined, dried using a sodium
sulfate column, and then concentrated to a final volume of 1.0 mL,
using a Kuderna-Danish apparatus with a Snyder column. Method 8270
uses gas chromatograph/mass spectrometer (GC/MS) capillary column
techniques to determine the concentration of the semi-volatile
organic analytes in the extract.
STUDY DESIGN
A prime contractor was selected to develop, prepare, and verify the
test sample series and to manage the performance of the study.
Participant laboratories were obtained separately on contract by the
QA Branch of EMSL-Cincinnati. These laboratories were shipped
samples, sample instructions, study instructions and written
analytical methods, for analyses within a set time frame.
The sample design was based on Youden’s non-replicate plan for
collaborative evaluation of precision and accuracy for analytical
methods. According to Youden’s design, samples are analyzed in
pairs, each sample of a pair containing slightly different concen-
trations of the constituents. Each analyst is directed to perform
single analysis and report the value for each analyte in the sample.
Analyses in reagent water evaluate the proficiency of the method on a
sample free of interferences whereas analyses in the other waters
reveal the effects of interferences on the method.
The results from the study are returned to the QA Branch for
evaluation using IMVS at the Agency computer center in Research
Triangle Park, North Carolina.
6— 152
-------
SELECTION OF STUDY PARTICIPANTS
In accordance with the standard competitive bid process, an abstract
of the scope of work was announced in the Commerce Business Daily.
Over 100 laboratories asked for the request for proposal (RFP), which
contained an abstract of the method, the scope of work, and the
evaluation criteria upon which the offeror would be evaluated. The
evaluation criteria for the technical proposal for this study were as
follows:
* The offeror must demonstrate the Suitability of the Project
Management Plan, which includes the experience of the
project manager in managing contracts of a similiar nature
as this contract, the suitability of the organizational plan
in which roles, responsibilities and authorities are clearly
identified. The offeror must demonstrate the ability of the
offeror to provide the required number of analyses within
the period of performance given in the contract.
* The offeror must demonstrate the experience of analysts
involved in the method study.
* The offeror must describe the facilities and instrumentation
which will be made available for this contract.
* The offeror must describe what efforts are to be made to
insure the quality, quantity and timeliness of the data.
After proposals were evaluated and ranked, technically acceptable
laboratories were sent a performance evaluation (PE) sample for
analyses as per the written method and were required to use the same
personnel and instrumentation specified in their proposal to complete
the analyses. The PE sample contained ten priority pollutants. The
offerors did not know the compounds or their concentrations. The
offerors’ data were evaluated based on statistics from an
interlaboratory method validation study of Method 625, a similar
CC/MS procedure. From laboratories performing acceptable, ten were
chosen based on competitive costs. These are listed in Table 1.
STUDY DETAILS
Bionetics, Inc. in Cincinnati, Ohio, was selected as the prime
contractor, and was responsible for the preliminary investigation of
sample design, sample stability, extraction efficiences and practical
concentration levels. Bionetics conducted developmental studies to
determine the number of compounds that could be analyzed practically,
in one mixture without chromatographic interferences, the stability
of various mixtures in different matrices and the extraction
efficiencies from different matrices.
6—153
-------
The goal of this study was to include as many analytes as possible
from the RCRA Appendix IX list yet still remain within resource
limits, Of approximately 120 analytes, there were 48 analytes for
which regression equations had been generated for bias and precision
in a Method 625 study. These were deleted because Method 625 differs
from Method 3510 and Method 8270 only in that packed column
chromatography is used in Method 625 whereas Method 8270 uses
capillary column chromatography. Each method uses mass spectrometry
for the determinitive step.
Of the 72 remaining analytes, 13 were deleted from the study due to
either instability, insolubility or poor recovery. Table 2 lists the
compounds and the reasons for deletion from the study. The remaining
59 analytes (see Table 3) were divided into three mixes which
exhibited no significant chromatographic interferences and were
stable in methanol or acetone. Each mixture was injected repeatedly
into the gas chromatograph/mass spectrometer at lower concentrations
to estimate the instrument detection limit. When a reasonable level
was established, these concentrates were spiked into the three
matrices (reagent water, ground water, and leachate) used in the
study to determine any adverse matrix effects on the chromatography
and extraction procedure. Spiking ampul concentrations and standard
solutions were stable for 90 days.
Each participating laboratory received copies of the analytical
methods, instructions for preparation of the spiking solutions,
standard solutions and quality control (QC) solutions, 54 anipuls of
spiking solution, six 1-gallon bottles of ground water, six 1-gallon
bottles of leachate, 12 ampuls of standard stock solution and six
ampuls of QC solution. Spiking and standard solutions were
heat-sealed in 5 mL glass ampuls. Each ampul contained approximately
2.5 niL of solution of which 1.0 mL was used to spike 1 liter of
water. The ampul concentrations were analyzed for accuracy against
standards prepared from the neat pure compounds, by Bionetics, prior
to distribution. After the data were received from the participating
laboratories, Bionetics again analyzed the ampul concentrations
against standards freshly prepared from neat materials to verify the
stability of solutions over the period of the study.
SAMPLE MATRICES
The ground water and leachate came from monitoring wells at
industrial/municipal waste dunipsites. The wells were purged and
samples were withdrawn. Waters were autoclaved to eliminate any
biological activity and thoroughly mixed to ensure homogeneity. The
waters were dispensed into bottles with Teflon lined caps and were
shipped on ice in coolers to maintain the sample integrity.
6— 154
-------
DATA HANDLING
IFIVS is used to evaluate the large number of data points in these
studies (approximately 11,000 data points in this study). The data
treatment tests for outliers using both Youden’s laboratory-ranking
test for laboratory outliers and the Thompson T-test for individual
outlier values. Statistical parameters included estimates of mean
recovery, single-analyst precision, overall precision, and tests for
matrix effects.
A questionnaire was sent to each of the ten participating
laboratories requesting information on the operating conditions of
the instrumentation, problems encountered with the method and any
other variables associated with the conduct of the method. Their
comments will be addressed in the final report.
STUDY REPORTS
An in-depth report and a project summary report, will be generated
from this interlaboratory method validation study. These differ in
the depth and detail in the: introduction, conclusions,
recommendations, descriptions, statistical treatment of data and
discussion sections.
At this time, the data on Methods 3510 and 8270 from the ten
participating laboratories are now being received. We anticipate
that all data will be received, reviewed for major errors, processed
through IHVS, a draft report generated and peer reviewed, and a final
report available September 30, 1987.
REFERENCES
1. On July 24, 1986, the USEPA proposed to amend its
regulations concerning groundwater monitoring at landbased
hazardous waste treatment, storage, and disposal
facilities. The amendment would require analysis of
groundwater for a specific list of chemicals, Appendix IX to
Part 264.
6—155
-------
Table 1: Participating Laboratories
ACZ Inc.
Steamboat Springs, CO
Cambridge Analytical Associates
Boston, MA
Lancaster Laboratories, Inc.
Lancaster, PA
James K. Montgomery, Inc.
Pasadena, CA
Pacific Analytical, Inc.
Carlsbad, CA
PEI Associates, Inc.
Cincinnati, 011
Science Applications International Corp.
La Jolla, CA
Southwest Research Institute
San Antonio, TX
Thermo Analytical, Inc.
Ann Arbor, MI
UBTL, Inc.
Salt Lake City, UT
6—156
-------
Table 2: 13 RCRA Appendix IX Analytes Not In Study
p-Benzoquinone No recovery at both 20 and 600 ppb.
Benzenethiol Unstable in spiking solutions, both
methanol and acetone.
Pentachioroethane Decomposes to tetrachloroethene
with less than 2% recovery at 600
ppb.
Resorcinol No recovery at both 20 and 600 ppb.
Flexachiorocyclopentadiene Unstable by thermal decomposition
and solvent reactivity.
p-Naphthoquinone No recovery at both 20 and 600 ppb.
Araniite No recovery at 50 ppb. 6% recovery
at 600 ppb. Four peaks are
produced from 96% pure material.
Hexachiorophene No recovery at 50 ppb. 18% recovery
at 600 ppb.
Dibenzo (a,h) pyrene Insoluble in spiking solvents and
commercially unavailable.
Dibenzo (a,i) pyrene No recovery at 50 ppb. Detection
limit approximately 150 ppb.
n-Nitrosodimethylamine Compound coelutes with solvent. 30%
recovery at 10 ppb.
1-Naphthy lamine 20% recovery at 10 ppb.
Phthalic anhydride No recovery at 100 ppb.
6—157
-------
Table 3: Three Mixes Containing 59 Analytes
Mix 1 in methanol:
n-Nitrosome thyle thylamine
n-Nitrosodiethylamine
Aniline
n-Nitrosopyrrolidine
a, a, Dimethyiphenethylamine
p- Chioroaniline
n-Nitrosodi -n-butylamine
o -Nitroaniline
m-Nitroaniline
2 -Naphthylamine
5 -Nitro-o- toluidine
p -Nitroaniline
1, 2-Diphenyihydrazine
4 -Aminobiphenyl
Nethapyrilene
Benzidine
p -Dime thylaminoazobenzene
3 , 3’ -Dimethylbenzidine
4,4’ -Methylene bjs (2-chioroanilifle)
3 .3, -Dimethoxybenzidine
2 -Methylnaphthalefle
6—158
-------
Table 3: Three Mixes Containing 59 Analytes
(continued)
Mix 2 in acetone
Methyl methane sulfonate
n-Nitrosomorpholine
p -Methyiphenol
2, 6-Dichiorophenol
1,2,4, 5-Tetrachlorobenzene
Isosafrole
Dibenzofuran
n- Nitrosodiphenylandne
Phenac e tin
Pronamide
2 -Acetylaminofluorene
Tris (2, 3-dibromopropyl) phosphate
Benzyl alcohol
Dihydrosafrole
1, 3 -Dinitrobenzene
Chlorobenzilate
Kepone
6—159
-------
Table 3: Three Mixes Containing 59 Analytes
(continued)
Mix 3 in acetone
2 -Methyiphenol
Ace tophenone
n-Nitrosopiperidine
Hexachioropropene
Safrole
2,4, 5 -Trichiorophenol
1 ,4-Dinitrobenzene
Pentach lorobenzene
2, 3 .4, 6 -Tetrachlorophenol
Dipheny laniine
Pentachioroni trobenz ene
2-sec Butyl 4,6-Dinitrophenol
7,12-Diniethylbenz (a) anthracene
3 -Methycholanthrene
Dibenzo (a,e) pyrene
Ethyl methane sulfonate
1,2 - Dibromo -3- chioropropane
1,2 - Dinitrobenzene
1, 3 , 5 -Trinitrobenzene
Diallate
4-Nitroquinoline-N- oxide
6—160
-------
DETECTING COELUTING COMPOUNDS IN GC/MS
Norman Low, Applications Chemist, Hewlett—Packard, Scientific
Instruments Division, Palo Alto, California
ABSTRACT
In recent years, more and more sophisticated analytical tools have been
applied to the analyses of solid waste. High resolution fused silica
capillary column gas chromatography/mass spectrometry has become a
standard tool for examining environmental samples. However, the
complex matrices often encountered makes it difficult even for such a
powerful technique.
The GC/MS target compound approach is commonly used to screen for a
list of suspected chemicals. This, by itself, is insufficient to
characterize the sample. To identify the non—target compounds, one
must perform a library search on the unidentified peaks. Peak
identification and selection is not necessarily easy to do. Automatic
peak selection may be performed on partially overlapping peaks but
completely coeluting peaks will be missed. A skilled chemist may be
able to spot coeluters by examining the mass spectral data. This
manual process is very time consuming as well as dependent on the
chemist’s ability. In very complex cases, compounds will remain
undetected.
We will survey the present techniques for co—eluting and overlapping
compounds along with their limitations. Then we will discuss another
software approach that can be used for these difficult cases. The
process of data examination can be used in either a manual inspection
or automated batch mode. Automatic screening makes it feasible to
process more files in a more consistent manner than can be done
manually.
INTRODUCTION
Solid waste samples can be very complex, requiring highly sophisticated
tools for analysis. For semi—volatiles analysis, fused silica
capillary column chromatography combined with mass spectrometry has
become the standard analytical technique in environmental labs.
In spite of the chromatographic and spectrontetric separation
capabilities of HRGC/MS, instances of coeluting compounds are
encountered more as the rule rather than as the exception. In this
paper, we will review the software techniques available for identifying
instances of coelution for both target and nontarget compounds.
6—16 1
-------
IDENTIFICATION AND QUANTITATICN OF TARGET COMPOUNDS
Many of the established GC/MS methods, such as 624 an 625 are designed
for target compound analysis (determining whether a compound is
present, and if so, in what quantity). Detection of these target
compounds in a complex mixture is relatively easy to do with current
software.
In the AQUARIUS automation software for the Hewlett-Packard RTE data
system, one prepares an identification specification for each target
compound including mass abundances and the expected retention time.
The software allows the operator to enter in the permissible acceptance
window for the mass abundances and retention time. This window is set
globally with the option of individual specification for each target
compound. The passing criteria for each component can be set in terms
of spectral match to the specifications, retention time, or a
combination of spectral match and retention time. The flexibility of
the system in setting the passing criteria allows one to be very
stringent or lenient in identification of target compounds.
One can also look for a specific compound without knowing the retention
time or even injecting the standard if there is a reference mass
spectra. Among the commands available is one which determines the
cross correlation between a reference spectrum and the spectra over a
chromatographic range. A plot is generated where the y—axis value
represents the correlation of the spectra with the reference spectrum.
The amplitude of the peak is an indication of the quality of the match.
Thus, one can locate a target compound or class of compounds within a
complex mixture.
These two methods, or variations thereof, are available on a variety of
mass spectral data systems and are routinely used by chemists in
environmental labs.
RECOGNITION OF COELIJTING PEAXS AMONG TARGET COMPOUNDS
Coeluting peaks are easily recognized when spectral comparisons are
displayed or printed out during target compound analysis. When one
encounters a component that has many more peaks than the reference
standard, even though it has been background subtracted or enhanced,
one presumes that another component is involved. This component may be
isolated by performing a subtraction of the standard or library
spectrum of the target compound from the sample spectrum. Multiple,
coeluting, target compounds may be successively removed to give a
residual spectra. A subsequent library search may be used in an
attempt to identify this residual spectra. In many cases, a reverse
library search may be sufficient to identify the components.
IDENT IFICATION OF NON-TARGET_COMPOUNDS
For high volume laboratories, it is desirable to reduce the amount of
time spent by an analyst to hunt for these coeluters. One can write a
6—162
-------
procedure for testing the cor 1 tarnirtat on index or some other value in
the library search results. If this value indicates the presence of
extraneous peaks in the sample spectrum, d spectral subtraction
followed by a library search of the residual spectrum is to be
performed.
In the EPA Contract Lab Program, the contractor is to identify a target
list plus the ten largest non—target peaks. This is typically done by
examining a total ion chromatogram of the run and labeling the target
compounds. Then, one selects the ten largest non—labeled peaks for a
library search. This is not a difficult process, but when one
considers many labs where there are a couple of mass spectrometers with
autosamplers running around the clock, the peak selection, library
search examination, concentration estimation, and report production can
easily overwhelm a single operator.
In addition to the time—consuming nature of identifying the non-target
peaks for a large number of samples, there is the problem of coelution.
Unless one carefully examines the target compound spectrum while
examining the chromatogram, one can very easily mark off a peak as a
target compound when it consists of a non—target compound in much
larger quantities than the target. One way to reduce errors is to use
some of the alternative peak labeling options in the RTE data system.
One can, for instance, label the target compound peaks with the
determined concentrations. When looking at a relatively large peak
with a small labeled concentration, one is alerted to the possibility
that there is another compound present. While this labeling will help,
one is nevertheless faced with visual examination of large numbers of
samples.
AUTOMATION OF NON-TARGET COMPOUND IDENTIFICATION
To what extent can the process of identifying the largest non-target
peaks be automated? After the qualitative and quantitative analysis,
one has the retention time of the target compounds along with their
starting and ending scan numbers or times. Thus, it is very easy to
remove the target peak with a tangent skim between the first and last
scans. After the target compounds are removed, a simple integration
will select the ten largest peaks for library searching. The chemist
would still have to use his judgment on the library search, but the
quantitation and generation of the Contract Laboratory Program report
form can be automated. This process is not just theoretically
possible, but has been implemented with the RTE mass spectrometer data
system. The procedure that performs these operations has been
available and used by many labs.
The procedure just described performs what an analyst does with visual
examination. However, we come back to the problem of the coeluting
peak. How does one handle the case where a relatively large non—target
compound coelutes with the target compound? In the above procedure, it
will be removed if it is within the beginning and ending scans of the
target compound.
6—163
-------
Let us consider the technique of spectral subtraction. Ordinarily, we
think of choosing a spectra before or after a peak and subtracting it
from the apex or an average spectrum separate partially coeluting
peaks. This process is usually done manually or, with some
compromises, automatically. While it gives relatively good spectra,
it doesn’t help us find the largest non-tarcet peaks.
The latest revision (Revision E) of the W [ F software has a new command,
BSB, which performs background subtraction in a different manner than
described. This command removes a reference spectrum from a portion or
all of a data file rather than from a selected mass spectrum. Thus, if
one removes the target compound spectrum from a retention time window,
all that remains would be the non-target compound spectra. If this
same process is repeated for all the non—target compounds, the
residual total ion chromatogram represents the non—target compounds.
From this point on, it is very easy to select the ten largest non—
target compounds for library searchiriq and quantitation.
[ Although the BSB command has legitimate uses as illustrated above, it
can be abused. Therefore, whenever it is used, a non—changeable flag
is set for the data file indicating that something has been removed.
This flag shows up in the annotation as “BSB” in any chromatograni
(total or extracted) or mass spectrum that is produced from this
modified file. The annotation will be displayed even on spectra from a
region for which this command has not been applied. If this command is
to be used, it is highly recommended that it be used on a copy of the
original data file. Once used, there is no way to reset the flag for
the data file.]
CONCLUSION
We have examined how coeluting peaks may be dealt with in CC/MS.
Naturally, the same techniques have some application for LC/MS as well.
The coelution problem may not be completely solved, but there are
sufficient software solutions to handle some difficult cases. The
selection and quantitation of the largest non—target peaks is no longer
a problem than must be done completely by the analyst. Although a
manual review is necessary (and desirable), identifying the peaks and
completing a quantitation report can be automated.
6—164
-------
THE IDENTIFICATION OF SELECTED SYNTHETIC SURFAC NTS FOR A
COMPLEX WASTE MATRIX USING THERMOSPRAY LIQUID CHROM(YIOGRAPHY-
MASS SPECTROMETRY
Paul C. Goodley, Applications Chemist, Hewlett—Packard, Scientific
Instruments Division, Palo Alto, CA
ABSTRACT
Synthetic nonionic surfactants are present in many manufacturing
processes which ultimately find their way into the environment.
Although surfactants are generally considered nontoxic, they have a
profound effect on the rates of absorption, leaching and/or movements
of other more hazardous materials.
Currently the surfactants are not monitored by GC/MS methods due to
their nonvolatile character. Also, the analytical isolation and
characterization of surfacants from complex matricies requires lengthy
work—ups.
This paper will describe an experimental approach which isolates and
characterizes selected nonionic surfactants from a waste matrix. The
analysis was performed rapidly using thermospray liquid
chromatography—mass spectrometry which is well suited for non—volatile
organic compounds.
6—165
-------
EVALUATION OF METHOD 3640 (GPO CLEANUP)
FOR APPENDIX VUI ANALYTES
Paul J. Marsden, Staff Scientist, S-CUBED, P.O. Box 1620, La Jolla, California
92o3e; James Longbottom, Section Chief, Organics, Environmental Monitoring
and Support Laboratory, U.S. Environmental Protection Agency, 26 W. St. Clair
Street, Cincinnati, Ohio 45268
ABSTRACT
Sample cleanup gel permeation chromatography (GPC) is a useful technique
for removing high molecular weight interferences from sample extracts. The
technique was originally developed for removing lipids from biological
samples, and has since been applied to a wide variety of environmental
samples as Method 3640 of SW-846. The experiences of several laboratories
using GPO for the cleanup of high concentration Superfund and RCRA wastes
demonstrate that Method 3640 may not be sufficiently rugged for application as
a generic cleanup procedure for all wastes. Work is not in progress to validate
use of Method 3640 for the analysis of Appendix VIII analytes in a variety of
wastes.
The GPC recovery of standards of most organic Appendix VIII analytes were
determined by using GC/MS, GC/ECD, GO/FPD, and HPLC-UV. In addition,
the GPC retention volume for each compound was also determined. Analytes
that gave poor recoveries on GPO when the system was calibrated by using
phthalate/PCP were identified. Recoveries of the same analytes were also
determined for standards spiked into hazardous waste in order to evaluate
method suitability for difficult sample matrices. Labeled bis-(2-ethylhexyl)
phthalate ( 3 H) and benzopyrene (140) will be used used as marker compounds
to monitor the effect of sample size on method performance.
Detailed guidance in the use of GPC for environmental samples is being
developed for laboratories because many of the problems associated with the
use of GPO appeared to be caused by inadequate training of instrument
operators. Specific guidance is provided on maximum sample size, sample
preparation techniques, and column packing procedures, and system
maintenance.
INTRODUCTION
Gel Permeation Chromatography (GPO) is a size exclusion separation
technique that is used to remove high-molecular weight interferences from
6—16 7
-------
sample extracts prior to GO or GO/MS analysis. GPC was first used for
environmental samples in the analyses of PCB’s and dioxins in fish tissue. At
that time, the protocol required the samples be etuted from the GPO column
using a solvent mixture of methylene chloride and cyclohexane. Method
performance studies were conducted at that time to establish the elution
pattern and the recoveries of a variety of pollutants using the
cyclohexane/methylene chloride solvent system.
In the last decade, GPO has come to be used as a generic technique for the
cleanup of hazardous waste samples. During the period, it has been observed
that recovery of a limited number of surrogate analytes was unchanged by
using 100 percent methylene chloride as an elution solvent. GPC cleanup
using methylene chloride as an elution solvent is now specified for the analysis
of hazardous waste samples in Method 3640 of the third edition of SW-846. By
virtue of its publication in SW-846, GPC has become the standard method for
the preparation of hazardous waste samples to be analyzed for the presence of
Appendix VIII compounds. Unfortunately, there is little data on GPO method
performance for the majority of Appendix VIII analytes when methylene chloride
is used as a solvent. Nor are there data available for analyte recoveries from
spiked sample matrices.
MATERIALS AND METHODS
All standards were obtained from either the EPA Pesticides and Industrial
Chemicals Repository, Research Triangle Park, North Carolina, or from
commercial sources (Aldrich, Sigma, Chem Service, etc). Methylene chloride
was HPLC or distilled in glass-grade. The GPO separations were
accomplished using an Autoprep 1002A from ABC Laboratories. The column
was 70 g of resieved Bio-beads SX-3 (200-400 mesh) packed using a 1:3
mixture of methylene chloride/cyclohexane. After packing, the column was
flushed with methylene chloride overnight. Each column was calibrated by the
fraction collection technique described in Method 3640; the dump and
collection times were set in order to eliminate 85 percent of the corn oil. All
standards and samples were run using automated collection, and collected
GPC runs were concentrated on a steam bath in 500 mL or 1-L Kuderna-
Danish (K-D) apparatus. Analysis was accomplished using GC/MS (Finnigan
4500, DB-5, Method 8270), GC/ECD or GC/FPD (HP5880; DB-5, DB-608, or
DB-210; Proposed Method 1618). HPLC studies used a Spectra-physics
terniary gradient pump, Model 8800, Spherisorb 5 pm ODS column, and a
Hewlett-Packard diode array detector, Model I 040A. Elution was with 0-100
percent aqueous acetonitrile.
6—168
-------
RESULTS AND DISCUSSION
Initial efforts in this study focused on determining the GPC retention volumes of
Appendix VIII analytes and the precision and accuracy of their recovery using
Method 3640. Mixtures of standards were prepared and run on the GPC using
two different collection techniques. First, 20-mL (four-minute) fractions were
collected and analyzed in order to determine the retention volume of the
various analytes (Figure 1). Second, recoveries of the analytes (Table 1) were
determined in triplicate using the phthalate/PCP calibration described in
Method 3640.
Figure 1 shows that only organophosphorus insecticides elute as early as do
phthalate esters on the GPC. Other classes of organic Appendix VIII
compounds elute after phthalates. It has been our experience that if the
laboratory does not need to determine phthalates or phosphates in a sample,
significantly better sample cleanup can be achieved by calibrating the GPC to
eliminate most phthalates.
Table 1 reports the range of mean recoveries and standard deviations for
individual compounds in a chemical class. These data show that the GPO is
capable of giving good-to-excellent recovery of Appendix VIII analytes, in terms
of both precision and accuracy. If there is a weakness of technique, it is with
strongly polar compounds, as evidenced by the lower recoveries of
phenoxyacid herbicides and aromatic amines. The data presented in this table
were collected using mixtures of standards. Work is in progress to establish
the recoveries of analytes spiked into hazardous waste extracts.
Table 2 summarizes some of the problems that have been observed using
GPO for the cleanup of hazardous waste samples. The problems have been
observed by our laboratory or by other laboratories that use GPC on a routine
basis for preparing samples. The first item listed in the table deals with:
(1) Acidic/corrosive samples. It has been our experience that
one sample with a high concentration of hydrochloric acid will
destroy a column. Washing the sample with acetate buffer or
aqueous sodium chloride reduces the acidity of the extract
and reduces column damage.
(2) Glass wool in the sample is a more insidious problem; small
amounts of glass fibers can severely damage or destroy the
23-port valve that allows semiautomated operation with the
Autoprep GPO. This problem seems to be worse with
industrial wastes, but can occur with any sample when too
much (or unrinsed) glass wool is used in preparing samples.
6—169
-------
PHTHALATE —
ORGANOPHO8PHATE
PESTICIDES
C CORN OIL
V/4 1 PAHe
CHLOROBENZENES
NITROSAMWIES, NITROAROMATICS
AROMATIC AMINES
NITROPHENOLS
CHLOROPHENOLS
ORGANOCHLORP4E
___________ PESTICIDE S/PC B ‘a
HERBICIDES
CCoIIect
30 40 50 60 70
TIME (minutes) C
Figure 1. GPO Elution Curves
-------
Table 1. GPC Recoveries of Standards
Range of Range of
Compound Class Percent Recovery Standard Deviation
Organochiorine Pesticides 84-108 1-15
Phenoxyacid Herbicides 67-80 9-23
Chlorophenols 77-102 1-3
Nitrophenols 77-118 1-7
Aromatic amines 69-97 1-8
Nitroaromatics 93-101 2-4
Nitrosamines 83-99 1-13
Chlorobenzenes 81-96 1-2
Phthalates 89-104 1-4
Haloethers 76-98 1-2
PAHs 79-102 1-13
Table 2. Problems Observed with GPC Cleanup of Hazardous Waste
Cause Symptom Solution
Acid/Corrosive Sample Permanently decreases POP Wash extract with aqueous NaCl
recovery, possibly breaks cal- or buffer prior to GPC
umn cross-linking
Glass Wool in the Sam- Clogs 23-port valve, may require Reduce use of glass wool in
pie factory repair sample preparation and rinse it
thoroughly before use. Centri-
fuge extract with celite prior to
loading the GPC.
Acetone in the Sample Creates dead volume in the cal- Ensure that all acetone is ex-
umn, reducing recoveries, changed to methylene chloride
prior to loading the GPC.
Extract Precipitates in One or more loops show high- Use co-solvents for preparing
the Sample Loop pressure problems, high carry- GPO extracts (butyl chloride,
over between samples. ethyl acetate or benzene).
Column Overloading High pressure shuts off pump, Split the extract between several
cross-contamination of samp’es. loops.
6—171
-------
(3) Acetone in the sample will cause the Blo-beads to shrink,
creating dead volume in the column and significantly
degrading the chromatography. As acetone/methylene
chloride (1:1) is a common extraction solvent for solids, it is
important to emphasize the need to remove acetone prior to
loading the GPC with Method 3640. Some extracts
precipitate in the sample loop after loading (a sample can be
held in the loop up to 24 hours in normal operation of the
GPC). We have found that precipitation can be reduced or
eliminated by using a mixture of methylene chloride with a
second solvent to load sample extracts into the loop. Work at
S-CUBED has demonstrated that ethyl acetate and butyl
chloride are suitable for this application; Jean Czuczwa of
Battelle has demonstrated that benzene can also be used.
(4) Finally, any chromatography system can be overloaded,
including the GPC. Experienced operators learn to split
extracts over several loops, based on their viscosity,
nonvolatiles residue, etc. S-CUBED is gathering data on
column capacity for different types of hazardous waste which
will be made available in a GPC guidance document.
The solutions to GPC problems presented here have not been fully tested and
should not be considered modifications to Method 3640. A proposed section
on sample preparation, as well as a guidance document on applying GPC
cleanup to samples, will be made available by S-CUBED to EMSL-Cincinnati
and OSW for review this year.
CONCLUSION
GPC is a valuable cleanup technique that can be used for the preparation of
extracts of a wide variety of hazardous waste samples. Many of the problems
that have been observed with the technique seem to be caused by improper
training of operators or by loading samples incompatible with the GPC. This is
an interim report on a study that will develop a guidance document for
laboratories that use GPC. The document will include specific requirements for
sample preparation, column capacity for different types of waste, and precision
and accuracy data for the recovery of analytes from the hazardous waste
extracts.
6—172
-------
NOTICE
The research described in this report has been funded by the United States
Environmental Protection Agency under Contract 68-03-3375 with S-CUBED, a
Division of Maxwell Laboratories, Inc., San Diego. This document has not been
subject to Agency review and does not reflect its views. Mention of trade
names or commercial products is for identification purposes on!y and does not
constitute endorsement or recommendation for use.
6—173
-------
USE OF WIDE-BORE CAPILLARY COLUMNS FOR
THE OC ANALYSIS OF ENVIRONMENTAL SAMPLES
Paul J. Marsden, Staff Scientist, S-CUBED, P.O. Box 1620, La Jolla, California
92038; Victoria Taylor, Staff Scientist, S-CUBED, P.O. Box 1620, La Jolla,
California 92038; Joan Fisk, Chief, Analytical Support, Office of Emergency and
Remedial Response, U.S. Environmental Protection Agency, 401 M Street,
Washington, D.C. 20460; L. Don Betowski, Analytical Chemist, Environmental
Monitoring and Systems Laboratory, U.S. Environmental Protection Agency,
922 East Harmon, Las Vegas, NV 89114.
ABSTRACT
Two GC methods for environmental samples utilizing wide-bore (0.53-,mm l.D.)
capillary columns have been demonstrated to be reliable, rugged techniques
that provide data of good quality. One method for the analysis of 20 single
component organochiorine pesticides, toxaphene, and PCBs was developed
and validated for the Office of Emergency and Remedial Response (OERR). It
is currently being considered for use as a Contract laboratory program (CLP)
method. The second method was developed for 26 organophosphorous
analytes during validation of the current SW-846 Method 8140. This study was
conducted as a task from the Environmental Monitoring Systems Laboratory, at
Las Vegas (EMSL-LV) for the OSW. Both methods were validated in separate
single laboratory studies at S-CUBED for the analysis of water, soil, and
hazardous waste samples.
These wide-bore capillary methods offer improved GC resolution over similar
packed column techniques. The improved chromatography results in better
separation of analytes from matrix interference, as well as more confidence in
the identification of analytes by retention times. The larger capacity of wide-
bore versus narrow bore capillary means that samples that would overload a
0.25 mm column give good chromatography on a 0.53 mm column. Minimal
hardware changes are required to install a wide-bore capillary column in a
packed column GC.
Summaries of both dual columns methods are reported along with the bias
and precision of each. The improved chromatography achieved with wide-bore
capillaries was demonstrated using chromatograms of mixtures of Appendix IX
analytes and by comparing capillary versus packed column chromatograms of
selected samples. In addition, the sample capacity of wide-bore capillary
6—175
-------
columns, in terms of nanograms of nonvolatile residue injected, was
determined for several types of environmental extracts.
INTRODUCTION
The state-of-the-art in chromatography has advanced beyond the packed
column technology specified in many analytical methods for the determination
of pesticides by GO and selective detectors. One major advance in GC
chromatography was the development of wide-bore (0.53 mm) fused silica
capillary column. These columns have efficiency and inertness comparable to
narrow bore capillary columns with capacity close to packed columns. The
use of these columns allows: more positive identification of method analytes
based on retention time, better separation of analytes from matrix interference,
and the ability to analyze more analytes in a sample that can now be done
using packed column techniques. At the present time several very similar
packed column GO methods have been developed for the analysis of
pesticides in environmental samples. These include Methods 608, 608.1 and
617 (Method 8080) for halogenated compounds and Methods 614, 622 and
701 (Method 8140) for organophosphates. Validation studies at S-CUBED
have demonstrated that it is possible to consolidate these methods into two
GC procedures by the use of wide-bore capillary. This approach is particularly
useful for multiresidue environmental analyses because a single extraction
sample preparation procedure can be used for both analyses.
MATERIALS AND METHODS
Standards for this study were supplied by the EPA Pesticides and Industrial
Chemicals Repository, Research Triangle Park, North Carolina, the
manufacturers, or were purchased from Chem Services, Westchester,
Pennsylvania. Solvents were pesticide grade or better.
Sample cleanup was achieved using an Autoprep 1 002A by ABC Laboratories
(Columbia, Missouri) and Dial bonded silica cleanup cartridges (Analytichem,
J.T. Baker or equivalent). GC analyses were performed on HP 5880
chromatographs equipped with electron capture (EC), flame photometric (FP)
or nitrogen-phosphorous (NP) detectors. Chromatographic separation was
achieved using one of several 30 m wide-bore (0.53 mm I.D.) fused silica
capillary columns including the DB-5, DB-608, DB-210 (J&W Scientific, Folsom,
California), SPB-5 and/or SPB-608 (Supelco, Inc., Bellefonte, Pennsylvania).
6—176
-------
The columns were plumbed into 1/4-inch packed column injector ports using a
dual column injection tee kit (Supelco). The columns were plumbed into the
detector ports using makeup gas adapters with argon/methane (P-5) for the
ECD, nitrogen for the FPD, and helium for the NPD.
RESULTS AND DISCUSSION
The procedures for the extraction, cleanup and analysis of samples is
summarized in Figure 1. The modular approach has been taken in developing
these protocols. The samples are extracted using one of the four standard
procedures, sonication or soxhlet for solids and liquid/liquid or separatory
funnel extraction for waters. Soil extracts are then subjected to GPC cleanup
using Method 3640. Although GPC retention volumes were somewhat different
for the two classes of compounds (Marsden and Longbottom, this
symposium), both were recovered using the standard phthalate/PCP
calibration. Extracts from all samples are exchanged to hexane and cleaned
up on Diol bonded silica columns. The organochiorine pesticides and PCBs
are eluted from the Diol cartridge using 10 mL of 1:9 acetone in hexane. While
this same solvent system elutes most organophosphates, recovery of all
analytes for Method 8140 requires 10 mL of 4:6 acetone in hexane. GC
analysis is accomplished on two columns for each set of analytes. The
organochiorine pesticides/PCBs are determined using DB-5 and DB-608 (or
SPB-5 and SPB-608) with a temperature program of:
T 140°, 0.5 minute
Ramp 1, 8°/mm to 180° C
Ramp 2, 3°/mm to 275° C
TF 275°, 10 minutes
The organophosphates are determined using DB-5 (SPB-5) or DB-210 with a
temperature program of:
T, 50 C, 1 minute
Ramp 1, 5°/minto 140° C
Hold 1, 140° for 10 minutes
Ramp 2, 10°/mm to 240° C
TF 240° C, 10 minutes
The recovery of the single component organochlorine pesticides are presented
in Table 1. These values are comparable with what is achieved with packed
column methods. lsodrin and HBB are not analytes of the CLP method are
included as method surrogates which are added to every sample, blank and
matrix spike QC sample in order to monitor method performance. These two
6—177
-------
compounds are a significant improvement over dibutyichlorendate that was
used previously as a surrogate in terms of both accuracy and precision of
recovery as well as similarity in analytical behavior with other compounds on
the CLP target compound list.
The recovery of organophosphates from solids using soxhlet extraction is
presented in Table 2. These values are somewhat better than can be achieved
using sonication. The values are somewhat lower and less consistent than
those reported for the organochiorine pesticides, primarily because phosphate
hydrolyze in environmental samples.
CONCLUSION
The use of wide-bore capillary GC analysis is a significant improvement over
packed column techniques. The use of capillary analytical procedures has
been validated for organochiorine and organophosphates in
industrial/environmental samples.
NOTICE
The research described in this report has been funded by the United States
Environmental Protection Agency as a Special Analytical Service of the CLP
and under Contract 68-03-3375 with S-CUBED, a Division of Maxwell
Laboratories, Inc., San Diego. This document has not been subject to Agency
review and does not reflect its views. Mention of trade names or commercial
products is for identification purposes only and does not constitute
endorsement or recommendation for use.
6—178
-------
Figure 1. Analytical Scheme
Sonicate - Method 3550 Water
Soxhiet - Method 3540 Cont. liq/Jiq - Method 3520
Sep. funnel - Method 3510
I GPO - Method 3640 I
I Solvent exchange to hexanel
Diol cartridge cleanupi
Organochiorines Organophosphate
1:9 acetone/hexane 4:6 acetone/hexane
______ 4r
I (C /F(’ fl An tysic I I (C/FPfl An tysis I
DB-5, DB-608 DB-5, DB-210
SPB-5, SPB-608 SPB-5
6—179
-------
Table 1. Recovery of Single Component Organochiorine Pesticides
Matrix Matrix
Compound Water Soil Compound Water
a-BHC 80 80 Endosulfan II 79
fl-BHC 62 77 4,4’-DDD 87 112
-BHC 63 60 Endosulfan sulfate 68 81
7 -BHC 87 95 4,4’-DDT 91 104
Heptachior 69 74 4,4’-MethoxychlOr 73 71
Aldrin 104 118 Endrin ketone 68 77
Heptachlor epoxide 74 98 Endrin aldehyde 55 49
Endosulfan I 101 85 a-Chlordane 82 78
0 Dieldrin 79 104 7 -Ch lordane 85 75
4,4’-DDE 73 67 Isodrin 97 94
Endrin 119 94 HBB 63 69
Average of six recoveries CRQL to I 20x GAOL.
-------
Table 2. Recover of 27Orqanophosphates by Soxhiet Extraction
Compound Recovery High
Azinphos methyl 110 6 87
Bolstar 103*15 79
Ch lorpyrifos 66*17 79
Coumaphos 89±11 90
Demeton 64±6 75
Dimethoate 48±7 98
Diazinon 96*3 98
Dichtorvos 39±21 71
Disulfoton 78-’ 76
Ethoprop 70*7 75
EPN 93±8 82
Fensulfothion 81*18 111
Fenthion 43*7 89
Malathion 81 8 81
Merphos 53 60
Mevinphos 71 63
Monocrotophos NR NR
Naled 48 NR
Parathion, ethyl 80*8 80
Parathion, methyl 41*3 28
Phorate 77*6 78
Ronnel 83*12 79
Sulfotep 72±8 78
TEPP 34 33 63
Tetrachlorvinphos 81±7 83
Tokuthion 40 6 89
Trich loroate 53 53
NR = Not recovered.
6—18 1
-------
A DATA BASE FOR ESTABLISHMENT OF PREANALYTICAL HOLDING TIMES
H. P. Maskarinec, Research Staff Member, J. E. Goodin, Research
Associate, R L. Moody, Research Associate, Analytical Chemistry
Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee
ABSTRACT
Experiments aimed at precisely determining the preanalytical
holding time for soil and water samples containing volatile organic
compounds have been conducted. Three water samples, distilled
water, surface water, and groundwater, as well as three soil
samples, were fortified with a wide range of volatile organics.
Aliquots of these samples were stored at three conditions for up to
one year. Four replicates were analyzed by GC/MS at logarithmic
time intervals. The objectives were to develop procedures for the
preparation of the samples, to establish the holding time for each
compound under each storage condition, and to develop or apply
preservation techniciues for the extension of the holding time.
The procedures for preparation of the water samples include the use
of Tedlar bags for spiking and dispensing. The procedure gives a
precision of less than 5 for almost all compounds. The procedure
also allows for reproducible production of numerous samples. Soil
samples were prepared by addition of Ca. 0.5 mL water (containing
the volatile organics) to a 1 g sample in a VOA vial. The VOA vial
is then directly attached to the purge unit. The precision of this
technique is dependent on the compound, but is generally about
10%. These techniques will be published and available for use in
the production of performance evaluation samples.
Surprisingly, most of the volatile organic compounds are stable for
at least eight weeks, and probably beyond. Those which are not
stable either undergo dehydrohalogenation (tetrachioroethane to
trichloroethylene) or bacterial degradation (aromatics). This can
be reduced or eliminated by the use of HC1 as a preservative. Data
on the stability of preserved samples will be presented. Results
on the stability of volatiles In soil will also be reported.
INTRODUCTION
Preanalytical holding times are defined as the time between sample
collection and analysis. The degree to which a sample changes or
is no longer representative is expected to be minimal during this
period of time. In the absence of a definitive data base
establishing this time period for individual analytes,
recommendations have been made under various regulations limiting
the holding times. These recommendations have been made based on
available information, experience, and intuition. This paper
6—183
-------
reports on the data base being established to support these
recommendations.
For this study, two matrices have been chosen, water and soil.
Analytes considered include a wide range of volatile organic
compounds and a limited number of semivolatile organic compounds
including four nitroorganics. Three samples have been chosen for
each matrix. Distilled water (drinking water), groundwater, and
surface water make up the aqueous samples, while three soils were
chosen to give a variety of soil types. Storage conditions were
chosen to provide an insight into the type of change which might
occur to the samples under normal sampling and transportation to
the laboratory, as well as storage in the laboratory prior to
analysis. Thus, for each individual compound, a preanalytical
holding time can be deduced based on storage condition and sample
type. This information is also valuable for the development of
performance evaluation samples for use by regulatory agencies and
quality assurance officers.
MATERIALS AND METHODS
Materials . All analytee were obtained from either the USEPA
repository (volatile organic compounds) or in the case of
nitroorganics from the U. S. Army Toxic and Hazardous Materials
Agency (USATHANA) as Standard Analytical Reference Materials
(SARMs). Volatile organics were obtained In solution and used as
received. Nitroorganica were dissolved in ethanol prior to
addition to the samples. All solvents were distilled—in—glass
grade.
Methods . Multiple aliquots of volatile fortified water samples
were prepared as follows. Four liters of the sample were added to
a Tedlar gas Bample bag (SKC, Inc. Eighty Four, Pa.). The methanol
solution of the volatilea was then added to the bag and the
contents mixed thoroughly. The samples were then dispensed into 40
raL vials using a length of Teflon tubing sufficient to reach the
bottom of the vial. Thus, while the samples were aliquotted, no
headapace was formed in the bag, and minimal mixing of the sample
with air occurred during the filling of the vials.
Soil samples containing volatile organic compounds were prepared in
a similar manner. The soils were all obtained after air drying. A
measure of the amount of water required for saturation was then
made. This amount ranged from 80—1002 by weight. Approximately
half of this amount was added to the soil three days prior to the
addition of the volatiles. This was done In order to activate any
microbial activity and to assure a steady state microbial
population on Day 0. The remaining water, containing the volatile
organics, was added from a Tedlar bag as described above.
6—184
-------
For the seuiivolatile organic compounds, addition of the solution
directly to the sample was performed. For the soils, prior
incubation with water was again used. For the water samples, 1
drain vials were used, and for the soil samples, 4OinL vials were
used. All samples were stored immediately after preparation.
For the analysis of volatile organic compounds, standard Contract
Laboratory Program (CLP) methods were employed. Gas
Chromatography/Mass Spectroinetry (GC/MS) was used exclusively. For
the analysis of the nitroorganics, standard USATHANA methods were
employed (1, 2). These methods utilize High Performance Liquid
Chromatography (HPLC) with ultraviolet absorbance detection. Water
samples were analyzed di’ectly, while Boil samples were extracted
in an ultrasonic bath with acetonitrile.
RESULTS AND DISCUSSION
Volatile organic compounds in water . Analysis of the three water
samples was carried out at 0, 3, 7, 14, 28, 56, and 112 days. At
each time point, four replicates were analyzed at two levels of
fortification. The results are summarized In Tables 1—6. Each of
the water samples behaves slightly differently. In the case of the
distilled (drinking) water sample, almost all of the components are
stable through 28 days. The exception is 1, 1, 2,
2—tetrachioroethane, which degrades to trichloroethylene. This is
seen much more quickly in the case of room temperature storage than
under refrigeration, although the refrigerated storage does not
prevent this degradation. Otherwise, there are no significant
changes through Day 28.
For the groundwater sample, no changes were noted through Day 28
which were significant, with the exception of ethyl benzene,
styrene and o—xylene. This is probably due to chemical oxidation,
although groundwater in general may be more likely to remain stable
than other water samples. This is a conclusion which must be
validated by the use of additional samples.
For the surface water sample, several significant changes were
noted. The aromatic compounds benzene, toluene, ethyl benzene,
xylene and styrene all appear to be degraded at rates far greater
than distilled water. Again, this is probably due to oxidation of
the compounds and subsequent degradation. This oxidation may
either be due to microbial or chemical reactions, although
microbial action is more likely to be the cause in this case.
In general, it is perhaps surprising to note that there appears to
be little or no loss of these compounds due to volatilization, at
least at refrigerator temperature. This is supported by the fact
that even the gases (bromomethane, chioroethane) are not lost at
rates greater than the less volatile compounds. This is an
6—185
-------
indication that the containerization of water samples for volatile
organic analysis using septum capped vials with no headspace is
sufficient as a sampling/storage technique.
All of the degradation which occurred during the original study was
due either to dehydrohalogenatlon or to microbial oxidation.
Therefore, the addition of HC1 to the samples was considered to be
an appropriate preservation technique for volatile organics in
general. This has been suggested in the past, particularly for
drinking water. A second experiment was performed in order to
evaluate the ability of acid preservation to extend the holding
time of volatile organics in these samples. The study was
repeated, at one concentration level, under refrigeration, at day
0, 14, 28, and 56 for the three samples. Data for the distilled
water and groundwater through Day 28 are presented in Table 7.
The extent of dehydrohalogenation of tetrachioroethane has clearly
been reduced, while no significant deleterious effects of the HC1
were notede For the groundwater sample, the data indicate a
significant reduction in the rate of degradation of the aromatics.
For both samples, higher levels of the early eluting components
were found on Day 0 than either Day 14 or 28. We believe that this
is due to a bias in the Internal standard rather than to
degradation, since the target level in this case was 50 pg/L, and
no change was noted from Day 14 to Day 28. The data for the
surface water are not yet available, although if the data support
the use of H 1 as a preservative for volatile organics, it would
appear that the preanalytical holding time for these compounds
could be increased to 28 days. This would have a dramatic impact
on the ability of the contract laboratories to handle large numbers
of analyses without exceeding the holding times.
Volatile organic compounds In soil . For the purpose of
establishing the holding times of volatile organic compounds in
soil, a preliminary experiment was carried out in order to
determine the proper containers for longer term storage. Three
types of bottles were used: 40 mL septum—capped vials, 120 niL
septum—capped vials (Shamrock Glass), and 120 niL wide—mouth vials
with Teflon cap liners (1 —Chem). These containers were chosen on
the basis of current use in the sampling of soils for volatile
organic compounds. Replicate samples were prepared for each type
of vial containing l00,u g/kg of a range of volatiles. The samples
were analyzed In triplicate at day 1, 3, and 8. Storage was at
refrigerator temperature. In the case of the 40 niL vials, the
entire 5 g sample was analyzed using a modified purge and trap
device capable of accepting the vial directly. The larger vials
were analyzed after weighing a 5 g aliquot into the purging
device. The results of thl8 study are given in Tables 8—10.
The results indicate that for all of the vials, recovery was not
quantitative even on Day 1. Not surprisingly, recovery dropped off
6—186
-------
with increasing molecular weight in all cases. The best recovery
was seen using the 40 mL vials. Recovery was essentially similar
with the two larger vials, and lower than the 40 mL vials. The
more volatile components were affected more by the transfer and
weighing Steps than the later eluting components. It is likely
that recovery can be improved by the heating of the purging device,
although this technique has problems as well. However, the losses
which apparently occur in the weighing and transfer steps are not
easily eliminated. These results recommend the use of preweighed
vials with a know mass of soil added, and no further operations on
the sample before purging.
With respect to the holding times, all samples showed degradation,
even on Day 3, and dramatic losses were seen on Day 8. The 1,
2—dichloropropenes degraded rapidly, and had disappeared by Day 8.
Many of the compounds showed significant deterioration in all
containers by Day 8. The losses are attributable in part to
microbial degradation, but are probably mostly due to
volatilization. In all cases, headspace will be present in a vial
containing soil. While the volatiles come to equilibrium rather
quickly in a sealed container, the caps are also permeable.
Therefore, gas exchange will occur between the headspace and the
outside of the container, causing the loss of volatile organics.
This effect is not seen when using the vials for water samples,
since there is no headspace, and diffusion is five orders of
magnitude slower in water than in air. A further indication of the
diffusion through the vial cap is the fact that methylene chloride,
not present In the original samples, had accumulated in the samples
by Day 8. This effect was again most dramatic with the wide mouth
bottles. These bottle caps actually tend to loosen as the
temperature is reduced in the refrigerator. These particular
samples were stored In a refrigerator containing methylene chloride
solutions so that the effects seen here are not unexpected. It
appears that additional development will be required in order to
assure that volatile organic analysis of soil samples i reliable
and meaningful.
The data set for the explosives in water is also complete, although
this data will only be summarized here. Aromatic compounds degrade
faster than aliphatics, refrigerated storage is essential for
maintaining these compounds, and only slight degradation 18 noted
for any of the compounds at refrigerator temperature after 28
days. Here again, preservation with HC1 may extend the holding
time well beyond 28 days.
An additional feature of this work is the ability to evaluate
laboratory performance over time with respect to the analysis of
these compound/matrix combinations. Initially, it was thought that
when lowered concentrations of a particular analyte were noted, the
holding time would be declared to have been exceeded. However, the
6—187
-------
combination of slight changes in the internal standard or
calibration with normal day to day variability in the methods, can
cause erroneous assumptions about the stability of the compounds.
In many cases, when one compound is found to have a lowered
concentration, all of the compounds integrated against that
particular internal standard are lowered by about the same amount.
These same concentrations may then be elevated at the next time
interval. It now appears that the best approach to determining the
holding times from a statistical standpoint may be to compare the
found concentrations with the known starting concentrations, using
all of the replicates to develop a precision statement on the
method. This approach Is currently being taken.
10
U LU
-------
Table 1. Volatile organics in water, low level in distilled water
Storage cond. 1 (4°C)
Storage Cond. 2 (25°C)
Name Day 0 Day 28
Day 0 Day 28
Methylene chloride 53 49
53 57
1,1-Dichioroethene 58 48
58 59
l,l-Dichloroethane 64 57
64 53
Chloroform 60 53
60 53
Carbon tetrachioride 51 50
51 53
l,2-Dichloropropane 50 49
50 47
Trichioroethene 66 66
66 91
Benzene 50 46
50 45
l,l,2-Trichloroethane 47 49
47 38
Bromoform 34 45
34 43
l,l,2,2-Tetrachloroethane 42 33
42 0
Tetrachloroethene 50 48
50 44
Toluene 50 50
50 49
Ch lorobenzene 52 53
52 54
Ethylbenzene 49 49
49 48
Styrene 52 46
52 47
Q-xylene 54 49
54 50
Table 2. Volatile organics in water, high level
in distilled water
Storage cond. 1 (4°C)
Storage Cond. 2 (25°C)
Name Day 0 Day 28
Day 0 Day 28
Methylene chloride 464 562
464 533
1,1-Dichioroethene 473 502
473 616
l,l-Dichloroethane 515 596
515 552
Chloroform 489 533
489 487
Carbon tetrachloride 446 481
446 453
l,2-Dichloroproparie 421 504
421 481
Trichioroethene 287 710
287 663
Benzene 400 455
400 432
1,l,2-Trichloroethane 440 550
440 281
Bronioform 619 678
619 631
1,1,2,2-Tetrachioroethane 531 85
531 2
Tetrachioroethene 370 371
370 294
Toluene 383 440
383 405
Chlorobenzene 382 416
382 387
Ethylbenzene 378 376
378 340
Styrene 394 361
394 341
o-xylene 398 411
398 379
6—189
-------
Table 3. Volatile organics in water, low level in ground water
Name
Storage
cond. 1 (4°C)
Storage Cond. 2 (25°C)
Day 0
Day 28
Day 0 Day 28
Methylene chloride
66
46
66 44
1,1-Dichioroethene
61
49
61 50
l,l-Dichloroethane
66
61
66 56
Chloroform
61
64
61 59
Carbon tetrachioride
56
50
56 47
1,2-Dichloropropane
34
61
34 56
Trichloroethene
108
90
108 92
Benzene
55
50
55 47
l,1,2-Trichloroethane
55
62
55 59
Bromoform
50
56
50 46
l,1,2,2 -Tetrachloroethane
48
58
48 58
Tetrachloroethene
53
69
53 67
Toluene
54
52
54 44
Ch lorobenzene
55
49
55 42
Ethy lbenzene
51
24
51 14
Styrene
63
12
63 2
Q-xylene
53
12
53 12
Table 4. Volatile
organics in
water, high level
in ground water
Name
Storage
cond. 1 (4°C)
Storage Cond. 2 (25°C)
Day 0
Day 28
Day 0 Day 28
Methylene chloride
396
474
396 508
l,l-Dichloroethene
383
667
383 617
l,l-Dichloroethane
389
727
389 748
Chloroform
385
705
385 725
Carbon tetrachioride
551
763
551 684
l 2-Dichloropropane
590
612
590 698
Trichioroethene
570
631
570 730
Benzene
570
560
570 585
1,1,2-Trichioroethane
584
647
584 808
Bromoforni
583
756
583 923
1,1,2,2-Tetrachioroethane
544
1190
544 980
Tetrachloroethene
531
577
531 521
Toluene
550
605
550 633
Chlorobenzene
547
492
547 526
Ethylbenzene
Styrene
o-xylerie
532
535
531
312
187
173
532 285
535 188
531 169
6—190
-------
Table 5. Volatile organics in water, low level in surface water
Storage cond. 1 (4°C) S
torage Cond. 2 (2.5°C)
Day 0 Day 28
Name Day 0 Day 28
Methylene chloride 69 62
69 62
l,l-Dichloroethene 53 53
53 50
1 .,l-Dichloroethane 64 57
64 54
Chloroform 69 62
69 59
Carbon tetrachloride 47 50
47 30
l,2-Dichloropropane 55 51
55 53
Trichioroethene 60 56
60 70
Benzene 56 50
56 49
1,1,2-Trichioroethane 59 53
59 55
Bromoform 55 48
55 45
1,1,2,2-Tetrachioroethane 54 52
54 31
Tetrachioroethene 57 52
57 50
Toluene 54 48
54 45
Ch lorobenzene 56 48
56 44
Ethy lbenzene 52 5
52 29
Styrene 58 0
58 15
o-xy lene 58 47
58 43
Table 6. Volatile organics in water, high level
in surface water
Storage cond. 1 (4°C)
Storage Cond. 2 (25°C)
Day 0 Day 28
Name Day 0 Day 28
Methylene chloride 556 603
556 509
1,1-Dichioroethene 420 543
420 419
1,1-Dichioroethane 608 650
608 545
Chloroform 586 608
586 508
Carbon tetrachioride 690 542
690 409
1,2-Dichloropropane 560 569
560 528
Trichioroethene 507 566
507 531
Benzene 480 555
480 482
1,1,2-Trichioroethane 651 622
651 609
Bromoform 775 739
775 701
l,1,2,2 Tetrach1oroethane 818 781
818 537
Tetrachioroethene 416 438
416 295
Toluene 473 487
473 404
Chlorobenzene 488 483
488 394
Ethylbenzene 427 260
Styrene 495 416
Q-xylene 483 475
427 299
495 334
483 383
6— 191
-------
Table 7. Volatile organics in HCL-treated waters
Water 1
(distilled
water)
Water 2 (ground
water)
Day 0
Day 14
Day 28
Day 0
Day 14
Day 28
Broniomethane
65±1
47±3
49±1
66±2
47±1
48±2
Chioroethane
78±1
59±2
63±3
78±2
58±3
66±2
1,1 -Dichioroethene
71±1
58±2
58±1
71±3
50±5
58±2
1,1-Dichioroethane
77 ± 1
64 ± 2
63 ± 2
74 ± 1
64 ± 2
64 ± 1
Chloroform
70±1
57±2
60±2
69±2
58±2
60±1
Carbon tetrachioride
63 ± 1
63 ± 1
59 ± 2
62 ± 1
62 ± 3
59 ± 3
1, 2-Dichioropropane
62 ± 1
59 ± 2
53 ± 1
62 ± 1
58 ± 1
54 ± 1
Trichloroethene
59±1
53±2
50±0
62±1
54±1
53±1
enzene
60±2
53±2
53±1
57±1
49±2
50±1
11,2-Trichioroethane
62±3
60±3
59±2
63±3
61±3
61±0
Bromoform
54±2
56±2
59±2
55±2
58±2
61±2
1,1,2,2-Tetrachioroethane
61 ± 4
61 ± 4
59 ± 1
64 ± 2
63 ± 1
61 ± 2
Tetrachioroethene
57 ± 1
49 ± 1
46 ± 1
57 ± 1
49 ± 1
45 ± 0
Toluene
57±1
50±1
49±1
55±1
48±1
49±1
Ch lorobenzene
54±0
50±1
49±1
55±1
51±1
50±1
Ethy lbenzene
54±1
48±1
46±1
51±1
45±1
46±0
Styrene
44±1
39±2
36±2
41±1
35±1
32±0
Q-xy lene
54±1
48±2
45±1
51±1
46±1
45±0
6—192
-------
TabLe 8. Volatiles in soil (storage in 40 mL septum-capped vials). n 3, CONC = gig/kg.
Number
Compound
Day
1
Day
3
Day
8
Conc.
Sid dev.
Cot -ic.
Std dev.
Conc.
Std dev.
].
t—l 2-Dichloroethene
8739
1.64
86.94
4.37
54.42
9.99
2
1,2-Dichloroethane
67.54
22.69
86.68
5.72
64.65
5.83
3
1,1,1-Trichloroethane
81.84
3.32
81.25
9.92
67.89
13.07
4
Bromodichieromethane
64.33
6.82
67.17
6.67
58.99
5.19
5
—1,3-Dichloropropene
44.21
1.07
21.73
11.01
1.26
0.61
6
c—1,3-Dichloropropene
53.34
0.66
19.43
12.00
1.48
1.00
7
Bensene
78,38
0,41
82.40
4.01
58.08
7.34
6
Bromoform
42.81
3.98
41.68
6.44
29.87
1.25
9
11,2 ,2-Tetrachloroethane
49.72
7.26
57.55
8.16
44.86
2.51
10
Toluene
65.11
8.95
74.15
2.62
45.31
5.83
11
Ethyl benzone
51.22
2.40
55.42
1.22
32.13
8.06
12
tlethylene chloride*
4.47
0.03
6.97
0.55
25.06
0,39
13
Acetofle**
189.42
47.65
137.99
21.69
77.95
10.70
*Not added to soil..
**Not, added, but found in blank.
6—193
-------
Table 9. Volatiles in soil (storage in wide-mouth bottles). n = 3, CONC ag/kg.
Number
Compound
Day
1
Day
3
Day
8
Conc.
Std dcv.
Conc.
Std dcv.
Conc.
Std dev.
1
t—1,2 —Dicbloroethene
256.43
409.45
12.82
7.31
2.97
1.93
2
1,2—Dicbloro.tbane
54.57
5.13
45.23
11.18
24.26
10.89
3
1 ,1,1—Trichloroethane
14.00
2.54
10.17
4.61
2.54
1.60
4
Bromodichloro.ethane
39.81
4.41
29.39
9.94
13.74
7.41
5
t—1 ,3-Dichloroprop.ne
26.78
1.96
17.68
6.67
2.34
0.56
6
1,3Dichloropropene
37.76
1.92
24.30
7.45
1.04
0.23
7
B.nz.ns
28.92
5.60
21.29
7.86
15.59
1.45
8
Brooform
39.33
0.16
37.28
3.64
17.98
3.78
9
l 1,2.2—T.trachloro.tbsn.
53.55
0.78
61.02
3.91
43.68
3.63
10
Toluins
27.76
3.40
29.21
5.86
27.65
3.88
11
Ethyl bsns.ns
23.05
2.41
22.58
6.24
10.62
3.22
12
Msthylen. chlorids
16.15
19.76
87.44
67.32
442.77
199.70
13
Aceton.**
115.63
25.43
149.06
44.98
165.14
30.20
*Not added to soil.
**Jot added, but found in blank.
6—194
-------
Table 10. Volatiles in soil (storage in 120 rnL septum-capped vials). n = 3, CONC = ag/kg.
Number
Compound
Day 1
Day 3
Day
8
Conc.
Std dev.
Conc.
Std dev.
Conc.
n = 1
1
t-1,2—Dichloroethene
21.27
0.32
25.92
0.73
17.56
2
1,2-Dichioroethane
54.98
4.62
68.09
9.10
49.19
3
1.1,1—Trichioroethane
14.21
0.28
17.63
0.43
16.36
4
Bromodichioromethane
39.68
2.49
44.60
2.73
39.56
5
t-1,3-Dichloropropene
28.29
3.15
24.62
1.53
1.99
6
c—1,3—Dichloropropene
39.58
3.07
30.70
3.52
0.56
7
Benzene
30.49
1.35
35.65
0.57
27.43
8
Bromoform
40.51
2.48
40.73
1.54
29.22
9
1,1,2,2-Tetrachioroethane
54.98
2.19
60.35
0.75
56.01
10
Toluene
72.93
5.61
38.34
2.57
28.18
11
Ethyl benzene
49.76
2.83
29.59
2.46
21.65
12
Methylene chloride*
5.76
1.86
4.12
0.32
18.04
13
Acetone**
122.82
24.30
116.22
11.93
47.90
*Not added to soil.
**Not added, but found in blank
6—19 5
-------
THE DETERMINATION OF CHLOROPHENOXYACID HERBICIDES
BY LIQUID CHROMATOGRAPHY USING CABBON-14 TRACERS
R. Merriweather, Research Associate, W. M. Caidwell, Lab Technician,
M. P. Maskarinec, Research Staff Member, J. E. Caton, Group Leader,
Analytical Chemistry Division, Oak Ridge National Laboratory, Oak
Ridge, Tennessee
ABSTRACT
The determination of 2,4—dichiorophenoxyacetic acid (2,4—D) and
2—(2,4,5--trichloro) phenoxypropionic acid (Silvex) in both solid and
liquid environmental samples by several liquid chromatographic
approaches has been studied. The extractive recovery was evaluated
for both a soil matrix and water using Carbon—14 tracers of both 2,4—D
and Silvex. For water, the extractive recoveries were evaluated as a
function of both pH and ionic strength. Recoveries from the soil
matrix were evaluated for both Soxhiet extraction and ultrasonic
extraction. This evaluation of the sample preparation methods with
the assistance of Carbon—14 tracers allowed the distribution of each
herbicide to be estimated in all phases of the procedure.
Several approaches to the liquid chromatographic determination of
these herbicides were studied. One approach employed a heavily loaded
(18% carbon by weight), completely endcapped C18 reversed phase column
eluted with an aqueous solution of 30% (v/v) acetonitrile that
contained 0.001 M concentration of a pH 4.6 acetate buffer. The
effluent from thu column was monitored at 280 nm. The minimum amount
detected by this approach is 100 ng of either Silvex or 2,4—D in the
injected sample. A second approach avoids the use of a buffered
eluent by employing a very high performance reversed phase column and
an aqueous methanol (27% v/v) eluent. This second column has been
monitored at multiple wavelengths with a diode array detector. The
sensitivity of detection can be greatly increased at wavelengths below
240 nm; however, very short wavelengths (<230 nm) may not be
practical for real environmental samples.
There will be some discussion of interferences and sample clean—up
procedures. However, an important advantage of these methods is that
most organic compounds that are either higher in molecular weight or
less polar than the herbicide acids will be retained longer by the
reversed phase column.
6—197
-------
NOVEL EXTRACTION SOLVENTS FOR ENVIRONNENTAL SP NPLES
Linda Sheldon, Ruth Zweidinger, Analytical and Chemical Sciences,
Research Triangle Institute, Research Triangle Park, NC
ABSTRACT
Extraction of organic compounds from environmental matrices occurs
when the solvent is brought into contact with the sample. If the
analyte of interest has a higher affinity for the extracting solvent,
it will partition into the solvent. The percent of analyte extracted
into the solvent phase (%E) can be calculated by
%E= 1OO
+
wo
where is the distribution coefficient of an analyte between the
solvent and the martrix and W and W are weights of the sample and
solvent, respectively. From his eqiS ation, it is obvious that high
analyte recoveries will depend upon either a large distribution
coefficient or a small sample:solvent ratio.
The choice of solvent is critical to liquid—liquid extraction (LLE)
procedures. As discussed above, the solvent must have a large
distribution coefficient for the compounds of interest. Extensive
listings of partition coefficients from water are available for
various solvent systems. Along with having a high extraction
efficiency for the analytes of interst, the extracting solvent should
be immiscible with the sample, not contain contaminants which might
compromise subsequent analysis, be chemically inert, be specific for
the compounds of interest, and be amenable to the analytical method of
choice. Since solvent evaporation may be employed to further
concentrate sample extracts, solvents with low boiling points are
preferred.
Solvents such as benzene and chloroform were used extensively in the
past, but now have limited applicability because they have been
identified as carcinogens or cocarcinogens. Nethylene chloride has
become the solvent of choice for most procedures. Unfortunately,
methylene chloride is not a universal solvent and does not have high
distribution coefficients for many polar compounds. Alternate
solvents and solvent systems have been tested for this purpose.
Nethyl—t—butyl ether has been demonstrated as a good solvent choice
for phenols and organic acids. Data for a variety of solvent systems
will be presented.
Although part of the research described was funded wholly or in part
by the U.S. Environmental Protection Agency through Contract Numbers
68—03—2704 and 68—03—2845, it has not been subjected to the Agency’s
required peer and administrative review and does not necessarily
reflect the views of the EPA, and no official endorsement should be
inferred.
6—199
-------
AN EXPERT SYSTEM ‘iX) AID IN USING USING SW-846
Albert D. Bethke, Senior Research Computer Scientist, Alvia Gaskill,
Jr., Robert Truesdale, Research Triangle Institute, Research Triangle
Park, NC; Nancy Rothman, ERCO, Cambridge, z i
ABSTRACT
An automated system is being developed to aid in the use of the manual
“Test Methods for Evaluating Solid Waste” (SW—846). This system will
assist users in choosing the appropriate methods from Sc —846 to meet
RCRA regulatory requirements, such as determining if a waste is
hazardous, analyzing a sample for specific compounds, analysis of
ground water monitoring samples, analyzing used oil to be sold for
burning in non—industrial boilers, etc.
This system will provide a basis for the interfacing of analytical
requirements, techniques and methods with regulatory requirements and
guidance to enable the regulated community to better understand the
analytical requirements and more easily interpret analytical data.
Additional benefits of the system include obtaining a greater degree
of uniformity in the selection of methods fot similar tasks, and
making the quality control requirements more accessible to the
laboratory chemist and the regulated community.
So far, a demonstration system has been developed. This demonstration
system handles only one of the above types of problems: analyzing
used oil to be sold for burning in non—industrial burners.
The user interacts with the system through a sequence of menus. As
the user answers the questions, the system provides guidance in
understanding the applicable regulations and in understanding the
tradeoffs involved in the various decisions he must make about testing
his used oil.
The system is designed to run on an IBN PC/AT and is written in
ARITY/PROLOG.
6—201
-------
LIQUID CHROMATOGRAPHY MASS SPCTROMETRY: AN EVOLVING TECHNIQUE
Drew Sauter, A. D. Sauter Consulting, Henderson, Nevada; R. K.
Mitchum, U.S. Environmental Protection Agency, Las Vegas, Nevada
ABSTRACT
Historically the U.S. Environmental Protection Agency has utilized
CC/MS as the primary technique for environmental analyses. Clearly,
CC/MS methods are limited to those analytes that are both volatile
and chromatographable. These requirements can severely hinder or
make MS environmental characterization of compounds of regulatory or
other interest either impossible or less than ideal. Because
accurate regulatory work or accurate characterization of
environmental phenomena can require characterization of labile
organic compounds or both labile and non—chromatographable
pollutants, GC/MS can give only a partial view of the analytes being
characterized.
Liquid chromatography/mass spectrometry offers one of the obvious
solutions to this dilemma. Since the mid—seventies, various LC/MS
interface designs have been proposed and tested which have given
varying degrees of success. The technology has not been widely
disseminated or accepted in the environmental monitoring community.
The purpose of this article is to discuss the potential of newer
LCJMS technology called, generically, particle beam liquid
chromatography—mass spectrometry (PB/LC/MS), to show why such
technology has the potential to enhance mass spectral based
environmental monitoring and to indicate where the implementation of
such technology could help fulfill scientific and regulatory
environmental needs.
INTRODUCTION
Particle beam liquid (PB/LC/MS) technology was introduced by
Willoughby and Browner (i). More recent efforts, Apffel and
Willoughby (2,3) have demonstrated significant important
development. These include production of NIH/EPA library matchable
electron impact mass spectra, and sensitivity in the nanogram range
for a wide range of non—volatile and volatile (or gas
chromatographable) analytes.
These properties coupled with the ease of interfacing PB!LC/MS
technology directly to quadruple mass spectrometers routinely used
for GS/MS analysis with little or no ion source modification will
result in rapid, cost effective deployment of such technology.
Although the practicality of this concept requires significant
further rigorous testing, it appears that such is the case.
6—203
-------
The liiportance of a rugged, routine LC/MS technique for
environmental analytical organic (and potentially inorganic arid/or
oreganometallic) chemistry is of practical significance to
environmental, biomedical, defense and drug studies related to
characterization of environmental contaminants, drug inetabolites,
characterization of chemical deterent agents, and analysis of
explosives. PB/LC/MS techniques also have application in routine
targeted compound analysis scenarios related to analysis of U.S. EPA
R RA Appendix VIII and IX lists, and the priority pollutants. All
of these applications have in common the analysis of either labile
or labile and non—chromatographable analytes to characterize the
phenomena or interest. The routine availability of such technology
could, therefore, have a major impact in the analytical chemistry of
environmental contaminants, and eventually reorient how such
characterizations are performed.
The purpose of this article is to briefly show recent applications
of PB/LC/MS to the appendix VIII and IX compound list and to discuss
the implications of such results for environmental testing in
general. While it is clear that such technology is still evolving,
it appears the refinement of LC/MS interfaces could be as
significant to the dissemination of LC/MS technology as the glass
jet separator was to the routine application of CC/MS technologies.
Because the availability of routine electron impact LC/MS technology
has the potential to analyze both non—volatile or labile analytes in
one MS experiment, LC/MS technologies could assist with providing a
more accurate approach to characterization of environmental
phenomena, hazardous wastes, hazardous waste sites and other
chemical mixtures or processes.
EIIfENTAL
All data reported were provided by R. C. Willoughby, Pittsburgh,
Pennsylvania, and A. I). Sauter, (under contract to the U.S. EPA), or
J. A. Apffel, Hewlett Packard in Palo Alto, California.
Experimental Parameters
Consisted of a 100 X 2.1mm 3um Spherisorb ODS2 with a mobile phase
of A(.O1M NH4OAC) and B(Acetonltrile) using a 0—902 B/S am gradient
at a .4 mi/sin flow rate. The MACIC helium flow rate was 1.2L/inin
with the desolution chamber operated at 40°C. The MS (Hewlett
Packard) was scanned at 50—500 amu/sec with the source temperature
at 330°C. The quantity injected was lOOng/cotaponent.
The Thermabeam experiments utilized a 15cm X 2.0mm C18 reverse phase
HPLC column operated at ambient temperatures. A gradient of 20% .1M
NH 4 OAC to 902 methanol at 42/ainate at a flow rate of .5m1/min was
utilized. The MS (Model EL750) was operated at 70eV, scanning the
mass range 70—550 amu with an ion source temperature of 200—350°C.
6—204
-------
Thermospray experiments utilized a 25cm X 4.6mm C18 reverse phase
column operated at ambient temperature. A solvent gradient of 10%
.1)1 NH 4 OAC to 80% methanol at 2%/mm with a flow rate of 2.ml/mln
was utilized. The MS (Model El—l000) was scanned from 70—550 amu.
RESULTS AND DISCUSSION
The thermospray technique employs a thermal nebulizer. Although
there are several different designs offered by instrument
manufacturers, they all have in common the entertainment of ions
formed via charge transfer from an ionic buffer. There have been
variations of the thermospray technique which allow electron
ionization as well as discharge ionization. Typically the technique
has applicability to those compounds with proton affinities greater
than ammonia derived from the ammonium acetate buffer commonly
employed.
Both Thermabeam and MAGIC use a nebulization technique, a
therinospray in the former and a gas nebulizer in the latter.
Ionization In both techniques occurs after desolution of the analyte
followed by electron ionization. A nebulizing Interface has also
been used in conjunction with an atmospheric pressure Ionizer using
a corona discharge as the ionization mechanism (4).
It Is not the purpose or intent of this article to make judgment on
the relative merits of these approaches, rather, it is our purpose
to demonstrate exciting recent results acquired by two separate
approaches and to demonstrate where such technology or related
technology is required to ensure the fundamental application to the
characterization of hazardous wastes, hazardous waste sites and
other environmental media. Obviously, such technology has apparent
merit to many aspects of analytical organic (and potentially
inorganic) chemistry, as well.
Because PB/LC/MS interfaces eliminate most of the LC solvent or
sample only a small portion of the solvent, subsequent ionization of
the analyte in the source can take place at pressures conducive to
the generation and observation electron ionization mass spectra. In
Figure 1 we show 70eV electron ionization mass spectra for the
selected Appendix IX compounds, warfarin, thiram, and
dichlorobenzidine, and the RIC for several other analytes, all
generated with the MAGIC interface. In Figure 2 the spectra of
propyithiouracil, 2,4- D and brucine as well as the RIG for other
selected Appendix VIII compounds were generated with the Thermabeam
Interface. We draw attention to the chemical and structural
diversity of the analytes for which apparent 70eV electron
Ionization mass spectra have been generated. Comparison of mass
spectra with the NIH/EPA library is shown for selected analytes,
Figure 3, as well as the difference spectra. Without more detailed
6—205
-------
statistical analysis, it is apparent that particle beam technologies
show good fragmentation spectra of numerous analytes.
For comparison, we show Thermospray mass spectra acquired in a
recent U.S. EPA funded project (5), Figure 4. This study performed
a very preliminary comparison of the particle beam LC/MS tecbni ue
and thermospray analysis of Appendix VIII listed analytes. The
difference in information content Is apparent between the two
techniques, compare Figures 2 and 4.
The difference between thermospray and particle beam LCIMS
technology can be seen in application to the Appendix VIII listed
analyte, aurainlne—O. In Figure 5, we show a 3—dimensional plot of a
particle beam experiment with reverse phase gradient elution LC
conditions. In this experiment, auramlne—O was identified at
nanogram levels, but another structurally related analyte was also
identified. Anramine—O can undergo hydrolysis to the ketone,
producing a species whose molecular weight is increased one amu
compared to auramine—0 268 vs. 267. The tttermospray spectra of a
freshly prepared solution of auraraine—O, Figure 6, is dominated by
the MW 1 species at mlz 268. While auramine—0 and the hydrolysis
produced auramine ketone were separated under the LC analysis
conditions figures, and hence could be detected in one LC/MS assay,
it is clear that fragmentation information is required to
unambiguously identify the two species when the occurrence of the
latter was unexpected. This example shows an important property of
PB/LC/MS and the value of electron ionization mass spectra compared
to techniques that generate pseudo molecular ion information. Due
to the lack of structural information contained in thermospray
spectra, collision induced spectra have proven valuable (6). A
comparison of thermospray and PB/LC/MS techniques must consider
other sensitivity benefits that thermospray ionization possesses.
It is clear that the work to date is preliminary and requires
further in-depth examination and comparison.
One of the most appealing aspects of environmental LC/NS is the
possibility of making Injection volumes a factor of 100 to 1,000
times greater than commonly employed in FSCC—GC/MS. Consequently,
given the nanogram range for full scan electron Ionization PB/LC/MS
spectra, the capability for large injection volumes, and other
analytical options (e.g., SI!1 or Nd), the issue of sensitivity does
not appear to be the limiting factor. Clearly, though,
significantly more data are required to evaluate the relative merits
of particle beam, thermospray and other LC/MS techniques as applied
to environmental problems or regulatory or more fundamental
Interest. Nevertheless, these data show considerable promise.
A primary prerequisite for the development of powerful LC/MS methods
is the capability to be utilized in the reverse phase gradient
elution LC domain. In FIgures 1 and 2, we show gradient elution
6—206
-------
WWARII
*3 *
*03
/
II i s?
I —
I II , ’
*50 200 250
Ii . . s(O.erg .
Figure 1.
MAGIC spectra of selected Appendix IX analytes
o c o Wac.4
*54
Si
*00 IS O 2B
— P1 9 /Ch$g.
I. P ’T* .DU0. Cil .
S. SITOVVCZI C
i. 4NIT enIo.Im
4. II fll* .1. *31
3. O 1.4-0
1. 0$X1IC
7. IJ6ICCS 1*
S. D1(lWfl. $TZIJ(ST .
1. ft 1WC-0
iS. W,11IC -0 ETO C
(All ........ts •i Ill •.l.
&..A/\
Fiqure 2. Thermabeam spectra of selected Apper 3ix VIII analytes.
* . 2C4
• 10000
$000
‘ £000
C
4002
2000
£ 5
S.0E4
• 0C4
U
3.SC4
gZ. 0r4
a
£10200
TIORAM
1MW — 240)
a
*20 240
L 70
I i
* .2C5
• i.0C5
•1
•.0C4
£.0C4
4.0E4
2.0C4
300 — - - *00 150
M ..(Ch .rg.
200 250
5. 0E5
4 . 0E5
0
3. E5
•0
C
- 2.0E5
1 • ø 5
0
250
TIC
01
I
Aiu
of
jj
DRTA:APPIX?.D
x
w
ii
1 2
Time (.(n.)
3 5
I.
I i L I , i I
I cHP cnvs
- • r — ;•. •i lI’ 111 • • i$l • i •S • 1 •
5.
I lT
jiI I!J_fl -) 1
.•. ...
I .,
S
S
6.
IItILLkl.. ,..l
I -.I • • ise ... M
II
ó Ti W’iu.
1.11 1.13 a.a 3.43 4.51 575 $3 5 US? lU
5cwTPC
tii • - i
*5.37 11.53
6—207
-------
PB/IC/MS data under the conditions noted. The observed peak shape
with detection limits in the nanogram range represent important
observations, as these data show good response and other analytical
characteristics for a diverse range of chemical species.
CONCLUSION
Having shown the ability to generate 70eV/El mass spectra for a wide
range of diverse species using PB/LC/MS and the potential of two
different PB/IC/MS interfaces to employ gradient LC conditions, the
Issues of sensitivity and ruggedness become important as do the
practicality of PB/IC/MS.
As discussed previously, and as shown In a recent study funded by
the U.S. EPA Office of Solid Waste (5), response characteristics
appear to allow the response factor equation to be employed. The
study used a prototype PB/LC Interf ace and it yielded nanogram
sensitivities for many Appendix VIII listed analytes which could not
be done by CC/MS techniques. In this work, Ion current was found to
be highly correlated with injected weight even when Internal
standards were not employed. Many of these experiments, however,
were conducted under flow Injection analysis conditions. In the
referenced study, and in subsequent recent work, it appears that for
a wide variety of analytes, low nanogram sensitivities are possible
using electron impact conditions. We would note, however, that the
lack of “real-world” data on actual samples is due to a variety of
factors foremost of which is the study of the rapid developing
technology. Nevertheless, data does exist which suggest that
PB/IC/MS is useful and powerful for the analysis of real world
samples. However, we are cautious about stating Its useability,
although the data of Willoughby suggests this technology represents
a major advance in the state—of—the—art of LC/MS and will have
significant impact in characterization or non—chromatographable
analytes of regulatory interest.
Given the previous data and Its qualification it is instructive to
consider exactly when, how and why such technology Is important in
envlroiuental chemistry. It is also of Importance to Identify
potential problems and limitations. In terms of application to
existing programs in the RCRA and Superfund area, It Is known that
such technology is important to accurately characterize hazardous
waste or hazardous waste sites at which labile compounds or their
degradation products are present. Obviously, such applications
could include what has been termed “exotic” RCRA Appendix VIII
compounds. In the Superfund area, the technique can have
application to the characterization of hazardous waste sites “known”
to contain nonvolatile, (I.e., non—gas chromatographable analytes)
species at high concentration levels, but that are not well
characterized due to current measurement methodology. Also,
application too difficult to analyze priority pollutants is expected
6—208
-------
LIBRMY SPECTRUM 24790
100
200
500
Figure 3. Library matched, Thermabeam spectrum of DES and O-Toluidjne
6—209
50 100
250 200 250
DIETHYL STILBESTEROt.
300 350 400
ACOUIRED SPECTRUM
HO&5 \
DIFFERENCE (F OUIRED-LIBR Y)
$.i! 1 . I
LIB V SPECTRUM 1737
OTOLU IDINE
P OUIRED SPECTRUM
IIII III1III III P ,I, 11111
O IFFERENCE ( COUIRED-LIBRMY)
• 1
,uIIIIyIuIuIJI•IIIyI• I 1 1
-------
I. PROPTLTHI 1L
I
1. ii.- l.
1.-
• -
. S JCIPC
— .. - - .-— —
IUILCIIfl INILYIIS i VOl AIIAiTtES.
I. fl .Tw C1 .
1 1 1110m .1. 1 17.111
3 3. 4-I11 S4 111J1C
4.
S. C7Q C Il
1. S l .1. IJS
.
S. V O1S .1.
S. l 1O DC
ff 11 ........t. si I II ..J.LI
S
11.17 13.15 ItS) 5.75 11.67 11.53 liii 11.36
1C V11r(
[ i%.jure 4. Therinospray analysis or Apperdix VIII anal.ytes
Yigur. . d ..ns onel plot ot *6Are .. fl.-U Pd/Id.. /$. . .ct.u.
Tm..sp.ay AASS 1PCCT1 o, N*A IN(.
.43
a.
o. YLov 1n •ction Th.r osa. ctrua 01 AUCU ItS
6—2 10
-------
by be straight forward (i.e., Benzidine). Implementation of the
technology would, of course, require further testing and method
development and validation, but given our observations, it would
appear as if existing systems with appropriate Ion optics and source
design could be retrofit for relatively small cost (compared to a
complete LC/MS systems), and given the national quadruple GC/MS
resource system, introduction of such technology could be made more
rapidly, than technology requiring special modification or complete
systems purchase. Perhaps, more Important, PB/LC/MS technology or
similar technology could provide the necessary tools to regulate a
wide range of analytes. If In fact, such technology proves to be
robust enough, then the “target compound approach” associated with
regulatory lists could be eliminated or significantly reduced.
DISCLAIMER
This article has not been subjected to Agency policy review, and
therefore does not necessarily reflect the views of the Agency.
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude for the assistance
of R. C. Willoughby, Extrel Corporation, Pittsburgh, PA 15238, and
A. Apffel, Hewlett Packard Co., Palo Alto, CA 94304.
REFERENCES
R. C. Willoughby and R. F. Browner. Analytical Chemistry, 56, 2626
(1984).
A. Apffel and B. Nordman, “Development of a Magic LC/MS System”
Presented at the 35th Annual Conference on Mass Spectrometry and
Allied Topics, Denver, CO, 1987.
R. C. Willoughby and F. Poeppel, “Particle Beam—Liquid
Chromatography—Mass Spectrometry (PB—LC—MS): Advantages and
Applications” ibid.
E. D. Lee, L.O.G. Weidoif and J. D. Henion, “Ion Spray Liquid
Chromatography, API Tandem MS of Peptides,” ibid.
A. D. Sauter and R. C. Willoughby “Particle Beam LC/MS Analysis of
Selected Appendix VIII.” Listed analytes, under contract to
Dynamac Corp., for the U.S. EPA, Report available at request from
author R. K. Mitchum.
L. D. Betowski and J. M. Ballard, Analytical Chemistry, 56, 2604
(1984).
6—2 11
-------
MINIMUM DETECTION LIMITS ND DATA Z NALYSIS
John Warren, Office of Policy, Planning and Evaluation, U.S.
Environmental Protection Agency, Washington, D.C.
ABSTRACT
Many of the Agency’s regulation require collection of chemical or
environmental data, then a suimnarization of this data to a mean or
variance, and finally the comparison of these estimates to some
standard in order to determine compliance status. Providing the
contaminants or variables in question occur at sufficiently high
levels that that measurement can be done with high precision, the
determination of cciiçliance or non—compliance with a regulation is
relatively straightforward. When data are recorded as being “Below
the Detectable Limit” or “Trace,” the calculation of meaningful
suitenary estimates becomes somewhat tenuous, owing to the absence of
numerical values.
This paper looks at the results of making simple approximations to
unrecorded data when calculating means and variances. The
consequences of setting all below—detection—limit values to a constant
(such as the detection limit itself) are discussed in the context of
comparison of data to a standard. The method due to Cohen for
normally distributed data will be den nstrated and the work of Gilliam
and Relsel for lognormally distributed data will be discussed.
6—2 12
-------
DEVELOPMENT OF ROBOTICIZED ANALYTICAL METHODS
John G. Cleland, Ph.D., Manager Industrial Applications, Research Triangle
Institute, Research Triangle Park, NC
ABSTRACT
The Research Triangle Institute (RTI) has undertaken a robotics-related task to
assist the Technology Assistance Branch, Office of Solid Waste, Environmental
Protection Agency (OSW, EPA). OSW is encouraging producers of analytical
instrumentation to develop robotic assisted methods as adaptations of SW-846 (Test
Methods for Evaluating Solid Waste Physical/Chemical Methods, 3rd Edition) for
determination of hazardous constituents. Laboratory robotics will improve the
precision and accuracy, overall analytical data quality, safety, and efficiency of
analyses. It has been demonstrated that rather than eliminating jobs and upsetting
the work environment, laboratory robots have enriched the work of technicians
and professionals by providing new intellectual challenges and eliminating
repetitive tasks. Although laboratory robotics have only recently begun to be
successfully applied, their impact on analytical methods development is certain to
be important.
Robotics can be defined as the study and use of machines capable of manipulation
and/or mobility with some degree of autonomy. The autonomy may be almost
complete -— as in the case of an industrial manipulator which follows a sequence
of preprogrammed moves, or limited, as with teleoperators used in nuclear and
undersea operations. Robots become more autonomous and flexible when they
incorporate sensory capabilities, such as machine vision or tactility, or are
controlled by computerized expert knowledge or heuristic decision-making
programs. Therefore, laboratory robotics utilize machine intelligence and
flexibility rather than simple, “fixed automation” instruments.
A few vendors are supplying laboratory robot systems for such operations as
sample weighing, transfer, separation, and mixing. Research areas in robotics
which will become more important for laboratory applications include:
development of sensors, improvement of speed and accuracy, development of
better internal models of the environment in which the robot works, interface
standardization (both for software and hard ware), incorporation of mobility,
improved robot teaching methods (including off-line programming), and
reformulation of control architectures and path control. Robots will eventually
become incorporated into laboratory information management systems (LIMS).
Mobile field sampling and analytical robots are also being investigated.
The initial RTI task is a step directed toward development of laboratory work cells
interfacing robots, automated instrumentation and personnel in the most efficient
manner. The task is coordinated with an interagency agreement between EPA the
National Bureau of Standards, which will serve as a technical expert responsible
6—213
-------
for standardization of roboticized analytical methods. Enlistment of a producer of
analytical instrumentation is also planned. The participation of such agencies as
the National Aeronautics and Space Administration will be solicited, e.g. related to
the transfer of such technology as automated laboratory designs for Space Station
and Space Shuttle experiments. The study will incorporate review and evaluation
of existing and planned robotic-related methods which may be important to
hazardous constituents determinations.
Initial emphasis will be placed upon development and demonstration of a flexible
robotic technique for a specific solid waste test method. Incorporation of
flexibility in this case will emphasize statistically-based experimental design. Test
method optimization by such techniques is ideal in a computer-based robotic work
cell. Analysis results may be continuously evaluated and the analytical plan
updated and optimally redefined using the robot, since it is capable of altering
function and sequence without human intervention. The human scientific
knowledge required for making such decisions will eventually be incorporated into
expert systems related to all the functions of a particular laboratory work cell.
6—214
-------
EXPERT SYSTEM FOR INTERPRETATION OF THE INFRARED SPECTRA OF
HAZARDOUS WASTE DRUM SAMPLES
Steven P. Levine, Director of Industrial Hygiene Program,
University of Michigan, Ann Arbor, Michigan; Ying Li—shi, Shanghai
Medical University, Department of Occupational Health, School of
Public Health, Shanghai, Peoples Republic of China; Sterling A.
Tomellini, Department of Chemistry, University of New Hampshire,
Durham, North Carolina
INTRODUCTION
In order to satisfy the requirements of hazardous waste analysis
at Superfund and at licensed disposal sites [ 1-6], a program for
automated waste mixture identification (PAWMI) through the
interpretation of the infrared (IR) spectrum of the waste mixture
was developed [ 7,8] and tested on hazardous waste drum samples
[ 9]. This approach, which utilizes the speed and sensitivity of
Fourier transform infrared (FT—IR) spectrometry meets many of the
requirements of a near real—time, principal component screening
technique for organic hazardous waste samples [ 1].
Two limitations of PAWMI were that once a training set, consisting
of a library of reference of spectra, was defined, the rules for
the inference engine (PAIRS) [ 10-16] had to be generated manually.
The second limitation was that the PAWNI compound identification
software only uses peak location information.
An approach to the automated generation of functional group
interpretation rules for PAIRS was previously developed (17).
Efforts by other investigators have been successful for the
interpretation of IR spectra using computerized interpretation or
matching procedures (18—27).
This paper describes a program for the identification of the
principal components of mixtures based on computer assisted
interpretation of the mixture’s infrared spectrum. This program
(intlRpret) , which was developed as a preliminary screening tool
for unknown organics handled on hazardous waste remedial action
sites, has five main subroutines: the interferograxn processing
and peak selection subroutine (PUSHSUB) [ 81, the automated
knowledge acquisition subroutine (AUTOGEN) [ 171, the system
optimization subroutine (STO), the interpretation subroutine
(PAIRS) [ 7, 10—16], and final processing subroutine to subtract
spectral similarity (PAIRSPLUS) [ 81.
Many of these subroutines are substantial modifications of the
programs pi-e iously reported [ 7,8,17]. Principal advantages of
this system compared to the previously reported PAWNI system are
speed (all spectral information is encoded automatically),
6—215
-------
flexibility (changes in the data base and in interpretation rules
are readily accommodated) and accuracy (interpretation is based on
peak position, frequency of occurrence and peak size, each of
which is weighted in an optimal fashion).
The method has been evaluated using the 62 most commonly
identified organic compounds on hazardous waste sites [ 181.
IntlRpret was designed to be automatic, self—training, and self-
optimizing so it could be operated on-site (in a mobile
laboratory) during a remedial action project, by personnel with
limited training. Other applications of the intlRpret technique
would include screening incoming organic waste at licensed
disposal facilities or for interpreting gas phase spectra obtained
during industrial hygiene air monitoring.
EXPERIMENTAL SECTION
Spectra were acquired on a Nicolet 20—SX optical bench. Each
spectrum was generated with a background and sample signal
averaging of 128 scans. The number of data points collected was
16,384, resulting in a nominal spectral resolution of 2 cm-i. All
programming and spectral analysis, including rule writing,
compiling and spectral interpretation, was performed with a
Nicolet 1280 computer equipped with a 160 Mbyte Winchester disk
system.
RESULTS MilD DISCUSSION
IntlRpret has five main subroutines the interferogram processing
and peak selection subroutine (PUSHSUB) [ 8], the automated
knowledge acquisition subroutine (AUTOGEN) [ 161, the system
optimization subroutine (STO), the inference engine (PAIRS) [ 7,
10-161, and the final processing subroutine which subtracts
spectral similarity (PAIRSPLUS) [ 81. Figure 1 is a flow chart of
the intlRpret process, where the logic of each of the five major
subroutines is diagrammed.
Because PUSHSUB, AUTOGEN, PAIRS and PAIRSPLUS have been explained
in detail in the above referenced papers, the reader will have to
study those papers in order to understand the details of the
operation of those subroutines. However, a summary of those
subroutines is given in this publication. In his publication,
emphasis is placed on describing STO, which is central to the
operation of the self—training, self-optimizing mode of operation
of intlRpret. In addition, the linkage of all of the subroutines
is explained.
This system is used in conjunction with spectral library SEARCH
programs to optimize the ability of the system to identify
unknOwnS. 6—216
-------
PTJSHSUB
In order to automate PAWMI, a peak selection of subroutine
PUSHSUB, was developed that does not require the operator to set a
peak selection threshold, and successfully follows non—linear
baselines [ 8]. Figure 2 shows an example of the operation of
PUSHSUB. In tracing A, the IR spectrum of a hazardous waste drum
sample is shown. Tracing B shows the spectral baseline
automatically established by PUSHSUB. Tracing C shows the
resulting spectrum, with the peaks of importance for the
identification of the principal components of this mixture
separated from the non-linear baseline.
STO
This subroutine accesses the peak tables generated by PUSHSUB.
The peaks in a spectrum that are chosen for the purposes of
decision making are called rule peaks. Not all spectral peaks are
rule peaks. Each rule peak is assigned a “goodness value” that
indicates the probable presence or absence of each compound in the
training set.
Three factors are used to weight the goodness values assigned to
each rule peak listed by AUTOGEN: kl (frequency of occurrence),
k2 (intensity), and k3 (frequency of occurrence X intensity).
These three factors are designed to follow the logic used by an
expert during the interpretation of the infrared spectra of
mixtures. In this respect, the underlying intellectual framework
is similar to that described in the work of McLafferty in which
Match Factors were automatically calculated for the interpretation
of mass spectra [ 28,29].
The factors ki, k2 and k3 are defined by the program. The
goodness available to each peak window is divided between ki, k2
and k3, with the default value for the constraints set equal.
These default values can be changed by the operator.
1(1, which is a measure of frequency of occurrence of peaks in the
training set within a given wavenumber window, essentially states
that a peak should be given added importance (or goodness) if it
is in a region of the spectrum in which there are few peaks in the
other spectra in the training set. Kl is equal to:
number of peaks present in the rule peak window
number of peaks in that window in all spectra in the training set
6—217
-------
K2 essentially states that added significance should be attached
to the presence of a peak that represents a large fraction of the
total peak intensity for the spectrum of a compound. 1(2 is equal
to:
rule peak intensity
total intensity for all peaks in the spectruR of that c pound
1(3, which is the cross—term between frequency of occurrence and
intensity, essentially states that a large peak should be given
added importance if it is in a region of the spectrum in which
there are few large peaks in the other spectra of the training
set. 1(3 is equal to:
rule peak industry
intensity of all peaks in that window in all spectra in the
training set
As stated earlier, the default values chosen for this study were:
window widths of 4-,’— 3, 5, and 10 cm—i; goodness values divided
between these windows of 50%, 30%, and 20%, respectively; and ki =
k2 k3. It is not known if these are the optimal values for this
training set, for all possible mixtures that can be prepared for
compounds in this training set, or for other training sets.
Studies are underway to define the optimal window width based on
experimentally determined peak shifts.
STO is a very significant departure from the practice previously
reported for PAWMI [ 7,8]. In that program, only peak position
information was used. Using the STO portion of intlRpret allows
the optimization of goodness values for each rule peak in each
training set.
AUTOGEN
The automated generation of rules for a defined training set is
essential to the success of this approach. The nature of these
rules is dealt with in great detail in references 7 and 10-16, and
is described in brief in the section of this text describing PAIRS
(below). Without AUTOGEN, PAIRS and PAWMI are hampered by the
potential for errors that always occurs when data is manually
encoded, and by the constraints imposed by the length of time it
takes to enter data for new or modified training sets. Because of
these problems, such a system is inherently inflexible. AUTOGEN
solves these problems.
PAIRS
As previously reported, PAIRS [ l0-l6 1 14 modified in the PAWNI
6—218
-------
program [ 7,8]. The mixture interpretation software uses peak
location information and is based on a three—level filter
algorithm designed to compensate for potential peak shifts in the
mixture spectrum. If a peak falls in a relatively wide frequency
window assigned to a certain compound, a percentage of the overall
goodness value will be added to the total. tiGoodness t is a
measurement of closeness of match between the spectrum of the pure
compounds used for rule generation and the spectrum of the unknown
compound(s). The goodness scale ranges from 0.001 for a complete
mismatch to 0.999 for a complete match.
In the intlRpret program, peaks in the library spectra are picked
by PUSHSUB, the goodness values are weighted by STO, and the three
level rules are written by AUTOGEN. A peak table is then created
for the unknown mixture by PUSHSUB. PAIRS accesses that table and
generates goodness values that indicate the probable presence of
compounds in the mixture of unknowns.
PAIRSPLUS
PAIRSPLUS was developed to limit the effect of spectral
similarity. A detailed description of PAIRSPLUS can be found in
reference 8.
An example of the spectrum of a four-component mixture that was
analyzed using intlRpret is shown in Figure 3. 1,1,1-
trichioroethane, chlorobenzene, toluene, and benzene were
correctly identified in this mixture.
LIBRARY SEARCH
Many hazardous waste samples, when analyzed by GC—MS show an
incomplete material balance [ 9]. This is due to the presence of
polymeric, thermally labile and/or highly polar components. Using
this FTIR technique, simple air drying of the sample will
frequently be sufficient to leave a residue containing only the
commercial polymer that makes up a bulk of the sample. Figures 4
and 5 show two examples of this. In each case, the analysis
performed by GC—NS showed only the presence of volatile solvents
comprising less than 30% of the sample weight. The dried specral
library file, when searched against the Aldrich library of 4,000
compounds and commercial mixtures, resulted in the identification
of Polyamide Resin and Igepal polymer as the principal components
of these hazardous waste drum samples.
6—2 19
-------
TRAINING SET OF PURE COMPOUNDS
rAc gui
[ er fero gram
Subtract low {Tspia 1
resolut;on from difference
high resolution 1 5Pect m
PUSHSUB spectrum
Peak select
and store in
format file
4r ______
Read the peak Report
table for rule
each pure peaks
compound
STO ____________
Optimize [ eport
weighting
L! f tors L ors
Enter rule peaks
and weighting
AUTOGEN factors into PAIRS SPECTRUM OF MIXTURE
L using CONCISE OF UNKNOWN COMPOUNDS
[ PUSHSUB
Inference engine:
I Identify components
PAIRS J of mixture and
Report I
L assign goodness values [ data
r TSubtract spectrai]
jsimilarity j
PAIRSPWS
Report
L data J
Figure 1. Flow chart of the intlRpret process, showing the logic of
each of the five major subroutines.
6—220
-------
w
C)
2
0
C l)
400
Figure 2. Spectral tracings showing the operation of PUSHSUB. A
spectrum of hazardous waste sample; B = baseline
automatically generated by PUSHSUB; C difference
between A and B, which is used to identify peak locations
for spectral interpreatation.
C
WAVEN UMBER
1120
6—22 1
-------
1.10
.88
w
0
z
.66
0
C l )
.44 —
.22
0. 1
2010 1830 1650 1470 1290 1110 930 750 570 390
WAVENUMBERS
Figure 3. Portion of the spectrum of the mixture chlorobenzene +
1,1,1—trichioroethane (TCE) + toluene + benzene in 1 : 1
0,5 : 0.1 ratio (w/w). Peaks used as PAWMI rule peaks but
not intlRpret rule peaks are shown by (A); peaks missed by
PUSHSUB are shown by (B); peaks heavily weighted in the
spectrum of benzene by intlRpret that fall within the
+/-3 cm-i window are marked by (C).
6—222
-------
DRIED FILE
POLYAMIDE RESIN
w
U
z
0
C’)
4000 400
Figure 4. Spectrum of hazardous waste drum sample after air drying for
five minutes (Dried spectral File), and closest library match
(Polyamide Resin).
3100 2200 1300
WAVENUMBER
6—223
-------
DRIED FILE
w
C)
z
0
400
Figure 5. Spectrum of hazardous waste drum sample after air drying for
five minutes (Dried spectral File), and closest library match
(Igepal).
IGAPAL
4000
WAVENUMBER
6—224
-------
LITERATURE CITED
[ 1] Puskar, M.A., Levine, S.P., and Turpin, R.: “Compatibility Testing
and Materials Handling” in “Protecting Personnel at Hazardous
Waste Sites”, Chapter 6, Levine 1 S.P. and Martin, W.F., Eds.,
Butterworths/Ann Arbor, Woburn Mass., 1985.
[ 2] Gurka, D.F., “Project Summary: Interlaboratory Comparison Study:
Methods for Volatile and Semivolatile Compounds”, Environmental
Monitoring Systems Laboratory, Las Vegas, NV, EPA—600/S4—84—027,
June 1984.
[ 3] Hallstedt, P.A., Puskar, M.A., and Levine, S.P., J. Haz. Waste
Haz. Mat., 1986, 3 (2), 221—232.
[ 4] Eckel, W.P., Trees, D.P., and Kovell, S.P. “Distribution and
Concentration of Chemicals and Toxic Materials Found at
Hazardous Waste Dump Sites”, Proc. National Conference on
Hazardous Waste and Environmental Emergencies, May, 1985.
[ 5] Mayhew, J.D., Sodaro, G.M., and Carroll, D.W. “A Hazardous
Waste Site Management Plan”, Washington, D.C.: Chemical
Manufacturers Association, 1982.
[ 6] “The Hazardous and Solid Waste Amendments of 1984”, Congr.
Rec., 1984, Oct 3, H11103.
[ 7] Puskar, M.A., Levine, S.P., and Lowry, S.R. Anal. Chem., 1986,
58, 1156—1162.
[ 8] Puskar, M.A., Levine, S.P., and Lowry, S.R. Anal. Chem., 1986,
58, 1981—1989.
[ 91 Puskar, M.A. , Levine, S.P. , and Lowry, S.R. Environ. Sci.
Technol., 1987,
[ 101 Woodruff, H.B. and Munk, M.E. J. Org. Chem., Vol. 42, No. 10,
1977.
[ 11] Woodruff, H.B. and Munk, M.E. Analytica Chimica Acta, 1977, 95
13—23.
[ 12] Woodruff, H.B. and Smith, G.M. Anal. Chem. 1980, 52, 2321-2327.
[ 13] Woodruff, H.B. and Smith, G.M. Analytica Chimica Acta, 1981, 133,
545—553.
[ 14] Tomellini, S.A., Saperstein, D.D., Stevenson, J.M., Smith, G.M.,
and Woodruff, H.B. Anal. Chem. 1981, 53, 2367—2369.
6—225
-------
[ 15] Tomel ljnj, S.A., Stevenson, J.M. and Woodruff, H.B. Anal. Chem.
1984, 56, 67—70.
[ 161 Tomellini, S.A., Hartwick, R.A., Stevenson, J.M., and Woodruff,
H.B. Analytica Chimica Acta, 1984, 162, 227-240.
[ 171 Blaffert, T. Anal. Chim. Acta. 1984, 161, 135—148.
[ 18] Zupan, J. and Munk, M.E. Anal. Chem. 1985, 57, 1609-1616.
[ 191 Trulson, M.O. and Munk, M.E. Anal. Chem. 1983, 55, 2137—2142.
[ 201 Frankel, D.S. Anal. Chein. 1984, 56, 1011—1014.
[ 211 Lowry, S.R. and Huppler, D.A. Anal. Chem. 1983, 55, 1288—1291.
[ 22] Jurs, P.C. and Isenhour, T.L. “Applications of Pattern
Recognition”, Wiley: New York, 1975.
[ 23] Rasmussen, G.T., Isenhour, T.L., Lowry, S.R. and Ritter, G.L.
Anal. Chim. Acta., 1978, 103, 213—221.
[ 24] de Haseth, J.A., Woodruff, H.B., Lowry, S.R. and Isenhour, T.L.
Anal. Chim. Acta, 1978, 103, 109—120.
[ 25] Saperstein, D.D. Appi. Spectrosc., 1986, 40 (3), 344-348.
[ 261 “Computer Supported Data Bases”, Zupin, J., Ed., Howard Ltd -
Wiley Co., NY (1986).
[ 27] Jurs, P.C. “Spectral Library Searching and Structure
Elucidation”, Chap. 16 in “Computer Software Applications in
Chemistry” Jurs, P.C., Ed., Wiley Co., NY (1986).
[ 28] Kwok, K-S, Venkataragahaven, R. and McLafferty, F.W. J. Amer.
Chem. Soc., 1983, 95, 4185—4194.
[ 29] Atwater, B.L., Stauffer, D.B., McLafferty, F.W. and Peterson,
D.W. Anal. Chem. 1985, 57, 899—903.
6—226
-------
ACKNOWLEDGMENT
The authors would like to thank Greg Kinnes for his help in prepar-
ing the mixtures and acquiring the IR spectra, and to Mary Weed and
Dave Hunsche for preparation of manuscript figures. In addition, Mark
Puskar developed the PUSHSUB, modified PAIRS and PAIRSPLUS programs,
and Steven Lowry had timely assistance in data interpretation.
CREDIT
This work was supported by grant 1-RO1-0H02066-O1 from the National
Institute for Occupational Safety and Health of Centers for Disease
Control.
6—22 7
-------
IMPROVING SONICATION TECHNIQUES IN CLP ORG INICS .P NP LYSIS
AND SOLID WASTE EXTRACTION
S. Berliner, Director, Technical Services, Heat Systems Ultrasonics,
Inc., Farmingdale, New York
ABSTRACT
EPA laboratories and contractors and others following EPA protocols
use sonication for analysis of extractable organics and pesticide/PCBs
in sediment/soil in the CL? (Contractor Laboratory Program).
Sonication is also used in evaluating solid waste by SW—846 Method
3550. Proper tuning of ultrasonic liquid processors and proper use of
new half wave extender tips will greatly improve reproducibility and
reduce costs to contractors and the Government. Thning the processor
is a means of assuring optimum operation at the highest efficiency.
Techniques for tuning virtually every model of ultrasonic processor in
current use are discussed. Use of half wave extender tips avoids
solvent erosion of the replaceable tip joint and allows top insertion
through the neck of flasks and bottles while lowering disposables
costs.
6—229
-------
INTRODUCING THE ThIRD EDITION OF SW—846: TEST METHODS FOR
EVALUATING SOLID WASTE
Margaret Layne, Environmental Engineer, Research Triangle Institute,
Research Triangle Park, NC; Denise Zabinski, Chemist, Office of Solid
Waste, U.S. EPA, Nancy Rothman, Senior Chemist, ERCO/A Divions of
SECO, Cambridge, MA; Alvia Gaskill, Senior Environmental Scientist,
Research Triangle Institute, Research Triangle Park, NC
ABSTR ACT
On March 16, 1987, the U.S. EPA announced the availability of the
Third Edition of the manual “Test Methods for Evaluating Solid Waste,
Physical,/Chemicai Methods” EPA publication SW-846 (B2 FR 8072). The
manual contains methods suitable for specified hazardous waste testing
and monitoring purposes, and provides a source of information on
sampling and analysis for compliance with the Resource Conservation
and Recovery Act (RCRA) regulat ons. This poster session presents an
introduction to the information contained in the manual, the
information required to use the manual, and an example of a decision
tree which can be applied to the analysis of a sample containing
organic compounds.
PURPOSE OF ThE MANUAL
Test Methods for Evaluating Solid Waste (SW—846) is intended to
provide a unified, up—to—date source of information on sampling and
analysis related to compliance with RCBA regulations. It brings
together into one reference all sampling and testing methodology
approved by the Office of Solid Waste for use in implementing the RCRA
regulatory program. The manual provides methodology for collecting
and testing representative samples of waste and other materials to be
monitored. Aspects of sampling and testing covered in SW—846 include
quality control, sampling plan development and implementation,
analysis of inorganic and organic constituents, the estimation of
intrinsic physical properties, and the appraisal of waste
characteristics.
The procedures described in the manual are meant to be comprehensive
and detailed, coupled with the realization that the problems
encountered in sampling and analytical situations require a certain
amount of flexibility. The solutions to these problems will depend,
in part, on the skill, training, and experience of the analyst. For
some situations, it will require a combination of technical abilities,
using the manual as guidance rather than in a step-by—step,
word—by—word fashion. Although this puts an extra burden on the user,
it is unavoidable because of the variety of sampling and analytical
conditions found with hazardous wastes.
REGULAIORY SIGY I FICANCE
SW—846 is a collection of methods suitable for specific hazardous
waste testing and monitoring purposes. The manual does not establish
testing requirements. Rather, the various testing requirements are
6—231
-------
established in the appropriate sections of the Code of Federal
Regulations.
EPA is not incorporating the Third Edition into the RCRA hazardous
waste regulations at this time because of the significant changes
which have been made in the methods. The Agency plans to replace the
existing methods with the revised versions in the near future, after
allowing time for comment in the revised methods. Until that time,
however, the methods described in the Second Edition of SW—846 must
continue to be used where the Second Edition is incorporated be
reference or mandated by particular regulation.
C TENTS
The manual is provided in four 4” D—ring binders. The first binder
contains Volume 1A, which includes Chapters One through Three.
Chapter One addresses quality assurance and quality control
considerations which apply to all methods described in the manual.
The chapter is not intended as a comprehensive guide to QP /QC
programs, but emphasizes that data generated by the analytical methods
contained in the manual are meaningless unless supported by
appropriate quality control procedures and documentation.
Chapter Two contains information designed to aid the experienced
analyst in selecting the appropriate methods from the maual, based on
characteristics of the sample and the objectives of the analysis.
Figures and tables are included to assist the user in identifying
cleanup, preparation, and determination methods for various
combinations of analytes and sample types. A table of containers,
preservation techniques, and holding times is also provided.
Chapter Three covers sample preparation and determinative methods for
metallic analytes. Preparation methods include acid digestion and
dissolution procedures, while the determinative methods are
Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP) and
Atomic Absorption (A1 ).
The second binder is labelled Volume lB, and contains Chapter Four,
Orgainc 2½nalytes. Sample preparation methods cover a variety of
extractions and cleanups, while the determinative methods include gas
chromatography, gas chromatography/mass spectometry, and high
performance liquid chromatography.
The third binder, Volume 1C, comprises Chapters Five, Six, Seven, and
Eight, addressing miscellaneous analytes, properties, and
characteristics. Miscellaneous test methods covered in Chapter Five
include WX, WC, cyanide, sulfide, sulfate, phenolics, oil and
grease, total coliform, nitrate, chloride, and radium. Chapter Six
addresses properties such as pH, specific conductance, cation exchange
capacity, liner performance and compatability, paint filter liquids,
and radioactivity. Chapters Seven and Eight discuss the
characteristics of hazardous waste, as defined by the Code of Federal
Regulations. Regulatory definitions are covered in Chapter Seven and
test methods appear in Chapter Eight.
6—232
-------
Volume Two is contained in the fourth binder, and includes chapters
Nine through Thirteen, which addresses sampling methods and monitoring
procedures. Design, development, and implementation of sampling plans
is discusses in emissions. Monitoring of groundwater, land treatment
facilities, and incinerators is covered in chapters Eleven, Twelve,
and Thirteen, respectively.
AVAILP BILITY
Test Methods for Evaluating Solid Waste, Physical/Chemical Methods,
EPA publication SW—846, is available from the U.S. Government Printing
Office, Washington, D.C. 20402, order number 955—001—00000—1, for a
cost of $110, which includes future updates.
6—233
-------
QUALITY
ASSURANCE
thairpersons
Duane Geuder Torn Logan
Chemist Engineer
Ofticeof ergency and Environmental Monitoring
Remedial. Response and Support Lab
U.S. EPA U.S. EPA
401 N Street, S.W. Research Triangle Park,
shingtcn, D.C. 20460 NC 27711
-------
U.S. ARMY [ OXIC AND HAZARDOUS MATERIALS AGENCY
INSTALLATION RES IORATION QUALITY
ASSURANCE PROGRAM
Kenneth T. Lang, Chief, Analytical Branch, Technology Division, U.S.
Army Toxic and Hazardous Materials Agency, Aberdeen Proving Ground, MD
ABSTRACT
The Army’s Installation Restoration Program (IRP) was established in
the mid—1970s to identify, evaluate and clean—up contamination
resulting from past waste disposal activities on property under Army
control. The U.S. Army Toxic and Hazardous Materials Agency
(USATHAMA) was assigned to execute this mission as the IRP central
manager. In order to assure that high—quality, verifiable data was
produced from chemical analyses of environmental samples, USATHNIA
developed and implemented a state—of—the—art Quality Assurance (QA)
Program. Since the implementation of the USATHANA QA program in the
mid—1970s, we have continued with a policy of reevaluation and
revision of the program’s concepts and methods. Our process of review
and analysis includes experts from industry,m national laboratories,
academia, and other government offices.
The main objectives of the USATHAMA QA Program are to: (1) provide a
consistent framework for the generation of good analytical data; (2)
require contractor laboratories to demonstrate their ability to
analyze for all of the compounds of interest in the appropriate sample
matrices over a range of concentrations; (3) require the laboratory
to analyze daily quality control (QC) samples, and to evaluate
laboratory performance on a daily basis; (4) require the laboratory
to provide USATHANA with daily QC charts on a weekly basis, and to
provide detailed accounts of problems encountered and corrective
actions taken; (5) provide the means (e.g., software) to perform
certification calculations, daily QC and data entry, and internal and
external audits of data, facilities, and sampling excursions; (6)
require documentation of all aspects of sampling and analysis; (7)
provide Standard Analytical Reference Material (SARM’s) which are NBS
traceable whenever possible; (8) use analytical methods based on
those of the U.S. Environmental Protection Agency (EPA), AOPiC, and
methods in the open literature; (9) develop methods for military
unique compounds; (10) participate in interagency cooperative studies
as the EPA/DOD /DOE Holding Time Study; and (11) re—evaluate and
improve the USATHANA QA Program.
Presented here will be a discussion on how USATHAMA. accomplished these
objectives and some major differences between USATHANA’S QA Program
and other programs in common usage.
INTRODUCTION
The main goals of any Quality Assurance (QA) program are: (1) to
produce high quality, verifiable data from chemical analyses, or put
another way, to ensure that one gets what is paid for (i.e., good
analytical data); and (2) to prove to oneself and others that the data
7—1
-------
is representative of the environment from which the samples were
collected. There is, of course, nxre than one way to accomplish these
goals, and reviewers of environmental data and regulators should keep
an open mind to this fact. To do otherwise would result in dismissing
a great deal of analytical data collected at great expense, and result
in delaying crucial remedial actions, as well as the additional costs
associated with resampling and reanalysis of samples from a previously
studied site. In addition, to impose a de facto Ql program without
any provisions for making changes or achTIv equivalency with the
existing regulatory requirements would stifle scientific debate on Q
issues, stagnate further develop nts in QA, and inhibit the peer
review process which is so important in the scientific field.
USN I’HN’W.A developed a state—of-the-art QT program when the Agency was
organized in the mid—1970s. This was about the same time that the EPA
protiLtigated the drinking water regulation and associated QA/QC
requirements. It was well before the passage of CERCLA in 1980 and
the CLP in 1981. From the beginning, we adopted a policy of continued
reevaluation within the Army, to debate and implement changes to the
Qh program. This process of review and analysis has expanded over the
years to include experts from industry, national laboratories,
academia, and other government offices.
(Xir main objectives in fornulating the UShTUMh Q1 program have been:
(1) to provide a consistent framework for the generation of good
analytical data. We have accomplished this objective by establishing
a formal QA program in which all QA program requirements are contained
within a single document; in addition, (2) we require the
laboratories to demenstrate their ability to analyze for all of the
conçounds of interest in the appropriate sample matrix; (3) the
laboratories are required to analyze daily quality control (QC)
samples, and to evaluate their performance on a daily basis; (4) we
also require the laboratories to provide USA’ mN1 with all of their OC
charts on a weekly basis, and to provide detailed accounts of problems
encountered and corrective actions taken; (5) we provide the
laboratories with tools such as software to perform certification
calculations and daily QC; and we have automated data entry/data
reporting to expedite the review process; (6) we require
documentation of all aspects of sampling and analysis; (7) we provide
standard analytical reference materials (SARM’s) which are traceable
to NBS whenever possible; (8) we use analytical methods published by
the USEPA, C, and in the open literature; (9) whenever methods are
not available, such as military unique compounds, we conduct research
and development efforts and ruggedness testing of new methods; (10)
we participate in inter—Agency cooperative studies such as the
EPA/’DCVJT)OE Holding Time Study and the Dh/EPA evaluation of field
portable instruments and methods; and finally, (11) we continue to
reevaluate the effectiveness of our Q1 program taking into
consideration new analytical, statistical, and software developi ents,
and changes in the resources available to us (i . e •, manpower and
lu get).
In developing the strategy to accomplish our goals, we considered
factors such as the individual program and site specific requirements.
7—2
-------
In addition, we considered the cost of performing analyses, since
every QC ample, blank, duplicate, etc., that is analyzed impacts the
total cost of an environmental study. We have striven to limit or
reduce the number of samples, or lots of samples, that will be
rejected due to poor lab QC, exceeding holding times, high lab blanks,
etc. In determining the number of QC samples to be run in each lot,
we evaluate the instrument and method limitations so we can optimize
the ratio of the number of environmental samples to QC samples that
must be analyzed. Further, we assess the intended use of the data to
determine the amount of QC required and the confidence level to which
the data will be reported. In the mid—l970s when our QP program was
being formulated, we elected to set the reporting limit at the 90%
confidence level where both the alpha and beta errors are 5% each.
Even though this was an arbitrary decision, we have elected to stay
with the 90% confidence level primarily because our experience has
shown that the data produced is technically acceptable for the
intended use, and at a reasonable cost. Finally, we have designed a
Q1 program that is manageable with the manpower that we have
available.
As mentioned previously, we continue to reevaluate and revise our Qt
program to make it among the best of the Q1 programs available. We
believe that we have a technically sound QA, program; however, we also
recognize that it can still be improved. There have been many changes
in the USATIWIA cY program since its inception in the mid—1970s, the
most significant being a trend toward tighter restrictions on the
laboratories operating under the program. When we began our program,
we recognized that there were differences between laboratories, such
as instruments, equipment, and personnel. We, therefore, designed the
USATU M Q program to be flexible enough to accomodate these
differences. Not long after establishing the original program, it
became apparent that some laboratories would take advantage of every
ambiguity in the Q program. Since that time, we have systematically
made the USATHAI4P Qf program less flexible. We have taken actions
such as requiring QC on a more frequent basis and spelling out the
exact QC required; we have required more documentation of problems
and corrective actions taken; we have improved our control charting
procedures to make them more responsive to our needs; we increased
the frequency of laboratory and field sampling audits; we began
holding quarterly QA/QC meetings with all laboratory Q personnel in
attendance. At these meetings, the laboratory QA personnel discuss
problems which they are experiencing with methods, instruments, QC,
etc., and they arrive at common solutions which are implemented by
each laboratory; more recently, we have required six of our
laboratories to use identical analytical methods where options that
are found in some methods, such as flowrates, column temperatures,
extraction solvent, etc., have been standardized. We have seen a
significant reduction in the variances among these laboratories. We
have now begun a study to evaluate a variety of commonly used methods,
standardized the options of each method, and write the methods in a
standard format. Each method will then be validated following the
USATHA1’IA certification procedure. Upon completion of each standard
method, our analytical laboratories will be required to follow the
method(s) explicitly. When a new method becomes available, it will be
7—3
-------
evaluated against the standard method, and depending on the result of
the evaluation, it may be adopted as the standard method.
Some of the specific aspects of the USAThAI’1 Q program include a
formal Qk program document which contains all of the program
requirements, except the analytical methods. The analytical methods
are maintained separately. The US THAMA Q program has existed as a
formal document for nvre than ten years, and it has been the subject
of extensive peer review during that time. The program was designed
to provide real—time information on the analytical process so that
deficiencies can be detected and corrected on—the—spot. The USAThMP
Q program has been highly automated in order to expedite the transfer
of data and its review. This program has been instrumental in the
early detection of deficiencies and their ultimate correction, and it
has been effective in the detection of deficiencies in laboratories
following other Q1’ programs.
A major aspect of the USAThMA Q1 program is that we require our
laboratories to certify that they can actually perform the required
analysis(es) using the equiçoent and personnel that will be used on
“real” samples. Laboratories must denonstrate their proficiency
before any field samples are analyzed. The laboratory certifies for
each compound of interest and in each sample matrix they will be
analyzing. Provisions are made for those methods where it would be
impractical to analyze for every compound (e.g., GC/?IS). Proficiency
is den nstrated over a range of concentrations rather than at a single
concentration. The method certification data established the baseline
that future laboratory performance must fall within (i . e., upper and
lower control limits) to be acceptable. We have observed improved
laboratory performance as a result of the certification process. We
also develop reporting limits for each method using the data collected
during certification. These reporting limits become the USATHAN1
Certified Reporting Limit (CRL) for each method. The CRL is the
minimum quantification level that USATHM1 laboratories may report.
The CRL should not be confused with the classical approach of
comparing signal to noise under ideal conditions (limits which are
difficult to achieve under routine laboratory conditions). The CRL is
determined by application of regression theory to target (known
analyte concentrations) versus found (values determined by actual
analysis of spiked samples) curves generated by the analysis of real,
spiked samples over a range of concentration . The technique is based
on principles described by Hubaux nd Vos . The reporting limits
derived from this procedure are nore representative of a laboratory’s
capabilities on a day—to-day basis and provide confidence that an
analyte present at a concentration above the reporting limits will
consistently be found.
Another major aspect of our QP program is the evaluation of a
laboratory’s performance on a regular basis. The laboratories are
1 Hubaux, A., and G. Vos, 1970, “Decision and Detection Limits for
Linear Calibration Curves.” Analytical Chemistry , Vol. 42, No. 8,
pp.849—855.
7—4
-------
required to analyze quality control (QC) samples on a daily basis.
The results of these analyses are plotted daily on QC charts. The
control limits used on the QC charts are those which were developed
from the laboratory certification data collected prior to the
beginning of routine laboratory analysis. Therefore, if a laboratory
uses its best personnel and equipment at the beginning of the
certification process, they will have to maintain that same level of
performance throughout the analytical performance period. Should an
attempt be made to make the control limits artificially wide,
certification will likely be denied based on acceptance criteria
developed by other laboratories using the same method(s). Control
limits are periodically revised based on laboratory performance as
more data is collected. Experience has shown that the control limits
generally become narrower due to improved laboratory performance as
more experience is gained with a particular method. When out of
control situations do occur, the laboratory is required to investigate
the probable cause(s), and to document the causes(s) and action(s)
taken. The control charts are submitted to USATHAI4P on a weekly basis
for review. These reviews verify that the laboratory is performing a
daily review of the analytical and QC data. If a laboratory’s
performance declines, or if they do not comply with the provisions of
the QT program, the laboratory will be decertified for the method(s)
in question. USATH1 N has decertified laboratories for poor
performance under the UShTHANA Q program.
Miother part of the USA TH1 NP Q program is the performance of
laboratory and field audits, and quarterly Q1 QC meetings of the
USATHAi,1A certified laboratories. Each laboratory performing work for
USATHAM is randomly audited to ensure that the OIVQC procedures are
being followed, that appropriate records are being maintained, and
that good laboratory practices are being followed. Particular
attention is given to those laboratories that have been experiencing
problems. Audits are also conducted of field sampling operations to
ensure that samples are being collected following the proper protocols
(i.e., sample storage conditions, proper collection bottles, sample
identification, etc.). In addition, we hold quarterly Q /QC meetings
with the Q personnel from each laboratory present in order to discuss
problems and find common solutions.
In summary, I would like to emphasize that we are striving to get high
quality data from our contractors. We have developed a QF program
that has worked for more than ten years and continues to work. We
have gained a lot of experience with QA/QC programs, and have
experience with laboratories working under their own or other Q1
programs. Our experience with some of these laboratories has been
disappointing in some instances since their performance has not been
what was expected given the programs being followed. We also realize
that we can continue to improve our program with assistance from other
experts in the field, and we hope that meetings such as this, and
those in the future will help us to accomplish this objective.
7—5
-------
AUTOMATED MAINTENANCE AND REPORTING OF
ANALYTICAL QUALITY ASSURANCE
ON A PERSONAL COMPUTER
RICHARD D. BEATY AND LEIGH A. RICHARDSON,
TELECATION ASSOCIATES, CONIFER, COLORADO
ABSTRACT
In order to support the validity of analytical results, a
certain amount of overhead work is required for quality
control. This overhead involves not only additional analyses
of spiked samples, duplicates, and quality control samples,
but also requires additional book1 eeping to track, document,
and report quality assurance data. TELECATION ASSOCIATES has
developed a software package for personal computers, which
aids in the bookkeeping chore and automates the generation of
quality control data and reports.
SMARTLOG CR) and SMARTLAB CR) are two laboratory data manage-
ment programs which vary in the amount of laboratory informa-
tion maintained in the data files. Both programs store
analytical results for all analyzed samples, including all
quality control work. After analysi5, data may be reviewed
and evaluated on the computer screen. Quality control data
is identified and processed according to the type of QC cal-
culation to be performed. The basic techniques of analytical
quality control include calculation of spike recovery, deter-
mination of agreement between sample replicates, and analysis
of quality control standards to verify the accuracy of the
analysis. The programs support all of these techniques,
generating both tabular and graphical output of results.
For laboratories involved in the EPA ’s Contract Laboratory
Program, SMARTLOG and SMARTLAB will offer automatic
generation of sample analysis and QC reports, compatible with
CLP specifications. A report generation package will report
sample data, automatically appending all appropriate flags
and producing the various required QC reports. Report formats
may be customized for individual purposes or to allow for
changes which may occur in the CLP reporting protocols.
INTRODUCT ION
The analysis of environmental samples involves the manipula-
tion of vast amounts of data. In addition to actual analyti-
cal results, information about the sample, including who
submitted it, where it came from, and what is to he deter-
mined, must be recorded. Finally, to completely document
sample handling, a “chain-of-custody, listing the movement
and treatment of the sample, must be maintained.
7—7
-------
In addition to the information which must be kept on each
actual sample, additional analyses must be performed to
verify the validity of sample analysis. While this quality
control work is necessary, if the laboratory is to have any
credibility regarding the reliability of its data, it must be
recognized that maintenance of a QC program will add
significantly to the analytical work load of the labOratory.
Figure 1 illustrates the additional analyses required to
fulfill the requirements set forth in the CL? protocol for
IC? analyses. The left column tabulates measurements which
relate directly to determining sample results. The right
column lists the quality control measurements which are
called for in the CLIP QA/QC procedures. As can be seen from
this table, a full 50% increase in the number of analytical
measurements is required to support the minimum standards of
quality control. It should be noted that this table assumes
20 samples of a similar type. If the 20 samples included
different types, such as ground water and soil extracts,
additional preparation blanks, spiked samples, and duplicate
samples would have to be run to verify the accuracy of the
different sample treatments.
ANALYTICAL HKASUREMKNTS )R SAMPLK ANALYSIS
AND QUALITY CONTROL
SAMPLIE MKASUR 1KNTS QC MEASUREMENTS
Calibration: blank mit Calib Verif: QC solution
3 standards Calib blank
Interf check
____________ Linear range
4 4
Analyai5: 10 samples Continuing Vent: QC solution
Calib blank
10 samples QC solution
Prep blank
Spiked sample
Dup sample
Lab control
__________ Interf check
20 8
Total: 24 12
QC Measure ent Overhead = 12/24 = 50%
Figure 1. Analytical overhead for Quality Control.
7—8
-------
Sinc the amount of work to maintain the necessary quality
control is substantial, it is worth evaluating methods which
will streamline the process. Nothing will shortcut the need
to actually make the quality control measurements. However
the work is not done when the measurements are complete.
Figure 2 itemizes the data handling which begins once the
data has been collected.
First, the data must be maintained by some kind of bookkee
ing system, either manual or computerized. The specifto
regarding the sample must be entered into the system.
followed by entry of the analytical results. Then for
quality control samples, the QC purpose must be identified.
In other words, a sample must be identified as a spike,
duplicate, etc. A calculation must then be performed.
depending on the nature of the quality control technique.
And finally the outcome of the QC measurement must be
reported. While we cannot reduce the task of generating QC
data, we can do things to expedite QO data management.
DATA
MAN
AGEMENT REQUIREMENTS
FOR
QUALITY
CONTROL
1. QC identification
and
login
2. Data entry
3. QC calculations
4. Reporting
GOAL:
to
expedite processing
of QC
information
Figure 2. Data management overhead for Quality Control.
SMARTLOG (R) and SMARTLAB CR) are two laboratory data manage-
ment programs written for the IBM PC-XT, AT or compatible
computers. Both programs store analytical results and offer
quality control routines, which automate the treatment and
reporting of QC data. SMARTLOG is designed to collect
analytical data, perform simple calculations, and print
reports. SMARTLAB, which maintains additional sample
handling and business information, functions as a mini
Laboratory information tianagement System. The programs may
be used separately or together, as indicated in Figure 3. To
illustrate how the quality control functions integrate into
the other laboratory functions, a description of each program
follows.
7—9
-------
LABORATORY SOFTWARE FROM TELECATION ASSOCIATES
St4ARTLOG: The PC Data Station
* Collect data automatically through direct
instrument interface
* Perform calculations for precision, sample weight
and dilution factor correction
* Print customizable reports
* Manage QC information
* Transfer data to another computer or LIMS system
including automatic transfer to SMARTLAB
SMARTLAB: Laboratory Information Management System
* Login sample information and results
* Track sample status
* Generate work sheets and status reports
* Manage QC information for evaluating
completed samples
* Report results and generate invoices
* Archive results and chain-of-custody
SMARTLOG/SMARTLAB: Data Station/LIMS system combination
* Collect data with dedicated SMARTLOG System
interfaced to analytical instrument
* Transfer data to SMARTLAB to merge with main
data base, when convenient
Figure 3. Features of SMARTLOG and SMARTLAB systems.
0LLECTING DATA WITH SNARTLOG
The main purpose of SMARTLOG is to provide a means of
collecting, storing, reviewing, and reporting the results of
laboratory analyses. For many analytical instruments, data
may be collected automatically by simple connection of the
instrument serial output to the RS-232 input of the computer.
Or data may be entered manually through the computer
keyboard. Once in the data base, results may be grouped by
test or by sample number and selected data isolated for easy
review. Figure 4 illustrates a SMARTLOG data base record and
the information which can be maintained for a single analyte
on a single sample.
7—10
-------
SANPLE ANALYSIS RESULTS FOR CADMIUM
Report Name: waterl Test: Cd
SAMPLE INFORMATION
Lab 1 704551 Date: 04/23/87 Sample type: water
Sample ID: ground water from well P87-7
Client Name: Peterson and Company
- INSTRUMENT EDITED DATA
105 1 weight:
2.4600 2.4600 Average I volume: alt.result:
1.9000 1.9000 2.1867 units: ug/L
2.2000 2.2000 SD
0.28024 1
CV 12.82 RESULT: 2.1867
* I
I I
QUALITY CONTROL
Trend: 67-7 QC: spike5l
Spike: 0
F2 list F3 find F5 prey F7 prt record F9 copy HOME
F4 update F6 next F8 prt report PlO menu END
Figure 4. Data maintained in SMARTLOG record.
After data is entered, you may add additional information
about the sample, such as sample identification, sample type,
and who submitted the sample for analysis. Up to ten
replicate determinations of an analytical measurement can be
entered into the data base. Where replicate measurements are
made, the average result, standard deviation, and coefficient
of variation are automatically calculated. Additional
mathematical manipulations can be made on each result, to
correct for such things as sample weight variations, or
dilution factor differences.
If data has been collected automatically from an instrument,
the results will be shown under the “instrument” section of
the screen. This data area is inaccessible to manual entry
or change. Automatically collected data can be edited in the
•‘edited data’ window to eliminate biasing effects of obvious
outlying results, while retaining the original instrument
data to support the validity and credibility of the analysis.
All functions of the software are controlled by selection of
a computer function key, the functions for which are always
labeled at the bottom of the screen (as shown in Figure 4).
Therefore, operation is self-prompting and quick to learn.
To print a report of the selected data, for instance, one
would simply depress the F8 function key.
The format of a data report is completely customizable. A
report generation utility guides the user through the
7—11
-------
creation of customized report formats, starting with
previously existing formats, which can be used as examples
or templates. Any number of customized report formats may be
created and stored for later use.
One of the most valuable functions of SMARTLOG is its appli-
cation as an interim data collection station, for subsequent
data transfer to a larger “Laboratory Information Management
System” (LIMS). Using SMARTLOG and an appropriately config-
ured personal computer, data from a laboratory instrument can
be collected and stored, until it is convenient to transfer
the data into the laboratory’s main data base.
The instrument/personal computer interface requires only
simple direct connection for serial data transfer. The
PC/LIMS connection may be made through a PC network, RS-232,
modem, or even by simple disk transfer. While a built-in
utility provides automatic transfer of SMARTLOG data to the
SMARTLAB PC LIMS system, data can be transferred to any
computer capable of receiving standard ASCII information.
RUNNING THK LABORATORY WITH SMARTLAB
SMARTLAB is a laboratory information management system
designed for the personal computer. SMARTL,AB tracks the
completion status of every sample, from login to archiving of
the results. SMARTLAB identifies what samples are in the
laboratory; what preparations have to be performed; what
analyses have to be run; and what samples have been corn-
p]eted. When a sample is completed and the results approved,
reports and invoices may be automatically generated. Old
sample data and complete chain-of-custody documentation can
be archived on floppy disk and accessed as needed. Figure 5
illustrates the main functions of SMARTLAB.
MAIN FUNCTIONS OF SMARYLAB LIMS
Samples in Progress
LOG IN NEW SAMPLES
ENTER SAMPLE DATA
SAMPLE PBOG ESS PORTS
Completed Saluple5
BILLING & CUSTOMER REPORTS
ACCESS SAMPLE ARCHIVES - --
Utilibi..
CONJIGURE SYSTh4
BACKUP SAMPLE FILES
Figure 5. SMARTLAB Main Menu.
7—12
-------
All of the functions of SMARTLAB are accessed by making a
selection from the main menu. The choices clearly designate
basic laboratory functions. There are no hidden commands to
memorize. Functions under a main heading are similarly
accessed by selecting from a clearly labeled menu. Further
control is exercised by selecting one of 10 labeled function
keys, similar to that already illustrated for SMARTLOG.
SMART LAB: ENTRY OF SAMPLE INFORNAT ION
Many details must be entered into a sample entry log, if
complete documentation and definition of the sample is to be
maintained. Complete name and address information for the
client, sample collection details, and of course, the tests
to be performed on the sample, must all be entered. An
example of the information to be entered for a sample is
shown in Figure 6.
SAMPLE LOGIN INFORMATION
Client Client Ajax Manufacturing Company
information: Quality Control Department
Address 1500 Broadway
Suite A
City Denver State CO Zip 80001
Contact Dan Wallace Phone (303)555-8500
Sample information: Description effluent Trend platte
QC Reference
Collection information: Location Platte River outlet iti
by Whom K. Jones
Preservation method refrigeration
Test information: SECTION - TEST - FEE ($) -
metals As 15.00
metals Pb 15.00
organic 2,4-D 25.00
organic endrin 25.00
wet F 15.00
wet N03 15.00
Figure 6. Client & sample information stored in SMARTLAB. —
7—13
-------
The goal for using a computer for laboratory information
management is to expedite the handling of data and informa—
tion. This includes information entry, as well as informa-
tion access. SMARTLAB provides two means to speed the
process of information entry. When multiple samples from
the same client are logged in at the same time, one key
stroke will copy information from a previous sample into the
record for each additional sample. And where a regular
client is concerned, login is even simpler. By typing in the
client ID, all of the client details are instantly looked up,
and automatically entered. Frequently encountered series of
analyses can similarly be identified and automatically
entered into the log. Even client-specific pricing will be
recalled and used at invoicing time.
SMARTLAB: DEThICHINING SAMPLE STATUS
One of the most powerful benefits of SMARTLIAB is its ability
to determine the status of samples in the laboratory.
SMARTLAB maintains cross-referenced sample indices, listing
the samples in a variety of useful orders. Samples may be
displayed in any of the menu selectable orders by simply
selecting the desired report from the Progress Report menu,
shown in Figure 7. This mode may also be used to generate
lists of work to be done in each analysis section of the
laboratory. Examples of some of the progress report options
are shown in Figures 8 through 11.
SAMPLE PROGRESS REPORTS
Printed Reports
Sample Analysis Backlog by: >Lab 1
Client
Test
Due Date
Anal Method
Sample Prep Backlog by: Prep Method
Screen & Printed Reports
Selected Test Status by: Lab
Client
Section
Test
Anal Method
Prep Method
CUSTOMIZED
Figure 7. SMARTLAB report selection menu.
7—14
-------
ANALYSIS BACKLOG BY LAB NUMBER
December 10, 1986
ore
ore
ore
ore
ore
ore
ore
ore
boiler water
boiler water
due date
11/29/86
11/29/86
11 / 29)86
11/29/86
11/25/86
11/25/86
11125186
11/25 / 86
11/25186
11/25/86
11/25/86
11/25/86
12/08/86
12/08/86
section test
metals Ca
metals Cu
metals Fe
metals Mg
metals Ag
metals Au
metals Cu
metals Mo
metals Ag
metals Au
metals Cu
metals Mo
metals Ca
metals Cu
lab * client description
610032 Western Utilities Company boiler water
boiler water
boiler water
boiler water
610033 Western Utilities Company boiler
boiler
boiler
boiler
Burrow Enterprises
Burrow Enterprises
11129186
11/29/86
11/29/86
11/29/86
metals Ca
metals Cu
metals Fe
metals Mg
wa t e r
wa t e r
wa t er
wa t e r
water
wa t e r
wa t er
wa t e r
wa t er
wa t e r
water
wa t e r
water
wa t e r
wa t e r
water
610103
610104
610116 Waterford Engineering Laboratory
610117 Waterford Engineering Laboratory
610118 Ajax Manufacturing Company
610119 Ajax Manufacturing Company
610120 Western Utilities Company
ground
ground
ground
ground
ground
ground
ground
ground
ground
ground
ground
ground
12/06 /86
12/ 06 /86
12/06 /86
12 /06 186
12/06 /86
12 /06 /86
12/06/86
12106/86
12/06 /86
12/06/86
12/06 /86
12/06 /86
12/07/86
12I07 /86
12/07186
12 /07 I86
12/07/86
12107186
12/07/86
12/07/86
metals
met a Is
organic
organic
wet
wet
metal s
met a is
organic
organic
wet
wet
metals
metals
organic
wet
met a 1 s
met a Is
organic
wet
As
Pb
2 , 4_D
endr in
F
NO 3
As
Pb
2 , 4_D
endr in
F
NO 3
As
Hg
chloroform
CN
As
Hg
chloroform
CN
effluent
effluent
effluent
effluent
effluent
effluent
ef f luent
effluent
Figure 8. List of incomplete semples sorted by laboratory number.
7—15
-------
ANALYSIS BACKLOG BY TEST AND DUE DATE
December 10, 1986
section test lab * client description due date
2,4_D 610116 Waterford Engineering Laboratory
610117 Waterford Engineering Laboratory
Ca
CN 610118 Ajax Manufacturing Company
Ag
As
Au
8
organic
met a 1 5
me t a 1 s
metal s
me t a 1 5
met a is
organic
wet
ground water
12I06/86
ground water
12/06/86
610103
Burrow Enterprises
ore
11/25186
610104
Burrow Enterprises
ore
11/25/86
610122
Geophysical Exploration, Incorpo
sediment
12107186
610123
Geophysical Exploration, Incorpo
sediment
12/07186
610124
Geophysical Exploration, Incorpo
sediment
12/07186
610126
Geophysical Exploration, Incorpo
sediment
12/10/86
610116
Waterford Engineering Laboratory
ground water
12/06/86
610117
Waterford Engineering Laboratory
ground water
12/06186
610118
Ajax Manufacturing Company
effluent
12/07186
610119
Ajax Manufacturing Company
effluent
12107/86
610122
Geophysical Exploration, Incorpo
sediment
12/07/86
610123
Geophysical Exploration, Incorpo
sediment
12/07/86
610124
Geophysical Exploration, Incorpo
sediment
12/07/86
610126
Geophysical Exploration, Incorpo
sediment
12/10/86
610128
Ajax Manufacturing Company
effluent
12/13/86
610129
Ajax Manufacturing Company
effluent
12/13/86
610130
Ajax Manufacturing Company
effluent •
12/13/86
610131
Ajax Manufacturing Company
effluent
12/13I86
610103
Burrow Enterprises
ore
11/25/86
610104
Burrow Enterprises
ore
11125/86
610122
Geophysical Exploration, Incorpo
sediment
12/07/86
610123
Geophysical Exploration, Incorpo
sediment
12/07/86
610124
Geophysical Exploration, Incorpo
sediment
12/07I86
610126
Geophysical Exploration, Incorpo
sediment
12110/86
610121
12/09/86
610125
12/12/86
610127
12I10I86
610032
Western Utilities Company
boiler water
11/29186
610033
Western Utilities Company
boiler water
11/29/86
610120
Western Utilities Company
boiler water
12/08186
610118
Ajax Manufacturing Company
effluent
12/07186
610119
Ajax Manufacturing Company
effluent
12I07/86
610128
Ajax Manufacturing Company
effluent
12113/86
610129
Ajax Manufacturing Company
effluent
12113/86
610130
Ajax Manufacturing Company
effluent
12/13/86
610131
Ajax Manufacturing Company
effluent
12I13/86
Brown
Farms
Brown
Farms
QC
chloroform
soil
soil
soil
Figure 9. List of incomplete samples sorted by test.
effluent 12/07/86
7—16
-------
SAMPLE PREPARATION BACKLOG
December 10, 1986
prep method test lab * client description due date
acetate leach B 610127 OC soil 12110186
610121 Brown Farms soil 12109186
610125 Brown Farms soil 12112186
Mn 610127 QC soil 12110186
610125 Brown Farms soil 12112186
610121 Brown Farms soil 12109!86
Mo 610127 QC soil 12/10 186
610125 Brown Farms soil 12/12186
610121 Brown Farms soil 12109186
N03 610121 Brown Farms soil 12109186
610125 Brown Farms soil 12I1 2I86
610127 OC soil 12I10!86
P 610121 Brown Farms soil 12I09I86
610125 Brown Farms soil 12112186
610127 QC soil 12110186
Figure 10. List of incomplete sample preparations.
bensene extract
distillation
HNO3 digestion
2 , 4_D
chloroform
endr in
CN
F
Ag
As
610117
Waterford Engineering Laboratory
ground
610116
Waterford Engineering Laboratory
ground water
12/06/86
610129
Ajax Manufacturing Company
effluent
12!13l86
610119
Ajax Manufacturing Company
effluent
12/07/86
610128
Ajax Manufacturing Company
effluent
12113I86
610118
Ajax Manufacturing Company
effluent
12I07186
610130
Ajax Manufacturing Company
effluent
12/13/86
610131
Ajax Manufacturing Company
effluent
12/13186
610116
‘Jatertord Engineering Laboratory
ground water
12106/86
610117
Waterford Engineering Laboratory
ground water
12/06/86
610119
Ajax Manufacturing Company
effluent
12/07/86
610128
Ajax Manufacturing Company
effluent
12/13/84
610131
Ajax Manufacturing Company
effluent
12113186
610129
Ajax Manufacturing Company
effluent
12I13/86
610130
Ajax Manufacturing Company
effluent
12I13/86
610118
Ajax Manufacturing Company
effluent
12/07/86
610117
Waterford Engineering Laboratory
ground water
12/06/84
610116
Waterford Engineering Laboratory
ground water
12/06/86
610103
Burrow Enterprises
ore
11125186
610104
Burrow Enterprises
ore
1 1I25/86
610126
Geophysical Exploration, Incorpo
sediment
12!10I86
610124
Geophysical Exploration. Incorpo
sediment
12/07/86
610123
Geophysical Exploration, Incorpo
sediment
12/07186
610122
Geophysical Exploration, Incorpo
sediment
12I07186
610130
Ajax Manufacturing Company
effluent
12/13/86
610118
Ajax Manufacturing Company
effluent
12/07186
610129
Ajax Manufacturing Company
effluent
12/13186
610122
Geophysical Exploration, Incorpo
sediment
12I07186
7—17
-------
ANALYSIS BACKLOG BY ANALYTICAL METHOD
Deceaber 10, 1986
F 610116 Vaterford Engineering Laboratory
610117 Waterford Engineering Laboratory
furnace AA As 610122
610119
610116
Pb
2. 4_D
Geophysical Exploration, Incorpo
Ajax Manufacturing Co pany
Vaterford Engineering Laboratory
anal aethod test lab B client description due date
Ag titration CN 610131
Ajax
Manufacturing
Company
effluent
610118
Ajax
Manufacturing
Company
effluent
610129
Aj&x
Manufacturing
Company
effluent
610128
Ajax
Manufacturing
Coapany
effluent
610130
Ajax
Manufacturing
Coapany
effluent
610119
Ajax Manufacturing
Coapany
effluent
OC
brucine
cold ,apor
flaae AA
Laboratory
Laboratory
NO 3
Hg
Ag
Au
Waterford Engineering
Brown Faras
Brown Faras
Waterford Engineering
Ajax Manufacturing
Coapany
Ajax Manufacturing
Coapany
Ajax Manufacturing
Coapany
Ajax Manufacturing
Coapany
Ajax Manufacturing
Coapany
Ajax Manufacturing
Coapany
10127
61011?
610121
610125
610116
6101 18
610119
610131
610130
610129
610128
610126
610124
610122
610123
610122
610124
610126
610123
Geophysical
Geophysical
Geophysical
Geophysical
Geophysical
Geophysical
Geophysical
Geophysical
12113186
12/07186
12113186
12113186
12113/86
12107/86
12106/86
12106/86
12110186
12106186
12109186
12/12/86
12106186
12/07/86
12107186
12113186
12/13/86
12/13186
12113186
12110186
12! 07186
12/07/86
12107/86
12/07/86
12107/86
12110/86
12/07/86
12107186
12/07/86
12106/86
12/06/86
12106186
12/06/86
12106/86
12/jo’s’
12112/86
12109/86
Exploration,
Exploration,
Exploration,
Exploration,
Exploration,
Exploration.
Exploration,
Exploration,
Incorpo
Incorpo
I ncorpo
Incorpo
Ineorpo
Incorpo
Incorpo
Incorpo
ground water
ground water
soil
ground water
soil
soil
ground water
effluent
effluent
effluent
effluent
effluent
effluent
sod iaent
so d a. n t
sediaent
sod iaent
sod iaeat
sed iaent
sediaent
sediaent
sod iaent
effluent
ground water
ground water
ground water
ground water
ground water
soil
soil
soil
CC
ICP
610117 Vaterford Engineering Laboratory
610116 Wat.rford Engineering Laboratory
610117 Vaterford Engineering Laboratory
610116 Waterford Engineering Laboratory
QC
B 610127
610125 Brown Faras
610121 Brown Faras
Figure 11. List of incoaplete analyses sorted by analytical aethod.
7—18
-------
SMARTLAB: CUSTOM I ZABLE REPORTS & INVOI CES
SMARTLAB will automatically find completed samples, print out
the results in a report, and if desired, automatically gener-
ate an invoice. The form of the sample report and invoice
are easily customizable to your preferred format. Figures 12
and 13 illustrate the standard format for analytical reports
and invoices.
GOLDEN ANALYTICAL SERVICES. INC.
P.O. Box 1486 * Golden, Colorado 80043
phone: (303) 555—2100
- SAMPLE ANALYSIS REPORT -
To: Burrow Enterprises
Route 2
Box 955
Cheyenne WY 89335
Attn: Ed Burrow (490) 888—5525
Our Lab No: 610101 Report Date: 12110186
Sample Description: ore
Your Sample ID: sample I
COLLECTION INFORMATION
Date/TimelLocation: 11115186 10:15 am Claim N5W
Collected bylFreserved by: E.B.
ANALYSIS RESULTS
TEST RESULTS TEST RESULTS
Ag 28 ozlton
Au 12 oslton
Cu 57 oilton
Mo 93 oLlton
All methods in accordance with Environmental Protection Agency
recommended procedures.
Submitted:
GOLDEN ANALYTICAL SERVICES, INC.
Ch cmi st
Figure 12. SMARTLAB sample analysis report
7—19
-------
GOLflEN ANALYTICAL SERVICES, INC.
P.O. Boi 1486
Golden, Colorado 80043
phone (303) 555—2100
To: Burrow Enterprises
Rout. 2
Boz tSS
Cheyenne WY 89335
Invoice date 12/10/86
Attu: Accounts Payable Invoite no. 144207
- INVOICE -
Lab S Sub*itted Saaple Identification Charge
Burrow Enterpr isis
610101 11119186 sample I $4796
610102 11119I86 saaple 2 $30.56
Total $78.52
= = = = = =
Ter*s: net 30
Thank you.
Figure 13. SNARTLAB standard invoice.
7—20
-------
In addition to the reports and invoices which go out to your
customers, SMARTLAB can generate useful in-house accounting
reports, such as the accounts receivable report, shown in
Figure 14.
GOLDEN ANALYTICAL SERVICES, INC.
- ACCOUNTS RECEIVABLE -
Samples Completed and Billed
invoice
client lab * mv 8 date charges
Ajax Manufacturing Company 610016 144204 11127186 $140.20
610114 144205 11/28/86 $48.00
Client Total $188.20
= = = = = = =
Burrow Enterprise5 610101 144207 12/10/86 $47.96
610102 144207 12110/86 $30.56
Client Total $78.52
= = = = = = = =
Western Utilities Company 610113 144206 11/28/86 $32.00
610110 144204 11/28/86 $32.00
610112 144206 11/28/86 $32.00
Client Total $96.00
Total Receivables $362.72
=
Figure 14. SMARTLAB accounts receivable report.
7—21
-------
SMARTLAB: ACCESS TO SANPLE DOCUMENTAT ION
The completion of a sample does not mean that there will
never be a need to review the results. Further, for environ-
mental and legal work, it is becoming more and more important
to document the movement and treatment of every sample, with
name, date, and treatment information (“chain-of-custody”).
SMARTLAB saves all completed sample information on the
system’s hard disk storage, with long-term archiving possible
on floppy disk. Figures 15 and 16 show examples of archived
sample results and chain-of-custody.
Ails Manufacturing Company
Quality Control Department
1500 Broadway
Suite A
Denver CO 80101
Joe Collier (303) 555—8500
TEST
As
Cr
Hg
Pb
Se
ch lorof ora
aeth chloride
t oluene
vinyl chloride
CN
F
RESULTS
38 ugIL
58.2 ugIL
0.4 ugIL
25 ugIL
83.8 ugIL
0.71 uglL
1.63 ugIL
5.1 uqIL
0.81 ugIL
26.4 ug/L
73 ug/L
Fzgure 15.
SMARTLAB saaple archive5.
SAMPLE ARCHIVES REPORT for:
Lab *
Logged in:
Co p1eted:
Description:
Saaple ID:
610016
10128186
1 1 119I86
effluent
Platte *1—861025
7—22
-------
As 0.5% HNO3
KLG 10128186
Cr 0.5% HNO3
KLG 10/28186
Hg 0.5% HNO3
KLG 10/28/86
Pb 0.5% HNO3
KLG 10/28/86
Se 0.5% HNO3
KLG 10/28/86
ref r i g era t i on
KLG 10128/86
ref r i g era t i on
KLG 10/28/86
ref r i g era t ion
KLG 10128/86
refrigeration
KLG 10/28/86
ref r i g era t i on
KLG 10/28/86
HNO3 digestion
RAN 10/28/86
HNO3 digestion
RAN 10/28/86
HNO3 digestion
RAN 10I28I86
HNO3 digestion
RAN 10/28/86
HNO3 digestion
RAN 10128/86
benzene extract
RAN 10/28/86
benzene extract
RAN 10128/86
benzene extract
RAN 10/28/86
distil lat ion
RAN 10/28/86
Figure 16
SMARTLAB chain-of-custody report.
Chain—of—Custody Report
Lab t: 610016 Ajax Manufacturing Company
Description: effluent Denver CO 80101
Sample ID: Platte $1—861025
COLLECTION INFORMATION
Date/Time/Location: 10(25/863 pm Platte River output *1
Collected by/Preserved by: K. Jones refrigeration
LABORATORY HANDLING INFORMATION
Sample Sample Sample
Test Preservation Preparation Analysis
Approval
ROB 11/19(86
furnace AA
OLE 10/28/86
ICP
DLB 10/28/86
cold vapor
DLB 10128/86
furnace AA
OLE 10/28/86
furnace AA
OLE 10/28/86
CC
DLE 10128/86
CC
OLE 10/28/86
GC
DLB 10/28/86
Ag titration
OLE 10/28/86
ion electrode
OLE 10/28/86
ch lo rot o rm
aeth chloride
to luene
CN
F
ROB
ROB
ROE
ROB
ROB
ROB
ROB
ROB
RDE
11/19/86
11/19 / 86
11/19/86
11/19/86
11 / 19186
11 / 19/86
11/19 / 86
11/19/86
11/19/86
7—23
-------
MAINTENANCE OF LABORATORY QC DOCUMENTATION
Both SMARTLOG and SMARTLAB offer similar QC capabilities. In
SMARTLAB access to QC routines is available to the user at
the time that sample results are approved, thus providing the
necessary information to make an informed decision on the
validity of the results. Standard techniques of quality
control addressed by SMARTLOG and SMARTLAB include: determi-
nation of spike recovery, repeatability of replicate
analyses, and agreement of results with known or certified
values. Tabulated and graphed quality control documentation
can be stored for later reference and printed for
distribution.
Control of the QC function is exercised by first identifying
the QC samples in the data base. This is done by providing a
“QC Ref erence name in the designated field of a sample
record. For spiked samples, a spike value must also be indi-
cated in the data field provided. In addition to standard
quality control procedures, long-term trends for designated
samples may be monitored by entering a “Trend name in the
data field provided for this purpose.
When the user wishes to process QC or Trend information, the
desired data is isolated from all other data in the data base
by selecting a specified QC or Trend name. Then by simply
depressing the labeled function key indicating the desired QC
technique, calculations are performed and the data is
tabulated in a standard format determined by the quality
control procedure selected. Figures 17 through 20 present
examples of printed output from the QC routines.
In addition to the standard QC displays, both tables and
graphs are customizable to accomplish various additional
graphing and tabulating functions, which might be required by
a particular laboratory. A graph generation utility aides
the user in designing custom graphs.
For laboratories involved in the Contract Laboratory Program
or who wish to report data in the CLP format, a new quality
control report option, which will automatically generate CLP
reports and quality control forms, is under development.
When complete, this option will automatically determine the
flags to be appended to data, as well as generate all of the
QC forms to be submitted to the EPA. Working prototypes
based on the current CL? protocols have been demonstrated.
Completion of this work is pending the establishment of the
new protocols by the EPA.
7—24
-------
GENERAL SOFTWARE TOOLS FOR THE LABORATORY
SMARTLOG and SMARTLAB are dedicated laboratory software
programs. A user may operate these programs fully, without
any knowledge of the software system used. However, for
those who wish to use it, additional capability exists.
SMARTLOG and SMARTLAB were written around Innovative
Software’s SMART System, an integrated software package
providing a word processor, data base manager, spreadsheet,
communications capability, and more. SMART was hand picked
as a basis for Telecation’s laboratory software, due to its
power, speed and programmability. As a result, all of the
functions of SMART are available to the SMARTLOG or SMARTLAB
user. In addition, many of the features of SMARTLOG and
SMARTLAB may be customized beyond the utilities provided, by
the experienced SMART user. All of the documentation which
accompanies the SMART system is provided to assist the user,
who wishes to take advantage of this capability.
7—25
-------
SPIKE RECOVERY
Figure 17.
Spike recovery table.
QUALITY CONTROL TABLE
Lead Precision
Figure 18.
Quality control table.
lab *
test
units
result
spike
Recovery % Recovery
610032
Ca
ugIL
1000.0
0.0
610033
Ca
ugIL
1090.0
100.0
90.0
90.0
610032
Cu
ug/L
20.0
0.0
610033
Cu
ug/L
48.0
25.0
28.0
112.0
610032
Fe
ug/L
30.0
0.0
610033
Fe
ugIL
74.0
50.0
44.0
88.0
610032
Mg
ug/L
2000.0
0.0
610033
Mg
ug/L
2487.0
500.0
487.0
97.4
lab *
test
units
result
Average
—
10 %
+
10 %
610227
Pb
ug/L
26.0
25.2
22.7
27.7
610228
Pb
ug/L
24.6
.
.
.
610229
Pb
ug/L
24.2
.
.
.
610230
Pb
ug/L
25.6
.
610231
Pb
ugIL
28.8
.
.
.
610232
Pb
ugIL
23.2
.
.
610233
Pb
ug/L
23.4
.
.
.
610234
Pb
ug/L
24.2
.
.
610235
Pb
ugfL
26.7
.
.
.
610236
Pb
ug/L
25.3
25.2
22.7
27.7
7—26
-------
SMARTLAB
Lead Quality Control
Precision Range: +/-u
Chart
10%
‘I
a’
-4
U.
C ,)
- .
1 )
‘.4 Q)
35.0
32,5
30.0
27.5
25.0
22.5
20.0
1
0
2
2
7
1
0
2
2
8
1
0
2
2
9
6
1
0
2
3
0
6
1
0
2
3
1
6
1
0
2
3
2
6 6
1 1
6
1
0
2
3
3
1
0
2
3
4
0
2
3
S
0
2
3
6
Lab#
-------
0
Telecation Associates
SMARTLAB
Lube Oil Trend Analysis
1.4
I L .
U,
U)
tZ)
400
360
320
280
240
200
160
120
80
40
0
1
0
2
3
7
6
1
0
2
3
B
6
1
0
2
3
9
6
1
0
2
4
0
6
1
0
2
4
1
6
1
0
2
4
2
6
1
0
2
4
3
6 6 6
1 1 1
0
2
4
4
0
2
4
5
0
2
4
6
Lab#
-------
LABORATORY AND FIELD AUDITS AS PART OF THE EPA
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY (HWERL)
QUALITY ASSURANCE PROGRAM
W. Burton Blackburn, Staff Scientist, S-CUBED Division of Maxwell Labora-
tories, La Jolla, California 92038-1620; Guy F. Simes, Quality Assurance Officer,
Hazardous Waste Engineering Research Laboratory, U.S. Environmental
Protection Agency, Cincinnati, Ohio 45268
ABSTRACT
Audfts are an important and integral part of the EPA HWERL Quality Assurance
Program. As part of this overall QA Program, audits are used to determine
contractor compliance with quality assurance plans and to assess the overall
quality of data collected during data gathering or data generation activities.
Additionally, audits are useful in evaluating the procedures used in collecting
and analyzing samples, in all facets of data quality, and in management of the
activity. Often, an audit of a sampling or analytical activity will reveal a problem
which otherwise may have gone undetected until the end of the project. In
such cases, on-site Corrective Action Recommendations are made.
Four different types of audits are performed under the EPA HWERL QA
Program: Management Systems Audits (MSA), Technical Systems Audits
(TSA), Performance Evaluation Audits (PEA), and Audits of Data Quality (ADO).
Such audits of HWERL projects are conducted by the Quality Assurance Officer
or a designated representative. Each type of audit has established Standard
Operating Procedures and has a well-defined direction in evaluating a
particular aspect of field and/or laboratory operations.
INTRODUCTION
Audits are an important and integral part of the EPA Hazardous Waste
Engineering Research Laboratory (HWERL) Quality Assurance Program. As
part of this overall QA Program, audits are used to determine the compliance of
contractors, in-house groups, and academic groups with previously approved
Quality Assurance Program and/or Project Plans. Also, audits can be
implemented to assess the overall quality of data collected during data
gathering or data generation activities, in evaluating sampling and analytical
procedures, and assessing the management of an activity.
7—29
-------
Four different types of audits are conducted under the HWERL QA Program:
(1) Technical Systems Audit (TSA) - a qualitative on-site
evaluation of all components of the measurement systems,
including personnel.
(2) Performance Evaluation Audit (PEA) - a quantitative evaluation
of critical measurements through an organization’s analysis of
Performance Evaluation (PE) samples.
(3) Audit of Data Quality (ADQ) - an on-site assessment of the
methods used to collect, interpret, and report the information
used to characterize data quality.
(4) Management Systems Audit (MSA) - an on-site evaluation of
an organization’s quality assurance management system.
In general, audits of HWERL projects are performed by the Quality Assurance
Officer (QAO), or by a designated representative.
AUDIT DESCRIPTION AND PROCEDURES
Technical Systems Audit
The Technical Systems Audit (ISA) is one of the most utilized types of audits of
the four types of HWERL audits. All aspects of sampling or data generation
can be evaluated and the TSA may be project-specific or organization-specific.
The TSA is particularly valuable in that it can be quickly implemented and
completed, thus identifying potential problems and recommending corrective
action while the subject project is still in progress. The TSA is conducted by
one or two scientists or engineers, depending upon the nature of the project.
These reviewers are generally chosen on a project-specific basis, based upon
a combination of their quality assurance and related technical expertise. The
QA Project Plan provides the basis for the audit, which generally takes one to
two days to complete.
Some questions which may be of relevance to a ISA include:
• Is the project adequately staffed with personnel of relevant
background to accomplish the objectives of the project? Is
project organization and management sufficient to meet these
same objectives?
7—30
-------
• Does the organization devote sufficient resources and personnel
(i.e., a Quality Assurance Officer) to an established Quality
Assurance Program?
• Is the equipment used on the project being used for the purpose
for which it was intended and is it being used properly?
• Are proper chain-of-custody and sample handling procedures
being followed?
• Are personnel involved with the project knowledgeable, properly
trained, and adequately supervised?
• Are the facilities being used sufficient to meet the needs of the
project?
• Are sampling and analytical equipment being maintained
properly, calibrated properly, and used in an appropriate
manner?
• Are quality assurance and quality control procedures outlined by
the organization in such documents as Work Plans and Quality
Assurance Project Plans being followed according to planned
procedures?
• Do standard operating procedures exist and are they up to date
and complete?
• Do data validation procedures exist and is there sufficient review
of data to minimize reporting of calculation errors?
• Is documentation of sampling and analytical procedures, data,
QAIQC information, and corrective action adequate and well
organized?
The audit proceeds in five phases:
(1) A pre-visit planning phase is begun where the reviewers
familiarize themselves with relevant documents (i.e., QA
Project Plan) and make contact with on-site project
management. A Pre-Visit Worksheet is sent to project
7—31
-------
management, which is returned to the reviewers prior to the
site visit. This worksheet provides the reviewers with current
information regarding project organization, personnel, and
equipment, as well as any changes to procedures given in the
QA Project Plan.
(2) The site visit begins with an initial meeting conducted by the
reviewers and attended by key project personnel. Review
procedures and objectives are discussed and an agenda is
agreed upon. Following this discussion, the QA Project Plan
is examined and questions or clarifications needed by the
reviewers are addressed.
(3) A facilities tour is conducted where project-related
measurement equipment is viewed and responsible analysts
are interviewed. In the case of a sampling effort review,
sampling equipment and locations are seen. If possible,
actual sampling and analytical work in progress is observed.
(4) A review of documentation and data handling procedures is
conducted. Usually, this is accomplished by selecting one or
more samples at random and tracing their progress through
sampling and analytical procedures and up to final data
reporting.
(5) Following an adjournment where the reviewers assimilate
notes and observations, a debriefing and summary meeting
is held. The reviewers discuss concerns and rank them in
order of importance (i.e., critical, major, and minor). Project
personnel are encouraged to engage in a constructive
discussion with the reviewers and offer clarifications or
additional information to the reviewers in response to the
noted concerns. In concluding this meeting, and in cons ider-
ation of discussions with project personnel, Corrective Action
Recommendations (CAR) may be made. These are
documented, on site, on a CAR Form (Figure 1). The HWERL
Technical Project Officer receives this form within three days
and is responsible for implementation and follow-up of any
identified corrective action.
7—32
-------
8.6.2
CORRECTIVE ACTION RECOMMENDATION (CAR) FORM FOR HWERL
Lab Workpian No: rr I I I I Project Category:________________________
QA I.D. No.:_________________________
Organization Reviewed:_____________________________________________________________________
Project Title:
_______________________________ Affiliation:____________________________
Review Type:__________________________ On-Site Location:_____________________
Rating:_____________________________________ Technical P0 :
I. Date Problem Identified:_________________________
Probtem serious enough that Part I I must be completed by the HWERL technical project officer and
submitted to the HWERL. QA Officer? Yes No
Nature of Problem:
Recommended Action:
Corrective Action Recommendation Form (Part I) has been Teviewed by the organization’s on-site
representative:
Signature: _________________________________________________ Date:___________________
(Organization’s On-Site Representative)
It. Date Corrective Action Taken:
Summary of Corrective Action:
QA Activity Implemented to Prevent Future Occurrences:
Effectiveness of Corrective Action:
SIgr ature (HWERL technical P0) : Date:_______________
HWERL (CAR)
(April 1987)
Figure 1. Corrective Action Recommendation (CAR) Form
7—33
-------
Within one week following the on-site portion of the TSA, the reviewing group
submits to the reviewed group a Draft TSA Summary Report detailing the
positive aspects of the review, any concerns ranked by order of importance,
and any corrective action recommendations, along with an audit rating (see
Table 1). The reviewed group is allowed one week to read and respond to the
TSA Summary Report. This response is included with the Final TSA Summary
Report, which is then forwarded to the HWERL Technical Project Officer with a
copy to the HWERL QAO. An important aspect of the HWERL TSA procedure
is that the review is conducted, corrective action recommendations made on
site via the CAR Form, and a report submitted to the HWERL Technical Project
Officer in about three weeks. This timeliness allows for prompt resolution of
any problems which may be present and gives the Technical Project Officer an
immediate assessment of the quality of sampling and analytical activities
associated with the subject project.
Performance Evaluation Audit
A Performance Evaluation Audit (PEA) is a quantitative evaluation of
measurement systems. Usually, only the most critical measurements, as
designated in the QA Project Plan, are evaluated. The PEA involves
measurement or analysis of a reference material having associated with it a
known value or composition. The value or composition of the reference
material is certified, or at least verified, prior to use. The USEPA Environmental
Monitoring and Support Laboratory in Cincinnati is active in preparation and
certification of these materials.
Among the questions that may be determined from a PEA are:
• What is the bias and precision of the measurement system at
the time of the audit?
• Under the ideal conditions presented in analysis of a
performance evaluation sample, are the results generated by a
given instrument and from a given method under control?
• If the results of previous audits are available, has data quality
significantly changed?
Ideally, the performance evaluation sample should be submitted as a blind
sample so that it receives treatment similar to other routine samples. The
sample may be analyzed with an audit team present so that the actual
7—34
-------
procedures used and the abilities of specific project personnel can be
evaluated. The PEA is conducted as follows:
(1) Submission of performance evaluation sample to the
laboratory.
(2) Results are returned to HWERL and subjected to a statistical
analysis involving between-laboratory analysis. Results are
considered acceptable if they are within defined limits (see
Table 2).
(3) If results are rated Conditionally Acceptable or Not
Acceptable, corrective action is required of the subject
laboratory. Such corrective action must be documented and
submitted to the HWERL Project Officer and QAO for
approval.
(4) HWERL then approves the corrective action or makes further
recommendations.
A PEA does not provide any information on overall project documentation or
implementation of quality assurance because it is very narrow in scope. It can
be very complimentary to a TSA because it provides quantitative data to
supplement the qualitative data generated during the TSA.
Audit of Data Quality
Due to several major weaknesses observed in the field application of Audits of
Data Quality (ADO), HWERL has not undertaken such audits during FY86 and
FY87. Instead, HWERL is involved in development of a different strategy for
ADQs. The positive, comprehensive approach developed by HWERL to
conduct Audits of Data Quality (ADO) is a structured two-tiered approach. The
first tier consists of a multipart questionnaire and rating scheme that enables
the qualitative assessment of the sufficiency of verifiable documentation of data
and their associated data quality indicators (e.g., precision, accuracy,
representativeness, completeness, comparability, limits of detectability). The
second tier consists of a two-part questionnaire and rating scheme that
enables quantitative evaluation of data quality and end-user data acceptability.
The HWERL ADO approach combines many of the advantages of other
approaches, including:
7—35
-------
• The use of three-valued logic (Yes/no/not applicable), to
minimize subjectivity.
• The use of weighted ratings, to emphasize the most critical
evaluators of data quality.
• well-defined point-scoring and rating schemes, to provide
unambiguous and comparable audit results.
• Well-defined acceptance criteria, to establish clear reference
points from which corrective action decisions or other
administrative judgment calls can be made.
• Independent rating of typical subsections of the questionnaires,
to enable detailed evaluation of project strengths and localized
identification of problem areas.
Management Systems Audit
The Management Systems Audit (MSA) is not presently performed by HWERL;
rather, it has historically been performed by either EPA’s Quality Assurance
Management Staff (QAMS) or Office of Environmental Engineering and
Technology Demonstration (OEETD). The audit is to investigate HWERL QA
procedures. An evaluation of the adequacy of the internal management
systems necessary for the successful implementation of a quality assurance
program is made. Since HWERL does not conduct these audits, detailed
procedures are not described here.
RESULTS OF FY86 AND FY87 AUDITS
Following the completion of a ISA or PEA, an evaluation rating will be
assigned. These ratings and the corrective actions required are given in Tables
1 and 2. It should be noted that HWERL policy dictates that measurement
work should not proceed for projects given Conditionally Acceptable or Not
Acceptable review or audit ratings. Only deficient measurement systems may
be stopped where other measurement systems within a project are acceptable.
Also, work may not be stopped at all if data quality would further suffer through
this action and corrective action can be quickly and easily implemented.
Table 3 gives a summary of the types of audits conducted in FY86 and FY87
and the ratings given.
7—36
-------
Table 1. HWERL Rating System for TSA
Explanation/Action Required
Rating by the Technical Project Officer Status of Measurements
Acceptable None Work Proceeds
Acceptable Minor deficiencies are revealed by the audit. Work Proceeds
with Minimum criteria are satisfied and good data
Qualifications quality seems likely; qualifications on the
possible limitations of the data are noted
and some corrective actions may be
recommended. The recommendations may
be implemented at the Technical Project
Officer’s discretion.
Conditionally Major concerns are identified by the audit. Measurement work should
Acceptable Corrective action* is required by the not proceed for Category I
Technical Project Officer to ensure quality or Il Projects.**
results
Not Acceptable Critical deficiencies are revealed by the audit Measurement work should
that require the Technical Project Officer to not proceed until deficien-
take substantial corrective action* before cies are resolved.
measurement work proceeds.
* Part U of the Corrective Action Recommendation Form is to be used by the Technical Project
Officer in documenting the resolution of cited deficiencies.
** Category III and IV Projects: Measurement work may proceed, but the deficiencies that were
cited in the audit report must be resolved within thirty (30) days. If the deficiencies are not
resolved during the 30-day grace period, then the measurement work should not proceed.
7—37
-------
Table 2. HWERL Rating System for PEA
Rating Explanation/Action Required by Technical Project Officer
Acceptable All critical measurements are within a 99 percent confidence
interval calculated from available performance evaluation data of
EPA and State laboratories.
Conditionally Eighty percent or more of the critical measurements are within the
Acceptable above-cited 99 percent confidence interval. Those critical
measurements which are outside of the confidence interval are
identified for corrective action at the direction of the Technical
Project Officer.
Not Acceptable Less than 80 percent of the critical measurements are within the
above-cited 99 percent confidence interval. The project, itself, is
flagged and those critical measurements which are outside of the
confidence interval are identified for corrective action at the
direction of the Technical Project Officer.
Note : Measurements which are not critical to meeting the project’s QA objectives and which are
not within the 99 percent confidence interval are flagged, but do not affect the rating.
Table 3. HWERL Audits in FY86 and FY87 1
Audit Type
TSA
Rating 2
Number Conducted A NQ
14 4 3
C/A
5
N/A
2
PEA
30 7 -
21
2
1 mrough June 1987
2 A - Acceptable
NQ - Acceptable with Qualifications
C/A - Conditionally Acceptable
N/A - Not Acceptable
7—38
-------
REVIEW OF AUDITS AND ANALYSES IN NEW JERSEY’ S
LABORATORY CERTIFICATION PROGRAM
Dennis M. Stainken, Robert L. Fisher, C. Don Bowyer, New Jersey
Department of Environmental Protection, Office of Quality Assurance,
Trenton, New Jersey
ABSTRACT
The New Jersey Department of Environmental Protection administers a
regulatory structure based on data of documented quality. One of
the means by which the State controls and administers quality
assurance activities is through the use of a State Laboratory
Certification Program.
Analytical data submitted to the State by permit holders must be
provided by Certified Labs. The current program registers 462 labs,
certifies by categories (e.g. microbiology, organics, inorganics,
wet/limited chemistry, radiation and bioassay), and covers
activities under NPDES, RCRA, Safe Drinking Water Act and various
State programs. Within categories, labs are also certified by
instrument where appropriate (e.g. GC, GC/MS). The Certification
program functions by use of audits and performance evaluation (PE)
samples, and through data review and complaint procedures from
permit bureaus.
The certification program currently registers 232 labs for drinking
water analyses and 381 labs for water pollution work with
approximately 147 labs using CC or CCIMS, 202 labs using AA, 27 labs
ICAP and 10 labs using HPLC. Many of the labs tend to provide
multiple services, and over 200 labs in the program offer
microbiological services.
Evaluation of PE samples has indicated that labs have the most
difficulty analyzing drinking water for Hg, CU d 3 , silvex,
dibromochioromethane, 2, 4—D, fluoride, Ag and Ba, while for water
pollution they met Me L, Chlorobonzene, Total Residual Chloride and
1, 2—Dichioroethane. Auditing has revealed that the most common lab
deficiencies are inadequate records, inadequate instrument tuning,
errors in Identification, problems with spikes, surrogates,
recoveries and minimum detection limit, and blank contamination and
inadequate standards.
INIRODUCTION
The New Jersey Department of Environmental Protection (DEP)
administers numerous regulatory programs to protect the environment
and public health. Effective administration of these programs is
based on data of documented quality. The Department annually
prepares a State Quality Assurance Program Plan which Identifies the
7—39
-------
processes which the Department quality assurance and quality control
activities will follow to achieve this goal. This Program Plan and
attendant work plan applies to all Department activities in
environmental monitoring and clean-up under the Safe Drinking Water
Act (SDWA), the Clear Water Act, R( A, C CJA, the Clean Air Act,
etc. The state QA program uses several elements to control QA
activities. These elements include the use of specialized QA/QC
requirements in permits, use of QA Program Plans and Work Plans,
administration of State analytical service8 contracts, use of lab
and field audits and performance evaluation (PE) samples, and a
State Laboratory Certification Program.
According to DEP’s regulations, analytical data submitted to the
State by permit holders must be provided by Certified Labs. This
laboratory certification program was established In 1981 (1) and
will be revised during 1987—88. The current certification program
registers 462 labs, and covers activities under NJPD , RQ .A, SDWA
and various State programs (e.g. A280 drinking water, ECRA).
Specific areas of RCRA programs currently requiring certification
include issues involving potable water, groundwater and NJPDES
permits. The new laboratory certification regulations under
promulgation will provide additional support for the RCRA programs
The laboratory certification program certifies by categories:
microbiology, organics, inorganics, wet/limited chemistry, radiation
and bioassays. Within categories, labs are also certified by
instrument (e.g. CC, GC/MS, HPLC) and by method regulations will
include several novel areas including: ames testing, sludge
analyses, several areas in microbiology (beach and pool monitoring),
potable water and E CRA programs (2).
Approximately 462 laboratories are in the program (Table 1) wIth
approximately 147 labs providing CC data, 91 labs providing CC/MS
data, 202 labs using AA, 27 labs using ICAP and 10 labs using RPLC.
Over 200 labs in the program provide microbiological services with
many labs providing multiple services. The certification program
Includes approximately 100 out of state labs including Canada. The
certification program functions by use of audits and PE samples,
with permit bureaus receiving QA data submitted within permit
requirements. Typically, results fro. more than 28,000 organlcs
analyses and 11,000 inorganics analyses are submitted annually
through permitting, compliance and monitoring activities.
7—40
-------
TABLE 1
CURRENT NUMBER. OF LABORATORIES IN THE NEW JERSEY
LABORATORY CERTIFICATION PROGRAM
Category Number of Certified Labs
Micro 201
Ltd. ChemIstry 411
Atomic Absorpt. 202
Gas chromatography 147
Radiochemistry 7
Bioassay 18
Program
Drinking Water 232
A280 46
Water Pollution 381
Radiochem 7
Bioassay 18
Laboratories entering the certification program submit a fee and
application which is reviewed for credentials, qualifications, etc.
Once the application is approved, a set of performance evaluation
(PE) samples are sent. After a lab has correctly analyzed the set
of PE samples, an on—site audit is conducted. When a lab has passed
these procedures, it is then certified.
Once ift the program, laboratories are audited (announced and
unannounced) approximately once per year with emphasis placed on
labs with inconsistent records, those handling high volume and/or
sensitive work, and those for which we receive complaints or faulty
data. These audits include verification of personnel, materials,
adherence to regulations, and includes tracking of randomly picked
actual sample numbers and data submitted by the lab to DEP, to
verify results. Depending on the certification category, labs must
also successfully analyze PE samples each year. Enforcement of the
lab certification regulations include the use of fines, suspensIons,
decertification or combinations of them.
The lab certification regulations stipulate the procedures and/or
methods to be used for analyses. For moat analyses, including
organics and inorganics, these methods are incorporated by reference
from the Federal procedures in 40 CFR part 136 and 141. Within the
context of the certification program methodologies, specific quality
control items are required. These include the maintenance of
control charts, spikes, duplicates, blanks, establishing a minimum
detection limit, calibrations, correcting baselines, linearity
checks of standards (and analyzing within the correct portions of
7—41
-------
the curve) and analysis of PE’s and standards. These and other
items described below are some of the items routinely evaluated in
auditing lab performance.
The Certification Program uses several quality control enforcement
mechanisms, routine and semi—routine, to enforce the program and
compliance of the laboratories in following prescribed methods and
procedures. The routine part of the program consists of monitoring
the analysis of PE samples for drinking water (SDWA and A280) and
water pollution (NJPDES). In addition, labs and permittees
participate in EPA’s periodic assessment of matrix blank samples in
1 4R studies. Routine on—site audits are conducted to verify
adherence to methods and regulations and to evaluate analytical
QA/QC records. The credentials of responsible individuals and
analysts are also reviewed and cross checked to be sure education is
appropriate and degrees are from accredited schools. Part of the
semi—routine enforcement involves investigating complaints from
personnel outside DEP and from inside DEP from permit bureaus and/or
data reviewers. These semi—routine procedures generally culminate
in an extensive on—site Investigation to verify adherence to the
regulations and verify analytical results. Verification of
analytical results involves taking sample data submitted to DEP and
tracking this data through all the steps from when the sample
entered the lab and was analyzed to when the data was forwarded.
The audits are extremely thorough and labs must support the data
generated.
The enforcement of provisions of these audits can culminate in a
variety of actions ranging from administrative consent orders (ACO),
fines, suspension and/or decertification depending on the severity
of the findings. The most common problems encountered with lab8 are
ranked in order of frequency: 1) failed PE’s; 2) audit
discrepancies (failures in adhering to methods and regulations); 3)
and credentials/personnel problems. Occasionally, fraudulent
activities are identif led concerning falsification of data, analyses
and records which will result in vigorous prosecution by the State.
The current number of laboratories applied or certified in the
program are listed in Table 2 by instrumental parameter. This
represents most of the analytical chemical work performed in the
program. The monitoring of performance evaluation samples for
drinking water and water pollution parameters has identif led some
characteristics of performance and discrepancies. Table 3 presents
a summary of results of PE sample evaluation for drinking water
parameters from 1982 to 1985. Although the number of labs in the
studies Increased over four years, the percent of labs with
acceptable results remained relatively constant from 1983—1985. An
example of some of the PE samples missed by laboratories is
represented in Table 4. Although some of the water study analytes
differ, two analytes (Bromodichioroethane — BCDN and Bromoform —
7—42
-------
CHBr 3 ) appear to be commonly missed in both studies. The rationale
for why analytes are not acceptably analyzed varies and ranges from
solvent contamination (e.g., methylerie chloride) to inadequate
laboratory analyses in determining MDLs, linear ranges, following
prescribed procedures, etc. Typical PE sample water study results
are provided in Table 5. This represents the percent of
laboratories achieving acceptable results for each analyte in
drinking water. In this study ) 75% of the analyses conducted by
labs were acceptable of almost 50% of the analytes. In Water Study
15 for water pollution PE analytes, 7 75% of the analytical results
were acceptable from 31 out of 46 analytes (67%). Although the
analytes, MDLS, procedures etc. vary between the PE water studies
and analytes, the number of laboratories generally performing
acceptably seems to average approximately 75—85%.
TABLE 2
CURRENT NUMBER OF LABORATORIES APPLIED OR CERTIFIED TO
SUBMIT ANALYSIS TO DEP PER INSTRUMENTAL PARAMETER
CC/MS 91
CC 147
AA 202
ICAP 27
Ion chromatography 91
HPLC 10
TABLE 3
SUMMARY OF DRINKING WATER STUDY PE RESULTS
WS—O1O WS—012 WS—014 WS—0l6
Study Year 1982 1983 1984 1985
Total Samples Run 1830 1830 2292 1964
% Acceptable 70 79 81 81
% With Error 50% 15 19 13 6
2 With Error ) 25% 41 41 28 30
2 With Error )l% 57 74 72 78
Labs In Study 77 93 95 94
7—43
-------
TABLE 4
EXAMPLES OF “PROBL 4” ANALYSIS IDENTIFIED IN
PE SAMPLE STUDIES DURING 1985
Drinking Water (WSO16) Water Pollution (WS015)
Z Acceptable Z Acceptable
Hg 55 l4eCl 53
CHCL 3 55 Qilorobeuzene 53
Silvex 58 Total Residual
DBQI 58 Qiloride 55
2,4—D 61 l,2—Dichloroethane 55
F 61 T N 65
CHBr 3 63 l,l,l—Tricb.loroethane 65
BDQ( 68 QIBr 3 65
Ag 72 Oil & Grease 68
Ba 73 BDC34 70
CCL 4 70
TABLE 5
N I JERSEY WSO16 RESULTS BY PARAMETER
(1985 DRINKING WATER)
PARAKETER Z ACCEPTABLE
As 74
Ba 73
Cd 88
Cr 79
Pb 84
Hg 55
Se 76
Ag 72
NO 3 75
P 61
ENDLtN 86
LINRANE 83
MAAmOXYCHLO R 89
TOXAPHENE 78
2,4—D 61
SILVEX 58
cHc l 3 55
HBr 3 63
BROM0DI HLOROMETBANE 68
DThROMOCHLOROMETHANE 58
7—44
-------
On—site audits constitute an important part of the compliance
program. Discrepancies identified in on—site audits fall into
several broad categories which include: lack of documentation,
improper records, incorrect lab practices, sample storage and
holding time excursion, inappropriate methods, equipment, lack of
experienced personnel and inadequate accredited credentials.Table 6
lists some of the common on—site audit discrepancies found in
auditing. The types and frequency of discrepancies found in audits
can be ranked in order of frequency from 1 (most common) to 7.
These are;
1. MDL — use of incorrect approach or failure to determine.
2. Incorrect adherence to method.
3 Calibration — failure to calibrate, maintain, verify purity
and concentration.
4. Control Qiarts — failure to establish, maintain, take
corrective actions.
5. Spike/recovery — inappropriate values or ranges relative to
analytical range.
6. Standards — lack of/or inadequate.
7. Linearity — failure to establish or work within linear
regions.
TABLE 6.
COMMON PROBLEMS ENCOUNTERED AUDITING
GC/MS not tuned, libraries & compounds mistakenly
identified
Problems with spikes, surrogates, recoveries, MDL
Sample preparation/extraction problems
GC columns incorrect or not monitored
Inadequate documentation
Standards not run
Inadequate personnel/lack of experience & credentials
Illustrations of some of the poorly analyzed or inappropriate
quality control data are provided in Figures 1—4. In Figure 1, the
lab submitted a quadratic fit calibration curve when a lInear curve
and linear regression was required. The chromatograms in Figures 2,
3 and 4 all illustrate poor resolution, peaks off—scale and elevated
baselines. Yet these chromatagrams were submitted as laboratory
standards.
Although most of the laboratories in the certification program are
in compliance in adhering to regulations and methodologies, results
of on—site audits and evaluation of PE samples indicates that some
are not or have periodic problems. The reasons why all laboratories
aren’t In full compliance vary and obviously may ultimately depend
on the individual laboratory, equipment, lab practices, type of
analysis, method, analyte and analyst. However, based on our
7—45
-------
findings, some consistent general categorization can be made. Many
of the labs in noncompliance simply fall to adequately follow
prescribed methodologies, either ignoring or short cutting various
analytical requirements, steps and quality control procedures. This
has been observed in issues concerning use of appropriate equipment,
materials, standards and establishing documented routine lab quality
control procedures. Problems in quality control procedures
generally Involve failure to establish standards, calibration
checks, cross checks and control charts as required per method.
Those labs issuing questionable data or failing PE samples have
generally suffered from errors In dilution and data transcription as
well as contaminated standards, failure to establish an adequate MDL
and quantifying in an inappropriate part of the calibration range.
Administering the New Jersey State Laboratory Certification Program
has demonstrated that three components are necessary to monitor
laboratory performance. These are the use of PE samples, audits and
review of data. On—site audits in which actual data packages
submitted to the State are fully validated are extremely effective.
The NJDEP is currently revising Its Laboratory Certification
Regulations (N.J.A.C. 7:18) which will include new areas for
certification including analyses for R RA programs. These new
regulations will be based on clear definition of methods and quality
control procedures required.
Please note that the interpretations and opinions expressed in this
paper are those of the authors and should not be construed as
official policy of the New Jersey Department of Environmental
Protection.
REF .ENCES
State of New Jersey, “Regulations Governing Laboratory Certification
and Standards of Performance,” N.J.A.C. et seq., NJDEP, Office
of Quality Assurance.
Hirst, R.R., R. L. Fischer, K. Stauber and D.M. Stalnken, 1986.
RCRA Laboratory Certification, Proc. 2nd Annual U.S. EPA
Symposium on Solid Waste Testing and Quality Assurance, July
15—18, 1986, Wash., D.C.
7—46
-------
L1Lt¼ LJ? r .i’ CC.LD U( POR Ai S
1 LYS1S..DAT
CQRRELc TiQ 1 COEFFICIENT r 2= Q. O2 —
p .(WRE T IC..EI.1 .
. I 4 ACTU L DATA
o 7
51070 i5 1 7 - 1
CONCEHTRAIIOH IH UG/L
£0
Figure 1. Example of Incorrect Linear Calibration
C
C
Figure 2. Examples of Poorly Resolved Mirex Standard Chromatograms
7—47
-------
U
Figure 3. Examples of Poorly Resolved Standard Chromatograms for VOA’s
a
f’ 1•
Figure 4. Example of Poorly Resolved PCB ( 242) Standard
7—48
‘0
‘a
I ,
‘0
“a
-J
n
‘I
I
I
•1
— ‘4
W I
*
1
-3
a r ‘4
a
S.
‘4 -. *
• S
a
* a
a
p
S
-------
LANDFILL CONSTRUCTION——QUALITY ASSURANCE BEYOND TESTING
B.L. Woodward, Geomembrane Specialist, CH2M Hill, 777 108th Avenue,
P.O. Box 91500, Bellevue, Washington 98009-2050
ABSTRACT
Prerequisites for quality assurance during construction of any land-
fill are selection of an appropriate site and provision of a high—
quality design. This discussion of quality assurance assumes that
these prerequisites are met, and is limited to the construction phase
of a landfill project. The most quantifiable portion of a construc-
tion quality assurance program is the physical testing of construc-
tion materials and completed work, in the laboratory and onsite.
This paper addresses the other aspects of quality assurance that are
less quantifiable, but are nonetheless essential to the construction
of a functional and safe landfill. Geomembrane construction is
stressed because it presents the most difficult quality assurance
challenges.
INTRODUCTION
The reason that geomembrane construction has been receiving so much
attention, is that, unlike most other types of heavy construction,
the smallest flaw or oversight can result in failure (leakage) of the
installation {1J. This makes quality assurance of paramount impor-
tance to successful construction. In addition, geomembranes are fre-
quently at greater risk of damage during installation than during
their service life, due to stresses caused by equipment handling dur-
ing layout, backfilling of trenches, movement of equipment over the
surface, temperature changes, and other temporary stresses. Vigi-
lance during construction is essential to assure the completion of
every detail of work, to prevent practices that may cause damage to
the geornembrane, and to expedite the repair of any damage that oc-
curs. To this end, EPA requires a construction quality assurance
(CQA) plan for projects under its authority {2]. A CQA plan should
be prepared and followed for the construction of every landfill re-
gardless of whether it is required by permitting agencies.
Testing of geomembrane material and seams, by both destructive and
nondestructive methods, must form a significant part of every CQA
plan, but testing alone cannot provide quality assurance. While the
strength of seams can be tested effectively in the laboratory on sam-
ples cut out of installed geomembranes [ 3], these tests can only be
representative. Each time a sample is removed for testing the lining
must be repaired. Commonly used nondestructive test methods, such as
the vacuum box, test the continuity rather than the strength of seams
and cannot be used in all locations [ 4].
7—49
-------
Geomembrane testing onsite is typically performed as part of the con-
tractor’s quality control program. While the contractor bears
responsibility for constructing the landfill as designed, quality
assurance personnel should be under the direction of the engineer,
representing the owner of a project. This is because quality as-
surance is inseparable from other responsibilities of the engineer
such as preparation of specifications, understanding of design re-
quirements as they relate to site conditions, and taking action to
remedy observed noncompliance by the contractor. Quality assurance
personnel should observe tests performed by the contractor and per-
form any additional tests required.
In addition to testing, quality assurance requires good project
specifications, full—time observation of construction, and harmonious
communication between the contractor and the engineer. Proper speci-
fications provide the basis for avoiding construction problems and
for correcting them when they occur by making clear to the contractor
what is expected. Full—time observation is essential for quality
assurance of work that cannot be tested or will be covered, and to
provide correlation for representative results of destructive seam
tests. Observation must go far beyond occasional checks on completed
work to include close observation of each step of the work as it is
done. Related to these tasks is the development of a good relation-
ship between the resident engineer (the observer) and the contractor.
This relationship should be pursued and, though not measurable, may
at times provide more quality assurance than do other factors. These
three aspects of construction quality assurance are discussed below.
SPECIFICATIONS
The primary purpose of specifications is to list material require-
ments and method of payment. However, many other conditions of the
work should also be set forth in the specifications. Current land-
fill designs typically include at least one geomembrane lining in
addition to a clay or clay substitute secondary lining, and the geo—
membrane material most frequently selected is high—density poly-
ethylene (HDPE) because of its chemical and biological resistance to
many types of leachate [ 41. Some of the following suggestions for
conditions to be included in the specifications reflect these common
design features, while others are general requirements useful for
many work elements, such as earthwork and leachate collection.
Geomembrane Specifications
In addition to listing detailed requirements for the geomembrane
material, the specifications should clearly state other basic re-
quirements. The intent of the work——to provide a watertight
facility——should be stated along with a definition of the word
watertight as, for example, “no leakage of solid or liquid waste
through the material.” The geomembrane installer should be the
material manufacturer, or a company licensed by the manufacturer, in
order to secure the most experienced and knowledgeable crew of in-
stallers with direct access to the manufacturer’s testing facilities
7—50
-------
and technical expertise. If the installer and manufacturer are not
the same entity, a manufacturer’s representative should be onsite to
provide technical supervision and assistance at all times during
installation of the geomembrane and overlying materials. If a leak-
age problem develops after construction, two—party responsibility
could easily lead to significant delays in repair work while the two
parties debate the source of the problem.
The specifications should also cover as many incidental requirements
as can be anticipated, such as protection of the geomembrane during
storage, from sunlight, soil, and other substances. Even if only a
short construction season is intended, conditions may change follow-
ing contract award, and material may be exposed to the elements
longer than expected. Even material with ultraviolet resistance
should be protected from long—term exposure. Allowable methods of
temporarily anchoring the geomembrane during layout before panels are
seamed, should also be specified to protect the material. If this is
not included the contractor may choose, for example, to use burlap
bags of gravel to hold the sheets in place; if one of these bags
breaks, one unrecovered piece of gravel could later cause a puncture
in the geomembrane when an overburden is applied. Old tires used as
anchors may have protrusions that could cause sciatches or punctures.
Tightly woven bags filled with sand are acceptable anchors.
The specification for seaming the geomembrane is critical to integ-
rity of the completed lining. Field seams on geomembranes vary for
different types of material; the most common seams are adhesive and
thermal [ 5]. HDPE is seamed by thermal methods, with heat and pres-
sure applied to the overlap (hot wedge method) or by the addition of
extruded HDPE along the overlap in addition to heat and pressure.
The latter type of seaming is called extrusion fusion welding and is
the recommended method because (a) the weld is visible to observers
(unlike the hot wedge seam), (b) the added resin increases the
strength of the seam, and (c) vacuum—box testing can be done in
planar areas without cutting through overlapped surfaces. Specifica-
tions for adhesive seams, required for PVC material, should specify
the appropriate amount of adhesive to add because too much can be
damaging to the material.
Seam specifications should also list appropriate overlap widths, and
should require that the welding equipment for HDPE seams be capable
of continuously monitoring and controlling temperature in the fusion
zone.
For locations where pipes within the lined area penetrate through the
perimeter membrane, special requirements should be specified to pre-
vent leakage. Such pipe penetrations typically occur where leachate
collection pipes exit the landfill. Specifications should require
that pipe boots be fabricated with the same quality seams as else-
where and dimensions should be shown on a drawing. Figure 1 is an
example of a pipe penetration through a single geomembrane and
soil—bentonite lining composite, which met the requirements for con-
ditions at one site. Other sites and project requirements would
7—51
-------
entail different design considerations. To obtain a tight seal where
the geomembrane is connected to the concrete transition block beneath
the pipe boot, specifications should indicate that the surface of the
concrete and both surfaces of the geomembrane be cleaned just prior
to the work and that neoprene adhesive be applied to both the con-
crete and the underside of the membrane. The correct amount of
tightening of the nuts on the anchor bolts should also be specified.
Nuts should be tightened sufficiently to deform the neoprene pad
uniformly along its length but not to compress the neoprene more than
to 90 percent of its original thickness. (See Figure 1.)
Poor weather conditions during seaming can greatly impair the in-
tegrity of seams. Precipitation can create voids in both adhesive
and thermal seams. Even small drops of water prevent adhesion or
lower the equipment temperature. No seaming should be performed dur-
ing adverse weather conditions. Adverse weather includes any amount
of precipitation, low temperatures (varies for different materials),
and high winds.
While quality assurance should be the engineer’s responsibility (on
behalf of the owner), quality control (QC) should be the contractor’s
responsibility. The contractor should control the quality of work
through whatever means are necessary, including inspection and test-
ing, and regardless of the amount of additional testing and observa-
tion performed by the engineer’s quality assurance personnel. The
specifications should state this clearly to eliminate any assumption
on the part of the contractor that a portion of work is of acceptable
quality simply because it has not been questioned by the engineer.
Defects found in completed work indicate a deficiency in the con-
tractor’s QC program.
The contractor’s quality control of the geomembrane installation can
be presumed to include a minimum amount of testing, such as non-
destructive tests of all seams (required by EPA) and strength tests
of sample seams made at the start of each shift [ 2]. The minimum
testing required to be done by the contractor should be specified,
along with the requirement to notifying the engineer prior to per-
forming the tests so that quality assurance personnel can observe
them.
Additional tests to be performed by quality assurance personnel
should also be described in the specifications. The contractor must
know the type and frequency of required testing programs and of tests
by others in order to price the work.
Specifications should state that all defects found during inspection
and testing of the seams by either the contractor or the quality
assurance personnel should be marked Immediately for repair. Patches
required for these repairs, and for repairs to the lining where
samples have been removed, should be large enough to provide the
minimum overlap and should have rounded corners.
7—52
-------
General Requirements (Conunon to Many Work Elements )
The specifications should list experience requirements for work re-
quiring special skills, such as geomembrane installation. Experience
should be required of both the installing company and the personnel
assigned to the given project. This will avoid the problem of a
project becoming the experimental arena for a contractor or subcon-
tractor new to the business. A lack of knowledge about the proper-
ties of geomembranes and the equipment and seaming methods required
for installation would make leak—proof installation virtually
impossible.
A schedule of work should be required to provide evidence that the
work has been planned and is proceeding as planned and to enable the
engineer to plan the quality assurance work. Different phases of
construction will affect the number and the experience of quality
assurance personnel required. A construction schedule will aid in
providing sufficient staff and equipment for this work. The con-
tractor should be required to update the schedule daily, if
necessary.
The contractor should also be required to submit shop drawings for
layout of the work not detailed in the plans. This forces the con-
tractor to plan the work and may allow the engineer to detect poten-
tial problems in the execution of the work. For example, geomembrane
seams perpendicular to the direction of slope should be avoided.
Unacceptable drawings should be corrected and resubmitted.
Provide specifications for each material required even when only a
small quantity may be needed, such as neoprene adhesive, mastic tape,
and stainless steel bands. If product brands or requirements are not
known, specify “as recommended by geomembrane manufacturer.”
A dimension should be listed for the measurable tolerance for every
pertinent item of construction, such as for thickness of material,
width of overlap, performance test results, and elevations. For ex-
ample, a clay lining 0.5 meters thick may have an allowable deviation
in thickness of 2 centimeters. This provision may prevent disagree-
ments during construction and will define the work for quality as-
surance personnel and the contractor.
When materials arrive on the site, each separate item should be
labeled showing brand, dimensions, date of manufacture, and placement
(if appropriate).
Problems that may come up should be anticipated and the specifica-
tions should provide an avenue for resolution, if possible. The
words, “or as approved,” should be inserted where appropriate to give
the contractor the opportunity to apply better ideas.
The method of work should be the contractor’s option, except when it
affects the quality of work, as described above. The contractor
should be able to plan and execute the work with as much freedom as
7—53
-------
allowable in order to do the work at the least cost and so that the
contractor, rather than the owner, will bear responsibility for
unsuccessful methods of work. For these reasons, the specifications
should generally avoid listing methods.
CONSTRUCTION OBSERVATION
During landfill construction, particularly when a geomembrane is in-
stalled, construction quality cannot be assured without observation
of the work. This is because much of the work is covered, and test
methods for geomembranes are insufficiently developed to enable
quality assurance through testing alone. The staff assigned to the
task of observation should be knowledgeable about the reasons for
design and about the products used; should be sufficient in number to
observe work at different locations as it occurs [ 4]; and should be
knowledgeable about the problems created by the use of marginal ma-
terials and inappropriate construction methods. This section
describes many examples of construction tasks that must be observed
during landfill construction to provide quality assurance.
The material beneath the geomembrane, usually clay or a low—
permeability substitute, should be checked for full depth and for
continuous protection of the surface from dessication and precipita-
tion [ 6]. The surface should be smooth and free of protrusions of
gravel. The backfill around concrete pipe—penetration blocks should
blend with the surface of the concrete to provide a compacted, uni-
form surface. Transitions for changes in slope must be smooth and
rounded.
Installation of the geomembrane seams and pipe penetrations must be
carefully observed. Grinding of seam edges in an HDPE lining and
cleaning of seam surfaces in any geomembrane must be done just prior
to seaming. The alignment of the weld area on an extrusion—welded
seam should not vary more than approximately 0.5 centimeters in
either direction. Any variation in excess of this amount, in the
shininess of the surface, or in the apparent thickness of the weld
area should be given particular attention during nondestructive
testing and perhaps tested destructively. Quality assurance person-
nel must be familiar with the appearance of seams that have been
tested with positive results in order to more easily detect even the
most minor deviations that may indicate a potential problem. If
copper wires are inserted in seams for later spark tests, the place-
ment of the wires must be observed.
Poorly constructed pipe penetrations are one of the common causes of
leaks in geomembranes. The surface of the concrete supporting block
must be smooth and clean in order to protect the lining and to pro-
vide a good seal, and corners should be chamfered. As shown in
Figure 1, the neoprene adhesive should be placed on both the concrete
surface and the geomembrane surface continuously around the perimeter
of area to be connected, and the butyl mastic tape should also be
placed on both adjacent, clean surfaces. The nuts on the adhesive
anchors should be carefully tightened as described above in the
7—54
-------
section SPECIFICATIONS. The pipe boot should fit snugly over the
pipe and surrounding area without folds or gaps, and all seams should
be made against a firm subgrade.
Thermal expansion and contraction of HDPE is a common installation
problem, even in moderate climates, producing folds of expanded mate-
rial during warm hours and bridging over bends and trenches in the
subgrade during cool hours [ 6]. This can create stress in the seams,
can pull overlapped edges apart, and can make backfilling difficult
because folds and bridging cannot be allowed beneath the overburden.
Because of this problem placement of both the lining material and the
sand cover (or trench backfill) should be done when the lining is at
neither temperature extreme. If the lining is placed during cool
weather numerous wrinkles are likely to appear during the warmer
times of the day. To remove the folds during placement of the over-
lying sand layer, some folds will probably have to be cut, over-
lapped, and seamed. Similarly, bridging due to contraction may
require repair by cutting the lining, adding material, and seaming
each edge.
Granular drain material typically placed over the geomembrane, some-
times with a geonet between, must be at least 1 meter deep in areas
where equipment will travel in order to protect the geomembrane.
Prior to placement of the sand layer, all stray gravel and other
debris must be removed from the lining. Damage may occur due to
stresses caused by the equipment traveling on steep access roads, and
sharp turns and stops of equipment [ 6].
The backfill in the geomembrane anchor trench must be carefully com-
pacted. Trench backfill more permeable than the surrounding material
may allow liquid to seep beneath the liner [ 7].
Leachate collection pipes must be cleared of all obstructions prior
to final placement.
Edges of nonwoven geotextiles placed over the granular drain material
should be overlapped sufficiently to prevent the occurrence of any
uncovered area through which the sand might be contaminated with
overlying fine material. The contractor should not be allowed to
heat tack the geotextile because heating can burn holes in it.
Quality assurance personnel must be alert to practices that may cause
damage to the geomembrane, such as careless operation of equipment
and placing sharp tools on the lining surface. All gouges,
scratches, and punctures must be marked immediately and repaired in a
timely manner.
CONTRACTOR RELAT IONS
Because of the exacting nature of many portions of landfill con—
struction work and to the high degree of quality assurance required,
the working relationship between the engineer and the contractor is a
factor that can greatly affect the performance and quality of the
7—55
-------
work. A good relationship between these parties can be deliberately
developed on any project, beginning with specifications that clearly
state the requirements of the work.
At the prebid meeting for a project, the engineer should make sure
bidders understand the purpose of the job, the potential problem
areas, and the importance of close coordination between various
operations, such as between geomembrane installation and earthwork.
The general contractor should be aware before bidding of the impact
that geomembrane construction requirements may have on other work.
An example of this would be placement of drain materials over the
geomembrane during weather conditions most favorable to the geomeni—
brane to avoid expansion and shrinkage problems.
At the preconstruction meeting these issues should be addressed in
greater detail, including the issue of coordination between the
quality assurance personnel and the contractor. The contractor
should be told whom to contact when the resident engineer is not
available. The resident engineer should have a detailed understand-
ing of the project and be capable of making decisions in a timely
manner. In turn, the contractor must also designate primary and
backup personnel to provide communication at any time. The contrac-
tor must clearly understand that he is expected to provide current
schedules (updated daily).
After construction begins, the following practices may alleviate
potential problems:
1. The engineer/quality assurance personnel should ask questions of
the contractor during observation to ascertain the Intended
methods and sequence of work. Information gained might enable
necessary changes in procedure to be made earlier, thus avoiding
delays that may otherwise be unavoidable. Contractors and engi-
neers are both more amendable to revisions when they can be made
before the work in question is performed.
2. Mutual respect between the engineer and the contractor Is essen-
tial to maintaining good communication. An attitude of superi-
ority on the part of the engineer may cause a reluctance by the
contractor to performing any Item of work not expressly defined
In the plans. The engineer should listen to suggestions made by
the contractor and be open to implementing them if, upon anal-
ysis, they are more efficient or effective than the specified
item.
3. The engineer should explain the logic behind the design and
specifications. Problems of nonconformance during construction
are frequently the result of contractors not understanding the
reasons for certain requirements they may perceive to be im-
practical. For example, a contractor may tack a geotextIle
overlap using heat if it has not been explained why this is not
allowed.
7—56
-------
4. Use appropriate language (i.e., diplomacy) in handling perceived
problems. A polite question or reminder by the engineer may
motivate remedial action by the contractor and avoid a contest
of wills. Written directives are necessary and are always
preferable to verbal threats and accusations when nonconformance
becomes a problem. An example of diplomatic communication would
be, “I noticed the liner is not resting on the subgrade in some
places around the pipe boot. When were you going to work on
that?” Serious contractual problems may result from careless
verbal directives.
CONCLUSION
Testing programs of construction materials and workmanship are
fundamental to quality assurance but they must be combined with good
specifications, observation of the work, and good contractor rela-
tions for the achievement of the highest quality landfill construc-
tion. Specifications should be thorough but also must be clear and
provide avenues for unexpected changes. Construction must be closely
observed to avoid the small errors that may later cause serious prob-
lems. Attention to the development of good communication between
engineer and contractor can prevent construction problems and will
help to assure the superior workmanship essential to landfill
construction.
REFERENCES
[ 1] McCready, A. A. “Preventing Geomembrane Failures,” Geosyn—
thetics 1987 Conference Proceedings . New Orleans. pp. 385 to
391. 1987.
[ 2] EPA, Minimum Technology Guidance on Double Liner Systms for
Landfills and Surface Impoundments——Design, Construction, and
Operation (Draft). 1985.
[ 3] Haxo, H. E., Jr. and M. J. Wailer. “Laboratory Testing of
Geosynthetics and Plastic Pipe for Double Liner Systems,”
Geosynthetics 1987 Conference Proceedings . New Orleans.
pp. 577 to 594. 1987.
[ 4] Fluet, J. E., Jr. “Geosynthetic Lining Systems and Quality
Assurance——State of Practice and State of the Art,” Geosyn—
thetics 1987 Conference Proceedings . New Orleans. pp. 530 to
541. 1987.
[ 5] Koerner, R. M. Designing with Geosynthetics , Prentice—Hall.
Englewood Cliffs, New Jersey. pp. 294 to 296. 1986.
[ 6] Buranek, D. and J. Pacey. “Geomembrane—Soil Composite Lining
Systems Design, Construction Problems, and Solutions,” Geo—
synthetics 1987 Conference Proceedings . New Orleans. pp. 375
to 384. 1987.
7—57
-------
[ 7] Pfalser, I. L. “High Density Polyethylene Liners: A User’s
Experience,” Geosynthetics 1987 Conference Proceedings . New
Orleans. pp. 554 to 564. 1987.
7—58
-------
1 LAYE1 5
G107EX7/LE (7yP
-GF INO OFF SECTION OF WLT
P O7P UOIt’K ASOVE 557 NUT
3/8w NUT
557 WA5HE
‘/8w 8U7YL frIAS7IC 7APE
- /8’ 557 ADHESIVE
ANCHOR OL7S ( ,‘rI’cc
TO //4” 2 8A77EN
HA/Al HOPE LINES
• FIELD FA ICA7ED
• - - &‘aT SLEEVE
-. N —. / SEE DETAIL ____
• N. — --. / —
N 7 —I/4” . 2’/Z” CLOSED CELL NEOP ’ ENE
N V 5POt’ E SC 42- M, o -i
• co.v7i”uOus eC7,4NGLE
—c öENEATH BATTEN oARS
ADHESIVE
Co Ic E TE: A64/N3T
C0M ACTW QR• ULQI5TU 5EQ
MAT IAL-
f ECr/OAJ
MIL
HOPE E’O 5T S
557 A ,j’JQ
I/4”x 2” CLOSED
CELL NEOPPENE
SPONGE-
E,ACI(P7LL A8OVE CONCRETE
AS SHOWN IN TYPICAL
TRENCH SECTIONS 511 24
HOPE PIPE SOOT____
SEE DETAIL ( 7 \
- -
NEOP ENE CAL/L/
-------
DEVELOPMENT OF A SPECIAL ANALYTICAL SERVICES (SAS) SOP FOR
LABORATORY PERFORMANCE ON VOLATILE METHOD AND TRIP (FIELD) BLANKS
ASSOCIATED WITH POTABLE AND LOW LEVEL MONITORING WELL SAMPLES
F. Genicola, Y. Lee, J. Rose, New Jersey Department of
Environmental Protection, Divisions of Hazardous Site
Mitigation/Environmental Quality, Trenton, New Jersey
ABSTRACT
The objective of this presentation is to suggest that volatile
organic analysis of potable and low level monitoring well water can
be improved by a Special Analytical Services (SAS) defined
algorithm. The SAS includes an initial demonstration and
continuing achievement in method blank and trip blank controls for
inethylene chloride (this study) and any other desired target
volatile analytes(s). The SAS has been shown to lessen false
positives for reported data than that being routinely reported by
Regular Analytical Services (RAS). The SAS was used for four
sampling episodes at three Superfund Sites and was found to give MB
and TB control wherein data could be reviewed by a modified Quality
Assurance, Data Validation— Standard Operating Procedures withir
the SAS. For Superfund Sites I—CFS and Il—FL, the SAS was able to
produce data, via modified EPA Method 601, which could be validated
by means of successful control of Method Blanks and Trip Blanks to
less than, or equal to, 1.0 ug/L. The sample data for Sites I and
II suggests that previously run Regular (Routine) Analytical
Services analyses were suspect (yielding false positives).
The SAS evolved during developmental research for HSL analytes by
Cryofocus Capillary GC/MS. The research was performed on a low
level monitoring well, A—i, at Superfund Site Ill—PR. The data
suggested that previous and the currently RAS split samples
generally were with flawed in the methylene chloride analytical
results for this low level monitoring well.
The SAS was shown to be instrumentation independent and successful
for the analysis of both potable and low level monitoring well
samples. In New Jersey, the Department of Environmental Protection
has defined Level II Action for 4.8ug/Lto 48ug/L methylene
chloride. Decisions involving alternate water supplies to
residences is possible for those potables having this level of
contamination. Current practice of RAS for MB and TB control is
inadequate.
FIRST USE OF SAS FOR CONFIRMATION OR NEGATION OF METHYLENE CHLORIDE
PRESENCE IN RESIDENTIAL WELLS ADJACENT TO SUPERFUND SITE I .
[ 5,6,7,8,91
A request to begin an initial laboratory demonstration, sampling
and monitoring analysis for Nethylene Chloride for low level
7—61
-------
presence equal to or greater than 5 /L in potable samples from
residence wells adjacent to the Superfund Site I—CFS landfill. [ 6]
BACKGROUND INFORMATION
Within written Personal Correspondence: Genicola to Morris, dated
October 21, 1986 [ 81, a procedure for the determination of
methylene chloride as a single monitoring analyte needing
confirmation was developed. The basis for the testing is that data
generated heretofore at this site could not be confirmed nor
negation of methylene chloride presence in drinking water samples.
The SAS, Special Analytical Services, procedure has evolved
emphasizing control of VOA glassware, special control of reagent
water used for method blanks, regular trip blanks and special trip
blanks, TBO and TBO’ [ fully under laboratory control and the basis
point for regular TB evaluation when validated]. Included in the
laboratory demonstration with is deliverable records of analytical
quality of reagent water before filling the TB for the field and
the extra trip blank basis points TBO and TBO’. The TBO and TBO’
are stop/go decision laboratory quality control previous to sample
analyses (Figure 1 and Logic Diagrams I and II).
The Department’s Bureau of Environmental Laboratories was selected
because an earlier study [ 13] had shown that Method Blanks used in
Cryofocus Capillary CC/MS of HSLs had Methylene chloride control at
0.2 to 0.3 ug/L on a regular basis. Additionally, the BEL had the
desire to successfully attempt and complete this study.
AUTHORIZATION TO PROCEED WITH SAS TASKS
The dates for the Initial Demonstration were December 8, 1986 in
which a 90 vial lot of 40m1 VOA vials was made available by BEL, an
initial demonstration of reagent water In Rm 109 [ Room contains GC
and CC/MS instrumentation and charcoal Filter for Preparation of
Reagent Water for Method Blanks 1. The reagent water shall have 1.0
ug/L level of Methylene chloride background or less, three Trip
Blanks including TB1, TBO(l) and TBO’(l) shall be filled and
analyzed as described In the above cited correspondence [ 8] except
that TBO shall be analyzed on December 12, 1986. All TBs are to
show 1.0 or less ug/L Methylene chloride contamination in order to
proceed with samples from Superfund Site I—CFS drinking water
weils [ 81.
Samples to be run shall include thirty—three (33) potable wells
which are to be sampled in duplicate and Iced , no synthetic
refrigerant, and stored In special small approximate 2.5 gal volume
potable shuttles (must only be used for potable water samples).
Analytical considerations shall include that the monitoring shall
be performed under EPA Method 601 but only for one analyte Methylene
7—62
-------
and Samples
Samples(4), MB
1 .0 ug/L.
FIGURE 1:
SAS
Blank
Control
SAS Blank Controls for
Methylene Chloride
MB Control
Before Fill
Established <1.0 ug/L
of TB(O), TB(0),TB(2)
Required Continuing Lob Achievements
TB(0) < 1 .0 ug/L , +4 hrs of Fill
TB(O) <= 1 .0 ug/L , Pre—Sarnples
Successful Blanks
Sequence: MB, TB(1),
MBs must be <
7—63
-------
Logic Diagram I : Initial Laboratory Demonstration of Special
Analytical Services for the Analysis of Methylene
Chloride. The Method Blank (MB ), Trip B1a Basis Points
(TBO and TBO’) and Trip Blank (TB) Controls. TBs are
F illy under the Laboratory Quality Control for the
Initial Demonstration.
SAS ln tIal D.mon.trotlon
Blank Control for M.thyl.ne Charlde
7—64
-------
Logic Diagram II : Special Analytical Services for the
Analysis of Methylene Chloride with Method Blank (MB),
Trip Blank Basis Points (TBO and TBO’) and Trip Blank
(TB). TBO and TBO’ are fully under the Laboratory
Control arid as such are Stop/Go Decisions after filling
with Reagent Water (MB analyzed). TBO is analyzed after
four (4) hrs. of filling. TBO’ is analyzed Just previous
to beginning sample analysis.
ethylene
SAS Analyst
With Blank Controls
7—65
-------
Chloride. Only one surrogate shall be employed and recovery versus
laboratory generated limits of acceptance must be met or repeated
successfully at laboratory expense. All data relating to
standards, method blanks and trip/field blanks must be presented in
acceptable deliverables format. Data reporting shall be presented
in IFB/CLP deliverable form with CLP data qualifiers by the
laboratory.
If possible sampling should be undertaken on Monday December 15,
1986 and Thursday December 18, 1986. Sample bottles for December
15, 1986 must be picked up on Friday December 12, 1986 and TB2,
TBO(2) and TBO’(2) filled with preanalyzed (less than 1.0 ug/L
methylene and chloride) reagent water . . . Sample bottles for the
December 18, 1986 sampling must be picked up at BEL on Wednesday
December 17, 1986 and TB3, TBO(3) and TBO’(3) filled with
preanalyzed reagent water (less than 1.0 ugiL Methylene Chloride).
TBO (2 or 3) respectively must show a less than, equal to 1.0 ug/L
methylene chloride contamination level as a stop/go situation.
Thus analysis must be completed before sampling is to begin (within
4 hours of filling the TBO (2 or 3). If unsatisfactory, the
laboratory must contact the SMO through whom the sampling team
shall be told not to proceed with continuing taking samples.
Sample analysis shall be rated against the individual shift method
blank (less than 1.0 ug/L) and the TBO’ (2 or 3) being under
control at 1.0 or less ug/L Methylene Chloride before proceeding as
a stop/go decision point. If unsatisfactory, the laboratory must
contact the SMO before proceeding with the analysis.
The laboratory shall run an appropriate number of method blanks to
ensure that during sample analysis, the methylene chloride
contamination is kept under 1.0 ug/L. Suggested frequency for this
study is every 5 samples; these blanks are billable. Standard
checks at 2.0 ug/mi shall be after every five samples and method
blank check. The laboratory shall reanalyze at its own expense
those samples affected by loss of control for method blanks or
standard checks.
PROTOCOL CONTROL OF METhOD BLANKS — METHYLENE CHLORIDE [ 1,2,3,4]
EPA Methods [ 1,2,31 and OERR CLP—IFB Contracts have a MDL defined
limit for the contamination allowable in a method blank with no
reporting suggestions for levels below the MDL for the former, and
5 X the DL(cRQL) limit (Figure 0—a) for contamination but with
defined reporting requirements below the CR.DL, the “J” qualifier to
any found analyte “above zero”.
EPA guidance for validation in the CLP is that a “common laboratory
contaminant” can not be confirmed for a sample unless the sample
has 10 x that found In any associated blank. [ 4]
7—66
-------
The EPA Methods have no reporting requirements below the MDL and
therefore the EPA CLP guidance document is ineffective for low
level samples and method blanks.
SAS CONTROL OF METHOD BLANKS [ 7,8]
The NJDEP SAS requires laboratory reporting for methylene chloride
to a CRDL(CRQL) of 1.0 U with the “J” qualifier for value one order
of magnitude lower, 0.1 to 0.9. Additionally, any sample having an
associated found in the method blank would be reported by the
laboratory with a “B” qualifier. [ Multiple qualifiers to numerical
data are allowed].
Quality Assurance, Data Validation would “negate” as present any
methylene chloride in a sample with an associated method blank
wherein the sample was under 3 X MB. Negated values are raised to
the value in the sample and becomes “UB” qualified. Sample values
greater than, equal to 3 X MB are “JB”(qualitatively possible).
Sample values equal to, greater than 5 X MB are validated as Value
B (i.e. 15B).
Assuming Method Blank is contaminated to just under the RDL of
5ug/L for CLP, to the MDL of 0.5 ug/L and 2.8 ug/L for the EPA
601,624 and to the CRDL of 1.0 ug/L for the SAS, Figure 0—b gives
the data validation possible.
RESULTS OF THE SAS FOR SUPERFUND SITE I—CFS [ 9,10]
The SAS analysis of thirty—three (33) residential wells adjoining
the Superfund Site I—CFS were performed on sampling episodes of
December 15, 1986 and December 18, 1986. Each sampling episode
required its own set of MB, TBO, TBO’ and TB which were designated
as TBO(2), TBO’(2) and TB(2) for the former and TBO(3) and TB(3)
for the latter.
All MBs and TBs, including the associated TB basis points, were
kept under 1.0 ug/L Methylene Chloride. The resulting sample data
had thirty samples data validated as LOU or l.OUB. Two samples
were validated as “qualified” present at 2JB each. Only one
“suspect” residence had a methylene chloride result at 5.5B.
Solvents were evident in residence via Chain—of—Custody notations.
Alternately, RAS analysis was performed on 15 of residence wells on
a “split” sample basis. The results of the CLP laboratory yielded
unacceptable MB and TB values for validation purposes versus the
NJDEP Level II Action for Methylene Chloride (Figures la, lb and
Tables la, ib).
Note should be made that for residence PA at Superfund Sites I—CFS,
the value reported by RAS was 180 ug/L where both split and
7—67
-------
Figure 0-a, Comparison of Method Blank Controls and the
QA Data Validation for Methylene Chloride in the
EPA CL?.
Blank
300 T
25O
2OO
Controls and
QA Data Validation
For RAS(CLP)
LEGEND
MB 25 ug/L ___ DL (CRDL,CRQL.MDL)
MB Allowable
QA Qualified Presence
___ QA Real Presence
ASSUMPTION:
Method Requirements
of
Protocols
ank Contamination
CRDL or 5 X the CRDL.
15O
MB=5ug/L
50
0
is
I — C
-------
Figure 0-b. Comparison of Method Blank Controls and the
QA Data Validation for Methyiene Chloride in
Three EPA Protocols and the NJDEP SAS.
Blank Controls and QA Data Validation
For RAS(CLP) RAS(601), RAS(624) and
SAS
60 T LEGEND
DL (CRDL,CRQL.MDL)
50 MB Allowable
QA Qualified Presence
40 QA Real Presence
-D
- 30
ASSUMPTION:
C
V
20 1 Blank Contamination is
Just under the CRDL or
10 MDL.
0
Method Requirements
of Protocols
7—69
-------
Figure la :
—j
a,
L.
0
C-)
C
-c
a)
30
25
20
15
10
5
0
Methylene Chloride Results, Method Blank
Control and Trip Blank Results using RAS
versus Split Sample SAS for Superfund Site I-
CFS. Special Attention should be Noted on
the TB2 Results for RAS Versus SAS. The
Latter has little or no Control Versus
Validation Control for the Former.
Methylene Chloride
Method Blank RAS versus SAS
Superfund Site CFS 121586
LEGEND
Sample
Method Blank
QA Validation
SC SP MU AB SC .)V Ml TB2
Residential Well Results
7—70
-------
Figure ib : Second Sampling Episode Results for the RAS
Versus SAS Split Samples at Superfund Site I-
CFS. Special Attention should be Noted on
the PA Residential Well Results which is a
Level III Action Site, if Verifiable. Refer
to Figure i.c for Invalidation by SAS.
Methylene Chloride
Blank Control RAS vs SAS
Superfund Site CFS 121886
200 LEGEND
180 SAMPLE
160 METHOD BLANK
QA Validaton
140
I-
0
100
a.,
8O
- 60t
40
2: 1 t -4- - 1-- 4
JO MASCCOPATOMUTB3
Residential Well Results
7—71
-------
subsequent SAS reported 1.0 UB and l.OU, respectively. The PA
sampling site has never seen excessive (NJDEP Level Ill—more severe
than Level II) methylene chloride. The results of the RAS were
quality assured as being “probably flawed” but could not be negated
as present since the 180 ug/L was more than 10 x MB and TB values.
Only through the KB being in excess of the EPA Method 624, could
the QA review question the PA Residence results (Figure ic).
Additionally, previously obtained data for the potable samples: SC
and AB was compared to the December 1986 samplIng episodes. RAS
data is shown to either be issuing “false positive” or ineffective
data for quality Insurance for low level samples. (Figure 2 and
Table 2)
SAS USE IN POTABLE AND MONITORING WELlS AT SUPFRFUND SITE I l-FL
(11]
Results of the SAS for sample taken in April 1987 were l.OU or
l.OUB for all confirmation samples. Previous RAS results with
uncontrolled MB AND TBs were not expected to be contaminated
environmentally with Methylene Chloride. Validation on RAS data
was not possible for the NJDEP Action Level II.
SAS DEVELOPMENT FROM RESEARCH IN HSL VOLATILE ANALYSIS BY CRYOFOCUS
CAPILLARY CC/MS [ 13)
Monitoring Well A—i (Lower Cohansey Aquifer) RAS Results in 1984
for the Superfund Site Ill—PR yielded Methylene Chloride at 89 ug/L
and 1, 2—Dichioroetbane at 50 ug/L. The RI/FS excluded the
Methylene Chloride for all samples.
Monitoring Well A—i was analyzed In 1986 using split samples to one
CLP (New England), one nonCLP (Mid—Atlantic), and the NJDEP—BEL.
The former reported Methylene Chloride at 7B along with Acetone
6Th, Benzene 2JB. The nonCLP laboratory found “ND” for all HSL
analytes. The Cryofocus Capillary results found 6.2 ug/L trans—l, 2—
Dichioroethene (a known contaminant for the PR site); no other
compound above 1 ug/L was found. Methylene Chloride was present in
the MB a at 0.2 ug/L (Figures 3a, 3b and 3c).
Most HSL analytea had five level standardization calibration from
0.5 ugfL to 20 ug/L (Figure 3d).
CONCLUSIONS
The analysis of low level analytes including, and In particular
Methylene Chloride, can be controlled through SAS laboratory
practice on a confirmational basis. It is possible to integrate
the SAS within an RAS method or in tandem for QA purposes. The
results from three Superfund Sites for potable and low level
monitoring well samples can be satisfactorily analyzed and data
7—72
-------
Table 10 SPLIT SAMPLE ANALYSIS
Comporison of Methylene Chloride Results, ug/L , using Regular Analytical
Services and Special Analytical Services for CFS 121556:
Residence
Lab
Sample
Results
Method
Blank
QA
Decision
SC
2
RAS
SAS
17.0
0.13U3
12.2
0.48
Negate, 17UB
Negate, 1.OUB
SP
3
RAS
SAS
15.3
1.8
12.2
0.48
Negate. 15UB
1.8JB
MU
4
RAS
SAS
23.2
2. OJB
12.2
0.48
Negate, 23UB
2.OJB
AB
8
RAS
SAS
9.14
1.OUB
(0.05)
12.2
0.92
Negate. 9.1UB
1.OLJB
Trip Blank
TB (2)
RAS
SAS
7.14
0.2JB
12.2
0.48
Negate, 7.1UB
Negate, 1.OUB
7—73
-------
Figure ib : Expanded View with PA results removed.
Methylene Chloride
Blank Control
Superfund
RAS vs SAS
Site CFS 121886
JO MA Sc CO TO ML)
Residential Well Results
7—74
LEGEND
SAMPLE
METHOD BLANK
QA Validation
-J
03
-o
0
L)
0 )
C
a,
30 T
TB3
-------
Table lb SPUT SAMPLE ANALYSIS
Comparison of Methylene Chloride Results, ug/L using Regular Analytical
Services and Special Analytical Services for CFS 121886:
Residence
Lob
Sample
Results
Method
Blank
QA
Decision
JO
19
RAS
SAS
8.27
I .OUB
(0.09)
14.4
0.87
Negate, 8.3UB
1 .OUB
MA
20
RAS
SAS
8.90
0.1 OJB
12.2
0.87
Negate, 8.9U8
Negate, 1.OUB
PA
26
RAS
SAS
178
1.OU
12.2
1.OU
180B
1.OU
MU
32
RAS
SAS
23.6
1.OU
(0.08)
14.4
LOU
Negate, 23.6UB
1.OU
Trip Blank
TB (3)
RAS
SAS
9.72
0.45JB
12.2
0.87
Negate, 9.7UB
Negate, 1.OUB
7—75
-------
Figure ic : Superfund Site I—CFS. Results of
Split RAS/SAS Samples and Confirmational
Sample for the PA Residence. SAS Results
for 121886 and 040787 were 1.OUB and 1.OU,
Respectively. The 180 ug/L would have involved
A NJDEP Action Level III, if confirmed.
Decision was made to invalidate the RAS result.
LEGEND
Sample
Method Blank
QA VaHdotiori
Methylene Chloride
Method Blank RAS versus SAS
PA Residence at CFS 121886 , 040787
200
180
160
140
120
100
80
60
40
20
0
PA (RAS) I PA (SAS) II
PA (sAs) I
Residential Well Results
0
C )
C
4,
4 )
7—76
-------
Figure 2 : Comparison of RAS/SAS Split Results with
Previous RAS Data for Residences Sc and AB
At Superfund Site I-CFS.
Residential Well Analysis
Comparing Split RAS/SAS with
Previous RAS Results
20 LEGEND
18 RAS(1) Sample Results
16 Method Blank
QA Decision
14
12
10 Sc 121586
8 121586
a)
081986
6
AB 121586
4 RAS(1) 121586
2 082086
020386
0
RAS(1) RAS(3)
SC SC SC AB AB AB AB
Residential Wells
7—77
-------
TabM 2 RESIDENTIAL WELL ANALYSIS
Comparison of Methylene Chloride Results, ug/L using Regular Analytical
Services and Special Analytical Services for CFS 121586 with previous
site sampling episodes using RAS:
Residence
Lab
Sample
Results
Method
Blank
QA
Decision
sc
121586
081986
RAS
SAS
RAS
17.0
0.13UB
5.4
12.2
0.48
4.96
Negate, l7tJB
Negate, 1.OUB
Negate, 5.4UB
AB
121586
082086
020386
RAS ’
SAS
RAS
RAS
9.14
1 .OUB
(0.05)
117
<2.8
12.2
0.92
20
.
<2.8
Negate, 9.1UB
1 .OUB
1178
.
NA
Trip
Blank
(Associated)
121586
081986
082086
020386
RASW
SAS
RAS
RAS’
RAS ”
7.14
0.2J8
5.3
11
<2.8
12.2
0.48
4.96
1.9
<2.8
Negate, 7.1UB
Negate, 1.OUB
Negate, 5.3UB
118
NA
RAS ’ I d-AUontic CLP Laboratory, CC/MS (624) RAS N. and CLP Laboratory,
CC/MS (624)
RAS Southeastern CLP Laboratory, CC/MS (624) Nei Jersey State Laboratory,
Ha ll (Mo fi.d 601)
7—78
-------
FIGURE 3a:
Superfurid Site III (PR) Results by Cryofocus Capillary CC/MS.
N1 1C T I *l l $1415 6 13$ TO 14
I .’134 7 : N .X:l 1 1 12
I YO t 11 .*V F t & I*.*T1L
t S 155
IN 7 , G 1.14 H S. 4.1 J b a S. i.e ,i S U 21. 3
5
34’
U. I
p
115?
!5S2
15 53
1 153
131S
I-
.
2 1 1
, •
411
—I
2 :31
3 ,13
4 4
6:25
9 :52
SC 6 13STO14
152.
1 S 14
sc
Tili
I RS
SIC.
N1 1C TAs 31L1638 SI
11’t8’96 12:56,55 C .1: 35LM18 56 $2
SNrLE: YO C P1L Y F )& U tMT1LE$
.s t W II 9 155
c 1.14 I. a: N •. 4.5 a ô a e. l.a j
a
U 25.
3
SIC
250
1:36
450
3:13
‘a.
4:49
6:25
7—78
-------
FIGURE 3b :
Surrogate Recoveries for 5 ug/L Spiking Level
Using Cryofocus Capillary GCIMS.
PRICE’S LANDFILL, Mcnitcring
Surrogate Recoveries from 5 ug/L Spike Levels
and Analyzed by Cryofocus Capillary GC/MS for
CLP Target Volatile Analytes. the HSLs”.
10
e
6
4
2
0
‘ ‘ie11 IA
5.63
4.75
4.23
4.45
4.32
4.02
7—79
-------
FIGURE 3c: Monitoring Well A I of the Lower Cohansey
Aquifer at Superfund Site III (PR).
Price’s
7
6
5
4
3
2
1
0
Landfill, Monitoring Well A-i
May, 1986
6.8 6.2 6.8 4.83
Samples from MW of L. Cohansey Aquifer
Cryofocus Capillary, Mass Sp.clrom.try R.sults
LEGEND
-J
a,
2 .
( ‘U
_ a
a.
I- =
— —
00
V Cñ
C
0
-------
FIGURE 3d : Response curves showing five point low level
Standardization used by SAS, Cryofocus Capillary GC/MS.
% EF.O4 :I6. I; P 5 31)
(*Js 2.425)
‘ 1
I —,
/
/
4
3.—
I. —
I .—
I, *1 I $ 131)
(Nit 2.311)
3.—
7.—
5.—
4.—
3.—
1.
I .—
NW ,
/
‘I
a
/
1•
/
1n.
1.426
SC*TT u
1.421
2 ATtERs
17.635
S
I.—
0 11F1.
1.413
2 WAITEt’
11141
S
1.314
/
/
• 1.$
/
/
/
/
7—81
-------
validated if MB and TB controls and proofs are achieved by the
laboratory and documented.
REFERENCES
Fed. Reg., 49, No. 209, P. 29 (EPA Method 601, Purgeable
Halocarbons), October 26, 1984.
Fed. Reg., 49, No. 209, p. 141 (EPA Method 624, Purgeables),
October 26, 1984.
“Volatile Organic Compounds in Water by Purge and Trap Capillary
Column Gas chromatography/Mass Spectrometry” (EPA Method
524.2), August 1986.
“Laboratory Data Validation Functional Guidelines for Evaluating
Organics Analyses”, EPA—CLP Organics Caucus, Atlanta Georgia,
March 19—21, 1985.
[ Basis document for Review of Data by the DHSM, NJDEP through
incorporation with modifications at Attachment XVI for X—312
Contracts and SOP for Data Validation by DHSM, NJDEP].
“Guidelines for the Professional Quality Assurance Data Validation
of Analytical Sample Deliverables”, DHSM (OQA), NJDEP,
February, 1986.
“Drinking Water Guidance, Interim Action Levels and Recommendations
of Responses for Selected Organics in Drinking Water”, Division
of Water Resources, NJDEP, January 1986.
Level II . . . Confirmation of Level II concentrations wherein
Methylene Chloride is listed as = 4.8 ug/L and 48 ug/L) the
Department shall require “recommend alternative water sources
and/or appropriate treatment techniques.]
“Outline of Proposal for SAS for Methylene Chloride”, October 1986,
Personal Communication: F. Genicola to N. Morris.
“Final Report on Special Analytical Services, SAS, for Methylene
Chloride Presence in Residential Potable Water Samples at the
[ Superfund Site I—CFS]. SAS included Method 601, Modified for
Special Method Blank Controls and Expanded Trip Blank
Controls: TBO(2,3) and TBO’(2,3)”, January 1987, Personal
communication: F. Genicola to M. Morris.
“The Quality Assurance, Data Validation Results for the Special
Analytical Services Determination of Methylene Chloride: for
the PA Potable Water Taken April 7, 1987”, April 1987, Personal
communication: F. Genicola to R. Kaiserman.
7—82
-------
“Quality Assurance Data Validation for the Superfund Site 1 1—FL
Resampling of Nearby Potable Wells on 24 February 1987.
Analysis of Methylene Chloride by DHSM SAS, Special Artalytical
Services Methodology”, March 1987, Personal communication: F.
Genicola to A. DeCicco.
“Remedial Investigation and Feasibility Study for [ Superfund Site
111—PR Number 1]”, [ Engineering Firm X I, Inc., February 1985,
Contract Number S84094.
F. Genicola, “Volatile Organics Analysis of Hazardous Substance
List compounds in Monitoring Well A—i for the Lower Cohansey
Aquifer by Purge and Trap/Cryofocus Precapillary/Capillary
thromatography/Nass Spectrometry using a Modified Target
compound Analysis”, May 1987 [ Unpublished Study Submitted to
Rutgers University’s Cook College, Department of Environmental
Science].
7—83
-------
DEVELOPI 1ENT OF STANDARDS FOR EPA HAZARDOUS WASTE METHODOLOGIES
Neil H. MosesEnan, Jack K. Crissman, Supelco, Inc., Bellefonte,
Pennsylvania
ABSTRACT
Every laboratory involved in environmental testing must maintain
appropriate standards for the various methods generated by the US
Environmental Protection Agency (EPA). It is impractical for each
laboratory to develop and test all the necessary standards, or for the
EPA to provide all standards to every laboratory. Therefore,
conuuercial standards are a necessity for any environmental analysis
program. This presentation will cover the steps involved in
developing high quality, readily available standards for many EPA
analyses.
To ensure that standards are of the highest possible quality, the
quality assurance program must combine thorough testing with
verification against EPA standard materials. The ideal multi-step
quality assurance procedure includes testing all raw materials by
several independent methods, determining the appropriate solvent for
each component and its compatibility with other potential components,
and verifying each compound against internal reference standards and
(whenever possible) external reference materials.
The methods used to evaluate raw materials will be described,
with data for selected compounds used as examples. Data from an
independent laboratory study will show the traceability of commercial
standards to standards from the EPA repository. Development of new
standards for the growing list of EPA test methods also will be
discussed.
INTRODUCTION
Recently the EPA has developed several new methods that require
capillary gas chromatography as the analytical technique. The
Superfund Contract Lab Program (CLP) analytical protocol for
semi—volatile priority pollutants, which was the basis for EPA Method
8270, is an example of these methods. The CLP protocol calls for
combining the acid and base—neutral fractions, prior to injection, and
calibrating the capillary GC/MS system at a minimum of five different
concentrations over a wide calibration range. Two additional new EPA
methods, 502.2 and 524.2, describe capillary GC analysis of volatile
organic compounds in water. Methods 502.2 and 524.2 extend purge and
trap—based analysis to more than 50 compounds, covering a range of
volatility from dichlorodifluoromethane to trichlorobenzene.
QUALITY CONTROL
To fill a need these new methods have created, Supelco has
developed a series of analytical standards specifically designed for
capillary analysis. In developing these standards, efforts were made
7—85
-------
to ensure the highest quality by employing a multi-step quality
assurance program. Table I shows the steps involved in this program.
SEMI - VOLATILE POLL tYTANTS
The standards for the semi—volatile priority pollutants,
introduced in December 1984, differed significantly from previously
developed standards in their fonnulation and testing. Because many of
the semi-volatile priority pollutants are unstable or insoluble, when
combined in the same mixture, we grouped the compounds into eight
stable mixtures containing two to sixteen compatible components.
Table II shows the make—up of these mixtures. The new standards
contained higher concentrations of each component (2000 pg/ml each for
the pollutants, 4000 Pg/ml each for the internal standards). This
ensures that the concentration of each component will be sufficient
for calibrations, even if all eight standards are combined.
The final step in the quality assurance program was a comparison
of these standards, by an independent laboratory, to standards from
the EPA standards repository. As an added check, 51 isotopically
labeled compounds and 2,2’—difluorobiphenyl were used as internal
standards. The first comparison was done in July 1984. The results
indicated that for 12 of the 66 compounds evaluatea, the discrepancy
between the commercial standard and the EP A repository standard
exceeded 25%. Because this comparison was made against repository
standards that were several years old, a second comparison was made in
November 1986 against recently prepared repository standards. In
addition, the newer repository standards more closely matched the
commercial standards in composition and solvent matrix.
In the second comparison, only 2 compounds showed a difference
that exceeded 25% between the commercially available standards and the
EPA repository standards. Table III shows the results of the second
comparison using the isotope dilution technique. Table IV shows the
same comparison, but using the six internal standard method used by
the Contract Lab Program (CLP). The compounds are listed in
decreasing order from the highest to the lowest percent difference.
The two compounds with the highest percent difference (benzoic acid
and 4—nitroaniline) also have the lowest response factors and are the
most difficult of the compounds to monitor.
VOLATILE POLL1J2M TS
In the past few months we have completed work on a series of five
standard mixtures, each containing 8 to 16 volatile organic
compounds. Included are all 58 compounds currently required in EPA
methods 502.2 and 524.2. Table V shows the composition of the five
mixtures. These standards were developed through a quality control
program similar to that for the semi—volatile pollutants. The
volatile pollutant standards, however, have not been compared to EPA
standards, because not all components are yet available from the
repository.
7—86
-------
CONCLUSION
Commercial standards have been developed for several new EPA
methods for semi—volatile and volatile organic compounds. These
standards have undergone a multi—step quality assurance program,
including purity determinations for raw materials, lot to lot
comparisons and, when possible, comparison to EPA standard materials.
Comparison of the coninercial semi—volatile pollutant standards to
standards from the EPA standards repository show there is excellent
agreement between these standards. Work is continuing at Supelco to
develop new standards for additional EPA methods, including Appendix
IX compounds and nitrogen and phosphorus pesticides.
TABLE I
1) Purity Determination for Raw Materials
a) Melting point for all solids
b) Refractive index for all liquids
c) IR spectra for all compounds
d) Capillary GC/FID for all compounds
2) Gravimetric Determinations
a) Accuracy to + 0.5%
b) Recording balances used for all weighings
3) Evaluation of Standard Mixtures
a) Comparison of two independently produced lots
b) GC or HPLC techniques used for evaluations
c) Six months stability determined for mixtures
4) Capillary GC/MS Comparison to EPA Repository Standards
a) Analysis done by an EPA approved independent laboratory
b) All analyses done in triplicate
c) Isotope dilution techniques used for quantitation
TABLE II
BASE-NEUTRALS MIX 1 BASE-NEUTRALS i1IX 2
Bis ( 2—chloroethoxy )rnethane 2-Chloronaphthalerie
Bis (2-chloroethyl) ether 1, 2—Dichlorobenzene
Bis(2—chloroisopropyl)ether 1, 3-Dichlorobenzene
4—Bromophenyl phenyl ether l,4—Dichlorobenzene
4-Chlorophenyl phenyl ether Hexachlorobenzene
Bis (2—ethylhexyl )phthalate Hexachlorobutadie ie
7—87
-------
TABLE II (continued)
Butyl benzyl phthalate
Diethyl phthalate
Dimethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phtha late
N—Nitrosodimethylaxfline
N-Nit rosodi -n-propylainine
N-Nitrosodiphenyl amine
POLYNUCLEAR AROHATICS MIX
Acenaphthene
Acenaphthylene
Anthracene
Benzo ( a)anthracene
Benzo( a)pyrene
Benzo (b) fluoranthene
Benzo(ghi)perylefle
Benzo(k)fluoraflthefle
Chrysene
Dibenzo(a, h )anthracene
Fluoranthene
Fluorene
Indeno(l, 2, 3—cd)pyrene
Naphtha lene
Phenanthrene
Pyrene
PHENOLS MIX
Hexachlorocyciopentaci le;ke
Hexachioroethane
1, 2,4—Trichioroheizene
Azobenzene
N it robe n ze ne
2, 4-Dinitrotoluerie
2, 6—Dinitrotoluene
Isophorone
C1U ORINATED PESTICIDES MIX
Aidrin
cl—BHC
13-BhC
Y-BHC (Lindane)
6-BHC
4,4 • —DL U
4,4’ -DDE
4,4’ —DDT
Dieldrin
Endosulfan
Endosulf an
Endosuif an
Endrin
Endrin aldehyde
Heptach br
Heptachior epoxidc
BEN ZIDINES MIX
4—Chloro-3-methyl phenol
2 -Chborophenol
2, 4-Dichlorophenol
2, 4-Diinethylphenol
2 ,4—Dinitrophenol
2—Methyl—4, 6—dinitrophenol
2—Nitrophenol
4—Nitrophenol
Pentachboropheno 1
Phenol
2,4, 6-TrichboropheflOl
HAZARDOUS SUBSTANCES MIX 1
BenzOiC acid
2 -MethylpheflOl
4—MethylpheflO l
2,4, 5—TrichioropheflOl
Benzidine
3,3’ —Dichlorobenzidine
INTERNAL STANDARDS MIX
Acenaphthene—dl 0
Chrysene—dl 2
1, 4_DichlorobeflZefle—d 4
Naphthalefle—db
Pery lene—d1 2
Phenanthrenedl 0
HAZARDOUS SUBSTANCES MIX 2
Aniline
Benzyl alcohol
4 -Chboroani line
Dibenzof uran
2-Methylnaphtha lene
2-Nitroaniline
3—Nitroaniline
4—Nitroanhlifle
I
II
sulfate
7—89
-------
TABLE III
SIMILARITIES BETWEEN EPA REPOSITORY AND COMMERCIAL STANDARDS
(ISOTOPE DILTUION METhOD)
Relative Response Relative Percent
Compound EPA Supelco Difference (% )
Benzoic acid * 0.10 0.13 30.49
4—Nitroaniline * 0.06 0.05 —25.92
4—Nitrophenol 0.90 1.11 23.93
4,6—Dinitro—o—cresol 0.82 1.00 21.39
Bis(2—chloroethyl)ether 3.95 4.64 17.42
Benzo(b)fluoranthene 1.04 1.22 17.34
Phenol 3.07 3.55 15.70
Dibenzofuran 0.95 1.08 14.59
Phenanthrene 1.02 1.17 14.46
2—Chiorophenol 0.97 1.11 14.34
4—Bromophenyl phenyl ether * 0.22 0.24 13.63
Dirnethylphthalate 0.89 1.01 13.41
2,4,5—Trichiorophenol * 0.20 0.22 13.11
Chrysene 1.09 1.23 12.86
Benzyl alcohol * 0.26 0.29 12.48
Di—n—butylphtha late 0.94 1.05 12.27
Benzo(a)pyrene 1.04 1.17 12.26
Hexachlorobenzene 1.24 1.39 12.21
2,4—Dinitrotoluene 0.92 0.80 —12.20
2—Nitroaniline * 0.28 0.32 12.16
Benzo(ghi)perylene 1.23 1.37 11.87
Naphthalene 1.00 1.12 11.76
1,2—Dich lorobenzene 1.54 1.71 11.40
2,6—Dinitrotoluene 0.89 1.00 11.38
4—Chloro—m-cresol 1.00 1.11 11.24
1,2,4—Trich lorobenzene 0.88 0.98 11.23
1,4—Dichlorobenzene 1.49 1.65 11.00
Mithracene 1.06 1.17 10.88
2—Nitrophenol 0.96 1.07 10.87
Nitrobenzene 0.91 1.01 10.55
Benzo(a)anthracene 1.06 1.17 10.47
Acenaphthene 1.04 1.14 10.13
Pentachlorophenol 0.89 0.98 10.08
Pyrene 1.01 1.11 10.00
Di—n—octylphthalate 1.01 1.11 9.85
N—Nitrosodi—n—propy lanhine * 0.38 0.42 9.83
Diethylphthalate 0.98 1.08 9.61
Bis(2—chloroisopropyl)ether 0.89 0.98 9.47
Hexachiorobutadiene 1.59 1.74 9.46
Acenaphthylene 1.99 2.17 9.11
Fluorene 1.15 1.25 8.84
7—90
-------
TABLE III (continued)
SIMILARITIES BETWEEN EPA REPOSITOi Y M D COMt iERCIAL STANDARDS
(ISOTOPE DILTUION METHOD)
Relative Response Relative Percent
Compound EPA Supelco Difference (% )
4—chioroani line * 0.52 0.56 8.56
Butyl benzylphthalate * 0.41 0.45 8.49
1,3—Dichlorobenzene 1.52 1.65 8.46
Bis(2—chloroethoXy)lflethafle * 0.56 0.60 8.19
Fluoranthene 1.00 1.07 7.84
2—Methylnaphtha lene * 0.88 0.95 7.73
2-chloronaphthalene 1.38 1.48 7.10
HexachiorOethafle 1.81 1.94 7.04
2,4—Dichiorophenol 1.57 1.68 6.67
Benzo(k)fluoranthefle 0.96 1.02 6.51
Isophorone 1.05 1.12 6.41
Bis(2—ethylhexyl)phthalate 1.05 1.11 5.87
2,4—Dimethyiphenol 2.03 2.15 5.80
N—Nitrosodimethy lamine * 0.32 0.34 5.75
Hexachiorocyclopentadiefle 4.15 3.94 -4.91
2,4,6—TrichlorophenO l 0.77 0.81 4.79
4—l4ethy] .pheno l * 0.42 0.44 4.51
2—Methyiphenol * 0.41 0.43 4.00
Dibenzo(ah)anthracene * 0.39 0.40 2.32
4-chiorophenyl phenyl ether 1.07 1.10 2.26
Indeno(1,2,3—cd)pyrene * 0.40 0.41 2.10
2,4—Dinitrophenol 0.96 0.95 -1.38
3—Nitroaniline * 0.17 0.17 0.10
Difluorobiphenyl (mt. std.) * 1 1 0
Aniline * ND 0.76
Benzidine ND 1.69
3,3 ‘—Dichlorobenzidine ND 1.01
N-Nitrosodipheny l amine ND 0.21
Azobenzene ND 1.11
*No isotopically labeled analog
ND - Not determined
7—91
-------
TABLE IV
SIMILARITIES BETWEEN EPA REPOSITORY AND COMMERCIAL STANDARDS
(SIX INTERNAL STANDARDS METHOD)
Relative Response Relative Percent
Compound EPA Supelco Difference (% )
Benzoic acid 0.08 0.11 28.49
4—Nitroaniline 0.09 0.07 —23.43
4,6—Dinitro-o—cresol 0.08 0.10 20.90
2,6—Dinitrotoluene 0.22 0.26 17.80
Benzo(b)fluoranthene 1.27 1.49 17.73
Di-n—butylphthalate 1.06 1.25 17.66
2,4,5—Trichiorophenol 0.27 0.32 17.57
Benzyl alcohol 0.77 0.90 17.50
2—Nitroaniline 0.39 0.46 16.43
4—Bromophenyl phenyl ether 0.20 0.23 15.57
N-Nitrosodi-n—propylamine 1.15 1.33 14.79
Anthracene 0.89 1.02 14.70
Bis(2—chloroethyl)ether 1.58 1.82 14.68
Phenanthrene 1.02 1.17 14.46
Benzo(k)fluoranthene 1.31 1.50 14.03
Dibenzofuran 1.46 1.66 13.81
2—Chiorophenol 1.30 1.47 13.49
Chrysene 1.09 1.23 12.86
Phenol 1.65 1.86 12.69
Benzo(a)pyrerie 1.04 1.17 12.26
2,4—Dinitrotoluene 0.33 0.29 —11.89
Naphthalene 1.00 1.12 11.76
Dimethylphthalate 1.15 1.28 11.61
Diethylphthalate 1.16 1.29 11.54
Di-n—octylphtha late 2.00 2.22 11.25
1,2—Dichlorobenzene 1.39 1.54 11.13
l,4—Dichlorobenzene 1.49 1.65 11.00
N—Nitrosodimethylainine 0.97 1.08 10.54
4—Nitrophenol 0.06 0.06 10.16
Acenaphthene 1.04 1.14 10.13
Bis(2—chloroisopropyl)ether 0.45 0.49 10.13
Hexachlorocyclopentadiene 0.26 0.24 -10.06
Hexachlorobenzene 0.21 0.23 10.01
4-Methyiphenol 1.27 1.39 9.22
Hexachioroethane 0.48 0.52 9.21
2—}lethylphenol 1.24 1.34 8.74
Dibenzo(ah)anthracefle 0.88 0.95 8.15
Butyl benzylphthalate 0.75 0.81 8.11
2—Chloronaphthalene 1.48 1.59 7.84
Benzo(ghi)perylene 0.92 0.99 7.70
7—92
-------
TABLE IV (continued)
SIMILARITIES BETWEEN EPA REPOSITORY AND COMMERCIAL STANDAM)S
(SIX INTERNAL STP .NDARDS METHOD)
Relative Response Relative Percent
Compound EPA Supelco Difference (% )
Indeno(1,2,3—cd)pyrene 0.90 0.97 7.69
l,3—Dichlorobenzene 1.43 1.53 7.48
Fluorene 1.21 1.31 7.47
Pentachiorophenol 0.07 0.07 7.37
Pyrene 1.54 1.65 7.26
Fluoranthene 0.79 0.84 6.98
Acenaphthy lene 1.65 1.75 6.67
Benzo(a)anthraCefle 1.18 1.26 6.56
4—Chioroaniline 0.43 0.46 6.12
Bis(2-chloroethoxy)methafle 0.47 0.49 5.73
1,2,4—Trich lorobenzene 0.31 0.32 5.49
2,4—Dinitrophenol 0.09 0.08 —5.29
2—Methylnaphtha lefle 0.74 0.78 5.22
Difluorobipheflyl (mt. std.) 1.38 1.44 3.93
3—Nitroaniline 0.24 0.25 3.79
2—Nitrophenol 0.20 0.20 3.41
Nitrobenzene 0.19 0.20 3.30
Isophorone 0.87 0.89 1.66
4-Chiorophenyl phenyl ether 0.60 0.61 1.62
2,4,6—TrichioropheflOl 0.39 0.39 1.61
2,4—DichiorophenOl 0.28 0.27 —1.60
2,4—Dimethyiphenol 0.34 0.34 —1.57
Hexachiorobutadiefle 0.17 0.17 1.56
4—Chloro—m—cresOl 0.32 0.32 1.54
Bis(2-ethylhexyl)phtha late 1.10 1.12 1.21
Benzidine ND 0.16
Azobenzene ND 1.12
Aniline ND 2.37
3,3’—Dichlorobenzidine ND 0.33
N—Nitrosodiphenylamine ND 0.13
7—93
-------
TABLE V
VOLATILES MIX 1
tert-Butylbenzene
sec—Butylbenzene
Ch lorobenzene
2—Chiorotoluene
4—Chiorotoluene
1, 2—Dichlorobenzene
1, 3—Dichiorobenzene
1, 4-Dichlorobenzene
Isopropylbenzene
n—Propylbenzene
o-Xylene
p-Xylene
VOLATILES MIX 3
VOLATILES MIX 2
Benzene
Bromobenzene
n—Butylbenzene
Ethylbenzene
p— Isopropyltoluerie
Naphthalene
Styrene
Toluene
1, 2, 3—Trichlorobenzene
1, 2, 4—Trichlorobenzene
1,2, 4-Trimethyibenzene
1,3, 5-Trimethylbenzene
m-Xylene
VOLATLIES MIX 4
1, 2—Dibromo-3—chloropropane
1, 2-Dibromoethane
1, 1-Dichioroethane
1, 2-Dichioroethane
1, 2—Dichioropropane
1, 3-Dichioropropane
2, 2-Dichioropropane
1, 1—Dichioropropylene
Hexachiorobutadiene
1,1,1, 2—Tetrachioroethane
1,1,2, 2—Tetrachioroethane
Tetrachioroethy lene
1,1, 1—Trichioroethane
1,1, 2—Trichioroethane
Trichioroethy lene
1,2, 3—Trichioropropane
Bromochioromethane
B romodic hioromethane
Bromoiaethane
Carbon tetrachioride
Chioromethane
Dibroinochioromethane
cis—1, 2—Dichioroethylene
trans—i, 2—Dichioroethylene
Trichiorof luoromethane
VOLATILES MIX 5
Bromoforiu
Chioroethane
Chloroform
Dibroinomethane
Dichiorodifluoromethane
1, 1—Dichioroethylene
Methylene chloride
Vinyl chloride
7—94
-------
Quality Assurance of Analytical Chemistry Through Auditing
Eugene J. Klesta, Manager, Quality Assurance/Quality Control; Mark F.
Marcus, Director of Analytical Programs, Chemical Waste Management,
Inc., Technical Center, Riverdale, IL.
ABSTRACT
Hazardous waste is an extremely complex and variable material. In
general, the waste generator is not concerned with the homogeneity or
the consistency of the waste. Therefore, the material which is
submitted to Chemical Waste Management for waste disposal evaluation
offers the analytical chemist a challenge not found in other
industries. Manufacturers of chemicals are concerned with the yield
and quality of their product. Product specifications nnist be met. The
waste resulting from the process or material which is “out of spec”
is set aside for proper disposal without regard for the “quality” of
these materials. Because the composition of the waste cannot be
controlled, it became essential to develop a stringent program to
control and assess the analysis of the waste. The guidelines set
forth by the Association of Official Analytical Chemists were used as
the basis for developing the Chemical Waste Management Qaulity
Assurance Policy.
The major concern of Chemical Waste Management is to provide valid,
defensible data in a timely manner. The terms used in this statement
require further explanation. Valid means that the precision and
accuracy of the analytical data are maintained for all parameters
being tested at our laboratories. To be defensible, the data and the
records of the data must be able to withstand scrutiny, of the
highest degree. The records must be completely traceable from the
start of the analytical process to the final report. There are many
factors which will be described later that also contribute to
defensibility. Briefly, these include: qualified personnel,
appropriate analytical methods, and proper equipment and facilities.
The demands of chemical processing hazardous waste and the costs
resulting from delaying transportation equipment were critical in
evaluating which quality assurance and quality control procedures
would be included in the overall plan. The processing of hazardous
waste like most other industrial processes requires a short
turnaround time for analysis. Therefore, generation of the data must
be done in a timely manner.
The Chemical Waste Management Quality Assurance Policy includes the
following principles:
(1) Defensible Documentation
(2) Analytical Methods
(3) Sampling and Sample Control Methodology
(4) Facility Adequacy
(5) Equipment Maintenance and Calibration
(6) Personnel Training
7—95
-------
(7) External and Internal Assessment
(8) Quality Control Policy and Procedures
These principles are briefly described as follows:
(1) Defensible Documentation
This principle requires the use of accountable documents, standard
operating procedures (SOP) for recording data, chain—of—custody
requirements, instrument parameters, and analytical methods. The
procedures for loss or destruction of data and archiving are also
included.
(2) Analytical Methods
A site specific methods manual nust be in place and include all
methods described in the Waste Analysis Plan. Standard methods are
used when applicable. A methods review ccminittee submits new methods
or modifications of existing methods to the Director of Analytical
programs for approval. The themical Waste Management nine—point
format is required.
(3) Sampling and Sample Control Methodology
Standard operating procedures are used to describe the appropriate
sampling techniques and equipnent. The sample chain of custody must
remain inviolate. Test samples are properly prepared and test
portions are properly taken for analysis. The written procedures
include labeling, sample containers, holding times, documentation
requirements, and safety concerns for the sampler.
(4) Facility Adequacy
Adequate space, proper safety equipnent, appropriate laboratory
furnishings, and sufficient instruments and supplies are maintained
at all laboratories. Sufficient hood space in proper working
condition is provided to protect all laboratory employees.
Housekeeping guidelines are used to maintain a proper environment for
performing analysis.
(5) Equi nnent Maintenance and Calibration
All maintenance and daily performance checks must be documented as
permanent laboratory records. Instruments which require calibration
are scheduled for service on a regular basis. Equi znent may be sent
out or have the proper service performed in the laboratory as
required. All calibration curves and printouts must be properly
documented and stored.
(6) Personnel Training
Written training protocols must be documented. New employees as well
as current ençiloyees shall be kept informed of methods and new
technologies. Attendance at symposia, training courses, and further
7—96
-------
formal education is encouraged. Training which has been completed
must be documented for each employee and kept as part of the training
records.
(7) External and Internal Assessment
The Manager of Quality Assurance/Quality Control and the Qaulity
Assurance Unit, which consists of several quality assurance and
quality control auditors, are responsible for external assessment
through the use of audits. The performance of the laboratory is
evaluated and reported to management. The site laboratory manager
internally assesses compliance to the QTVQC policies and procedures
on a regular basis. Review of documentation, use of blind duplcates
and quarterly reference materials are used to evaluate individual
performance in the laboratory.
The use of audits and their results will be described below in more
detail. The external assessment of the Chemical Waste Management
laboratories has had the greatest effect on the quality assurance of
the analytical chemistry within those laboratories.
(8) Quality Control Policy and Procedures
The Quality Control Policy and Procedures apply to the data
generation processes which occur within the laboratories. Typical
quality control concepts deal with a product manufactured with “zero
defects” or with a service provided with “every customer satisfied.”
In the analysis of hazardous waste, the product is valid, defensible
data. The precision and accuracy of that data are a result of the
quality control principles that follow.
(a) Instrument Performance Paranienters
A daily check must be performed on all equipment and instrumentation
to verify that performance is within a set of criteria statistically
pre—established. The time spent to perform this check must be
reasonably short. The data is recorded in bound notebooks to
substantiate instrument performance.
(b) Contamination Evaluation
The analyses of method blanks for each parameter tested are performed
with each batch of samples. The results are documented to show that
cross—contamination of samples and contamination of the reagents have
not occurred. The method detection limits and the resulting limit of
quantification are used as criteria for the acceptability of the
blank results.
(C) Quality Control Check Samples
Each analytical test performed in the labortory is required to have a
quality control check sample which is analyzed daily. Statistical
process control charts are generated from the data for the check
samples. The control charts are used to maintain precision of the
7—97
-------
analytical process. After reference laboratory analysis of the
quality control check sample, the accuracy of the analysis can also
be monitored. Control limits of plus or minus three standard
deviations of the mean are established as acceptability criteria.
(d) Dupliction
Every tenth sample for each parameter is analyzed in duplicate. Two
test portions are taken and carried through the analytical procedure.
The acceptance criterion for the relative percent error hs been set
at a maximum of twenty percent. Control charts are generated from
this data.
(e) Fortification
Every tenth sample for each parameter is fortified with a known
amount of the analyte being tested. The fortifcation is made to the
sample which was duplicated. The accuracy of the anlytical test is
determined by calculating the percent recovery of the fortification.
The acceptance criteria have been set at 80% to 120% recovery.
Control charts for each parameter are generated from the recovery
data.
(f) Reference Materials
Each quarter, standard reference materials from the EPA, NBS, or
standard service organizations are analyzed for each parameter
available. The performance must be within the acceptance range stated
on the standard material.
(g) Round Robin Analysis
Each quarter, a set of samples are sul*uitted to the laboratories for
anlaysis. These round robin samples are formulated by the quality
assurance unit and the results are compiled and distributed to the
laboratory managers. All results must be within two standard
deviations of the mean to be acceptable.
(h) Reference Laboratory Evaluation
Each month, the labortory must suhuit a sample which has already been
analyzed to the Technical Center for interlaboratory duplicate
analysis. The quality assurance unit serves as the mediator between
the laboratories to address any discrepancies and to compile the data
to determine any systematic bias in analysis.
Evaluation of laboratory performance is determined through a rigorous
audit program. No specific audits have been developed; namely, the
Quality Assurance Audit and the Quality Control Audit.
Quality Control Audit
The quality control audit is in the format of a matrix. The
paramenters tested by the labortory form one coordinate of the
7—98
-------
matrix. A section of the QC policy is reviewed through a series of
questions which form the other coordinate. Each question is answered
with a numerical score. If all of the details are in place, then a
five is awarded. When some portion of the required information is
missing, a three is awarded. If the procedure is not in place, then a
score of one is given. Some questions may be answered as “riot
applicable.” The total number of questions answered are summed, and
the average is determined by dividing the sum by the number answered.
When all of the parameter columns are completed, then the overall
average of the columns is determined by summing the average scores
for each column and dividing by the number of columns. This process
is completed for each part of the quality control policy described
earlier. The score for the audit is derived by suinming the overall
average for all of the parts (a) through (h).
The individual sections of the quality control audit (a through h)
can be averaged for all laboratories audited to determine which areas
need further attention. The quality assurance unit will then provide
additional training and technical assistance to improve the specific
areas of concern. The overall progress of the program is monitored by
plotting the average scores for each section through time.
The scores for the audits are compiled and evaluated semi—annually.
When the data for the last year and a half are plotted together, the
progress is clearly shown.
Quality Assurance Audit
The quality assurance audit consists of a series of questions
covering each of the major principles of the Quality Assurance
Policy. The questions are all written in a form for which the correct
answer is “yes.” The other choices for each questions are “no” and
“not applicable.” The number of questions answered “yes” is divided
by the number of questions answered as applicable. This result
expressed as a percentage becomes the score of the audit. Acceptance
criteria have been established to determine if a particular
laboratory is performing up to company standards.
The quality assurance audit is summarized. From this summary, action
items are established and discussed with the laboratory manager and
appropriate facility management personnel. Resolution of these items
moves the laboratory toward the objective of the program. The quality
assurance practices improve through time as the audit process is
repeated. The auditor will review the preceding action items to be
sure that all of them have been addressed since the last audit.
The quality assurance audit scores are plotted in descending
numerical order. Each succeeding round of scores is added to the same
graph. The average score has increased and the slope of the line has
decreased. The positive impact that auditing has on the quality
assurance of analytical chemistry is most significant.
7—99
-------
PREPARATION OF NATURAL MATRIX TYPE SAMPLES
FOR PERFORMANCE EVALUATION OF RESOURCE CONSERVATION
AND RECOVERY ACT (RcRA) CONTRACT LABORATORIES
Harold A. Clements, Chief, Raymond E. Loebker, Chemist, and Donald
L. Kiosterman, Chemist, Evaluation Section, Quality Assurance
Branch, Environmental Monitoring and Support Laboratory —
Cincinnati, U.S. Environmental Protection Agency, Cincinnati, Ohio
45268
ABSTRACT
The Quality Assurance (QA) Branch develops, analyzes, and
distributes natural matrix samples to evaluate the performance of
laboratories conducting analyses for the Office of Solid Waste
(osw). In these studies, each laboratory is requested to analyze an
organic sample and an inorganic sample using the analytical methods
specified by OSW.
The sample designs and reference values are established by multiple
analyses in the QA Branch laboratory and contract referee
laboratories. These data are evaluated statistically and must be
judged acceptable before samples are distributed to the 50—60 RCRA
contract laboratories. These data are also furnished to OSW for use
in evaluating the results reported by the contract laboratories.
INTRODUCT ION
The Quality Assurance (QA) Branch of the Environmental Monitoring
and Support Laboratory — Cincinnati (EMSL—Cincinnati) provides QA
support for the quality control efforts of laboratories within the
United States which conduct water and waste analyses under the Safe
Drinking Water Act (SDWA), the Clean Water Act (CWA), the Resource
Conservation and Recovery Act (RCRA), and Comprehensive Environ-
mental Response Compensation and Liability Act (CERCLA) and
regulations.
Under RCRA, the QA Branch develops, analyzes, and distributes solid
matrix samples quarterly to evaluate the performance of laboratories
conducting solid waste analyses. In these studies, each laboratory
is sent two samples, one each for organic and inorganic analyses
using the methods specified in Office of Solid Waste (osw) Method
Manual 846, Test Methods for Evaluating Solid Waste, 1982. The
methods for trace element analysis includes an acid extraction
(1310), followed by measurement using the Inductively Coupled Plasma
Method (6010). The extraction procedure simulates the leaching of
waste in a sanitary land fill. The recovery of the metals at the
extraction pH 5 + 0.2 is comparable to conditions in a sanitary land
fill. Because the samples are natural matrix samples and
7—101
-------
2
theoretical values are not calculable, their reference values are
established by multiple analyses in the QA Branch laboratory and
contract referee laboratories. These data are then evaluated
statistically and must be judged acceptable by the QA Branch before
the samples are approved for distribution in a formal study to the
50—60 RCRA contract laboratories. The reference values and
statistical estimates are then furnished to OSW for use in
evaluating the results reported by the participating laboratories.
SAMPLE DESIGN MID PREPARATION
Base materials for preparation of the organic and inorganic samples
are industrial wastes or soils such as electroplating waste,
wire—coating waste or creosote—contaminated soil added to infusorial
earth, kaolin or garden soil. Stock solutions of analytes prepared
from reagent—grade chemicals are added to the base materials (called
spikes) to produce the desired levels for each contaminant.
PREPARATION OF INORGANIC SAMPLE
Approximately 75 lbs. of relatively dry base material are passed
through a 4 i n mesh screen and transferred to a 3.5 cu. foot
cement mixer. Use of the small cement mixer is ideal; it is
open at the top for easy access and is tilted at a 450 angle
for tumbling and mixing. Stock solutions of chemicals are
spiked into the base material to adjust analyte levels. Spikes
are first dissolved in hot or cold reagent water before being
slowly mixed with the tumbling base material. The amount of
spike is based on the desired “analyzed” target value and
experience gained in previous use of the matrix. Note that the
R RA method for metals incorporates a weak extraction using
acetic acid which may not recover 100X of a metal. For example,
an analyzed value of 230 mg/L barium required an initial spike
of 846 grams of BaCl 2 with 75 lbs. of base material and a
total of 8 liters of water. Furthermore, analyses for some
metals exhibit a “threshold” point before detection. In these
cases (i.e., Ba, Cr, Pb), doubling or tripling the spike may
yield little or no response until the threshold is reached. On
the other hand, compounds like sodium arsenate, cadmium nitrate,
mercuric nitrate, and sodium selenite show the effect of
addition innediately without building to a “threshold” level.
7—102
-------
3
In preparing the samples, the chemist needs information on the con-
centration of various compounds. This is obtained by analyzing
spiking, mixing and analyzing again. The target concentrations are
generally the desired levels set by Office of Solid Waste and
Emergency Response (OSWER), i.e., barium at a level of 100 mg/L.
Table 1 summarizes the final composition and analyses of a single
batch of an inorganic solid matrix sample for RCRA.
TABLE 1: INORGANIC SOLID MATRIX SAMPLE
ADDED AS COMPOUND WATER FOUND
ELEMENT COMPOUND ADDED, g ADDED, mL mg/L*
As Na 2 HAsO 4 7H 2 0 300. 2000 7.
Ba BaC 1 2 846. 3000 107.
Cd Cd(N0 3 ) 2 4H O 34. 1000 3.
Cr K 2 Cr 2 O 7 367. 3600 0.1
Hg Hg(N0 3 ) 2 1120 21. 400 0.5
Pb Pb(N0 3 ) 2 582. 3000 1.5
Se Na 2 SeO 3 167. 1000 9.
*Results of analyses by inductively—coupled plasma after spike.
Experience in preparing this type sample indicates that the pH
generally stays above 5.0 units. The process of adding more base
material and water to reduce an overspiked value and/or adding more
of a compound to raise an analyte value is continued until the
target values are reached. If the sample is too viscous for good
mixing, water is added and mixing continues. Dry solids adhering to
the bottom or sides are broken up and mixed into the slurry.
Once the analyses confirm that the desired range of concentrations
has been reached, a small subsample is taken to verify that the pH
remains above 5.0 units. If a pH adjustment is necessary, a small
amount of ammonium hydroxide is added.
7—103
-------
4
When the analyzed values are on target, the consistency is not too
viscous, the sample appears well mixed, and the pH is acceptable,
mixing is continued at least overnight. While the mixer is still
rotating, a long handled scoop is used to remove approximately 200
grams of solid sample and then to place it in each of a series of 8
oz. glass jars. About an inch of air space is left to facilitate
mixing prior to removal for analysis. The spillage is wiped off the
outside, the cap is placed on tightly, and plastic tape is then
wrapped tightly around the cap to prevent loosening.
About 65 samples (200 + 50 grams) are prepared from each batch, and
labeled. The order in which the samples are filled is noted on the
labels and analyses of the 1st and last sample serve as an indica-
tions of batch homogeneity.
PREPARATION OF ORGANIC SAMPLE
Industrial wastes and soils are also used as base materials for the
organic sample. Some samples contain adequate organics without
spiking, such as creosote—contaminated soil. Industrial waste or
soil may be diluted with infusorial earth if necessary. Organic
compounds from the Appendix VIII list are added to the base material
to obtain the desired levels.
The key steps required in the preparation of a typical organic
sample are described below:
The amount shown for each of the compounds in Table 2 is weighed out
and dissolved in methylene chloride and added to the base material.
TABLE 2: PREPARATION OF AN ORGANIC SOLID MATRIX SAMPLE
ANALYTE
AMOUNT FOUND
ANALYTE ADDED, mg ug/g
acenaphthene 229 97
2,4—dimethyiphenol 583 70
p—chlorophenol 938 104
2,4,6—trichlorophenOl 344 120
4—bromo—phenyl phenyl ether 476 132
p,p’DDD 345 98
bis (2—ethyl hexyl) phthalate 350 121
isophorone 470 150
7—104
-------
5
The selected volumes of analytes are added to 1500 grams of a 50/50
mix of infusoria]. earth and kaolin. Water and additional base
material are added to bring the total weight of the sample to 1775
grams.
The sample is placed in a 3 liter ceramic container, sealed, and
mixed on a ball mill for three to four hours. The container is
opened periodically and the solids cleared from the sides of the
container. Mixing continues for approximately 4 hours. At this
point, at least 2 subsamples are withdrawn and analyzed to verify
homogeneity. When the analyses indicates a homogeneous mix,
approximately 20 grams are dispensed into each of seventy 2—oz.
bottles with teflon lined caps, and sealed.
ANALYSES OF PREPARED SAMPLES
The QA Branch analyzes the samples at time zero, two weeks and six
weeks. Concurrently, samples are sent to three referee laboratories
for analyses as unknowns. The OSW specified methods are used to
analyze the inorganic and organic samples. The data are compiled
and the mean recoveries and standard deviations of each analyte
determined. These statistical results are then provided to OSW for
use in establishing acceptance limits for data from the 50—60
participating RCRA laboratories. The calibration standards required
in the organic analyses are provided from the USEPA Repository of
Toxic and Hazardous Materials to ensure comparable standards among
all laboratories.
STUDY DESIGN
The samples are mailed to the RCRA laboratories with instructions
for sample handling and specifying the OSW methods to be used for
the analyses. Since some organic methods permit use of more than
one solvent for the extraction, the instructions specify one solvent
for use in the study. The instructions also include a listing of
the compounds to be measured in this sample. The complete package
of organic and inorganic samples with instructions for handling and
analyses) and the calibration standards are mailed to the
laboratories specified by OSW. In about four weeks, data from these
laboratories are returned to OSW for evaluation of laboratory
performance.
7—105
-------
A FIELD AUDIT PROGRAM TO ENSURE THE QUALITY OF
ENVIRONMENTAL MEASUREMENTS
William F. Lowry, Audit Program Manger, Stephen A. Borgianini,
Christine M. Andreas, New Jersey Department of Environmental
Protection, Division of Hazardous Site Mitigation, Environmental
Measurement Section, Trenton, New Jersey
ABSTRACT
The word audit, by definition, means to examine with intent to
verify. In recent years, field audits at hazardous waste sites have
become a tool which may be used to insure quality control at a site
under reniediation. A Quality Control Field Audit is the primary
method utilized by the State to insure contractor compliance with
the terms and conditions of the contract and also to determine the
adequacy of field operations in generating data representative of
environmental conditions. NJDEP/DHSM has developed stringent
guidelines for contractors either working directly for the
Department or planning to submit data to the Department for review
and validation. NJDEP/DHSM has been operating a Field Audit Program
for over four (4) years with audits being conducted at over five
hundred sampling events. The Field Audit Program has taken a very
aggressive approach to Remedial Investigations (RI) at Superfund
Sites since New Jersey has initiated RI’s at over ninety NPL sites.
By use of a detailed field auditing form, the auditor can highlight
certain deficiencies which may affect data quality and usefulness.
Conversely, the forms may be used to Indicate complete compliance by
the contractor. In most cases, the auditor and contractor will be
able to develop altered field procedures as necessary, to ensure the
quality of the samples and subsequent data generated.
If deficiencies noted in the field are deemed major by the auditor,
and no satisfactory alternative can be arranged between the auditor
and the contractor, the power to stop work until the problem is
mitigated is invested in the auditor. Should the contractor choose
to continue with field activities, he is informed it is at his own
risk.
Typically, an auditor is present on the first or second day of
sampling conducted for environmental analysis. The auditor may
return to the site further into the project to witness sampling of
various media. If an incidence of non—conformance is noted early in
field activities, he may return at a later date, unannounced, to
insure the contractor is in compliance with the revised protocol.
Contractor non—compliance falls into two categories: (1) minor
infractions, and (2) major infractions. Minor infractions are those
problems brought immediately to the contractor’s attention in the
7—107
-------
field. The impact to data generated is greatly minimized if the
improper procedure can be quickly corrected. Major infractions are
those procedures which substantially deviate from the approved work
plan or DUSM Field Procedures. A major infraction would be
something that potentially causes data critical to the evaluation of
a project to become qualified or suspect.
Hence, the Field Audit program is a valuable mechanism by which the
State can assess contractor compliance with approved field methods
and to minimize the need for resampling at sites or data
qualification due to contractor non—compliance.
INTRODUCTION
The word audit, by definition, means to examine with intent to
verify. In recent years field audits have become an essential
mechanism for ensuring that quality control requirements and
procedures are strictly adhered to during the collection of
environmental measurements. The New Jersey Department of
Environmental Protection (NJDEP) has operated a Field Audit Program
for the past four (4) years with over five hundred (500) sampling
episodes having been audited by trained personnel. The
justification for allocating resources to operate such a program is
derived from NJDEP’s aggressive commitment to insuring the validity
of analytical data resulting from the collection of environmental
measurements at various hazardous waste sites. The audit program is
based upon the premise that one must insure the quality and
representativeness of samples generated in the field in order to
certify that corresponding analytical data is in fact representative
of site conditions. A second important premise is that sampling
techniques utilized during several sampling events must be
consistent for the duration of a project to ensure that data from
one event is comparable to another in terms of how samples are
collected, handled, and packaged prior to shipment to the laboratory
for analysis. An audit program also provides oversight in the field
to insure contractor compliance with the terms and conditions of
NJDEP approved site specific documents which may include Quality
Assurance Project Management Plans, Field Sampling Plans, and Health
and Safety Plans. These documents provide a bench mark for traIned
auditors upon which contractor performance Is evaluated. Thus,
review and familiarity with these documents by the auditor Is a
prerequisite to conducting a Quality Control Field Audit.
This paper will describe the primary components of NJDEP’S Quality
Control Field Audit Program and focus on the application of field
audits to Remedial Investigations at Superfund sites conducted in
accordance with the provision of CERCLA.
7—108
-------
PROGRAM ORGANIZATION
NJDEP’S Quality Control Field Audit Program operates out of the
Bureau of Environmental Measurements and Quality Assurance (BENQA)
within the Division of Hazardous Site Mitigation. BENQA is
responsible for all aspects of data collection, handling and
validation. The Bureau is charged with assuring that all data
collected by and presented to the NJDEP Hazardous Waste Programs is
of a known and verifiable quality and Is representative of actual
environmental conditions under Investigation. The Bureau Is further
organized into two major groups; the Environmental Measurements
Section and the Quality Assurance Section. A primary component of
the Environmental Measurement Section (ENS) Is the Quality Control
Field Audit Program and Technical Review Service.
At present this program consists of one full time program manager
and four full time auditors to oversee sampling activities at New
Jersey Superfund sites. An additional pool of trained personnel is
available within EMS when temporary increases in field activities
occur. Program personnel work in concert with NJDEP Site Managers
and Technical Coordinators with respect to review of technical
documents, approval of specific sampling procedures, and also
recommend changes in procedures to Incorporate the most current
QA/QC protocol into the site evaluation/remedlation strategy. The
auditor also interacts with the Quality Assurance Section staff to
provide valuable information that may be used to answer questions or
clarify issues related to analytical data quality and validation.
PERSONNEL TRAINING AND QUALIFICATIONS
Members of the EMS audit team should possess the appropriate
education, training and field experience prior to conducting field
audits individually. Each member of the audit team should satisfy
the following requirements.
1. A minimum of one year field experience in the collection of
environmental measurements following NJDEP methodologies and
use of approved sampling equipment.
2. A minimum of one year experience performing technical
document reviews and preparation of comment memorandums.
3. At least six months field training with an experienced
auditor in conducting field audits, and preparing QC Field
Audit Reports.
4. Demonstrated ability to accurately identify problems with a
contractor’s performance and maintain a professional and
objective attitude when interacting with contractor personnel
to mitigate observed problems.
7—109
-------
5. Enrollment in the NJDEP Medical Surveillance Program.
It is important to ensure that auditors are familiar with problems
that routinely occur in the field and that they are able to
recommend the appropriate corrective measures. This ability is
essential to prevent unnecessary delays and prevent actions from
occurring which are out of compliance with the approved project
plans.
ADMINISTRATIVE AND OPERATIONAL PROCEDURES
To properly manage an audit program day to day operations must be
organized and set forth in writing. To meet this objective NJDEPS
audit program has developed several procedures and administrative
requirements to promote efficient operations.
All NJDEP Site Managers and Technical Coordinators must provide the
Audit Program Manager with the applicable technical documents for
review at least two weeks prior to scheduled field activities. This
allows the auditor sufficient time to become familiar with the
objectives of a particular sampling event and an opportunity to
recommend changes in sampling procedures if deemed necessary.
These documents are reviewed for Individual technical merit and
evaluated against those procedures found in the NJDEP Field Sampling
Procedures Manual, July 1986 Edition. These plans are also reviewed
in terms of their applicability to site specific conditions and are
evaluated to determine If the specific objectives of the sampling
episode will be met through implementation of the various documents
as written.
Copies of the final revised documents are also reviewed to determine
if comments from the auditor were incorporated. Participation In
the review process is an important quality control function designed
to prevent plans of questionable integrity from being implemented
which may result in the generation of less than acceptable
analytical data.
All requests for document reviews and field audits are received via
a one page Work Request Form which specifies who the requester is,
the nature of the task, and other relevant Information Including the
due date. AU work request8 are then entered into a tracking system
to maintain up to date information on current work load, staff
assignments, and project status. This tracking system is also used
to plan weekly field schedules, record outputs (Field Audit Reports,
Document Review Memorandums, etc.) and as a source of information
for preparation of monthly reports on program activities.
Additional procedures with corresponding flow charts have also been
developed to display the flow of information, personnel
7—110
-------
responsibilities, and docunientatiori required for completing all
document reviews, field audits and audit reports.
Another important element of the program is the actual format used
to document contractor performance in the field. Detailed reporting
forms are utilized by the auditor to record information related to a
particular sampling event. These contractor evaluation forms
consist of the following.
1. Cover Sheet Page
This form asks for a brief synopsis of the actual audit in
terms of the site name and location, the contractor’s name,
who requested the audit, what the relevant plan and documents
are, highlights of problems encountered if any, matrices
witnessed and a recommendation for qualification of
analytIcal data due to collection procedures if warranted.
2. General Information Page
Here the names of personnel on site, their affiliation an
phone number are recorded. Also Included are spaces to
record the regulatory program being applied to the project,
the project phase, weather conditions, level of personnel
protection utilized and any other comments the auditoc feels
are warranted.
3. QAJQC Information Page
On this page the auditor is asked to record the name of the
laboratory performing the analysis, specific analytical
parameters requested, and indicate if Chain—of—Custody was
initiated properly for all samples.
Other items requested include specif lea on sample preserva-
tion technique, decontamination of sampling equipment, use of
QA/QC samples (duplicate field blanks, trip blanks) descrip-
tion of sample shuttles and type, frequency of use and
calibration methods used for air monitoring instrumentation.
The audit forms also include separate pages which must be
completed for each sample. Separate forms for aqueous and
non—aqueous samples are used providing 8pecific information
on sample locations, sampling method, collection time, sample
appearance and other relevant details of a specific sample.
Another report form Is used when observing the installation
of monitoring wells as part of a groundwater investigation
program.
7—111
-------
By using a consistent reporting format and following specific
operational procedures, audit personnel are provided basic
tools for implementing the program in a timely and efficient
manner.
PROCEDURES TO ADDRESS CONTRACTOR NON—CONFORMANCE
When conducting a Quality Control Field Audit, the auditor may
witness a situation that may cause the contractor to deviate from
the approved sampling plan or the auditor may actually witness
sample collection and handling procedures that will potentially
compromise the Integrity or representativeness of a particular
sample or sample shipment. Recognition and mitigation of these
problems is perhaps the moat important element of a QC Field Audit
Program.
NJDEP is committed to immediate identification and subsequent
correction of any problems in the field on a real time basis. In
this situation the auditor is empowered to request the contractor to
stop work immediately, Identify the problem observed and recommend
specific actions necessary to permit the sampling episode to
continue. By recommending immediate corrective action the auditor
can ensure that resampling of a particular location Is accomplished
if the initial procedure witne8sed was performed incorrectly. This
type of corrective action is preferred over simply recording the
deviation in the audit report and allowing a sample of questionable
Integrity to be submitted for analysis.
Here again specific procedures to guide the auditor with respect to
communicating with contractor personnel and NJDEP management staff
have been developed to accomplish problem resolution. These
procedures are separated into two categories. The first deals with
mitigation of minor infractions or non—conformance Items. These are
problems brought immediately to the attention of the contractor and
are easily corrected. Negative impact to samples generated can be
readily eliminated and procedures are corrected to achieve the
desired result. Examples of minor infractions routinely observed
are as follows:
1. Failure to change disposable gloves between procurement of
individual samples.
2. Insufficient well evacuation.
3. Incorrect decontamination of sampling equipment.
4. Improper operation of sampling equipment.
The second category addresses major infractions or non—conformance
items. These include events or procedures that substantially
deviate from approved plans and sampling procedure or will otherwise
result in Increased project coats not previously approved. A major
infraction would be something that causes subsequent analytical data
7—112
-------
critical to the environmental evaluation of a project to be
qualified or suspect. This type of infraction has the potential to
result in the auditor requesting total or partial rejection of
samples scheduled for shipment to a laboratory or a recommendation
to the Quality Assurance Section Chief that analytical data
generated be qualified or rejected.
Examples of major infractions may include:
1. Failure to provide required QA/QC samples.
2. Lack of field equipment decontamination.
3. Substantial changes in sample location, matrix or frequency.
4. Use of non—approved sampling equipment or methodologies.
In addition to monitoring contractor performance the presence of an
auditor on site has other benefits.
In the event citizens or media representatives arrive on site these
individuals communicate directly with an NJDEP representative who
can briefly explain what is happening on site and refer individuals
to the Department’s Press Office, Office of Community Relations or
other appropriate officials for further information. The presence
of an auditor during the sampling of residential potable wells also
provides an opportunity for homeowners to speak directly with a
government official who can answer questions and/or direct them to
the appropriate NJDEP office for further information.
Other benefits of the program include the generation of a Quality
Control Field Audit Report to NJDEP Site Managers which provides
valuable information within two weeks of the sampling event.
Without the benefit of the report, management staff might wait
substantially longer periods of time until the results of the site
investigation are reported formally to NJDEP by the contractor. The
audit report can also be used to verify or challenge the findings of
the contractor’s report with respect to the collection and handling
of samples and the evaluation of on site conditions.
The findings contained within the audit report may also be used to
assist in the review and validation of analytical data. In addition
to laboratory performance, the methods employed to collect, handle,
preserve and package environmental samples may have a significant
impact on data quality. By monitoring these activities in the field
the audit program assists in assuring that remedial decisions are
based on data generated from representative samples.
Potential cost savings are also realized by preventing samples of
poor quality from being analyzed and preventing subsequent remedial
decisions from being based on data of questionable integrity.
7—113
-------
PROGRAM SUMMARY AND FUTURE OBJECTIVES
The Quality Control Field Audit Program has gained respect and
recognition as an essential component of the Department’s Quality
Assurance Program. Field audits initially were employed as a tool
for evaluating contractor performance in the field when Remedial
Investigations were first conducted at Superfund Sites.
Since that time New Jer8ey has initiated Remedial Investigations
(RIa) at over ninety (90) NPL sites and has proposed an aggre8sive
schedule for conducting RIs at many additional sites in fiscal year
1988. The audit program managed by BEMQA has also been utilized by
other regulatory programs within NJDEP when collection of
environmental measurements by contractors are required to evaluate
the known or suspected presence of hazardous materials on site.
The audit program has several program development items on Its
future agenda. The first is the development of a contractor
education program. This program will consist of seminars for
professional firms having been awarded contracts by the Department
to conduct remedlation activities at hazardous waste sites. These
seminars will focus on applicable QA/QC requirements associated with
the collection of environmental measurements and generation of
analytical data. The Intent of the8e seminars is to provide an
opportunity for DEP and contractor personnel to discuss current
policies and requirements prior to the preparation of site specific
proposals. Since QA/QC policies are frequently revised as new
technical approaches are developed, It is Important for all
contractors working for the Department to be familiar with up to
date requirement s.
A second important objective is the continuance of a cooperative
working relationship with the IJSEPA Region II Monitoring Management
Branch. For the past year BEMQA 8taff have been meeting regularly
with our EPA counterparts to discuss various QA/QC issues involving
both field and laboratory performance standards. Our mutual goal Is
to establish consistent policies between the two agencies and to
maintain a forum for policy review and evaluation.
Finally, the audit program Is committed to the continued objective
and consistent evaluation of professional firms contracted by the
Department to conduct hazardous waste site Remedial Investigations.
Through these efforts we will continue to assist in assuring that
the Department Is provided with accurate, measurable data which
provides an Important basis for responsible decision making.
Acknowledgements: The authors of this paper are staff members of
New Jersey Department of Environmental Protection, Division of
Hazardous Site Mitigation.
7—114
-------
ACKNOWLEDGEMENTS
The authors of this paper are rtaff ae bers of New Jersey Dep: r ent
of Eaviron ent Protection, D1vi ion of Hazardous Site Mttjgatton,
Note: These procedures are repre eotative of the rypes of coctraic
which are in effect rs ol: Mar i987 These conirols are sebje t to
revision and expansior
7—115
-------
PERFORMANCE AUDITS RECOMMENDED FOR VOLATILE AND SEMI—VOLATILE
ORGANIC MEASUREMENTS DURING HAZARDOUS WASTE TRIAL BURNS
R. K. M. Jayanty, J. M. Allen and C. K. Sokol, Research Triangle
Institute, Research Triangle Park, North Carolina; D. J. Von Lehmden
and T. J. Logan, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina
ABSTRACT
The measurement of air toxic organics requires sophisticated
sampling and analysis systems. Agency personnel responsible for the
use of air toxic measurements need to be concerned about the
accuracy of such measurements. The application of quality assurance
practices is important to the generation of high quality measurement
data. One such practice is the performance audit, which involves
providing “unknown” or “blind” samples for measurement. When
conducted simultaneously, during a hazardous trial burn test, the
performance audit provides an assessment of the measurement accuracy.
The Quality Assurance Division of USEPA’s Environmental Monitoring
Systems Laboratory has a program to develop selected volatile and
semi—volatile organic audit materials to federal, state, and local
agencies or their contractors for use in performance audits to
assess the accuracy of measurement methods during hazardous waste
trial burn tests. Research Triangle Institute (RTI), under contract
to the USEPA, has responded to this need through development of gas
cylinders containing 27 gaseous volatile organic compounds in five,
six, seven and nine component mixtures at ppb levels and 5
semi—volatile organics spiked on XAD—2 cartridges.
Studies of all organic compounds (both volatiles and semi—volatiles)
were performed to determine the stability of the compounds and their
feasibility as performance audit materials. Results indicate that
all of the selected organic compounds are adequately stable as
reliable audit materials.
Performance audits have been conducted using the audit materials to
assess the accuracy of the measurement methods. To date, 110
performance audits have been initiated using ppb level audit gases.
The audit results obtained with audit gases during hazardous waste
trial burn tests were generally within ± 50 percent of the audit
concentrations. A limited number of audit results were obtained
with the spiked cartridges. Results were generally within ± 35
percent of the audit concentrations. The list of volatile and
semi-volatile organics, RTI measurement procedures, stability data,
audit procedures and results of representative performance audits
will be presented.
7—117
-------
INTRODUCTION
The measurement of air toxic organics requires sophisticated
sampling and analysis systems. Agency personnel responsible for the
use of air toxic measurements need to be concerned about the
accuracy of such measurements. The application of quality assurance
practices is important to the generation of high quality measurement
data. One such practice is the performance audit, which involves
providing “unknown” or “blind” samples to laboratories for
analysis. When conducted simultaneously during a hazardous waste
trial burn test, the performance audit provides an assessment of the
measurement accuracy.
The Quality Assurance Division of the U. S. Environmental Protection
Agency’s (EPA’s) Environmental Monitoring Systems Laboratory has a
program to develop selected volatile and semivolatile organic audit
materials and provide them to Federal, State and local agencies or
their contractors for use in performance audits to assess the
accuracy of measurement methods used during hazardous waste trial
burn tests. Research Triangle Institute (RTI), under contract to
USEPA, operates a performance audit program through development of
gas cylinders containing 27 gaseous volatile organic compounds in
five—, six—, seven—, and nine—component mixtures of
parts—per—billion (ppb) levels (7 to 10,000 ppb) and 6 semivolatile
organic compounds are shown in Tables 1 and 2 respectively.
The gaseous organic compounds in Table 1 are purchased in compressed
gas cylinders from commercial suppliers. These cylinders, along
with an appropriate delivery system, are used directly without
dilution In the performance audits. The semivolatile compounds,
shown in Table 2 are spiked onto XAD—2 cartridges at RTI.
Before being used as an audit material, the contents of each
cylinder undergo a series of analyses at RTI. The cylinder contents
are analyzed after initial receipt from the manufacturer to check
the accuracy of the reported concentration of the test compound.
The cylinder contents are then analyzed periodically over the course
of a year and later on a yearly basis to estimate the compound
stability. Similarly, representative samples of spiked cartridges
are analyzed initially and then periodically to determine the
compound stability. All concentrations are measured by using gas
chromatography (GC) with flame ionization detection (FID) and/or
electron capture detection (ECD).
7—118
-------
TABLE 1. PPB-LEVEL ORGANIC GASES IN REPOSITORY
Group 1:
5 Organics in N 2
Group 2:
9 Organics in N 2
Group 3:
7 Organics in N 2
Group 4:
6 Organics in N 2
Carbon
Trichloroethylene
Vinylidene
Acrylonitrile
tetrachioride
1, 2—Dichioroethane
chloride
1, 3—Butadiene
Chloroform
1,2—Dibromoethane
F—113
Ethylene oxide
Perchioroethylene
F—12
F—114
Methylene
Vinyl chloride
F—li
Acetone
chloride
Benzene
Bromomenthane
Methyl ethyl
ketone
1,1, 1—Trichloroe—
thane
Acetonitrile
l,4—Dioxane
Toluene
Chiorobeuzene
Propylene oxide
Ortho—xylene
Ranges cylinders
Ranges cylinders
Ranges cylinders
Ranges cylinders
currently
currently
currently
currently
available:
available:
available:
available:
7—90 ppb
7—90 ppb
7—90 ppb
7—90 ppb
90—430 ppb
90—430 ppb
90—430 ppb
430—10,000 ppb
430—10,000 ppb
TABLE 2. SEMIVOLATILE ORGANIC COMPOUNDS IN THE
AUDIT REPOSITORY
o chlorobeuzene
o Toluene
o O—Xylene
o Pyridine
o 1,1, 2,2,—Tetrachioroethane
o Nitrobenzene
7—119
-------
EXPERIMENTAL METHODS
Instrumentation
Analysis of Audit Cylinder Cases : Analysis of ppb—level audit
mixtures are performed on a Hewlett—Packard 5880A gas chromatograph
(CC) equipped with flame ionization and electron capture detectors.
The electron capture detector (ECD) is used principally for
measurement of all the halogenated organic compounds except vinyl
chloride. A fixed volume of each gas in a sample loop is injected
onto the CC column through a six—port gas sampling valve mounted
near the injection port. After separation of the organic compounds
from nitrogen, a GC response is produced and electronically
integrated. CC conditions for all analyses of ppb—level audit gases
have been previously described(l).
Preparation and Analysis of Spiked Audit Cartridges : The
preparation of the cartridges, 50 g of XAD—2 resin is loaded into a
glass source assesament sampling system organic vapor trap and the
trap is connected to a Nutech Model 201 Method 5 sampler and room
air was sampled through the resin at a flow rate of approximately 1
ft. 3 fmin. During the approximately 30 mi i i. that room air is being
drawn through the resin, a known amount of semivolatile standard mix
is injected onto the resin via the flash evaporator. As nitrogen
flows through it, the flash evaporator is gradually heated to 250°C,
and the volatized compounds are absorbed on the resin.
For analysis of the cartridges, the spiked XAD—2 cartridges are
soxhiet extracted with 300 mL of methylene chloride over
approximately 20 h. A Kuderna—Danish evaporator is then used to
concentrate the extract to 10 niL. The concentrated extracts are
analyzed on a Perkin—Elmer Model 3920B CC with a flame ionization
detector.
Standardization and Measurement
Calibration of the CC involves measurement of known concentrations
of gases in nitrogen. Permeation tubes purchased from commercial
suppliers or the primary standards that are prepared and analyzed by
the National Bureau of Standards (NBS) are generally used as
calibration standards for all the ppb—level inulticomponent organic
mixtures.
The calibration standard mix for the semivolattle compounds is
prepared by weighing known amounts of neat liquids into a volumetric
flask which is then filled with methylene chloride to the mark.
7—120
-------
A multipoint standard curve is generated for those compounds that
exhibit a linear FID response. For those which show evidence of
non—linearity, at least two calibration standards are prepared each
time a sample is analyzed. Both standards are prepared within 20
percent of the expected sample concentration.
RESULTS AND DISCUSSION
Performance Audits
As stated previously, the EPA (through RTI) supplies audit materials
for performance audits upon request from Federal, State or local
agencies of their contractors. The contractors must, however, be
performing hazardous waste trial burn tests on behalf of EPA or one
of these public agencies to qualify for the performance audit. The
performance audit should be conducted simultaneously with the actual
planned test. Performance audits prior to trial burn tests to asses
the proficiency of the measurement system (including the sampling
and analytical personnel) are also encouraged.
When a request is received for an audit of the volatile organic
compounds, the cylinder pressures are measured, and the cylinder and
glass manifold delivery system (when volatile organic sampling train
is used for sampling) are shipped by overland carrier to the audit
site. General instructions for conducting the audit are included
with the audit materials. The audit results are reported to the
agency (Federal, State, or local) coordinator requesting the audit.
There is no charge to the user except the cost of returning the
audit cylinders/audit cartridges.
To date (May 1987), 130 performance audits have been initiated with
ppb-level audit gases. Of these 130 audits, 65 audits have been
conducted to assess the accuracy of measurement methods (both
sampling and analysis procedures) used during hazardous waste trial
burn tests. The results obtained for some of the performance audits
are shown in Table 3. The results of audits for the VOST methods
are usually within the ± 50 percent limit stated in the VOST
protocol(2). A limited number of audits were conducted with the
spiked XAD—2 cartridges. An example of audit results obtained with
the spiked cartridges Is shown in Table 4. Results were generally
within ± 35 percent of the audit concentrations measured by RTI.
The principal results of these audits have been to ensure that
analyses are performed properly, to detect problems that could be
corrected, and to document the accuracy of the measurement systems
used during the hazardous waste trial burn te8ts.
7—121
-------
Table 3. Example of Performance Audit Results
(RCRA Testing)
Auditee
Rh-measured results
Group Audit Measured Audit concentration accuracy
No. No. system audited material (ppb) (%)a
Group 1 5 VOST collection; Carbon tetrachioride 11.0 + 0.3 -31
GC/MS analysis Chloroform 45.0 + 0.9 +220
Vinyl chloride 20.5 T 0.7 +7
Group 2 37 yOST collection; Acetonitrile 7.9 + 0.6 +380
GC/FID analysis Trichloroethylene 9.0 + 0.4 +40
Freon-li 9.4 T 0.3 +3
1,2-Dibromoethane 10.2 T 0.4 +1
1,1,1-Trichioroethane 10.1 + 0.5 +8
Group 3 81 VOST collection; Freon-114 8.5 + 1.3 +131
GC/MS analysis Vinylidene chloride 10.9 1.0 -16
Acetone 27.6 5.7 -18
Toluene 30.8 + 3.6 +78
1,4-Dioxane 28,8 8.9 -79
Chlorobenzene 31.4 3.6 +49
Auditee concentration-RI! concentration
% accuracy = X 100
RI! concentration
Table 4. Example of Audit Results
(Spiked onto XAD-2 Cartridges)
RI! audit Auditee
Sample concentration concentrationa Percent
No. Compound (ug) (ug) accuracyb
1 Toluene 175 121 -31
Chlorobenzene 168 126 -25
Tetrachioroethane 162 124 -24
2 Toluene 172 120 -30
Chlorobenzene 165 126 -24
Tetrachloroethane 159 123 -23
IJncorrected for recovery efficiency.
Auditee concentration-RI! concentration
accuracy = X 100
RI! concentration
7—122
-------
Stability Studies
The data collected over a period of time from the measurement of
cylinder concentrations are used to estimate the stability of the
cylinder gases. Cylinder gas stability data are important for
several reasons. First, the gaseous compounds in the cylinders must
be stable to be considered by EPA as reliable audit material.
Second, if organic gases in the cylinders are stable, other
investigators may more readily use cylinder gases as calibration
standards and/or quality control check samples. Finally, if
cylinder contents are stable, future EPA regulations may include
performance audits as a means of assessing the accuracy of the
measurement systems.
The term ‘stability” as it pertains to organic gaseous compounds, is
defined here as the absence of detectable changes in concentrations
over time for a given cylinder at a specified concentration level.
The cylinder gases are analyzed initially by the manufacturer
facility before shipment to RTI. The gases are then analyzed at RTI
upon receipt from the specialty gas vendor to corroborate the
vendor’s analysis. The gas mixtures are again analyzed at 2 months,
6 months, 12 months, and annually thereafter following the initial
RTI analysis to determine any change in the gas mixtures. The
stability data obtained to date for all the ppb—level organics are
summarized and published in a separate report(l). An example of
stability data for ppb organic gases in compressed gas cylinders is
shown in Table 5. tin examination of the stability data for many of
the organics in the ppb—levei. cylinder gases show that most varied
by less than 10 percent over a two—year period. Ethylene oxide and
propylene oxide at low concentrations (10 ppb—levels) were found to
be unstable and therefore those compounds are not recommended for
audits.
A limited amount of stability data for the XAD—2 spiked cartridges
has been obtained. The data obtained for 6 weeks show that the
seinivolatile compounds spiked onto XAD—2 cartridges were stable over
that period. Further stability studies are in progress.
SWIMARY AND RECOMMENDATION
Compressed gas cylinders containing 21 gaseous volatile organic
compounds at part—per—billion levels and XAD—2 cartridges spiked
with five semivolatile organic mixtures have been used successfully
in audits to assess the accuracy of measurement systems, especially
those used during the hazardous waste trial burn tests. Audits have
not yet been initiated with the six group 4 gaseous volatile organic
compounds (acrylonitrile, 1,3—butadiene, ethylene oxide, propylene
oxide, methylene chloride, and 0—xyleae).
7—123
-------
Table 5. Example of Stability Study
(Group I Compounds)
I Compound
Carbon Perchioro- Vinyl
Activity tetrachioride Chloroform ethylene Benzene chloride
Manufacturer Analysis
Theoretical ppb (9.5) (39.6) (9.4) (18.9) (20.0)
1st Analysis Date 11/17/83 11/17/83 11/17/83 12/06/83 12/06/83
ppb (7.7) (42.1) (9.2) (15.8) (18.3)
2nd Analysis Date 12/16/83 12/16/83 12/16/83 12/15/83 12/15/83
ppb (7.1) (41.2) (7.7) (19.6) (19.2)
RTI Analysis
1st RTI Analysis Date 1/06/84 1/06/84 1/06/84 1/23/84 1/23/84
ppb (9.3+0.3) (40.0+0.8) (9.4+0.2) (19.6+0.6) (20.2+0.6)
2nd RTI Analysis Day* 59 59 59 A9 49
ppb (9.3+0.3) (38.7+0.8) (10.3+0.2) (20.0+0.6) (19.0+0.6)
3rd Rh Analysis Day* 238 238 238 226 226
ppb (9.5+0.3) (40.6+0.9) (9.5+0.2) (19.4+0.6) (20.6+0.7)
4th RTI Analysis Day* 404 404 404 375 375
ppb (9.6+0.3) (37.1+0.8) (9.1+0.2) (19.4+0.6) (20.4+0.7)
5th RTI Analysis Day* 872 872 872 850 850
ppb (10.5+0.3) (40.5+0.9) (11.2+0.3) (21.7+0.7) (19.4+0.6)
of days after 1st RI! analysis date.
7—124
-------
Stability studies for all 27 volatile organics and 6 semivolatile
organics have been performed to determine the feasibility of using
them as audit materials. Results indicate that all of the organic
compounds tested (except ethylene oxide and propylene oxide at the
10 ppb—level) are stable enough to be used as reliable audit
materials. One hundred thirty performance audits to date (May 1987)
have been performed using ppb—level audit gases and the results have
generally been within ± 50 percent of the audit concentrations
measured by RTI.
It is recommended that a performance audit using these organic
mixtures be conducted during each hazardous waste trial burn test as
a routine quality assurance procedure. Any federal, state and local
agencies and their contractors planning hazardous waste trial burn
test may request a performance audit by contacting Mr. Robert L.
Lampe (for cylinder gases) and Ms. Ellen Streib (for semivolatile
organics) of the EPA, Environmental Monitoring Systems Laboratory
Quality Assurance Division, Research Triangle Park, North Carolina
27711.
ACKNOWLEDGEMENTS
This project was conducted by the Research Triangle Institute,
Research Triangle Park, North Carolina, under Contract Number
68—02---4125 for the Quality Assurance Division, Environmental
Monitoring Systems Laboratory of the U. S. Environmental Protection
Agency. The authors gratefully acknowledge the technical assistance
of Mr. J. Albrltton and R. Wright of RTI. Mention of trade names or
commercial products does not constitute endorsement or
recommendation for use. This paper has been reviewed in accordance
with the U. S. Environmental Protection Agency’s peer review and
administrative review policies and approved for presentation and
publication.
REFERENCES
J. M. Allen, C. K. Sokol, R. K. N. Jayanty, C. E. Decker, and D. J.
Von Lebmden, “Status Report Number 3, Stability of
Parts—Per—Billion Hazardous Organic Cylinder Gases and
Performance Audit Results of Source Test and Ambient Air
Measurement Systems,” EPA Contract No. 68—02—4125, Research
Triangle Institute, Research Triangle Park, NC, December 1986.
E. N. Hansen, “Protocol for the Collection and Analysis of Volatile
POHC’s Using yOST,” EPA—600/8—84—007, U. S. Environmental
Protection Agency, March 1984.
7—125
-------
DESI AND USE OF LABORAWRY QC PROGRAMS W SATISFY DQOS FOR
ENVIRONMENTAL MEASUREMENT SYSTEMS
E. P. Brantly, Environmental Scientist, N. Messner, Research
Environmental Scientist, L. Myers, Research Statistician, B. Price,
Research Statistician, Research Triangle Institute, Research Triangle
Park, North Carolina
ABSTRACT
A protocol has been developed for design of laboratory QC programs for
use in environmental measurement systems whose purpose is to test
compliance with regulatory standards. The data quality objectives
(DQO’s) which are the basis for evaluating the overall measurement
program are bounds on error rates in determining compliance or
noncompliance with the standard. The QC program can be tailored to
the individual laboratory according to its own bias and precision. In
addition to their use for system performance checks, QC data are used
to correct measurements for recovery and to estimate false positive
and false negative rates associated with the compliance/noncompliance
determinations. This process is adaptive: error rate estimates are
periodically updated, and the QC program reevaluated and revised
(intensified or relaxed) as appropriate.
The use of the protocol is envisioned for GC/MS and other analytical
chemistry measurement systems used by 0 5W. The protocol will be
described in sore detail and illustrated in specific contexts.
7—126
-------
RCR7 EXPERIENCE IN SOUTHEAST FLORIDA
Charles Ouseph, CHMM, Florida Department of Environmental Regulation,
West Palm Beach, Florida
ABSTRACT
The paper highlights some of the areas where refinements or changes,
both regulatory and technical, appear needed in the nation’s current
program of managing Hazardous Wastes. Based on our experience in
implementing the federal hazardous waste program at the State level,
some of the difficulties encountered are highlighted in order to
demonstrate the need for refinements or changes in the existing
hazardous waste regulations. Some areas of the hazardous waste
regulations, as they currently exist are likely to be inappropriately
applied and thereby not meeting the intent of the law. Such areas
include test methods, sampling, regulatory interpretation, and
enforcement. Some of the paradoxical situations encountered during the
past five (5) years of field experience in managing the Federal
hazardous waste program in Southeast Florida are illustrated with
actual case histories (without naming the companies).
INTRODUCTION
To ensure proper management of HWs, Congress passed in 1976 the first
landmark piece of federal legislation known as RCRA. It was amended
once in 1980 and recently in November 8, 1984. The 1984 HSWA
significantly expanded the scope and detailed requirements of RCRA with
approximately 87 deadlines (up to November 8, 1992). A number of
“hammer” provisions are included as statutory requirements to go into
effect automatically in case EPA fails to issue the required
regulations by certain dates. Because of the hammer provisions, EPA
often had to move fast to complete the rule making process and to come
out with the regulations that will satisfy the critical concerns of
both the environment and the industry.
The materials presented here are limited to those regulations finalized
by EPA under RCRA Subtitle C: Sections 3001-3019 as codified in Volume
40 of the CFR, Parts 261 through 265. Now that both the regulators and
industry have used these new and old regulations for a period, it is
time to reflect on whether these regulations together accomplish the
congressional intent, or do they need refinements, especially to
address situations perhaps not anticipated when they were written. It
is in this frame of reference that I have attempted to summarize both
my RCRA experience in Southeast Florida and what I learned working with
industry and other real life situations in trying to do the best I can
to meet the intent of the regulations. The views expressed here are
strictly the author’s own and are not necessarily those of the Florida
Department of Environmental Regulation.
7—127
-------
HAZARDOUS WASTE FROM NON-SPECIFIC SOURCES ($ 261.31)
[ FR Vol. 50, December 31, 1985, P. 53315—201
Spent cleaning fluids containing 10% (by volume) of one or more of
the listed solvents (FOOl to F005) have to be managed as a listed
hazardous waste and any spill has to be cleaned up in accordance with
RCRA regulations. But the same cleaning fluid as raw material when
spilled is considered a non—hazardous waste spill, and therefore unless
the spilled material is characteristically hazardous, RCRA regulations
do not apply. Obviously, this is not the intent of the law. Also,
there is no way the inspector can distinguish between the two spills
even with laboratory testing.
Conversely, there is a spent cleaning fluid containing 10% (by
volume) of one or more of the listed solvents FOOl to F005. It is not
a hazardous waste unless shown to be characteristically hazardous. If
there is a spill of this material, the clean—up doesn’t have to meet
RCRA standards, and sometimes no clean-up is required. Yet a spill
involving far less quantity of a listed waste solvent (because of the
application of the mixture rule 261.3(b) (2)) requires extensive clean
up of both soil and affected ground water to background levels (which
often means no detection) or requires delisting ( 260.22). Clearly
such a scenario doesn’t meet the intent of the law. The inspector as
before, has little choice but to accept the operator’s statements.
SATELLITE ACCUMULATION RULE 262.34 (c) )
If a 55 gallon and a one gallon container are full at the same
satellite area, the rule says that the 1 gallon container (representing
the amount in excess of 55 gallons) becomes subject to full regulation
within three days of accumulating the excess amount. However, I feel
that the larger quantity posing the greater threat should be regulated.
In case of accidents involving satellite areas, the rule doesn’t
require notification under 5265.56(d) or (j). The rule also exempts
satellite areas from such things as no smoking signs, spill control
measures, personnel training involving use of emergency equipment as
well as implementation of preparedness and prevention procedures.
I feel that a 55 gallon spill in a satellite area can pose a serious
threat to human health, and therefore personnel should be knowledgeable
in HW handling as well as in proper fire and spill control measures.
One way to accomplish these objectives would be to have the existing
contingency plan include remedial actions to be taken in case of spills
or fires at satellite areas.
In accordance with 262.34(c), all wastes in excess of 55 gallons (or
in excess of 1 quart of acute HW) are to be marked with the date the
7—128
-------
excess amount began accumulating and this excess amount must be removed
within three days. For the inspector, verification of compliance here
is not very practical.
WASTEWATER TREATMENT UNIT EXEMPTION
In accordance with 264.l(g)(6), TSD standards do not apply to
“wastewater treatment units” as defined in § 260.10. To retain this
exempt status, the tank must be part of a wastewater treatment facility
regulated under section 402 (for industrial wastewater discharges) or
307(b) (for POTW discharges) of the Clean Water Act. It must (a) treat
or store an influent hazardous waste or (b) generate, accumulate, treat
or store a hazardous wastewater treatment sludge. Also, any mixture of
sewage and other wastes passing through POTW is not regulated as
hazardous wastes [ 264.4(1)].
An interesting scenario here is that there are privately owned sewage
treatment works operating as efficiently as any POTWs, and yet, these
privately owned works unlike POTWs are not exempt from RCRA.
NOTICE IN THE PROPERTY DEED [ 264.116 & 119]:
EFR Vol. 51, May 2, 1986, P. 16433—341
Prior to the 5/2/86 FR changes (p. 16422-59) regarding closure,
264.120 required a notice in the property deed once on-site
contamination due to past }IW disposal has been established. The law
also provided deletion of such notation once contamination is cleaned—
up.
The regulatory changes (effective 10/29/86) appears to have modified
this process. 264.116 states that a survey plat must be submitted “no
later than the submission of the certification of closure.” “No later
than” implies that the survey plat must be submitted on or before the
submission of certification of closure. It is the owner/operator (ole)
who decides when to file the plat, and if so, the 0/0 could wait until
the last minute before filing the plat with local land use authorities.
Often closures take several months and sometimes years to complete the
closure. During this long interval, potential property buyers cannot
be made aware of hazardous waste contamination if such information is
not in the survey plat and/or notice in the deed.
If the regulatory agency cannot require evidence of significant
contamination entered into the title deed, then how can the government
reasonably assure that potential buyers are notified in advance (and
before a violator “skips town”)?
Subpart G ( 264.119) seems to provide such notices and plats only
after closure and assumes that a disposal facility will always complete
7—129
-------
closure prior to selling the property. While most companies will
complete cleanup, we have encountered some that attempt to evade the
law, delay notifying potential buyers, sell before (or in lieu of)
cleanup, etc.
If a facility is certified closed, chances are that such facilities are
the “good guys” operating under the law, and for the most part,
contaminants at levels of concern may have been already removed. If
that is so, as in the case of clean closure, then why bother to have an
entry in the title deed or plat? Once HW concerns are established, it
is prudent to have entry as soon as possible in the title deed or
property plat for protecting the innocent buyers from the polluters who
are eager to sell and leave town. The whole concept seems to suggest
that perhaps the closure may not be clean and so be correspondingly
forewarned.
EXTRACTION PROCEDURE ( 261 Appendix II )
Based on the prescribed test procedure under RCRA, the EP toxicity
result is expressed in units of mg/i. But actually, the EP leached
quantity is milligrams per 50 grams of test sample. But the results
from contaminated soil, sediment, or sludge sample when analyzed for
total metals or pesticides are customarily expressed as mgs/kg. If the
total numbers are high enough to be of environmental concern, the next
question is “how much of it is leachable”? Therefore, to get a proper
perception of leachability against the background of their
corresponding totals present, the EP test data has to be multiplied by
20. Otherwise the EP test data is a watered down number for
ieachability, and this may give a false sense of security.
Examples :
(a) I have a sludge sample with a total lead content of 100 mg/kg.
If I am told that it is 50% leachable or soluble, it means 50 mg
lead/kg sludge is leachable. Yet the EP number under RCRA will
be approximately 2.5 mg/i and I have observed “informed” people
concluding that 97.5% is not leachable and tied up in the soil by
their cation exchange capacity!
(b) We had one facility with a closed out drainfield where the EP
toxic values for all cored soil samples showed levels at or below
the GW standard for chromium. Yet the on—site GW consistently
showed high levels of chromium (by as much as 100 - 400% above GW
standard) even after a year. The problem had to be corrected by
removing soil from the drainfield.
Another intrinsic problem with the EP test procedure is that it doesn’t
take into account the amount of water or other liquids present in the
test sample. It is suspected that samples collected on behalf of the
7—130
-------
PRP are in such a way to deliberately include a greater proportion of
liquid in sediment or sludge samples in an effort to bring down the EP
values. Even when sludge samples from a wastewater tank or sediments
from a canal bottom are collected and analyzed independently by two
chemists, the EP test data can show significant differences depending
on how the samples were collected. Similarly, when contaminated soil
samples are collected, one during dry season and another immediately
after a heavy rain, the EP test data between the two test samples would
show significant variation.
Another limitation of the procedure is limiting the pH during
extraction to a rigid value of 5 which may be a conservative value for
the EP study most of the time. I have come across at least 3 sites in
Southeast Fidrida where the natural GW pH was observed to be around 4±.
A proper pH for EP study pertaining to those sites should approximate
the natural GW value. But the EP procedure under RCRA doesn’t give
that flexibility to change the prescribed pH from 5 in order to reflect
the non—typical field conditions.
Despite the little problems described above, the EP test procedure is a
very useful tool in evaluating the environmental threats posed by
various hazardous wastes. It is my opinion that the usefulness as well
as precision and accuracy of EP test data for samples (such as sludge
and contaminated soil) can be enhanced by expressing leachability as
mgs/kg of dry solid.
NEW SMALL QUANTITY GENERATOR RULES
[ FR Vol. 51, March 24, 1986, p. 10146-78 )
Based on one year of past experience in implementing the new SQG rules,
following are some thoughts on bottlenecks as I have observed.
(1) MANIFEST EXEMPTION PURSUANT TO RECLAMATION CONTRACT
[ p. 10155—8; 262 Subpart B]:
The “ tolling arrangement ” exemption (e.g., absence of a TSD
“sign—off” manifest copy) does not provide a means for ensuring
the generator that the wastes were actually received at the
recycling facility. The mere presence of a contract does not
guarantee that the wastes will be properly managed. The
transporter/recycler could mis—manage or even not deliver the
waste and the generator would not be aware of the fact. In
addition, the presence of a contract does not necessarily mean
the generator is still using that service. The generator could
also be mis—managing the wastes, but could tell an inspector that
the waste is being handled in accordance with the contract.
Since the generator would not be required to keep even a log
(only the transporter/recycler must keep a log), an inspector
would not be able to verify the pickup of waste and the inspector
7—131
-------
would also be unable to verify that the waste actually reached the
designated facility without inspecting the TSD facility (which is
not feasible for every shipment even if the TSDF is located in the
district, and certainly not very practical if the TSDF is located
out—of—state). I believe that the “tolling” exemption should be
modified, at a minimum , to include a requirement for:
(a) the generator to keep a log showing the same information
required of the transporter’s log; or
(b) the transporter providing the generator a receipt for
having picked up so much of a certain waste.
In suimnary, the “tolling” exemption needs to be improved to eliminate
the possibility of abuses, and to make it easier for an inspector to
verify proper waste management. It should also be noted that
implementation of above suggestion would not place a burden on small
companies, but rather would protect them.
(2) ACCUMULATION LIMIT
[ p. 10161; . ‘ 261.34(d)]
The greatest paradox in implementing the new RCRA program exists here.
For example company “A”, a SQG producing on the average 4 drums a
month, has on site about 25-30 drums for about 200 days, and meets only
very minimal requirements (which doesn’t even include a written
contingency plan) to protect human health and environment. Across the
street company “B” producing on the average 6 drums a month is being
cited for storing without a permit about 6—8 drums for about 100 days,
even though this facility is better operated and meets more regulatory
requirements (including an elaborate and government approved
contingency plan) to protect both the environment and human health.
Obviously company “A” with four times the amount of waste sitting twice
as long than company “B” (coupled with the fact that company “A” meets
only very minimal requirements in terms of personnel training,
contingency plan, etc.) poses much greater risk. It makes the
inspector very uncomfortable to cite violations on company “B” when
company “A” goes free.
The new rules would also be a departure from previous accumulation
strategy in which exceeding a limit would subject the company to the
next higher level of regulation. For example, a SQG would be subjected
first to generator requirements, and then only to possible TSDF
requirements. The new rules would subject an SQG to a storage permit
if > 6000 kg or 180/270 days is exceeded, without going through a
generator phase. CESQG who accumulates > 100 kg HW, instead of meeting
the SQG requirements, is simply being exempted from regulation until
the 1000 kg limit is exceeded. In the case of acute HW, it is either a
CESQG (< 1 kg/mo.) or a generator ( > 1 kg/mo.), but can never be SQG.
In other words, an incremental, step—wise level of regulation is not
maintained, and the absence of this simple logic makes the inspector’s
job more difficult.
7—132
-------
It should be noted that it is the 6000 kg quantity , (and not the
180/270 day time) that is crucial in considering risk. On p. 10161,
EPA states that “the Agency could see no substantive difference in
potential risk” when comparing 180 versus 270 days. If so, some of the
small companies are losing out on the regulatory breaks. The
legislative intent of HSWA clearly was to give certain regulatory
relief to small companys for their survival while enhancing the
protection of the environment. EPA appears to identify small companies
as those generating < 1000 kgs/month and this alone can hardly be a
criteria to define “small companys”. Some small companies by the very
nature of business they are in (such as electroplating) historically
produce a larger quantity of HW, but may have only very few employees
and limited financial resources. Whether these small companies, though
generating > 1000 kgs/month, be allowed to store the same amount of HWs
for the same duration as with SQGs is worth considering (especially
when these small companies satisfy a higher level of regulatory
requirements for the protection of human health and environment).
( 3) MULTIPLE WASTE STREAMS (p. 10161)
A statement in the third column of p. 10161 states “...generators that
have multiple waste streams which are managed at different facilities
may actually be subject to different accumulation time limitations for
the different waste streams.” Such a situation would make an
inspector’s job more difficult to verify and enforce compliance. If
complex regulations and even more complex exemptions make an
inspector’s job more difficult, then the vast universe of companies may
be even more confused. The inspector would have to:
(1) determine the nature and quantity of waste (taking into
account the numerous exclusions for the purpose of quantity
determination as described on p. 10151—3),
(2) determine the accumulation practices and times of each
stream,
(3) verify the designated facility of each stream,
(4) verify the designated facility type and its distance from
the SQG facility ( > or < 200 miles).
I believe that it is a nightmare to the Inspector and possibly to some
facilities also.
(4) CONTINGENCY PLAN (p. 10164)
If a SQG is big enough to accumulate 6000 kgs of HWs, then it must be
large enough to meet the full generator requirements including full
CP. EPA states that the Agency “was careful to modify the standards
7—133
-------
only where administrative requirements not essential to the
substantive functioning of the standards were irivolved,!i and the EPA’s
final rules are “sufficient to protect human health and the
environment.” These rules seem to imply that a SQG accumulating 6000
kg does not pose a sufticient risk to require a full CP, but a
generator accumulating a much lesser amount poses enough risk to
require a full plan! It is like saying that 6000 kg HW is less
dangerous than 1000 kg HW.
A full cP is more than just “administrative” in nature; it contains
valuable information for use in an actual emergency. In other words,
by not meeting all or nearly all of the substantive requirements of a
full CP, than the “substantive functioning of the standards” will not
be met. This part of the rules is an example where EPA has tried to
balance the various pressures (Congress, Industry, etc.). However, I
believe that EPA’s logic in this instance seems inappropriate. One way
to bring a measure of logic into the implementation and continued
compliance under the new RCRA program may be to increase the
generators’ ability to store the same quantity of waste for the same
period as with SQGs (or vice versa).
The present watered down CP required of SQG5 may not be of much use in
case of an actual emergency. It doesn’t have even an evacuation plan!
It is my experience that the vast majority of SQGs and even some CESQGs
have voluntarily developed CPs, and submitted them for the department’s
review and comments. Developing a CP often seems more difficult than
it really is. Even if difficult, it is basically a one—time effort and
that it can help save lives in an emergency. My efforts have made the
preparation of C? simpler by developing a concise, easy to follow
guidance which the inspectors hand out routinely to facilities. Most
facilities including many SQGS and CESQGs, were able to develop a good
contingency plan for use by their employees.
(5) PERSONNEL TRAINING (p. 10164)
This rule merely states that the SQG “must ensure that all employees
are thoroughly familiar with proper waste handling and emergency
procedures.” No other criteria or records as in 265.16 are required.
How will the inspector verify compliance of such a vague rule? How
will SQGs demonstrate compliance without records? The personnel
training requirements for SQGs need to be more specific and
enforceable, at least to the same level required under the “Right—to--
Know” rule.
(6) CONTAINERS (p. 10165)
EPA has exempted SQGs producing flammable wastes (F’.P. 140°F) from
the 50—foot buffer zone requirement, until EPA promulgates the final
buffer zone rule. EPA provided this exemption because the rule “would
put many small businesses in a situation in which it would be
7—134
-------
impossible to comply.” This is also sometimes true of generators (who
may also be small companies but happen to generate 100 kg/month). If
SQGs are to be exempt (even with 6000 kys), then so should generators.
But since EPA has not exempted generators, if an SQG had enough space
to store 6000 kg with adequate aisle space, then the facility is
probably large enough to meet the 50—foot rule. In addition, NFPA
rules will require compliance, so the SQG will probably not avoid the
50—foot rule.
In the interim, I believe in having the 50 ft. rule that makes both
engineering and regulatory sense, while also ensuring fairness to both
SQGs and generators; and also for reducing the risk of fire and thereby
help protect lives at and near facilities with SQG status who may store
up to 6000 kgs with only a minimal contingency plan. A practical way
to reconcile the interests and concerns of various segments of the
society may be for EPA or the authorized state to have the option (on a
case by case basis) of giving a variance from the 50 ft. rule based on
(1) the recommendation of the State Fire Marshal, (2) consideration of
improved design of 11W storage area (such as erecting a fire wall),
and/or (3) stipulating the method of storage (including maximum
quantity stored) until the SQGs (and for that matter large quantity
generators also) meet the minimum protective distance for storing
flammable HWs.
There appears to be a belief that bigger the size of a company, the
larger the quantity of HW generated and vice versa. The truth is that
a giant corporation may produce very little of HWs while a small
company may produce large quantities of HWs. The quantity of HW
generated is basically -a function of the nature of the business a
particular company happens to be in, and of course, it also depends on
the kind of industrial process employed. In the eagerness to provide
regulatory relief to SQGs, it appears that protection of human health
may have been compromised.
(7) CESQGs ( (l00 kg/month ) [ 261.5(g) (2)]
An inconsistency exists in this section. The CESQG will be subject to
increased regulation, only if he accumulates ) 1000 kg. But, a SQG has
to meet increased regulation generating and storing ). 100 kg. In other
words, the SQG could have less waste on site, but be subject to more
regulation. This would be analogous to not requiring SQGs under
existing rules to meet generators rules for accumulating ‘>1000 kg.
(8) EPISODIC GENERATORS [ p. 10151, 10153—4]
In spite of the best efforts to help CESQGs and SQG5, all of the
regulatory breaks disappear when they become “episodic generators” at
any time (e.g., emptying a tank, clean up of 11W spill) and they then
must meet full generator standards as if they are large quantity
generators. For instance, once a full contingency plan is developed,
7—135
-------
it does not appear to matter whether the SQGs legally require a full CP
later on.
I have also observed that there is a tendency for many SQGs not to
disclose episodic generation, because the inspector is told that there
was never any need to empty the tank (such as the rinse tank, wash
tank, etc.) even though the same tank may have been in service for the
past several years. When the SQG is permitted to store up to 30 drums
on—site (with its numerous excTThsions in counting HW quantity) for as
long as 270 days with practically no record keeping, basically the
inspector has to accept what the operator tells him.
(9) REGULATORY IMPACT (p. 10172)
EPA states that the cost of these new rules is not “major” (i.e. < $100
million per year). Since EPAs total cost estimate is only $58.9
million, per year average cost estimates per firm are low, and
therefore annual recurring cost is very low. Hence, there is
considerable leeway to fine tune or modify the current regulations with
a view to eliminate some of the obvious inconsistencies existing on
many CESQGs vs. SQG5 vs. generators requirements, all without reaching
the “major” category. Certain obvious paradoxes can be eliminated by
taking an approach based on incremental, step-wise level of regulation
with each change in the level of generator category, and that will make
the inspector’s job easier, and will also reduce confusion and thereby
increase compliance in the regulated community.
(G) PESTICIDE SPILL
A pesticide drum accidentally spills in a golf course grounds. Since
this pesticide is referenced in . 261.33, excavation had to be
conducted till all of the spilled material was removed. The
contaminated soil resulting from clean—up is a listed hazardous waste
and therefore requires manifesting to a TSD facility. While the
excavated material was waiting to be manifested, the same spilled
material was being applied by a licensed pesticide applicator to the
turf grass for the control of insects.
In the above scenario, I believe that it is sensible to have the
excavated soil applied to the turf grass with the application rate
adjusted to yield the proper pesticide dose rate per acre. However,
whether such application constitutes disposal or beneficial use appears
uncertain under the present regulations.
Also, in the case of a spill as above, the clean-up has to be such
that none would be detected or be at the background level in the
remaining soil after excavation. Any level of a listed commercial
chemical in soil at the spill site will remain as a listed HW, and it
can only be re—classified as non-hazardous through the delisting
7—136
-------
procedure (260.22). However, if the residual soil after excavation
shows a level at or below the drinking water standard (or other
conservative standards such as G’J standard, one in a million cancer
risk, etc.), it may be unnecessary to excavate and manifest huge
quantities of soil in order to remove the last ppb or go through the
time consuming resource intensive delisting procedure.
(H) HAZARDOUS WASTE DISPOSAL IN GARBAGE DUMPSTERS BY CESQG5
Is it legal? Yes. Proper? No. The current HW regulations do not
prohibit CESQGS from disposing their HWs in garbage dumpsters. In fact
261.5 allows disposal of CESQG wastes at facilities permitted by the
state to accept municipal or industrial waste. Since permitted
landfills are typically authorized to accept municipal and industrial
wastes, a CESQG sending its hazardous waste to such a landfill is not
in violation of 261.5, even though such practices clearly pose great
risk to both human life and environment. Therefore, rule making by EPA
is needed to plug this regulatory loophole. Given below are three case
histories to highlight my concerns:
(a) A testing laboratory has placed ignitable, toxic, and
possibly reactive chemicals in a trash container (along with
other regular trash such as waste paper) destined for
disposal in a local landfill. Local fire department brought
this improper (if not illegal) HW disposal practice to the
attention of the State environmental officials. The
laboratory analysis of a discarded foam material found in the
trash container showed 13% (by weight) toluene. Also, the
adjoining tenants in the industrial plaza complained about
severe odor problem to local fire and county health
departments. Fortunately the problem was amicably resolved.
Had this practice continued, it is probably a matter of time
before a fire or explosion occurs, and if that happens
especially in a plaza, there could be significant loss of
life and property.
(b) A major department store in a shopping center dumped
significant amounts of discarded pool chemicals (hypochlorite
and inuriatic acid) into their dumpster. Fortunately a
passerby soon retrieved most of these chemicals for use in
his pool. In fact, the same person as a concerned citizen
alerted the incident to regulatory agency personnel. if it
were not for the quick retrieval of the chemicals, a fire may
have resulted or toxic chlorine gas could have been generated
if the hypochlorite and acid mixed. It also poses a health
threat to garbage collectors and landfill operators. It also
increases the potential for spontaneous fire to both garbage
truck and the landfill.
7—137
-------
(c) Recently, the Department staff responded to fire in a garbaqe
truck carrying trash picked—up mostly from small industrial
plazas with some residential garbage. When smoke profusely
started coming out of the truck, the driver had the good sense to
empty the truck right on the street, thereby avoiding a
potentially dangerous explosion. The trash spontaneously ignited
in the street with flames reaching about 100 ft. high.
Fortunately, nobody was hurt (this time). There was a strong
solvent odor area wide and many partially burned 5 gallon and 1
gallon solvent containers were retrieved. Analysis indicated a
number of hazardous chemicals (mostly flammable solvents).
Despite the best efforts by the landfill operators to prevent HWs from
reaching their landfills, the dumping of liWs by CESQGs still continues.
Even where such practices pose a great risk to human health and
environment (e.g., hazards in the landfill and prior to being delivered
to landfill, potential damage to landfill liners that protect GW,
etc.), the regulators cannot cite a violation under RCRA. Most CESQGs
are short on floor space, generally located in populated areas, and
operate on shoe—strings. Until HW collection centers are established
within reasonable distances and where CESQGs (and even home owners) can
take their hazardous wastes at an affordable rate, the cumulative
effect of all these HWs ending up in landfills across the nation will
be felt increasingly with time. We have the option to pay now or pay
dearly later on for both the proper closure of landfills as well as for
the clean up of groundwater.
(I) CONCLUSION
During the past five years, I have been directly involved in the field
implementation of federal HW regulations at the state level in
Southeast Florida involving compliance inspections, permitting,
enforcement, emergency response operations as well as in RCRA and
CERCLA clean—ups. I have come across many situations where a strict
application of the HW regulations may not be appropriate. For example,
if somebody spills a bucket of listed solvent in the Potomac River,
will any of us dare to say that the Potomac River is all HW, even
though technically that may be correct because of the mixture rule
26l.3(b)(2), and furthermore, would anybody go to the extent of
expecting a closure under 264 Subpart G. The moral of the story is
that having a good dose of common sense, pragmatism, and even more
important keeping in mind the intent of the law both by the regulators
and the regulated will go a long way in making the new RCRA program a
success, despite all the growing pains.
I could point out only a few problem areas within a short period of
time. Part of the problem is that EPA was and still is under tight
time constraints to meet the various deadlines of HSWA for producing
regulations. Therefore, EPA has to rush through rule making process
without the needed time to think through the problem and to broaden
7—138
-------
the public’s participation. I believe that the rule making process as
well as the fine tuning of existing regulations could be improved by
the following steps:
(a) Increased participation of field personnel (engineers,
scientists, and technicians) from both private sector
(including consultants) and regulatory agencies in the early
stages of writing the regulation.
(b) Be prepared to modify regulations when its application in a
particular field situation doesn’t make sense or does not
meet the intent of law.
Cc) All regulations (or at least the significant and perceived
controversial regulations) when introduced for the first time
may be interim regulations. Use it for a period of time (say
two or three years), and then in the light of that experience
and comments received, incorporate any changes if required,
and then adopt the interim as final regulations.
Cd) An incremental, step—wise level of regulation is to be
applied depending on the maximum amount of HW generated per
month or the maximum quantity stored at the facility premises
anytime. In other words, the regulatory requirements should
increase (from CESQG to SQG to generator and TSDF) in
proportion to the increase in potential risk.
(e) Regulations have to be written in plain English . For clarity
of meaning, avoid very long sentences. Use of these simple
concepts could reduce confusion at all levels and help
increase regulatory compliance. It would also reduce the
need for interpretative and guidance memos both from the
federal and state environmental officials.
In the final analysis, the regulators, the private sector, and the
public at large have to work together and certain compromises/trade-
off s have to be accepted on the basis of risk—reward ratio. Until
realistic options are provided and the public gets educated,
regulations alone will not solve our 11W problem. Ultimately everybody
has to ask and answer this one important question. “What kind of
quality of life do I want and at what price?”
ACKNOWLEDGMENTS
The author acknowledges the help of DEE staff in the preparation of
this paper, particularly Chris Johns with extensive file research,
Donald White for his thoughtful review, Marianna Smith for the
excellent word processing, and Scott Benyon for his encouragement and
financial support to present this work in the EPA Third Annual
Symposium on “Solid Waste Testing and Quality Assurance,” July 13—17,
1987 at Washington, D.C.
7—139
-------
REMEDIAL INVESTIGATIONS GUIDANCE STRATEGY
Linda S. Grayson, Christine H. Andreas, Stephen A. Borgianini,
Charles Elmendorf, New Jersey Department of Environmental
Protection, Division of Hazardous Site Mitigation, Trenton,
New Jersey
ABSTRACT
As an alternative to the conventional approach of initiating a
Remedial Investigation based on limited Information gathered from a
site investigation, the Division of Hazardous Site Mitigation has
developed a Remedial Investigation Guidance Strategy (RIGS). The
RIGS calls for a phased approach that will allow a progressive site
assessment and selection of the best possible remediation scheme.
Current RI’s are often based on a limited number of samples
collected during a site investigation. Instead of focusing on
specific site conditions and contaminants these RI’s are broad in
scope. Samples collected often generate data of little real value
in making decisions regarding remediation. One cause of this
phenomena is that the environmental systems affected are not
adequately defined. Also, conventional parameters (pp + 40, HSL)
are Included for analysis, though analysis for only site—specific,
clean—up driving contaminants would be more valuable and more
economical. A more focused, phased approach will allow for the
generation of more meaningful, useful data as a means to the end of
choosing a design for clean—up.
The Phased RI must be built around information collected during the
Site Investigation and the Pre—RI activities. The Pre—RI, an
es8ential preliminary study that will determine the future course of
the remedial, process, initiates the investigation by evaluating the
potential for contaminant migration.
Phase I of the RI will be a systems defining phase. In order to
assess and predict the behavior and movement of contaminants at a
site it is critical to explore the particular environment setting of
concern. Phase I Includes: topographic, geotechnical, radiation
and biological systems studies, and detailed studies of the site’s
physical feature8 (i.e. air, surface water/sediment, subsurface
elements including soil gas and groundwater). The emphasis will be
on Identifying site characteristics such as groundwater flow and
geological features. By installing two inch PVC wells,
investigating groundwater flow, performing gamma logging and
analyzing samples for surrogate parameters such as TOX and TOC, a
broad overview of the site can be achieved.
Phase II of the RI will be similar to the current approach but will
have the advantage of information gathered in the Pre—RI and,
7—141
-------
Phase I. Stainless steel wells can be logically placed based on
Phase I findings and their numbers aad the expense limited.
Similarly, the locations of the collection of other samples can be
logically chosen. Analysis should be for extended parameters (pp +
40, HSL) to Investigate the full extent of the problems on site.
The purpose of measurements made during Phase II is to determine the
variability, distribution and concentration range of contaminants.
Based on this information, monitoring will continue in Phase III.
Additional wells, borings and samples will be placed as needed, but
analysis will be limited to the parameters identified as
contaminants in Phase II. These are the contaminants that will
drive the clean—up, influence the decision £ or a remedial design,
and determine the scope of future site activities.
INTRODUCTION
The Remedial Inveatigation/Feasibility Study (RI/FS) was established
in 1980 as the means to achieving the end goal of choosing
cost—effective, health protective, clean—up designs for CERCLA
sites. In its Infancy, it was thought that the Remedial
Investigation (RI) would be of short duration, and relatively
inexpensive. However, the data generated from RIs of this scope was
inadequate to characterize the site. Consequently, RIs evolved into
lengthy and costly studies in an effort to gain enough Information
about the site to promote a responsible and appropriate clean—up
choice. Though additional time and resources have been allotted in
attempts to assure complete assessment of the site, RIs still are
of ten not Implemented as efficiently as they could be.
The current RI/FS approach is initiated with limited background
Information. The New Jersey Department Of Environmental Protection
(NJDEP) is recognizing that additional information gathered upfront
could improve the efficiency of the RI/FS by reducing the associated
costs and shortening the schedule. Often RIs are initiated based
only on the data generated from the CERCLA funded Site Investigation
(SI) and the Hazard Ranking System (HRS). In these CERCLA
activities the focus Is on source identification. The site’s
physical features and environmental setting are not adequately
addressed at this stage. Additionally, the RIs are broad in scope
rather than being focused on a particular contaminant, matrix or
area of concern. Quite a bit of time and resources are being
expended In the RI stage of the remedial project to make up for this
deficiency in site specific information. Once obtained, this
specific information may lead to changes in the RI scope in
mid—course, since the scope was based on previously available
limited or faulty information.
The guidance that is available on how to perform an RI/FS specifies
what should be included In the investigation but does not detail the
7—142
-------
RI process. Ideally, certain measurements should be performed first
In the study to allo i a characterization of the site’s physical
systems and migratory pathways. Then, with this information known,
samples to define the problem ‘an be collected in the appropriate
areas. Recently some investigators have recognized the logic of the
sequential gathering of data and organized their work plans so that
they conform with this procedure. A comprehensive guidance strategy
that defines a more systematic, phased RI/FS is needed.
The Remedial Investigation Guidance Strategy (RIGS) outlined here is
intended as an alternative to the current, accepted RI/FS approach.
RIGS promotes the generation of site specific information before the
RI begins, then allows a focused, phased approach once the
particular site’s features and problems are identified. In this
way, data gathered will prove more useful in narrowing down remedial
alternatives and ultimately in making a final decision regarding a
design for clean—up.
RIGS is applicable to all types of hazardous waste sites, whether
privately or publicly funded. By detailing a focused mechanism of
site study, RIGS allows a cost effective assessment of the site
conditions. Of concern to government agencies, RIGS assures
accountable spendIng of public money, reinforces decisions made in
the public’s best interest and, most Importantly, builds in
mechanisms for maximum protection of the public’s health.
OVERVIEW OF RIGS
RIGS consists of three phases, each of which play a critical role in
the overall strategy. During Phase I Information about the site’s
physical systems is gathered. Measurements made in Phase II are
intended to define the variability, distribution and concentration
range of contaminants present. Phase III may be used to perform
limited confirmatory sampling and treatability studies. However,
before the RI can begin its scope must be identified through the use
of a Pre-RI. The purpose of a Pre—RI is to gather site specific
information that will dictate the site’s fate in terms of what kind
of remedial investigation will be performed, if any. Figure 1 shows
how information gathered in the Pre—RI is used.
Sampling performed during the Pre—RI is intended to establish a
baseline of data for future monitoring of the site’s migratory
pathways. Possible receptor populations and environments are
considered. Through an Indepth file search and on—site
investigation, each potential hazard that the site poses is
prioritized and samples are collected accordingly. An approximate
time frame of six (6) months is anticipated for completion of the
Pre—RI. All the data gathered during the Pre—RI is compiled and
included in a Pre—RI Final Report. This Final Report will include a
recommendation of one of the four options shown In Figure 1.
7—143
-------
Figure 1 :
PRE—Rl and RIGS
Pre—RI
1. File search
2. Field investigation
Pre — RI Final Report
recommending one of these options:
Pre—RI indicates need
for immediate remedial
measures to mitigate
threat to public health
and environment
immediate
final
remedial
measure
immediate
interim
remedial
measure
focused,
phased
remedial
investigation
Pre—RI indicates
prob’em at site
is wide—spread,
affecting more
than one matrix
or area.
full—scale,
phased
remedial
investigation
Pre—RI indicates
no environmental
problem exists
on site.
Pre—Rl indicates
problem at site
is limited to one
matrix or area.
7—144
-------
If the Pre—RI reveals that no environmental problem exists on—site a
reconunendation of no—action is submitted. Any concern that caused
the site to be considered in the first place must be addressed
before a no—action alternative to a remedial investigation is
accepted. The Pre—RI report must detail the specific results each
step of the Pre—RI yielded.
Recommendations for remedial measures, including interim or final
actions, will be made if an immediate action is necessary to
mitigate a threat to public health or the environment. An immediate
interim remedial measure, such as fence installation, lagoon
stabilization or the removal of leaking drums, will help to secure
and preserve site conditions until a remedial investigation can
begin. By proceeding with an immediate final remedial measure, such
as the removal of intact drums or waste piles and limited sampling
of associated soil/water, the discrete hazards posed by the site are
addressed without the need for an extensive site study.
In some instances, the 8ite specific problem will be limited, by
either matrix, area or some other factor. Here, a focused, phased
remedial Investigation will be recommended in the Pre—RI report. A
focused RI may be recommended, for example, at a site where the
contamination is limited to groundwater, or only In a part of the
site where underground tanks are found.
Finally, the Pre—RI will allow successful identification of those
sites that require the full—scale application of a phased RI, sties
where environmental problems affect more than one matrix and/or are
spread throughout the site.
In addition to determining the course of remediation the Pre—RI
affords several advantages. The Pre—RI can preclude time delays.
The document “Guidance on Remedial Investigations Under CERCLA”
(EPA, 1985) points out that, under the current system, an RI does
not have to proceed to completion but may terminate at any level
provided that sufficient data has been obtained for selection of an
alternative. This provision is an effort to preserve the necessity
of timely action. However, through the use of a Pre—RI, a site with
limited problems may never even get to the full scale RI phase and
thus will not be subject to the delays a8sociated with project
bidding and contractor mobilization.
Along with saving time, ultimately the Pre—RI can save money. By
providing the necessary up front information the Pre—R1 allows a
scope of work to be formulated that is appropriate for the needs of
the site. In doing so, changes in scope once the project is
underway are less likely. So, costs associated with this
re—grouping are minimized, as are the costs of filling data gaps.
Finally, the Pre—RI study provides a baseline for consistency.
Assessing all sites in the same manner assures that each will be
7—145
-------
dealt with in the appropriate way. Each site will be Investigated
to the same degree before the RI Is initiated to decide on the need
for an RI and what the RI’s scope should be.
Once it has been determined that a site requires a remedial
investigation the phased approach can begin. (The scope of a
full—scale, phased Ri Is detailed here. For a focused RI, the
phased approach would be the same but the investigation is limited
to the matrix or area of concern.)
PHASE I
The purpose of the measurements made during Phase I of the RI is to
define, specifically, the site’s physical characteristics and
Identify sensitive environments. It Is necessary to explore the
particular environmental setting of concern in order to assess and
predict the behavior, movement and fate of contaminants. With this
information, an evaluation can be performed of the system’s
potential to allow contaminant migration, its degree and direction.
By defining the site’s physical systems early in the investigation,
more costly measurements (e.g., samples collected for Priority
Pollutant + 40 analysis), taken later in the Investigation, can be
collected in the most logical locations.
Phase I includes various physical studies. A topographic study will
be performed to provide a detailed map which will be used as a
reference for all subsequent operations at the site. A grid will be
established, if it is not still available from the Pre—RI field
investigation, to facilitate field measurements. Along this grid a
thorough radiation survey will be performed, If the need is
indicated by Pre—RI results. Similarly, soil gas measurements will
be taken on a grid to help establish the existence of a contaminant
plume or surficial soil contamination.
Geophysical techniques will be utilized to locate buried objects,
help define horizontal extent of plumes, and aid In determining
geologic etrategraphy. By Installing two (2) inch PVC piezometers
during Phase I more Information about the site’s geology can be
determined through borehole gamma logging.
These piezo.eters can be used to determine groundwater level, flow
and general quality. Analysis of samples collected from these
piezometers will be for surrogate parameters (I.e., TOC, TOX, COD,
TDS, specific conductance, pH, ..) to allow a general Indication of
water quality to be ascertained.
By looking at the biological systems within and adjacent to the site
area during Phase I, identification of critical, sensitive
environments can be achieved upfront and a determination of a need
for immediate action can be made. Also, conclusions regarding the
7—146
-------
extent of contamination can be drawn based on the patterns of
disruption that may be present.
The studies just outlined represent examples of Phase I measurements
that focus on defining the site’s physical characteristics. To date
a generic scope of work for Phase I has been prepared to offer
guidance and define the specifications required to meet the
objectives of each study. The information gathered in Phase I is
compiled and a scope of work for Phase II is prepared based on Phase
I and the Pre—RI. With this more complete picture of the site
specific setting and contaminant problem the scope of Phase II can
be focused and detailed.
PHASE II
The purpose of measurements made during Phase II is to determine the
variability, distribution and concentration range of contaminants.
Based on the findings of the Pre—RI, and utilizing data gathered
during Phase I, the scope of work drawn up for Phase II will be very
site specific. Stainless steel wells can be logically placed based
on soil gas data, gamma logging data, geotechnical data, and ground
water quality data from the piezometers. The number of wells
needed, and the associated expense, can be limited by the benefit of
more up—front information. Similarly, the location8 of other
sampling points can be more efficiently chosen.
In some ways Phase II of RIGS is similar to the current approach in
remedial Investigations. The goal of Phase II is waste type,
source, and concentration range definition, and determination of the
areal extent of contamInatIon. The mechanism for acquiring such
information is the collection of environmental and waste samples,
and the analysis of those samples for an extended range of
parameters (e.g. Priority Pollutants + 40, Target Compound List +
30). The quality of this data is critical and is assured through
the requirement of EPA Contract Laboratory Program (CLP) laboratory
deliverables. However, RIGS differs from the current approach in
that the use of the data is not the same for both. Data collected
in Phase II of RIGS will be subject to methods of data reduction.
This factor must be considered during sampling plan development for
Phase II of RIGS.
The Phase II investigation is intensive in nature to assure that any
and all contaminants are identified, their concentration range and
distribution known. Consequently, large amounts of data will be
generated. This data must be reduced using a standardized,
statistically — based method that takes Into account the
complexities Inherent to environmental data. Through data
manipulation, contaminant prioritizatlon can be achieved, allowing
the determination of limited parameters for the next phase of
sampling and/or to facilitate cleanup design. One mechanism of data
7—147
-------
reduction includes ranking the contaminants according to the number
of times each is found. Each contaminant’s contribution to the
total problem is determined by calculating the percent occurrence of
each. Ey cumulative addition of these percentages new- values are
given that represent the additive problem posed by the various
contaminants. Once a reinediation level has been determined (e.g.
remedlate to 90% clean), based on associated risks and costs, these
percentages will be useful to show the means of achieving the set
goal. Therefore, in preparing the Phase II sampling and analysis
plan, familiarity with the data reduction mechanism is essential. A
basic overview of the spatial relationship of non—parametric
variables, as presented in the RIGS guidance document, will be
helpful in planning Phase II activities. In addition to data
reduction, cartographic representation should be achieved by use of
computer mapping systems that employ mechanisms for accurate
depiction of spatially correlated variables (e.g. kriging).
Three—dimensional, mapping can be used to give an overview of site
trends, while contour mapping, with confidence limits defined, can
be valuable in showing clean and dirty areas. Such mechanisms can
be invaluable when utilized in making remedial decisions as they
allow an objective asseBBIIIeflt based in science. When the assessment
is based on a decision making model the process ultimately makes the
decision rather than individuals. Phase II will permit a
determination of the contaminant(s) of concern and an accurate
depiction of the areal extent of the problem, and allow
recommendations to be made for the limited Phase III scope of work.
PIIASE III
Activities to be performed in Phase III will be determined jointly
by Phase II results and the needs of the ongoing feasibility study.
Phase II may dictate that additional sampling and analysis for
extended parameters, with CLP reporting, is required to fill data
gaps. In addition, sampling and analysis only for the
contaminant(s) of concern, which will drive the cleanup, may be
needed to delineate a clean/dirty line.
To supplement the ongoing Feasibility Study, more sampling may be
required along with beach and pilot scale treatability studies.
Sampling and analysis and treatability studies performed in Phase
III are intended to fill data gaps, confirm conclusions regarding
the contamination and supply information needed to select a design
alternative. While Phase II of the Feasibility Study plays a major
role in Phase III RI activities, the activities of the RI’s first
two phases play a major role in progression of the Feasibility
Study. The Interaction and concurrent progression of the Remedial
Investigation and Feasibility Study Is shown in Figure 2.
7—148
-------
Phase I of the RI and FS progress concurrently. The Feasibility
Study begins by identifying potential treatment technologies and
applicable or relevant and appropriate requirements (ARAR), using
the Pre-R1 results along with Phase I results as they become
available. For example, if Phase I identifies a radiation problem,
suspected from the Pre—RI file search but not found during the
Pre—RI field investigation, communication of this information to the
parties performing the Feasibility Study is essential. The
potential treatment technologies must be expanded to include
radiologically contaminated material and ARARS that relate to such
materials must be identified.
Phase II of the FS is directly dependent upon Phase II of the RI.
In fact the data reduction and mapping which occur in the second
phase of the RI may reduce the scope of the second phase of the FS.
The nature of the information provided by Phase II of the RI may
allow a realization that certain alternatives will be more viable
than others and allow treatability needs to be more accurately
identified. Once Phase III of the RI Is complete Phase III of the
FS can begin. At this point, all the data will be available that is
required to confidently choose a protective, cost—effective cleanup
design.
Each Phase of RIGS, in association with the Pre —R1, allows specific
determinations to be made as part of a consistent, cost effective
and technically sound overall approach. The Pre—RI defines the need
for further action and scopes the initial Phase. Phase I provides
information about the physical environment where the site is located
and whether the contamination on—site is migrating, both of which
are critical for future actions. During Phase II the contamination
itself is characterized, allowing the remainder of the study to
focus on the specific problem. Phase III provides the opportunity
for confirmatory sampling, filling data gaps and treatability
studies.
RIGS promotes the generation of critical data early in the
Investigation and the focusing of the study as more is learned. In
this way, an applicable design alternative can be chosen in a
timely, cost effective manner.
REFERENCE
EPA Guidance on Remedial Investigations Under CF.RCLA .
EPA/54016—851002, June 1985.
Note: The authors of this paper are staff members of the New Jersey
Department of Environmental Protection (NJDEP), Division of
Hazardous Site Mitigation and Division of Hazardous Waste
Management. The options and opinions expressed herein represent
those of the authors and not of NJDEP.
7—149
-------
flpur. 2: RI/PS INIERAc ’noN
FEASIBIUTY STUDY
*
a
REMEDIAL INVESTIGATION
PHASE II
— Ds?!i . variability. dtsblbutio and
conc.nbotlon rang. of contarninents
through sampling of all medic for
ext.nd.d parameters (pp+4.O , TCL.)
eased on statistical oonsid.ratlons.
reduc. data to Identify major
contaminant, of conoerrt map
aff.ct.d areas
PHASE III
— Perform additional sampling for
sxt.nd.d parameter, to fill dote
gaps as Identified in Phase II of RI
— Perform additional. •uppl.rnental
sampling. as rt.ed.d. for Iimlt.d
porom.t.rs d.ntffi.d In Phase if
of RI
— — Pwforrn bench and pflot seal.
bwatobillty •tudi.s ld.ntlfl.d
c n..d.d In P1,0.. II of PS
• Adapted from an attachment to on EPA memorandum from .1. Winston Porter
,*: lrite.irn Guidance on Sup.rfund S.l.ctlon of R.rnedy.
7—ISO
Pre—Ri Final Report
with Phased RI/PS
Scope of Work
dsveiopsd. Including
Data Qualify
Obj.ctiv.. for
Phase
PHASE I
— Identity potential treatment tsch,,o4o ,
(!n luding innovative options).
contolnm.nt/dlapoect requirement,
for residuals or boated wasts and
related ARA s (1.... land ban)
- Aasembl. treatment/disposal omoln-
odor.. Into alternative.
— D.velope a range of alternative, attain-
ing various Ivels of performance
— D.fln of work for Phea. II of PS
-. Assess potential application of beat-
ability studies, ineluding mats$ale
handling
PHASE I
— Defin, physical nature of envlronm.ntal
ttlng by performing th. following
studies: topographic. g .otsehnlca l
radIological, biological, and d.tail.d
studies of th. sites physical
feature, (l.a.. air, surface, watsr/
sediment. subsurfac, element. md.
sail ga. and groundwater).
Analyze samples for surrogate
parameters to achieve a brood
overview of .it.
— Define Scope of work for Phase II
of RI. Including Data Quality
Objectives (OQO)
PHASE U
— Rosad on Phase ft RI results, screen
alternatives to narrow th. field to
b. analyzed in detail
— Identify need for scientific beotablflfy
studl.e and addltiønal Information to be
attained through sampling to allow
oemprehen&ve assessment of all
alturnatives .dll being considered
PHASE III
— D.vslope general performance
crtt.r$c to each elturnotlv .
— Analyze relative costs, long
and short term effectiveness
and lmpl.mnentablilty
— Y.rffy/compare protectiveness
protection of public health and
environment. complianc, with
AltAR’s. roduetlon of mnobillty/
to*lctty (attoinrnent of
prsferencs for p.rrnon.nt
solution. Involving treatment)
and other statutory factors
(considering waivers as
necessary)
al.cUon of remedy
Select a remedy that is
protective of human health
and the environment
select a cost effective remedy
Select a remedy that will attain
State/Federal ARAR s upon
completion
Select a remedy that use.
permanent solutions and
alternative technologie*
to U,. maximum extent
possible
Consider use of Interim
responses
-------
DEVELOPING AN ENVIRONMENTAL LABORATORY ACCREDITATION SYSTEM
John W. Locke, Executive Director, American Association for
Laboratory Accreditation
ABSTRACT
This paper describes the development of a laboratory accreditation
program to recognize the competence of testing laboratories which
test waste water, solid waste, hazardous waste, and drinking water.
The need for a program which combines these testing skills is
summarized and a description of how these skills have been combined
is presented.
The approach used by the American Association for Laboratory
Accreditation (AALA) is feasible, in spite of the complex nature of
the testing which must be accomplished. A number of laboratories
have already been assessed and there is great interest in continued
development of the program.
International criteria for assessing the competence of testing
laboratories developed by the International Organization for
Standardization (ISO) and U.S. Environmental Protection Agency
Quality Assurance (EPA QA) Project Plans are used as the basis for
defining competence. Each laboratory is visited by an assessor and
examined to see if it meets these criteria and to determine if it
has the personnel, equipment, testing procedures and quality
control/quality assurance programs necessary to perform the EPA
tests competently. Any deficiencies are noted and the laboratory is
required to attend to those deficiencies before accreditation is
granted. Performance samples are required in certain testing areas,
depending on the availability of programs.
This paper concludes with a discussion of problem areas being
addressed by AALA and recommendations for the kind of technological
support needed.
INTRODUCTION
Performing tests on samples of the environment is a growing
business. Testing is the key to making decisions about drinking
water —— ground water —— waste water —— solid waste —— toxic
substances —— and pesticides in individual communities and industry
facilities throughout the nation.
The Environmental Protection Agency (EPA) is developing new
standards and test methods and defining new levels of performance
which must be obtained in order to meet the requirements of our
environmental laws. It is complicated work, made even more
complicated because the standards and test methods are continuously
7—151
-------
changing. On top of these ever changing tast methods is the
uncertain quality of the test data obtained by laboratories. Those
who are responsible for environmental monitoring must select
qualified laboratories to do the work, and the data generated by
these laboratories must be accurate.
Citizens concerned about their environment continuously ask: Where
can I find a competent laboratory? States responsible for
environmental control are concerned: Are the data being generated
good data?
EPA has provided some assistance and guidance for those primacy
states who must develop their own lists of approved testing
laboratories. For example, it has developed performance evaluation
(PE) programs for water supply and water pollution studies in
support of state (or regional) drinking water and wastewater
programs. Participation is available every six months.
Laboratories seeking approval by the states must register with a
state or regional office responsible for preparing the list of
“Approved Laboratories”. In order to retain its approval, a
laboratory must obtain test results within specified ranges on two
sequential semiannual performance evaluation tests. Some states
have developed their own performance evaluation programs in which
their approved laboratories must participate.
This approval Is called laboratory certification in some states; the
International terminology used for this kind of recognition is
called laboratory accreditation. EPA seems to be adopting the
performance evaluation approach as the basis for all of its testing
approval processes, but without trying to coordinate the
requirements of the different environmental testing areas.
THE BASIS FOR AN ACCREDITATION SYSTEM
A fundamental question is: Does this EPA guidance provide a
reasonable model for a national laboratory accreditation system?
Two aspects come into question in particular. One relates to the
efficiency of the system across environmental testing areas. The
other relates to the adequacy of relying only on performance tests
for judging the competence of laboratories. In practice, we are
seeing EPA and various state environmental departments accredit
laboratories for narrowly defined fields of testing and very narrow
testing procedures. Accreditation Is granted for drinking water ——
and waste water —— and groundwater —— and solid waste —— and
Superfund site testing. Often the same testing equipment and
facilities are involved —— but each area is handled separately and
there is little if any communication among approvers. A duplicative
and confusing mixture of regulations, fees, and inspections has
arisen and Is growing. Some states are trying to put their own
programs together. But approval by one state often does not mean
acceptance In another state.
7—152
-------
EPA management espouses the principle that the only proof of testing
capability is a de ionstration of a laboratory’s ability to obtain
accurate test results. But, it only has proficiency tests for about
one—third of the parameters of interest. Where is the proof of
capability for the other parameters?
Performance evaluations run every six months certainly do not
provide sound statistical evidence that a laboratory can continually
perform competently. The use of on—site assessment of laboratories
wherein peers examine the laboratory operations and evaluate them
against widely accepted criteria is frowned upon by some EPA
managers as a paper exercise. Yet, can laboratories be required to
participate in performance evaluation for each of the hundreds of
parameters involved when we know that the limited frequency of
performance evaluation does not statistically prove their accuracy
in testing?
A combination of techniques must be employed to arrive at the most
effective mechanism for approving laboratories. This combination
will include on—site visits to laboratories to evaluate them against
accreditation criteria, participation in performance testing, and
establishment of sound quality control procedures —— including
control charts —— In the laboratory.
THE ANERICAN ASSOCIATION FOR LABORATORY ACCREDITATION
AALA is a nonprofit, membership, professional society whose sole
purpose is to recognize competent testing laboratories. Membership
is available to all persons and particularly to those interested in
developing a basis for eliminating or substantially reducing the
increasing number of unrelated, narrowly based, substantially
duplicative accreditation schemes that are developing.
The American Association for Laboratory Accreditation (AALA) has
embarked on a program for accrediting environmental laboratories
based on the combination of features described above.
Internationally developed criteria for accrediting testing
laboratories, International Standards Organization (ISO) Guide 25,
have been adopted as the basis for accreditation. This guide sets
requirement for laboratory organization, quality system, staff,
testing and measuring equipment, calibration, items to be tested,
records, and test reports.
ASSESSING THE LABORATORY
Table 1 presents a check list to guide the assessors in evaluating a
laboratory’s ability to meet the general criteria. Key to the
evaluation is the laboratory’s quality manual. This manual normally
goes to the assessor before the visit to the laboratory so that the
assessor can be familiar with what to expect during the visit to the
7—153
-------
laboratory. The quality system in the laboratory is flexible as it
must be in order to reflect the individual style of management of
each organization. AALA does provide the laboratory with a
guideline developed by the International Laboratory Accreditation
Conference (ILAC) that identifies all items which should be
addressed in the manual. The laboratory must be practicing quality
control as described in its manual, otherwise the assessor will cite
a deficiency.
The assessor’s evaluation of the laboratory’s actual testing
operation is focussed on the particular testing equipment and
process, rather than on particular areas of the environment because
the measurement process is often comparable, whether a test is for
drinking water, waste water, or solid waste. Care is taken to
ensure that the laboratory procedure reflects the requirements for
accuracy for each area of the environment (i.e. the accuracy
demanded for drinking water testing for one parameter may be much
greater than for the same parameter in solid waste testing).
Performance testing (called proficiency testing) is required. The
results of proficiency testing are reviewed by assessors for clues
to determine deficiencies and unsatisfactory performance.
INTEGRATING TESTING REQUIREMENTS
EPA and many of the states are organized in response to specific
legislation: drinking water; clean water, pesticides; and solid
wastes. Laboratories are not typically organized along lines
specified by EPA’s enabling legislation. Rather they have
biological sections and chemical sections, and in each of these
sections they may have units which focus on testing technologies
such as radio chemistry or spectroscopy. The laboratory
accreditation system must be able to accredit laboratories as they
are organized to provide the relevant data.
In order to do this, AALA has developed an assessor checklist for
each testing technology, and has set requirements based on the most
stringent requirements in EPA’s regulations. Table 2 presents a
description of how the checklist is broken down by testing
technology. For each testing technology, the checklist is further
broken into elements shown on the bottom of Table 2.
Each of the relevant EPA regulations has been consulted and the
specific requirements reviewed. The references include the
following EPA guidance:
Manual for the Certification of Laboratories Analyzing Drinking
Water, Criteria and Procedures
Manual fot Chemical Analysis of Water and Wastes
Contract Laboratory Program, Inorganic and Organic Analysis
Test Methods for Evaluating Solid Wastes (SW846)
Standard Methods for the Examination of Water and Wastewater
(13th — 16th editions)
7—154
-------
To be accredited for a given parameter, the laboratory must meet the
most stringent requirement in these references. Table 3 presents an
example in one testing area, Ion Chromatography.
DEFICIENCIES
At the conclusion of an assessment, the assessor prepares a report
of findings, identifying deficiencies (i.e., deviations from the
criteria and test method procedures for which accreditation is
requested) which in the opinion of the assessor the laboratory must
correct in order to be accredited. The assessor holds an exit
briefing with top management of the laboratory, going over the
findings and presenting the list of deficiencies (deficiency
report). The authorized representative of the laboratory or his
designee is asked to sign the deficiency report to attest that he
has reviewed the deficiency report with the assessor. The
laboratory Is requested to respond within one month of the date of
the exit briefing. It is entirely possible that the laboratory will
disagree with the findings that one or more items are deficiencies.
In that case, the laboratory is required to write to AALA
Headquarters explaining why it disagrees with the assessor.
ACCREDITATION DECISIONS
Assessor reports, communications from the laboratory, and results of
proficiency testing are submitted to the members of the AALA
Accreditation Council for evaluation and vote on accreditation. Any
negative votes are reviewed by AALA staff with the Council member
and with the laboratory for resolution. If resolution Is not
possible, the laboratory is not accredited.
If accreditation is granted, the AALA staff prepares and forwards a
certificate and scope of accreditation to the laboratory for each
enrolled field of testing and special program. The laboratory
should keep every scope of accreditation available to show clients
or potential clients the testing technologies and. test methods for
which it is accredited. AALA staff also uses the scopes of
accreditation to respond to inquiries and to prepare the AALA
Directory of Accredited Laboratories.
APPEALS PROCEDURE
The AMA staff advises the laboratory of its right to appeal adverse
accreditation decisions to the whole Accreditation Council. If not
satisfied with the Council decision, the laboratory may make a
further appeal to the Board of Directors. All decisions by the
Board are final. Details of the appeals procedures and the
laboratory’s right to a hearing are contained in the AALA Bylaws.
7—155
-------
UPDATING ACCREDITATION
Accreditation is granted for two years. However, before the second
year of accreditation, each laboratory must pay Interim fees and
submit updated information on its organization, facilities, key
personnel, and results from proficiency tests. Any changes in
location, ownership, management and supervisory staff, authorized
representative, or major facilities of the laboratory must be
promptly reported to AALA headquarters.
Well in advance of the expiration date of its accreditation, each
accredited laboratory is sent a renewal questionnaire. A successful
on—site reassessment must be completed before accreditation is
extended for another two years.
AALA conducts a full on—site reassessment of all accredited
laboratories every two years. Reassessments are also conducted when
evaluations and submissions from the laboratory or its clients
indicate significant technical changes in the capability of the
laboratory have occurred.
CONTINUING DEVELOPMENTS
AMA has formed an Environmental Advisory Committee to continue the
development of this program. This committee will be responsible for:
o Recommending changes to the program, including changes In
scopes of accreditation now offered to successful
participants if deemed necessary;
o Reviewing current assessor checklists and recommending
changes;
o Updating the list of test methods and parameters covered by
the program;
o Identifying appropriate proficiency tests in which
laboratories would be required to participate;
o Suggesting potential assessors and overseeing assessor
training programs, recognizing that assessors are the
fundamental link in the implementation of the program.
Additional areas of environmental concern may be considered for
addition to the program. These could include air monitoring, indoor
air quality, radon and other specific parameters of public interest.
The key to the successful implementation of the program is
Involvement —— Involvement of all those who would be willing to work
to see a national accreditation system capable of meeting their
7—156
-------
specifications developed. You are invited to join the Advisory
Committee and to join AALA. Our only purpose is to formally
recognize competent testing facilities.
American Association for Laboratory Accreditation (AMA)
656 Quince Orchard Road 11704
Gaithersburg MD 20878
(301) 6701377
7—157
-------
Table 1
AALA ASSESSOR CHECKLIST: GENERAL CRITERIA
Each entry below represents a criterion statement from ISO/IEC Guide 25.
Record coimuents for any entry on your own sheets. Any entry checked MnoM
(indicating a deficiency) must be explained.
Laboratory Name:
Address:
Yes No N/A
3. Organization:
Satisfactory organization structure
Can perform representative tests
No undue pressure observed
Staff responsibilities are clear
Identify technical manager: __________________________
Adequate security for client data
4. Quality System:
4.1 Laboratory operates internal quality system
Quality manual available for staff use
Quality manual maintained regularly
Identify quality manager: _________________________
4.2 Quality manual contents:
Organization charts
Staff duties pertaining to quality
General quality assurance procedures
Q/A procedures specific to each test
Proficiency/reference materials use
Feedback & corrective action program
Technical complaint handling procedure
4.3 Quality system periodic reviews recorded
5. Staff:
5.1 Necessary educ., training, knowledge, & experience
5.2 Job descriptions for senior technical positions
5.3 Adequate supervision
5.4 Technical/quality staff backups/deputies
5.5 Qualifications/training/experience recorded
6. Testing and Measuring Equipment:
6.1 Equipment available for scope of tests
6.2 Equipment maintained & instructions available
6.3 Overload & mishandling procedures available
6.4 Equipment records maintained:
6.4.1 Name of equipment item
6.4.2 Manufacturers name/type/serial number
6.4.3 Dates received/placed in service
6.4.4 Current location (where appropriate)
6.4.5 Details of maintenance
7—158
-------
ASSESSOR CHECK LIST: GENERAL REQUIREMENTS - page 2
6.5 For measuring equipment: yes No N/A
6.5.1 Date of last calibration & reports
6.5.2 Maximum time between calibrations
6.6 Calibration labels used
7. Calibration:
7.1 Prog. for initial & periodic calib. established
7.2 Traceable measurements (where applicable)
7.3 Reference standards for calibration only
7.4 Reference standards calibrated appropriately
7.5 In—service test equipment checks (where relevant)
7.6 Reference materials traceability
8. Test Methods and Procedures (use field checklists as applicable):
8.1 Equipment operating instructions up—to-date
8.2 Methods are as required by specification
8.3 Non—standard test methods fully documented
8.4 Calculations & data transfers checked
8.5 Data processing accuracy checks
9. Environment:
9.1 Equipment protected & environment monitored
9.2 Access to test areas controlled (as needed)
9.3 Adequate housekeeping
10. Handling of Items to be Tested:
10.1 Sample identification procedures adequate
10.2 Bonded storage available (if needed)
10.3 Sample protection procedures adequate
10.4 Rules for receipt, retention, disposal
11. Records:
11.1 Records adequate to permit repeat of test
11.2 Records & reports secure
12. Test Reports:
12.1 Work in laboratory covered by test reports
12.2 Each test report contains:
Name & address of laboratory
Unique identification (including pages)
Name & address of client
Test item identification & description
Date of receipt of test item & test
Test results relate to tested item — — —
Identity of test method used — — —
Description of sampling procedure — — —
Any deviations, additions, etc. to test
Identity of any non—standard test used
Results and any failures identified
Measurement uncertainty (if relevant)
Signature and date
Statement regarding reproduction of report
12.3 Report format OK?
12.4 Supplemental procedures
7—159
-------
Table 2
ENVIRONMENTAL ASSESSOR CHECKLISTS
Specific Criteria by Testing Technology:
Microbiology
Radiochemistry
Gross Alpha, Gross Beta, Liquid Scintillation, Proportional
Counters
Atomic Absorption/Inductively Coupled Plasma Spectrophotometry
Visible Spectrophotometry
Automated Spectrophotometry
Gas Chromatography (Drinking Water)
Gas Chromatography
Gas Chromatography/Mass Spectrometry
Ion Chromatography
High Performance Liquid Chromatography
Titrimetry
Gravimetry
Miscellaneous, Electronic Probes (pH, Fluoride Specific Ion’ Dissolved
Oxygen)
Thin Layer Chromatography
Turbidity
Chemical Oxygen Demand
Biochemical Oxygen Demand
Carbonaceous Biochemical Oxygen Demand
TOC
TOX
MBAS
RCRA Tests (Flaninability & Corrosivity only)
Cyanides
For each of these we have broken down the requirements into the following
elements:
1.0 Organization and Personnel Requirements
2.0 General Facilities
3.0 Instrumentation
4.0 Documentation
5.0 Analytical Methodology
6.0 Quality Assurance
7—160
-------
Table 3
Specific Criteria & Assessor’s checklist
ION CHROMATOGRAPHY
1.0 Organization and Personnel
1.1 Is use of the method (300.0) restricted to use of personnel
experienced in IC?
1.2 Are users experienced in interpretation of ion chromatograms?
1.3 Before using this method, does each user demonstrate the ability
to generate acceptable accuracy and precision using a laboratory
control standard?
1.4 Is 1.3 documented for each user?
2.0 General Facilities
3.0 Equipment
3.1 Ion Chromatograph
3.1.1 Anion guard column
3.1.2 Anion separator column
3.1.3 Anion suppressor column
3.1.4 Detector — Conductivity cell
3.1.5 Strip chart recorder
3.2 Are interferences solved by sample dilution or spiking?
3.3 Is suction taken to eliminate the water dip?
3.4 Is the action documented?
3.5 Are samples and solutions containing particles over 0.20 microns
filtered?
4.0 Documentation
5.0 Analytical Methodology
5.1 Is reagent water free of anions of interest in use?
5.1.1 Is the reagent water quality documented?
5.2 What is the element solution?
5.3 What is the regeneration solution?
5.4 Are diluted working standards prepared weekly except nitrite
and phosphate?
5.5 Are nitrite and phosphate standards prepared fresh daily?
5.6 Is sample holding time determined by the anion that requires
the most preservation time and shortest holding time?
5.7 Are holding times documented?
5.8 Is system calibration checked daily?
5.9 Is same size loop used for standards and samples?
5.10 Is spiking done to produce adequate resolution?
6.0 Quality Assurance
6.1 Are Calibration curves for each analyte prepared with a
minimum of three concentration levels and a blank?
7—161
-------
6.2 Is one of the points of the calibration curve near the method
detection limits?
6.3 Are retention times recorded and documented during
calibration?
6.4 Is the working curve verified daily?
6.5 If more than 20 samples are run, is the working curve
verified?
6.6 If results vary by more than + 10 , is the test repeated
using fresh calibration standards?
6.7 Does the laboratory maintain performance records to define
the quality of data?
6.8 If method modifications are made, does the user repeat 1.3?
6.9 Are a minimum of l0 of all samples spiked to monitor
performance?
6.10 I sample operator precision determined?
6.11 Is method performance calculated?
6.12 Does the laboratory maintain separate accuracy statements of
performance for water and wastewater?
6.13 Does the laboratory demonstrate by analysis of reagent water
that all glassware and reagent interferences are under
control?
6.14 Does the laboratory analyze field duplicates?
6.15 Does the laboratory perform quality control check sample
analyses?
7—162
-------
A PEE—REMEDIAL INVESTI(ikTION STUDY AS AN ALTE N TI\TE
APPROACH IN THE SITE REMEDIATION PROCESS
Christine N. Andreas, Linda S. Grayson, Stephen A. Borgianini, Charles
Elmendorf, Division of Hazarc3ous Site Mitigation, Environmental
Measurements Section, New Jersey Department of Environmental
Protection, Trenton, NJ
BSTHACT
In and of itself, a Site Investigation (SI) as currently applied under
the RCRA 3012 & CERCLA 104 programs does not provide sufficient
information to structure subsequent remedial activities at hazardous
waste sites. Site Investigations, by virtue of their limited scope
and purpose, provide limited information about the extent of hazards
posed by specific site conditions. The result of employing data
generated from current SI practices could lead to unnecessary time
delays, cost overruns, confirmatory resampling, additional sampling,
and in the worst case reimplementation of the RI.
As an alternative to proceeding from an SI to a full scale RI/FS, the
Division of Hazardous Site Mitigation has developed a Remedial
Investigation Guidance Strategy (RIGS) that utilizes a Pre—RI Study
that will determine the course a Remedial Process will follow. A
pre—RI study must be performed before any remedial activities can
occur.
Sampling during the pre—Ri is intended to better define site
conditions as well as define migratory pathways that exist on—site.
These pathways include air, surface water and sediment, and
groundwater. The possible receptors in all cases are the human
population. In assessing the air route, both on and off—site
locations will be considered. Contamination may also be present in
the surface water and sediment on—site. Other influences may cause
contamination to move to off—site surface water bodies. This portion
of the pre—RI investigation will further investigate the nearest
streams, rivers, marshes, lakes and bogs. A limited number of surface
water and sediment samples may be collected for analysis. 1 nother
major migratory pathway to be investigated is groundwater. As
groundwater is utilized by most municipalities to supply the demand
for potable water, it must clearly be addressed. Consideration will
be given to the saturated as well as unsaturated zones.
Mechanisms and locations for sample collection and measurement are
dependent on site specific conditions and parameters selected for site
characterization. For example, a series of biased samples reflecting
natural and man—made influences will be collected and analyzed for
selected parameters. Air, surface water and sediment, and groundwater
systems will be surveyed. In addition, anthropogenic routes of
*The options and opinions expressed in this paper represent those of
the authors and not of the N.J. Department of Environmental
Protection.
7—163
-------
contaminant transport will also be investigated. This includes
vehicular transportation, as well as collection systems (i.e.
stormwater, wastewater).
Also to be addressed during the pre—RI are areas of highest probable
contamination; these areas will be assessed but not necessarily
evaluated through field samples. Included in this survey are (1)
areas of known disposal; (2) waste arid waste treatment systems; (3)
storage facilities; (4) on—site laboratory facilities; (5)
production facilities; and (6) miscellaneous items such as
transformers and/or asbestos. Based on site specific characteristics,
the need for sampling during this phase will be determined.
A time frame of approximately three to four months is projected for
completion of the pre—RI. Utilizing the results of this preliminary
sampling, an interim report will be generated. This report will
recommend one of the following: (1) no action; (2) final or interim
remedial measures; (3) initiation of a focused RI; or (4)initiation
of a full scale RI/FS. The information obtained during the pre—RI
will be instrumental in planning further remedial activities.
INTRODUCTI
In and of itself, a Site Investigation (SI), as currently applied
under the RCRPi 3012 and CERCLA 104 programs does not provide
sufficient information to structure subsequent remedial activities at
hazardous waste sites. Site Investigations, by virtue of their
limited scope and purpose, provide limited information about the
extent of hazards posed by the site. Current Remedial Investigations
proceed directly from an SI to a full—scale RI/FS. Activities for the
RI are based on the limited data obtained during the SI. The time lag
which exists between obtaining SI data and initiating the RI may be
considerable. In an attempt to provide sufficient information to make
subsequent decisions about remedial activities at a site, NIJDEP,
Division of Hazardous Site Mitigation (DHSM) has developed a pre—RI
study. The pre—RI will attempt to fill the data gaps which exist and
time delays which arise when proceeding from SI to the RI. The
results of employing data generated from current SI practices could
possibly lead to unnecessary time delays, cost overruns, additional
sampling, confirmatory resampling, and in the worst case,
reimplementation of the Remedial Investigation (RI) study.
The pre—RI study will determine the potential for off—site migration
of contaminants.. It will expand the information known about the site
which was provided by the SI. The pre-RI study is intended to be a
mechanism for determining the need for implementing a full scale RI/FS
at a hazardous waste site. The data generated, and information
obtained during the pre—RI, in conjunction with SI data, will indicate
whether there is a need to assess only one migratory pathway and focus
the remedial investigation or whether there is need to implement a
full—scale RI to address a number of possible contaminant systems.
The pre—RI will be used to develop the Scope of Work (SCX ) for the
next phase of remedial activities as well as prioritize the actions
which will be conducted at the site. In addition, the pre—RI
7—164
-------
investigation will assess any immediate threat the site poses to the
general public and the environment. This will be done by thoroughly
addressing the potential for off—site contamination; migratory
pathways will be reviewed and evaluated. The protection of the public
health and the environment is of the utmost importance. If there is
an impending risk posed by existing site conditions, immediate actions
will be taken to stabilize and/or remediate the situation. Any
actions conducted at the site will be consistent with recommendations
made for further remedial activities. And lastly, the pre—RI will
provide a broad data base from which future actions may be decided. A
large number of field measurements and limited analytical data, used
in conjunction with past data and file information, will be used to
formulate an accurate picture of existing site conditions. From this
information, a well defined Scope of Work for subsequent remedial
activities may be developed. NJDEP—DHSM has projected a time frame of
six (6) months and $50,000 to complete the pre—RI utilizing in—house
personnel. While NJDEP—DHSM is proposing to use in—house personnel to
conduct the pre—RI, contractors may also be engaged to perform the
study. However, the uncertainties associated with engaging a
contractor may increase the overall cost and increase the time frame
proposed to complete a pre—RI study. Hence, the pre—RI study should
be used as a foundation for any future remedial activities at the
site. when used to its fullest potential, the pre—Ri will prove an
invaluable asset to the existing Remedial Investigation process and as
such should be utilized to refine the Remedial Investigation strategy.
Regardless of the funding source (Superfund, Spilifund, Private
party), a pre—RI should be performed to guide future RI activities at
the site. Performing a pre—RI has numerous advantages. First, there
will be standardization of procedures and practices which will be
carried out at all sites. The same criteria will be used to evaluate
all sites; the results will direct future work. Second, significant
cost savings may be realized if the pre—RI results indicate that the
site requires a focused RI/FS that is not warranted. The individuals
performing the study will be evaluating the data and developing the
Scope of Work for the next phase of the RI. This progression insures
consistency which is a critical factor in the environmental evaluation
area.
The pre—RI study, which precedes RIGS, is intended to fill the data
gaps which are not answered by the SI. The SI provides primarily
source data, while the pre—RI will provide data indicating the
potential for off—site migration of contaminants. The pre—RI will be
conducted between these two activities and will provide valuable
information on impending threats which have arisen at the site since
the initial SI. Information gathered during this study will be used
as the basis for planning future remediation activities. The purpose
of this paper is to briefly describe and outline the strategy of the
pre—RI and activities to be considered in the study.
The overall strategy of the pre—RI may be summarized as follows.
Sampling performed during the pre—RI is intended to better define site
conditions as well as define migratory pathways that exist on the
site. These pathways include air, surface water and sediment, and
7—165
-------
GENERIC SCHEDULE OF WORK FOR PRE — RI
— V m • i • a 2 5 5 5 5 5 5 N N N N
PRE-RIACTIVITIES 111111111111111111111111111
Referral receipt X
Pile review W X X X X
File sources checklist x X
File suary x x x
Pile hazard rank in chart x X
Procure funding cc X X X X X X X X X X X
Obtain enforcement clearance 5 * X X X X X X
Conduct initial site visit X
Interim site visit X X
Site visit prior to sa ling X X
Site hazard ranking chart x x
Preliminary site map x x
Final ut. map
EMS brt.fing x
Prepare Interim Pre—RI
Report S Work Plan X X X X X
Review interim report N C X X X X
Engage contractors 5$ X X X
Laboratory procurement x
Field saspling preparation x
Initiate field activities x x x x x x
AnalyatasM s X X X X X X X
Data Validation X X X
Final Pre—RI Report S
Recommendations x x x
* On—going throughout the project.
*1’ May be deleted if not required and will uignificently reduce
the timeframa.
5CC Priority turn—around tim. will reduce the tim, fram, but will Figure 1
Increase overall coat.
-------
groundwater. Mechanisms and locations for sanpie collection and
measurements are dependent on site specific conditions and parameter
selected for site characterization. A series of biased samples,
reflecting natural and man—made influences, will be collected and
analyzed for specific parameters. The contaminant transport systems
mentioned above, as well as anthropogenic routes of transport will be
investigated. Based on file searches and site investigations, areas
of the site will be prioritized based on environmental risk. Also to
be addressed during the pre—RI are areas of highest probable
contamination; these areas will be evaluated although not necessarily
evaluated through field samples. Included in this survey are (1)
areas of known disposal; (2) waste and waste treatment systems; (3)
storage facilities; (4) on—site laboratory facilities; (5)
production facilities; and (6) miscellaneous items such as
transformers and/or asbestos.
The pre—RI can be subdivided into a number of activities which
correspond to a projected time frame. Figure 1 outlines these
activities and indicates the overall time table for completion of a
pre—RI study. A time frame of six months has been projected for
completion of activities and preparation of the final pre—RI report.
Figure 2 details the pre—RI progression during the course of the
study. As evident in both figures, a number of items may be performed
concurrently. When a pre—RI is conducted at a site, a file
investigation and site visit should occur within the first week.
Depending on past activities at the site, the files may provide
valuable information which will be useful in guiding sampling
activities. In addition, a site visit should always be conducted
early in the project. Any voids or questions which arise during the
intial file search may be readily answered by a site visit. Where
files are sparse and/or conflicting reports exist, a site visit may
clarify these items. A preliminary site map should be developed and
used as a base map to indicate all pertinent structures, waste areas
and past and future sampling locations. The site map should he drawn
to an appropriate scale. An additional site inspection should be
conducted prior to writing the interim report to insure an accurate
and detailed account of existing site conditions.
The interim pre—RI report will summarize the existing file information
about the site. It will include site history, background, previous
analytical data, enforcement actions, previous remedial activities and
all other pertinent information which will help to determine future
activities. In addition, the interim report will include a summary of
the site inspections highlighting the areas of concern. A preliminary
site map will also be included. Based on the files and site
inspections a Scope of Work (S ) will be developed for the remaining
field activities at the site. The site specific sc will be used as
the basis for developing the Field Sampling Plan (FSP).
One final site visit should be conducted just prior to initiation of
field activities to prepare for actual field sampling. While the file
search and site visit are being conducted, attempts should be made to
procure funding and obtain the appropriate enforcement clearance.
These activities should be initiated as soon as possible due to the
7—167
-------
Generate
_____ File
Summary
Report
Complete
Relative
Hazard
Evaluation
Chart
PRE—RI
PROGRESSION
Obtain
Enforcement
Clearance
Complete
File Source.
Checkflat
File
Search
Figure 2
Perform Engage
Site Contrao—
tori.
NI
*f Referral 1
Prepare
Pre—RI
Interim
Report
and
Sow
Review
Report
Conduct
Site
Inspection
Procure
Funding
Approve
Interim
Report &
Sow.
Initiate
Field
Sampling
Activities
4
Samples
sent to
lab.
Await
analytical
results.
Develop
Site
Map
Field ’ I
I Sampling I I Procure I
Prep Laborotory
Generate
Site Inspection
Report
‘1’
Conduct
Interim
Site Visit
Send data
for
validation.
Complete Relative
Hazard Evaluation
Chart
I
Prepare
Final
pro—RI
report
with
recommend —
otleri.
-------
additional time which may be incurred for completion of these
activities.
Based on the file search and site investigation, completion of two
checklists are required. While the file seach is on—going throughout
the project, the initial thrust of activities is during the first few
weeks of preliminary activities, prior to actual field work. A File
Sources Checklist was established and must be completed indicating
those sources which have been utilized. This is to include
interviews, enforcement actions, newspaper articles and any other
information which should be considered in the interim report package.
In addition, a Relative Hazard Evaluation Chart must be completed.
There are two categories: one based on files and one based on the
site inspection. A numerical system has been utilized to help
prioritize different areas of the site. A low priority rates a one, a
medium priority rates a three and a high priority rates a seven.
Field ratings are selected based on the following criteria:
Low (1) — Lack of reference; documented release which has been
remediated.
Medium (3) — Consistent enforcement actions; citizens complaints;
water quality monitoring data; underground utilities
as conduits.
High (7) — Documented release which has not been remediated;
documented release which has been stablized but not
remediated; impending structural failure of container or
vessel; documented fish kills or other biota adversely
impacted; potential for subsurface gas production;
documented reports of discolored soils/stressed
vegetation; sampling data indicating contamination.
Site investigation ratings are selected based on the following
criteria:
Low (1) — No visual evidence of a problem. No reading on field
instrumentation.
Medium (3) — Sporadic readings on field instrumentation; permitted
discharge pipes; runoff pathways; seepage along a
bank; underground utilities as conduits; stressed
vegetation; odors; unusual physical characteristics
(oil slicks/foam).
High (7) — Obvious signs of release (documented/ undocumented);
potential release probable; areas suspected of
undocumented release(s); sustained instrument readings
above background; uncontrolled releases; non—point
source discharges; evidence of subsurface gas
production; aquatic stress; discolored soils.
The two ratings are added together in an attempt to indicate the areas
7—169
-------
of greatest concern. An additional column is provided to indicate the
route which will be affected. I n overall score of seven (7) or better
indicates an area of high priority. The overall priority rating
should be considered when making Scope of Work recommendations in the
Interim Report.
Upon completion of the aforementioned tasks, a pre—RI interim report
must be generated. This report is a written summary of the field
inspection and file investigation. It will include the File Sources
Checklist and Site Prioritization Chart which have been completed for
the particular site. The report will prioritize the potential routes
of exposure (air, surface water and sediment, soil, groundwater) based
on site conditions and file data. Lastly, the report will provide
recaiinendations and the Scope of Work (SCM) for the pre—RI field
investigation. Included will be a site specific sampling plan
detailing the field activities to be performed in the next segment of
the pre-RI. Upon completion of the Interim Report a review period may
be required. During this time, any necessary modification may be
incorporated.
The study will include investigating the release potential to all
environmental media at the site. Objectives of the air study include
(1) assessing radiological conditions; (2) characterization of
off—site migration of air—borne contaminants; and (3) continuous
determination of ambient air conditions on—site. A series of
stationary monitoring units will be located upwind, downwind and
on—site in an attempt to characterize air—borne contamination that are
migrating off—site. In addition, a radiation survey will be conducted
on—site. Based on information known about the site, grids will be set
up ranging form 25’ x 25’ to 200’ x 200’. A survey will be conducted
along the horizontal and vertical transverses. And lastly, ambient
air monitoring will be conducted at the site during all site
activities. Instrumentation will include on OVA, PID, explosimeter
and other applicable monitoring equipment. The surface water and
sediment study will (1) attempt to determine whether on—site
contaminants are migrating off—site and concentrating in sediments;
and (2) assess direct discharges from the site which are entering
adjacent water bodies. Surface water samples will be collected only
when there is direct evidence of a current discharge from the site.
This may include a running leachate seep, a discharge pipe coming from
the site, runoff or similar. Sediment samples will be collected
upstream, adjacent to and downstream from the site. If there are
no surface water bodies, sediment samples are not required. Field
measurements such as DO, pH, conductivity, temperature and flow rate
will be determined. The objective of the soil/groundwater study is to
determine whether the unsaturated zone is the medium for contamination
to enter the groundwater. If monitoring wells exist on site, samples
will be collected. Depending on site conditions, production wells
and/or potable wells may be sampled. Gamma logging will be performed
if continuous boring log information is not available for the current
wells/piezometers. In addition, a soil gas survey of the site will be
performed. Using 100’ x 100’ grids, eight (8) data points per acre
will be sampled. The grid size may be modified based on site
conditions. Surface soil samples for chemical analysis will be
7—170
-------
collected if site conditions warrant. This may be reathred if
stressed vegetation or leaking tanks or drums are present. On a site
where none of the conditions mentioned above exist, test pits, soil
borings and piezometer installation may be required. Careful
consideration should be given to selection of these alternatives as
they will significantly increase cost and may not provide sufficient
information from which to make subsequent decisions. And lastly,
consideration will be given to the Universal Soil Loss Equation to
determine the gross amount of soil loss migrating off—site which may
potentially carry contaminants.
Samples collected for chemical analysis (Target Compound List + 30 or
Priority Pollutant List + 40) will be limited; selection will be
based on requirements established for each area. Field measurements
with portable instruments will dominate the pre—RI study. A soil gas
and radiation survey will be conducted on site. Dissolved oxygen, pH,
conductivity and temperature measurements will be collected where
applicable. All measurements will be performed in the field by
trained DHSM individuals. Field collection of data will significantly
limit analytical costs while maximizing data generated.
Once the interim report is accepted, field sampling activities may be
initiated. The laboratory and any necessary contractors should be
engaged and field sampling preparation commence. Actual field
activities may take from one to six weeks depending on the Scope of
Work and personnel allotted for field activities. Barring prolonged
periods of bad weather, field activities should not delay the project.
Once field activities are complete, analytical results may take up to
eight weeks. This time frame has been built into the generic schedule
but may be reduced significantly if priority turn—around is requested.
This will however significantly increase the overall project cost.
Best judgment and the critical nature of a project umst be taken into
account when making this decision.
Upon receipt of analytical results, data will be validated by the
Quality Assurance Section of DHSN.
Based on the analytical results, field measurements, field searches
and site inspections, EIIS will generate a final pre—RI report. This
report will make recommendations for the next action to be undertaken;
it will provide the Scope of Work for phase I of the proposed
three—phase RI. Figure 3 shows how the information gathered during
the pre—RI will be used. The no action alternative will be
recommended when all information indicates that there is no
significant environmental problem which warrants additional action.
Immediate final remedial measures will be recommended if the site
poses an impending threat to the population or environment surrounding
the site. This action might be recommended if intact drums are
present. Immediate interim remedial measures may be recommended if
there is a lagoon on site which requires stabilization until further
remedial alternatives can be determined. The interim remedial measure
alternative will more likely be followed up by a focused RI or
full—scale RI depending upon overall site conditions. The immediate
7—171
-------
Figure 3
PRE—RI and RIGS
Pre—RI
1. File search
2. Field investigation
Pre — RI Final Report
recommending one of these options:
Pre—RI indicates
no environmental
problem exists
on site.
Pre—RI indicates need
for immediate remedial
measures to mitigate
threat to public health
and environment
immediate
final
remedial
measure
immediate
interim
remedial
measure
Pre—RI indicates
problem at site
is limited to one
matrix or area.
Pre—RI indicates
problem at site
is wide—spread,
affecting more
than one matrix
or area.
r
focused, full—scale,
phased phased
remedial remedial
investigation investigation
no
further
action
1’
1’
7— i /
-------
final alternative may require limited sampling during remedial
activities and no further action. The pre—RI measurements/analysis
will provide sufficient information to determine whether a focused RI
(only a groundwater study) or full scale RI is warranted at the site.
Depending upon the scope of the problem, immediate remediation may
occur under the supervision of the NJDEP or USEPA. However, if the
scope of work identified in the final report is beyond the
capabilities of the DHSM staff, a full—scale or focused RI will be
initiated.
In summary, the pre—RI strategy was developed to fill the gap that
exists between the current SI process and initiation of the RI/FS.
The information obtained during the study will provide a broad data
base from which future remedial decisions will be based. Regardless
of the funding source, a pre—RI investigation should be conducted on
all sites where a full—scale Remedial Investigation is being
considered. The results of the pre—RI may indicate conclusively that
a full—scale RI/FS is not warranted. Rather, a focused RI/FS may be
more appropriate. Thus, a significant cost savings may be incurred.
In addition, the pre—RI is intended to be conducted by in—house
personnel in an attempt to maintain consistency from site to site.
The underlying thrust of any pre—RI study is to obtain enough data and
information about the site to develop the Scope of Work for the next
phase of remedial activities. This will insure that the next action
taken will protect the public and the environment from any existing
hazard. Use of the pre—RI strategy will assure accountable spending
of public money, reinforce decisions made, insure maximum protection
of public health, and most importantly, maximize the efficiency of
ensuing remedial activities.
ACK(JdqLEDGEMENTS
The authors of this paper are staff members of the New Jersey
Department of Environmental Protection, Division of Hazardous Site
Mitigation, Trenton, NJ.
7—173
-------
SAMPLING AND
FIELD METHODS
thai rpersons
Billy Fairless David Bennett
thief CP / qSV Chief, Toxics
US. EPA Integration Branch
25 Funston Road Hazardous Site
ansas City, Kansas 66115 Evaluation Division
U.S. EPA
401 N Street, S.W.
Washington, D.C. 20460
-------
USING BARCODES AND PORTABLE COMPUTERS FOR SAMPLE TRACKING
Dennis Hooton, Chemist/Quality Assurance Coordinator, Midwest Research
Institute, Kansas City, Missouri
ABSTRACT
Midwest Research Institute is currently doing work in support of EPA’S
hazardous waste “Listing” and “Relisting” programs. Because of the
demanding Q1 requirements in maintaining the integrity of samples and
legal defensability of the data, new tracking procedures have been
developed to meet these objectives in an efficient manner.
A chain—of—custody/sample tracking system was developed to improve
quality assurance and management of project resources. The objectives
of this system were to identify incoming samples received at MRI,
expedite sample transfers under traceability and chain—of—custody
criteria, and to monitor work—in—progress. Ideally, it was the intent
that this system reduce the time required for effective documentation
and remain flexible to meet changing quality assurance and project
requirements. By incorporating barcodes for sample labeling and
data—entry, and using “lap—top” computers programmed to read barcoded
information, we were able to achieve a versatile and portable sample
tracking systems.
Barcoding, a technology that has been time—tested for about 20 years,
provides an accurate and fast way to record information. Preprinted
barcoded (and human—readable) labels are used to encode samples and
make them traceable to the source and history of each sample. Exact
replicas of these labels are affixed to corresponding forms, lab
glassware, data charts, in other words, anywhere where sample
identification is needed. Encoding samples this way also allows
confidential business information to be more easily protected in the
performance of day—to—day activites. Laboratory worksheets that contain
lists of barcodes and task descriptions are used to enter routine
information such as sample status and type of analysis. Code 3 of 9
barcode format allows alphanumeric encoding so that analysts,
equipment, etc., can be easily identified.
Portable lap—top computers offer a flexible and completely portable
means of recording sample analysis information and a convenient way for
reporting that information. These computers can record information by
scanning barcodes with an attached light pen, directly connecting to
other instruments (such as a balance) through RS—232 ports, and manual
entries typed on the key board. Customized software allows you to
“prompt” for the required information and automatically document exact
dates and times tasks are performed. The information that is collected
can be printed, then signed by the analyst to validate its accuracy; or
the information can be sent to a host computer using the computer’s
built—in modem via standard telephone lines.
Finally, by using the electronic files received by the host computer to
8-1
-------
automatically update the sample database, many different types of
reports can be generated to monitor sample holding times, document
sample traceability, and better manage project resources.
INTRODUCTION
Midwest Research Institute is a research and development organization
located in Kansas City, Missouri. One major area of effort is the
investigation of environmental problems associated with hazardous
waste, including sampling and analysis tasks involving identification,
characterization, and incineration of hazardous waste.
Sample tracking has become very important to these types of research
because of the need to produce and document data that is both legally
defensible and of known quality. Today’s research requires proof that
samples were uniquely identified, that chain—of-custody handling
procedures were followed, and that critical tasks were performed within
specified holding—times. In addition to these concerns, as an
organization there is always a need to efficiently manage resources and
control the documentation process.
SYSTEM RE JIREM TS
The technologies needed to set up computerized sample tracking were
found to be both available and inexpensive.
The system that evolved was comprised of two basic elements: the use of
barcoded information for fast, accurate data entries, and the use of
“lap—top” computers for portability and flexibility.
Preprinted barcode labels are coninercially available in just about any
size, material, or format imaginable. High—quality labels provide very
accurate scanning and direct computer entry of information (estimated
to be 20 times faster and more reliable than recording information
manually). For our application, six—replicate labels in sequentially
numbered series are used to identify samples. These labels are also
used to reference containers, forms, data charts, and laboratory
notebooks relevant to that sample. This number serves as a reference
point that uniquely identifies a sample from the time of collection
through data reduction and reporting.
The Radio Shack Model 100 portable computer was chosen for the sample
tracking system because it is completely portable and programmable. It
becomes an “electronic notebook” for recording and transferring data
without manual entries. Information is captured by simply scanning a
barcode with a wand. Data are printed in real—time or saved into a
file. Time and date of critical tasks are automatically recorded using
the computer’s built—in clock/calendar. This computer can also be used
to print customized barcodes for routine data entries, creating a
“bar—code” worksheet.
Code 3—of—9 format is used to code both alpha and numeric characters,
giving the system the flexibility to use descriptive identifiers.
8-2
-------
A microcomputer with a modem is used to receive and store the data
files that are imported from the satellite “lap—top” computers.
The microcomputer is used to effectively manage the system by
maintaining inventories, providing current status information,
monitoring holding time performance, and generating reports for project
management.
Multiple software programs are used to make the sample tracking system
actually run. Among these are commercial programs available to read and
write barcoded information, communication programs for transferring
data files electronically, word processing software for review and
verification of data files, and data base management software for the
import and management of data.
Simple programs were written in BASIC to provide flexibility for the
system, not unlike using a typed form for manually recording data.
These programs allow for “prompting,” time/date labeling, multiple
choice selection, and rejection of inappropriate information.
EXAMPLES :
This system was used to document chain—of—custody transfers by using
the following procedure:
Samples are labeled with a unique bar—coded identification number. Lab
worksheets were prepared for the analyst with barcodes describing each
sample split and analysis step for the particular phases of the sample
analysis.
Cards were issued to each analyst with their name and (barcoded)
initials to use as an “electronic signature” in documenting sample
transfers.
The chain—of—custody transfer is initiated by scanning the split code
and analysis code from the lab worksheet, then scanning the “electronic
signature” cards of the persons relinquishing and receiving the sample.
By scanning the barcode of each sample transferred, the identification,
time, and date of the transfer is printed immediately by the computer
in a chain—of—custody report. The analysts review the computer report
and verify its accuracy by both signing and dating the document. This
becomes the official chain—of—custody document that is archived for the
project records.
P 1 n identical data file is stored in the computer for exporting to a
microcomputer’s data base, where the status of the samples can be
reviewed, holding times can be monitored, and complete sample histories
may be listed.
Another example of how this system can be incorporated into other
laboratory situations is the interfacing of the “lap-top” computer with
electronic balances for automatic weight recordings. This extension of
the computer system reinforces the idea of an “electronic notebook”
8-3
-------
that changes the ways and speed in which data are recorded.
Robotics also incorporates barcode technology as a means by which
computers can read and verify sample identification and allows for
orderly processing of sample analyses.
S RY
The benefits of computerized sample tracking are improved accuracy,
efficiency, and flexibility.
Barcodes offer error—checks on reading, are simple to use, and reduce
the likelihood of transcription errors. Computerization reduces written
forms, standardizes and speeds data entry, and provides a tool for
resource management and real—time documentation. Flexibility is
achieved by using a low cost and expandable system, a system that can
be customized to specific project needs and changing quality assurance
requirements.
8-4
-------
EVALUATION OF A PROTOTYPE FIELD—PORTABLE X—RAY
FLUORESCENCE SYSTEM FOR HAZARDOUS WASTE SCREENING*
G. A. Raab, D. Cardenas, S. J. Simon, Environmental Programs, L. A.
Eccles, Advanced Monitoring Systems Division, Lockheed Engineering
and Management Services Company, Inc., Las Vegas, Nevada
ABSTRACT
A prototype field—portable X—ray fluorescence system developed by
EPA and NASA was evaluated at a site contaminated with Pb, Zn, and
Cu. The objective of the field test was to evaluate the
effectiveness of the instrument as a field analytical tool for
locating hot spots and as a preliminary screening device where
immediate data feedback aids in decision—making in the field.
By making use of an analytical method designed specifically for the
XRF system, all routine field measurements for Cu, Zn, and Pb were
made on site by placing the probe on the surface of the ground (“in
situ ” measurements). Subsequently, confirmatory samples were
collected and analyzed in the laboratory with an Inductively Coupled
Plasma spectrometer (ICP) while adhering to EPA Contract Laboratory
Program (CLP) protocols.
The quality assurance consisted of measuring NBS standard reference
materials to verify the data measured in the field and in the
laboratory in addition to duplicates, blanks, and replicate sample
analysis.
The analytical results were plotted in the sampling grid. One can
immediately locate the hotspots for Cu, Zn, and Pb on site. The
instrument detection limits for Cu, Zn, and Pb are 250, 200, and 70
ppm, respectively. Comparison of the XRF results with the ICP
results showed an overall mean percent error (MPE, which means lack
of precision and bias incorporated into one term) from NBS
concentrations of only a few percent for Cu, Zn, and Pb. Precision
and accuracy of the in situ measurements were within plus or minus
10 percent of the true value when compared to the samples analyzed
in the laboratory.
This document is a preliminary draft. It has not been formally
released by the U.S. Environmental Protection Agency and should not
at this stage be construed as present Agency policy. It is being
circulated for comments on its technical merit and policy
implications.
8-5
-------
INTRODUCTION
The Environmental Protection Agency, Environmental Monitoring
Systems Laboratory, Las Vegas asked Lockheed Engineering and
Management Services Company, Inc. (LENSCo) to field test and
evaluate the performance of a field—portable X—ray fluorescence
system for making in situ measurements. In situ measurements are
those measurements made by the XRF probe while in direct contact to
the ground surface. These measurements are conducted without any
sampling and sample preparation other than clearing the ground
surface to expose the soil. An in situ XRP measurement represents
the data obtained from only the exposed ground surface and does not
reflect any of the subsurface. The contaminants in a hot spot would
not register a response to the X—rays if the hotspots were covered
with as little as 1 cm of uncontaminated soil. Therefore, caution
must be exercised in the use of data obtained from in situ
measurements. The use of in situ measurements with the XRF system
would allow technicians to immediately locate surface hotspots of
lead, copper and zinc on the national priority list (NPL) sites and
other sites. The objective of this report is to describe the steps
necessary to complete the field test, review the analytical work,
and assess the instrumental performance. These steps are as follows:
o design a sound in situ analytical method for a field—portable
X—ray fluorescence system prior to the field test
o analyze each of the 40 samples in situ with the
field—portable XRP system at 40 locations on a 60 foot by 150
foot grid with sample intervals at every 15 feet.
o subsequently collect confirmatory surface soil samples from
the same locations.
o analyze the samples in the laboratory following the ICP CLP
protocol (an in situ measurement by a field—portable XRF
system has no homogenization technique as a part of sample
preparation. The only preparation necessary for an XRF in
situ analysis is to clear a flat surface on the soil.
Therefore, the sample area of the in 8itu measurement cannot
be considered homogeneous. Acceptance of an In situ XRF
measurement dictates the acceptance of a certain amount of
error in measurement’s accuracy. To validate the in situ XRF
measurement, the sample area of the in situ analysis
represents one sample and the volume of sample collected
represents another. The values of these two samples will
closely approximate one another but are technically not the
same. However, th’ difference in two values should fall
within the acceptance range for the overall inaccuracy of the
XRP In situ measurements. For the intents and purposes of
this report we will assume the in situ sample area and the
8-6
-------
collected sample containing the same area to be one and the
same sample.)
o compare the XRF results with those obtained from the ICP.
The EPA has recently expressed more interest In XRF systems than In
previous years because the use of microprocessors and
state—of—the—art technology have made the equipment smaller and thus
portable. Such field—portable XRF systems have been used to
delineate hazardous waste site hotepots for priority metals in the
field (Chappell et al., 1986; Mernitz and Olsen, 1985; Furst et al.,
1985; and Kendal et al., 1984). With immediate data feedback from
the field—portable XRF system, all samples can be collected with the
knowledge of their approximate concentration. This leads to a
decrease in the number of unnecessary samples which would be
analyzed normally. The XRF field data allows an analyst in the
laboratory to calibrate his laboratory instrument to the proper
concentration the first try; thus decreasing the number of attempts
at bracketing the correct one. Another use is as a laboratory
analytical instrument to screen samples of unkno .rn concentrations
quickly providing the analyst lth an approximate concentration.
All of these applications of the XRF systems net an overall decrease
In time and in money spent.
Furst et al., 1985 described three levels of analytical requirements
for establishing the extent of environmental contamination. The
first or highest level of analysis is used to develop data for
litigation and regulatory enforcement (see Figure 1). This level
demands the most rigor in sample preparation and in8trument time as
well as the highest degree of precision and accuracy. The second
level of analysis Is used to evaluate and assess average contaminant
exposures to people and animals. The data from the third level of
analysis is used for screening in order to obtain a preliminary
profile of sites. This data can be used for decision making while
in the field. Third level data may be used also to select which
samples should be sent to the laboratory for first level analysis
following the Contract Laboratory Program (CLP) analytical
protocols. This report discusses the results obtained by using a
portable XRF system under the third level of analytical requirements.
The area that the EPA selected for the field test was the Smuggler
Mountain NPL site in Aspen, Colorado, northeast of the Aspen city
limits. The site was li8ted on the NPL June 10, 1986. The Smuggler
Mountain mine produced lead, silver, and zinc ores. The site is
located on one of the slopes of Smuggler Mountain; some of the
mining, milling, and smelting was located here. These slopes are a
mixture of native soils, mine tailings, and other mine wastes. Much
of the surface Ms been subjected to reworking by prior and recent
construction projects. Several such projects used mine tailings as
fill.
8-7
-------
Three Levels of Analytical Requirements for Metals
c 3
co
Degree of Analytical Requirement
(Precision) (Accuracy) i (IDL)
Purpose
LEVEL I :
Very High
(±5%) (±10%) (ppb)
Litigation and Regulatory
Enforcement
LEVEL U :
Moderately High
(±10%) (±15%) (ppm)
Evaluate and Assess
Average Pollutant
Exposure to Humans
and Animals
LEVEL Ut:
Moderate to Low
(± 10%) ± 50%) ( ppm x 1000)
Screening, Preliminary
Evaluation, and On—Site
Decision Making
Figure 1. Three levels of analytical requirements for metals.
-------
Martin Marietta Aerospace people brought a prototype XRF system to
field test at the Smuggler Mountain site. The first prototype
system had evolved from technology used in the Martian Viking
lander. This system had to be redesigned to measure metals in
contaminated soils because of the changes in Its intended usage.
Prior to the field test, the software was not programmed for
efficient evaluation of soil samples and field application.
Reprogramming the software took place subsequent to the field test.
The Martin Marietta field—portable XRF system consists of three
units (see Figure 2): (1) the sensor head (when filled with liquid
N 2 , weighs 32 pounds), (2) the Canberra main unit analyzer (16
pounds), and (3) a Gridcase 2 portable computer (12 pounds). The
filtration wilt shown in Figure 2 was designed for laboratory use
only to preconcentrate metals in water samples. The cooled
semiconductor detector has excellent energy resolution and is
capable of simultaneous detection of a wide range of X—ray
energies. The cooled semiconductor detector decreases the dead time
response to the X—rays It senses, thus decreasing the length of time
needed for analysis. A typical analysis with the system lasts
between 120 and 300 seconds. The X—ray tube uses a molybdenum
target operating at 30,000 volts to produce a wide enough spectrum
to fluoresce the priority elements. The detector is a semiconductor
made of lithium drifted silicon. The detector must be cryogenically
cooled and must have a continuous supply of liquid nitrogen.
CONCLUSIONS
o The XRF system produced data of known quality from 229 in
situ measurements (defined as measurements made by placing
the probe on the ground surface and by analyzing the same
surface without moving the probe). The XRF field results on
the NBS standards compared relatively well with the certified
NBS values of the same standards.
o Field personnel can greatly decrease the time spent on site
by making In situ measurements. If necessary, the technician
can collect a confirmatory sample after each XRF analysis.
o The detection limits are low enough for obtaining data when
third level requirements are necessary for analytical work on
hazardous waste site Investigations.
o The NBS standards were adequate for quality control and
quality assurance. These standards were SRM 1633a, coal fly
ash; SRN 1645, river sediment; and SRM 1648, urban
particulate.
o The instrument uses cryogenics to cool the silicon—lithium
detector which requires a Dewar container filled with liquid
8-9
-------
FILTRATION ENCLOSURE
SENSOR HEAD
X-RAY TUBE
AND
DETECTOR
CANBERRA PHA
MAIN UNIT ANALYZER
LI
—i--I ’
Figure 2.
X—ray fluorescence subsystem identification.
r]I
iii
I
I
• /
GRIDCASE 2 COMPUTER
8-10
-------
nitrogen. Even though the Devar container will last 8 hours
before a refill is necessary, maintaining a continuous supply
of liquid nitrogen in the field can be difficult in some
locations.
o The overall advantages of all X—ray fluorescent systems
include: minimal sample preparation time, rapid turnaround
time for analyses, multi—element analytical capability,
nondestructive analyses, and sample size required for
analysis is small or possibility of surface analysis without
the need for sampling at all. These advantages make the XRF
system very cost effective.
RECOMMENDATIONS
o A field methods manual should be written for field XRF
measurements and should Incorporate field sampling, sample
preparation techniques, and analysis.
o Characterized spike soils should be pre red for the XRF
systems as reference standards. These would be spiked at
concentrations ranging from 5 to 10 times the IDL, to the
highest likely to be encountered; or at levels considered to
be a public health hazard.
o A procedure for primary calibration, field update of
calibration and of QAIQC checks of the Instrument accuracy
and precision should be worked out.
o The investigation of a sample preparation method for non in
situ measurements should be tested. This should include
examination of both the pelletizing and fusing techniques,
and the use of loose soil. Indications are that one would
obtain different levels of precision.
o The Martin Marietta XRF system should be compared with other
commercially available field XRP systems. The detection
limits, precision, and accuracy of each instrument could be
determined side by side in the laboratory with the ICP or A.AS
by using rigorous QAIQC protocols, characterized samples, and
certified spiked standards.
o Once the instrument detection limits are established for XRF
field—portable systems, the initial steps may be implemented
in developing these Instruments for characterization of
uncontrolled hazardous waste sites with respect to specific
toxic metals.
8-11
-------
FIELD TEST
Personnel from EPA, Martin Marietta, NASA, LENSCo, and Camp Dresser
and McKee, Inc. were present during the field test at the Smuggler
Mountain NPL site (Figures 3 and 4). The objective of the field
test was to assess the performance of the Martin Marietta
field—portable XRF system on an N?L site with regard to effectively
identifying hotspots of lead, copper, and zinc. This was
accomplished by analyzing in situ and by subsequently collecting
surface soil samples (see Section 5, Procedures for exact method) on
site at 40 locations on a 60 feet by 150 feet grid with sample
intervals at every 15 feet (Figure 5). Technicians analyzed each of
the 40 samples in situ for total Pb, Cu, Zn, and Fe with the
field—portable XRF system and processed the data for on site
decision making.
NBS standard reference materials (SRM) were used as quality
assurance/quality control standards. The NBS standards were
incorporated to give us reference data of known quality. Those used
were SRN 1633a, coal fly ash, SRM 1645, river sediment, and SRM
1648, urban air particulate. The NBS standards were analyzed in the
field with the XRF system. These values were compared against the
SRM—certified values.
A field method was designed specifically for the use of a
field—portable XRF system for this field test. This field method is
described later under “Analytical Procedures and Sampling Protocol
for the Field—Portable XRF System and the ICP.” The same samples 1
were then collected for later confirmatory analysis in the
laboratory. To further corroborate the XRF field data, we used the
known and accepted CLP methods of preparation and analysis. Because
the CLP method of sample preparation uses a relatively weak
extraction, we also used a Parr bomb method to provide a stronger
extraction method that was likely to more closely approach the NBS
certified values. The Parr bomb method was performed on 13 selected
samples but under CLP instrumental requirements for the ICP.
The quality assurance (QA) procedures developed for the XRF field
analysis allowed for the proper verification of the data. The
verification establishes the quality of the data. To evaluate the
reproducibility of measurements (i.e. QA procedures) and to minimize
statistical error, the blank, the instrument calibration standard,
and the same aliquot from each of the three NBS standards were
analyzed seven times each without disturbing the sample.
The results showing high concentrations of Cu, Zn, and Pb on the
sampling grids (Figures 6, 7 and 8) can be processed immediately on
a computer from this data produced by the XRF system. The
capability of immediate analytical results offers many advantages.
The data from the Initial screening allows field crews to:
8-12
-------
Figure 3. Contour map of Aspen, Colorado.
8-13
-------
Figure 4. Closer view of site location. Sampling grid is found
within the square located by arrow.
Centennial
ndomlniums
8-14
-------
til 4 DOS
03
-J
c - f l
SMUGGLER MOUNTAIN SITE
Figure 5. Sampling grid on NPL site.
WELLING
SPRING
-------
TRAILER COURT
n.d.
485
n.d.
n.d.
n.d.
S
n.d.
.
318
.
313
.
912
S
16 0 13p3
1056
S
n•d.
n.•d.
n.d.
.
n.d.
S
n.d.
S
n.d.
S
n.d.
.
n.d.
.
n.d.
S
n•d. nd.
ned.
n d.
ned.
n.d. n.,d.
n.d.
.
n.d.
S
373
S
n.d.
.
253
S
WASTE SITE
TENNIS COURTS
Figure 6. XRF values for Cu plotted on sampling grid.
726
S
SMUGGLER
MOUNTAIN SITE
XRF Data for Cu
pg/g
n.d.= not detected
8-16
-------
TRAILER COURT
AN
U.S. EPA AIR SAMPLER
& POWER SOURCE
33,16
n.d.
629
.
362
n.d.
207
n.d.
361
421
.
3711
.
52 0 4195
2
WASTE SITE
218
260
.
343
294
n.d.
n. d.
n. d.
454
n. d.
n. d.
n.d.
244
n. d.
354
453
.
228
n. d.
536
224
459
S
n. d.
.
253
.
SMUGGLER
MOUNTAIN SITE
XRF Data for Zn
jig/g
n.d. not detected
TENNIS COURTS
Figure 7. XRF values for Zn plotted on sampling grid.
8-17
-------
TRAILER COURT
I
2 8
296
.
161
S
174
12 , 38 15, 24
19,654 19. 37’
14,943
S
172
133
128
.
128 156
.
1 8 128 292
S
1 8 1 9 327
S
132 1 6 2 4
2 4 131 310
.
233 246
S
267 399
S
200
S
343
.
1
253
S
SMUGGLER
MOUNTAIN SITE
XRF Data for Pb
pg/g
WASTE SITE
TENNIS COURTS
Figure 8. XRF values for Pb plotted on sampling grid.
8-18
-------
(1) identify the hotspots,
(2) restrict the investigation to the contaminated area,
(3) either go to the next level of investigation or stop at the
screening level, and
(4) make decisions based on the data base which is generated from
the analyses.
The next level of investigation above screening requires
confirmatory samples be taken and analyzed. These can be analyzed
in the field. If this were in an area being investigated for the
first time, the hotspot would be identified and the locations
registering as background (non—contaminated) by in situ XRF
analysis need not be sampled for confirmation. In our case, the
intensified effort would be restricted to the area marked “waste
site” in figures 6, 7, and 8. This would decrease the number of
samples, which might otherwise be analyzed, and thereby reduce
overall analytical costs.
As more samples are analyzed, the data base for the area grows. The
initial screening provides the foundation from which the
investigation proceeds. The subsequent measurements then provide
data of higher quality from which plots can be generated. Field
crews can make decisions based on these plots and this data base.
It should be noted that the Martin Marietta XRP field—portable
system was required to perform in adverse weather conditions. It
began snowing during the field test and the XRF system did not
malfunction. The authors feel that the system performed well under
these conditions.
PROCEDURES
X—RAY FLUORESCENCE PRINCIPLES
The X—ray fluorescence (XRF) system used in this study is energy
dispersive by design. All XRF depends on either an electron or
X—ray beam bombarding the sample. The X—rays produced by an X—ray
tube impinge on the electron clouds or orbitals in the atoms within
the sample. When the X—ray displaces an electron from an inner
orbital of an atom, such as the k—shell, a vacancy is created. This
causes an instability below the electrons in outer shells. As the
outer electrons seek stability by filling the inner shell vacancies,
a cascade of electrons spontaneously follows. Energy is released or
emitted for each shell vacancy that is filled. This emitted energy
is characteristic of the atom from which it was produced. This
emission is call fluorescence. All elements excited by the X—rays
fluoresce simultaneously to produce a spectrum of characteristic and
8-19
-------
backacattered radiation. It is this spectrum that the detector
senses and counts. The whole spectrum data is transferred to the
analyzer where the software deconvolutes the peaks for the desired
elements.
THE PROBLEM OF RESOLUTION WITH VARYING PARTICLE SIZE
The bulk density of the sample and the particle size distribution
affect the characteristic X—ray intensity. When dealing with varied
distribution of particle sizes, an accurate analysis of these
particles is difficult if there Is no attempt to make all the
particles the same size, either by segregation or reduction.
The varied particle sizes have an effect on X—ray absorption and
enhancement especially during in situ emission analysis. Collecting
and grinding the sample to 0.075—mm (200 M) will generally solve
the problem, but this is not always a practical solution in the
field. It also defeats the purpose of using the field—portable XRF
system with an in situ technique, i.e., a fast turn—around time. We
can address this problem in a sample preparation step by using a
mortar and peatel to break up the soil samples and by grinding the
sample to approximately sand size assuming the sample is dry. Even
though this approach does not entirely solve the problem, it does
reduce the effect to an acceptable constant error while keeping
sample preparation time to a minimum.
ANALYTICAL FROCEDURES AND SAMPLING FROTOCOL FOR THE FIELD—PORTABLE
XRF SYSTEM AND THE ICP
Field personnel analyzed and collected 40 soil samples according to
the following method:
XRF ANALYSIS :
a. The analyst will make seven replicate measurements, rotate
the sample 90°, stir each standard, and make seven more
replicate measurements on the following:
o a blank sample made of silica sand,
o the instrument calibration standard,
o the three NBS standards.
b. Clear the area for analysis by removing rocks and debris from
the surface, mark the area, place the instrument on the
designated area, and activate the X—rays.
8-20
-------
c. The analyst will make single measurements on the following:
o the soil sample,
o standard with concentrations closest to the soil values,
o the instrument calibration standard.
d. The analyst will make a duplicate analysis on every tenth
8011 sample without moving the probe.
e. Record all data pertinent to the sample, its 1ocatioi In a
logbook, and store the analytical data obtained from the
instrument on a diskette.
f. After the analyst finishes measuring the soil sample in situ ,
collect the sample for later corroborative analysis.
g. Collect the sample using the following steps:
o Use a clean trowel to take a 200 cc sample. Penetrate a
2” x 2” area on the surface to no more than 2 inches
deep. Be sure to I nclude the area which has been
analyzed in situ .
o Empty the sample into an acid—washed 150 mm x 25 mm
plastic petri dish, being sure to fill the petri dish to
aproximately two—thirds full) seal the petri dish Inside
a plastic zip—lock bag, and seal the bag with tapa.
o Clean the trowel by wiping the blade with a paper towe .
and then rinsing with distilled water.
h. Take samples with contamination levels ranging from
approximately five times the IDL for Cu, Zn, and Pb up to the
maximum values found on sIte.
I. Use NBS standards for the reference calibration.
j. Use silica sand for the blank sample. The grains of silica
sand will closely approximate the grain geometry of the soils.
The sample preparation for the ICP analyses was done by both the
standard CLP method and the Parr bomb method which is described
below. The CLP method and QAIQC protocol for analysis may be found
in Exhibit D of the Invitation for Bid for the Contract Laboratory
Program (U. S. EPA, 1984).
8-21
-------
ICP ANALYSIS
Digestion of soil by a generic Parr bomb method (adapted from
Bernas, 1968, Buckley and Cranston, 1971; and Dolezal et al., 1969):
a. Dry at 60°C and homogenize the sample.
b. WeIgh 0.5 g of soil and place it in the Parr bomb.
c. Add 5 niL of concentrated nitric acid (HNO 3 ) and 2 niL of
concentrated hydrochloric acid (HC1).
NOTE: o Do not add more than the 2 niL liCi prescribed.
Too much can generate enough chlorine gas to
cause the Parr bomb to explode.
o Do not add any soil with carbonates; the
evolution of CO 2 could cause the Parr bomb to
explode.
o Do not add the filter paper if It is a cellulose
base. This could cause the formation of
nitrocellulose which is explosive.
d. Seal the Parr bomb and place it an oven at 120°C for 2 hours.
e. Remove the bomb from the oven and allow it to cool to room
temperature.
f. Open and rinse the contents into a filter funnel feeding Into
a 100—niL volumetric flask. Bring the flask up to volume with
DI water. The digest is ready for analysis.
RESULTS MID DISCUSSION
DISCUSSION OF SOFTWARE
The application of the XRY system changed throughout its
development, and the programming of the software did not keep pace.
The software was not modified for field use or for soils analysis
prior to the first field test. For the field test at Smuggler
Mountain, the XRF system was calibrated in the laboratory with
spiked soils for the metals of interest and was checked against
certified NBS standards, but the software was set up for metal
alloys. The data and the samples were collected in the field. Then
the data was taken back to the Martin Marietta where the software
was reprogrammed to process the data for field use and soils
analysis.
8-22
-------
The software program for peak deconvolution caused many of the
non—priority metals to drop out. While beyond the scope of this
study, the authors feel that the software should be investigated to
optimize the deconvolution of the peaks.
COMPARISON OF XRF RESULTS VERSUS ICP (PARR BOMB AND CLP) RESULTS
To evaluate the performance of the field—portable XRF system,
techtiicians analyzed the NBS standards and 40 soil samples from
Smuggler Mountain on a Perkin Elmer ICP II by using two sample
preparation methods. The first preparation method is an extraction
procedure (Parr bomb described previously) and the second
preparation method is the standard CLP instrument extraction
procedure. Technicians analyzed the Parr bomb digest also according
to CLP instrument requirements. We did not use a total dissolution
procedure because most such procedures use hydrofluoric acid (HF).
It is necessary to neutralize the HF with large quantities of boric
acid. The higti amount of dissolved solIds tends to clog the
nebulizer, thus affecting precision by causing drift.
The analytical results of the NBS standard reference materials (SR 1)
for four metals by XRF and ICP (both methods) are listed In Table
1. EPA requested that Pb, Cu, and Zn be examined. We added Fe to
show how the XRF unit responded to a constant high value whIle
analyzing the other metals in varying amounts.
When we compare the values overall, the XRF analyses tend to be
higher then the ICP analyses. The analysis of the urban
particulate, SRM 1648, showed the be8t results. The XRF results are
between 95 percent and 105 percent of what the NBS baa certified as
present in the SRN 1648 with the exception of Zn. Since Martin
Marietta designed the XRF system as a semi—quantitative instrument
for the field, these analyses are excellent within the limitations
of the standards. When the individual Pb values for the different
standards from Table 1 are plotted, the XRF results of the NBS
standards show excellent concurrence with the ICP, especially for Pb
(Figure 9).
The detection limits for the XRF system were determined by Martin
Marietta. Three 300—second spectra were collected for each element,
and the ratios of the net K and the L peaks to net backscatter
calculated. The instrumental detection limit was calculated from
the formula:
IDL = ‘J3S Cb/C 5
where S is the quantity (in micrograms) of metal present in the
sample, Cb is the background counts under the peak, and C 9 the net
sample counts. The peak background is calculated, and the net
sample counts are calculated by subtracting the background level
8-23
-------
TABLE 1. COMPARISON OF XRF AND ICP RESULTS FROM NBS STANDARDS
El ements
Analysis Fe Cu Zn Pb
Number (all units are in ug/g)
Coal Fly Ash
SRM 1 94,000 118 220 72.4
1633a 2 91,555 178 261 152.4
3 38,780 75.6 103 34.4
4 17,944 54.0 84 n.a.
River Sediment
SRM 1 113,000 109 1720 714
1645 2 313,686 183 1551 735
3 84,410 109 1632 688
4 n.a. n.a. n.a. n.a.
Urban Particulate
SRM 1 39,100 609 4760 6550
1648 2 41,252 584 2212 6247
3 21,746 550 4486 5986
4 20,857 432 3443 4192
No. 1 Certified by National Bureau of Standards.
No. 2 Martin—Marietta XRF unit. The Martin Marietta values for the SRM 1633a
and 1648 are the averages of 7 replicates; for the SRM 1645, the values
are the averages of 14 replicates.
No. 3 Perkin Elmer ICP II, Parr Boii Method. The values are the averages of 3
replicates.
No. 4 Perkin Elmer ICP II, CLP Methods. The values are the averages of 3
replicates.
n.d. = not detected.
n.a. = not analyzed.
8-24
-------
Pb Data: XRF and ICP vs NBS value
0 XRF
+,cP
3
01
1
-S
(0
C)
- c i
1
(p
(I )
(I- I
(p
-S
U,
V)
( ,1
C)
(3
-S
-I.
(3
‘3-
C)
0
C,
(3
-s
0
:3
I ’ ,
S ..
CU)
V
OC
(a)
00
4 -
C
0
a
C
0
0
7
6
5
4
3
2
I
0
0 2
4
(Thousands)
NBS Certified Concentration ( jg/g)
-------
from the total number of counts in the peak. The detection limits
for the XRF (the IDL for iron was not included) and ICP instrumental
detection limit (IDL’a) are listed in Table 2 and are compared In
the CLP contract required detection limits (CRDL). The CRDL’s
represent a level one requirement regime for analytical work, The
XRF IDL’s might represent a proposed level three requirement regime
for hazardous waste site investigations with field—portable XRF
systems.
In comparison of the IDL’s for Cu, Zn, and Pb against the values for
the soil samples In Table 3, what immediately becomes apparent is
that 29 of the values for Cu measured by the Martin Marietta XRF
system are less than or equal to the instrumental detection limit.
Any values below five times the IDL should not be used. The
variation of precision in this range is, as a rule of thumb, plus or
minus the IDL. Between 5 and 10 times the IDL, the variation in
precision ranges from 10 percent to 20 percent of the amount
present. Above 10 times the IDL, the variation in precision is less
than 10 percent of the amount present.
When a sample Is wet, the bulk density changes, greater absorption
of X—rays takes place, and the XRF values are lower. Compounding
the absorption effect are the grain geometry and particle size
distribution which also affect the absorption of X—rays (Rhodes and
hunter, 1972). These uncontrolled parameters account for some of
the XR.F results which are lower than the CLP results throughout
Table 3.
Some 13 samples were selected to be analyzed by the Parr bomb
method. Although the Parr bomb method is a strong extraction
procedure, the nitric and hydrochloric acid extract did not dissolve
all the metals because the silicates are not soluble in these
acids. The silicates often entrap some of the metals within them.
The CLP method is simply an HN0 3 extraction procedure, and thus, we
expect the recoveries to be lower than a total dissolution or XRF
analysis. The XRF analysis is comparable to a total dissolution in
that the measured response of the fluorescence accounts for all
elements with an atomic number greater than Z = 9 up to Z 92.
This range covers most elements occurring in a natural soil even
with contaminants.
A problem developed with use of the Parr bomb while analyzing SRN
1633a and 1645. The addition of 2 mL of RC1 caused the generation
of too much Cl 2 gas. The Parr bomb was not sealed tightly enough
and allowed the pressure to vent. This leakage strongly affected
some of the Parr bomb results and accounts for the Parr bomb values
for routine samples being less than the CLP values in Table 3,
especially with reference to the iron results. The iron combines
with the chlorine to form iron chloride which is volatile at 120° C
and leaked out with the Cl 2 gas. The standards SRM 1633a and 1645
8-26
-------
TABLE 2. COMPARISON OF THE MARTIN MARIETTA XRF SYSTEM AND THE ICP
INSTRUMENT DETECTION LIMITS AGAINST CLP CONTRACT REQUIRED
DETECTION LIMITS (CRDL)
I CP
Martin Marietta Perkin Elmer CRDL
Element XRF System 1 (ppm) ICP 2 (ppb) (ppb)
Cu 250 8 25
Pb 70 35 (5)*
Zn 200 17 20
Cr 1000 5 10
Ni 300 30 40
As 150 80 (10)*
Se 140 170 (5)*
Ag >1000 5 10
Cd ND 4.8 5
Sb ND 50 60
Ba >1000 20 200
Hg 80 50 (O.2)*
Ti 75 160 (1O)*
Data provided by Martin Marietta.
2 IDL t s qualified for CLP.
* These metals are not required for analysis on the ICP under CLP
protocols. The values in the parentheses are the measured IDL’s for
the atomic absorption spectrometer to show its capability.
ND = not detected.
-------
TABLE 3. COMPARISON OF XRF AND ICP RESULTS FROM SOIL SAMPLES
Elements
Sample Analysis Fe Cu Zn Pb
No. No. (all units are in mug/g)
G—1 1 25,511 253 323 253
2 n.a. n.a. n.a. n.a.
3 13,461 17.1 364.9 315.1
G—2 1 23,396 b.d.l. 232 183
2 n.a. n.a. n.a. n.a.
3 15,669 27.5 374.3 282.2
G—3 1 24,114 b.d.l. 257 224
2 n.a. n.a. n.a. na.
3 19,016 20.6 434.3 416.3
G-4 1 33,460 n.d. 294 310
2 n.a. n.a. n.a. n.a.
3 17,331 18.7 410.6 356.4
G—5 1 32,159 b.d.l. 454 246
2 n.a. n.a. n.a. n.a.
3 14,460 15.7 323.6 342.3
6—6 1 43,749 b.d.l. 453 309
2 n.a. n.a. n.a. n.a.
3 15.596 18.6 369.0 301.1
6—7 1 22,317 b.d.l. 228 200
2 n.a. n.a. n.a. n.a.
3 16,872 17.7 442.7 503.5
6—8 1 21,237 n.d. 248 417
2 19,410 59.8 1,617.3 3334.0
3 13,359 22.9 676.1 1234.5
6—9 1 70.532 1,303 4,195 19,379
2 23,758 55.3 4,517.6 11,492
3 27,119 67.8 6,042.5 11,791
6—10 1 57,729 1,056 2,978 14,943
dup 52,947 919.1 3,022.2 14,840
2 32,620 100.0 6,544.6 18,539
3 27,259 73.4 6,134.8 12,875
Footnotes at end of table.
(conti nued)
8-28
-------
TABLE 3. (Continued)
El ements
Sample Analysis Fe Cu Zn Pb
No. No. (all units are in uglg)
G—11 1 82,978 1,660 5,240 19,654
2 23,422 42.4 5,265.9 11,615
3 36,969 80.7 9,383.5 15,294
G—12 1 52,830 912 3,711 15,524
2 27,606 77.8 6648.4 11,619
3 26,673 102.7 6347.0 4,135.1
G—13 1 30,009 174 259 190
2 11,405 13.0 241.9 165.1
3 12,413 14.2 263.5 176.1
G—14 1 39,972 313 421 267
2 n.a. n.a. n.a. n.a.
3 15,099 17.1 344.1 226.8
G-15 1 38,180 208 354 233
2 n.a. n.a. n.a. n.a.
3 14,157 15.9 293.6 176.9
G—16 1 35,277 163 168 137
2 n.a. n.a. n.a. n.a.
3 21,159.3 23.2 102.7 19.1
G—17 1 35,331 151 153 166
2 n.a. n.a. n.a. n.a.
3 25,141 27.2 243.6 137.6
G—18 1 31,529 158 343 327
2 n.a. n.a. n.a. n.a.
3 17,234 22.3 422.5 327.6
G—19 1 39,512 373 536 292
2 n.a. na. na. n.a.
3 15,069 21.8 451.3 812.7
G—20 1 38,321 373 449 343
dup 33,990 304.3 417.0 281.9
2 n.a. n.a. n.a. n.a.
3 14,003 21.4 459.3 396.2
Footnotes at end of table.
(conti nued)
8-29
-------
TABLE 3. (Continued)
Elements
Sample Analysis Fe Cu Zn Pb
No. No. (all units are in ug/g)
G—21 1 45,242 b.d.1. bd.l. 133
2 n.a. n.a. n.a. n.a.
3 22,018 23.6 102.0 15.4
G—22 1 36,814 b.d.l. 260 156
2 na. na. n.a. n.a.
3 14,868 18.9 329.2 436.6
G—23 1 36,202 n.d. 165 121
2 n.a. n.a. na. n.a.
3 20,862 22.9 124.5 32.9
G—24 1 38,416 n.d. b.d.l. 129
2 n.a. n.a. na. n.a.
J 21,580 24.4 165.3 61.5
G—25 1 33,462 n.d. 244 132
2 n.a. n.a. n.a. n.a.
3 20,231 22.5 111.7 17.3
G—26 1 28,503 318 361 214
2 n.a. n.a. n.a. n.a.
3 14,765 20.1 390.2 281.3
G—27 1 31,856 b.d.l. 207 174
2 na. n.a. n.a. n.a.
3 13,543 18.9 388.5 212.5
G—28 1 37,600 b.d.1. 362 296
2 12,813 15.0 275.8 194.2
3 14,903 20.5 375.7 314.6
G—29 1 56,334 1,531 4,215 14,639
2 22,712 68.8 4,315.7 6,833
3 26,774 90.9 8,467.3 8,924
G—30 1 57,050 726 3,316 12,138
dup 58,350 742.9 3,489.6 11,927
2 25,613 73.8 4,497.4 8,031
3 20,075 81.4 4,780.1 7,235
Footnotes at end of table.
(conti nued)
8-30
-------
TABLE 3. (Continued)
El einents
Sample
Analysis Fe
Cu Zn
No.
No.
(all units are in
Pb
6—31
1
2
3
47,989
25,092
27,010
556
68.0
79.9
1,332.5
5,347.1
6,164.1
5,707.9
7,090.8
1,016.3
6—32
1
2
3
34,662
9,521
9,897
b.d.l.
11.4
11.2
424
234.5
253.4
241
240.6
269.4
6-33
1
2
3
29,650
7,902
12,007
485
8.4
13.9
629
175.3
282.6
248
159.6
255.4
6—34
1
2
3
20,454
46,904
9,916
b.d.1.
41.3
11.2
b.d.l.
209.8
244.2
183
178.5
164.3
6—35
1
2
3
21,212
n.a.
14,970
n.d.
n.a.
17.2
b.d.l.
n.a.
366.1
161
n.a.
286.4
G—36
1
2
3
37,457
n.a.
25,310
n.d.
n.a.
37.5
b.d.l.
n.a.
143.5
138
n.a.
47.8
G—37
1
2
3
40,321
na.
20,846
b.d.l.
n.a.
19.4
b.d.l.
n.a.
119.0
128
n.a.
43.9
6-38
1
2
3
35,042
n.a.
21, 7 c*U
n.d.
n.a.
20.8
b.d.l.
n.a.
180.2
129
n.a.
100.2
6—39
1
2
3
36,218
na.
24,390
b.d.l.
n.a.
21.5
b.d.l.
n.a.
108.5
128
n.a.
28.7
6—40 1 27,283 b.d.l. 218 172
dup 30,954 n.d. 94.9 151.7
2 n.a. n.a. n.a. n.a.
— — — 3 - 15,246 16.5 324.0 456.3
No. 1 Martin-Marietta XRF system.
No. 2 Perkin Elner P—2 ICP, Parr bomb method.
No. 3 Perkin Elmer P—2 ICP, CLP methods.
dup = duplicate analysis of No. 1.
n.d. = not detected.
n.a. not analyzed.
b.d.l. below detection limit.
8-31
-------
were reanalyzed, and the following changes were made: the sample
amount was changed to 0.25 g, the amount of HNO 3 was changed to 3
mL, and HC1 was changed to O.5mL; and special attention was given to
the sealing of the Parr bomb itself. The results show that the
reanalysi8 of SRM 1645 recovered concentrations very close to the
NBS certified concentrations. Recoveries for SRII 1633a, the coal
fly ash, show low concentrations but they are still closer to the
NBS certified concentrations than the CLP recovcred concentratIons.
COMPARISON OF TWO FIELD—PORTABLE XRF SYSTEMS
We sent 13 soil samples to the Kevex Corporation for analysis with
their field—portable XRF system, X—site 9900 and Analyst 6700.
Kevex agreed to analyze the samples at no costs Because Kevex
analyzed these samples without contractual requirements the analyses
were not verifiable. Therefore, the data must be accepted only as
approximate values. In spite of this deficiency, we gain enough
insight from the data to warrant its inclusion here (Table 4). The
samples were shipped as loose Boils sealed in petri dishes. Kevex
analyzed the samples by placing the probe of their x—site 9900 onto
the samples in the petri dishes. No sample homogenization or
preparation took place. Overall the Fe values of the two
instruments are close enough for semi—quantitative work, but the
values for the priority metals Cu, Zn, and Pb are diverse for the
two instruments and need further investigation. The authors suggest
that further comparative work in the laboratory with rigorous QA/QC
would determine which XRF system is better suited for field work.
STATISTICS
The replicate precision on the standards of the Martin Marietta XRF
system ranged from 1.2 percent RSD for Zn to a maximum of 34.4
percent for Pb on the NBS SRM 1648 (Table 5). The second column for
each element represents the 90° rotation in the same horizontal
plane after the seven—replicate analysis. The difference in the two
sets of analyses could reflect the effect due to surface morphology
from different areas Within a sample which can affect the X—rays the
detector senses.
The duplicate precision on routine samples of the Martin Marietta
XRF system ranged from 0.88 percent RSD to a maximum of 10.19
percent RSD on the three samples run (Table 6). The relatively high
percent RSD’a which appear with sample G—20 could occur due to the
counting statistics and being close to the instrument’s detection
limit. The more controlled studies need to be done in this area.
In Figure 10, the bar graph is a plot of the percent relative root
mean square deviation verses the elements Fe, Cu, Zn, and Pb for
each method. The data from Table 1 was used to calculate the mean
percent error (MPE) using the formula:
8-32
-------
TABLE 4. COMPARISOk OF TWO FiELD PORTABLE XRF SYSTEMS
El ements
Sample Analysis Fe Cu Zn Pb
No. No. (all units are in ugig)
G—1 1 25,511 253 323 253
2 25,052 111 n.d. 220
G—3 1 24,114 b.d.l. 257 224
2 29,473 n.d. 64 268
G—7 1 22,317 b.d.l. 228 200
2 24,326 45 n.d. 536
G—14 1 39,972 313 421 267
2 28,127 n.d. 175 340
0—16 1 35,277 b.d.l. b.d.l. 137
2 22,019 n.d. n.d. n.d.
0—19 1 39,512 373 536 292
2 33,488 n.d. 670 453
0-22 1 36,814 105 260 156
2 30,882 n.d. n.d. 621
0—25 1 33,462 n.d. 244 132
2 36,179 n.d. n.d. 268
G—27 1 31,856 b.d.l. 207 174
2 32,121 171 n.d. 302
0—31 1 47,969 556 1,333 5,708
2 30,904 n.d. 4,873 6,028
0—34 1 20,454 174 198 183
2 22,339 52 422 275
0-37 1 40,321 39 103 128
2 39,575 230 8,704 11,427
No. I Martin—Marietta XRF.
No. 2 Kevex Corporation XRF.
n.d. = not detected.
b.d.l. = below detection limit.
8-33
-------
TABLE 5. REPLICATE ANALYSES OF THE NBS SRM 1648
(Data supplied by Martin Marietta)
Gross backscatter (counts per second)
Flux Fe Fe Zn Zn Pb Pb
Factors XlOO X100 X1000 X1O X1000 X1O
A—i 3.464 6.150 2.818 1.853 5.795 1.563
A—2 3.289 5.999 2.203 1.854 2.870 1.629
A—3 3.225 5.909 2.820 1.873 4.857 1.541
A—4 3.401 5.818 2.951 1.871 2.333 1.640
A—S 3.206 6.429 3.909 1.830 3.469 1.642
A—6 3.316 6.371 2.591 1.813 3.810 1.596
A—7 3.684 5.887 3.523 1.843 2.578 1.564
mean 3.369 6.080 2.974 1.848 3.673 1.596
s.dev. 0.166 0.243 0.572 0.022 1.265 0.041
%RSD 4.927 3.997 19.23 1.167 34.44 2.569
TABLE 6. COMPARISON OF PERCENT RSD OF DUPLICATE ANALYSES BY XRF
Samples
Elements IDL G-10 G—20 G—30
Fe 57,729 38,321 57,050
52,947 33,990 58,350
%RSD 4.32 599 1.13
Cu 250 1,056 373 726
919 304 743
ZRSD 6.95 10.19 1.15
Zn 200 2,979 449 3,316
3,022 417 3,490
ZRSD 0.73 3.70 2.55
Pb 70 14,943 343 12,138
14,840 282 11,927
%RSD 0.35 9.76 0.88
8-34
-------
MEAN PERCENT ERROR FOR NBS STANDARDS
k\\ PARR BOMB
LLi
a-
I ,-,
a:
0
a:
a:
I 1 J
I.-
z
U
C-)
a:
‘ Li
a-
70
60
50
40
30
20
10
0
0,
Fe Cu
V/I
XRF
Zn
Pb
CL-P
Figure 10. Bar graph of the mean percent error from NBS concntrations.
-------
MPE’ J’ 1 xioo
where:
X 1 recovered concentration for a method
NBS certified concentration
The NBS certified value is assumed to be the “true” concentration.
Each method of analysis recovered concentrations equal to or less
than the true concentration for each element. The deviation from
the true concentration of each element represents the effect of both
bias and lack of precision or NPE for each method. What Is
Interesting to note is that while the Parr bomb method shows the
lowest !4PE of the three methods, the XRF 10 within a few MPE of the
CLP for Fe, Cu, and Pb. The Zn values for XRP exhibited a lower
observed MPE than the CLP values. The overall precision and bias of
the results from the CLP and XRF methods are within acceptable
scientific limits (±lOZ).
ABBREVIATIONS
AAS atomic absorption spectrometry
CL? Contract Laboratory Program
cR.DL contract required detection limits
DI water deionized water
HF hydrofluoric acid
HC1 hydrochloric acid
HNO 3 nitric acid
ICP inductively coupled argon plasma spectrometry
IDL instrumental detection limit
MPE mean percent error
NBS National Priority List
NPL National Bureau of Standards
ppb parts per billion
ppm part. per million
QA/QC quality assurance/quality control
RSD relative standard deviation (sample standard
deviation divided by the mean times 100)
SEN standard reference material
X I S X—ray fluorescence
8-36
-------
A (NOWLEDGEMENTS
1 The authors would like to extend thanks to C. A. Villa for his
editorial comments, to S. 0. Garcia of the Graphics Department at
LENSCo; to Warren C. Kelliher of National Aeronautics and Space
Administration, Langley Research Center and to L. A. Eccies and
K. W. Brown of the EPA for providing both support at the NPL site
and background information on the project; to T. F. Staible and D.
Duster of EPA, Region 8 — Denver, to T. S. Dunlop, Division of
Environmental Health, City of Aspen, Colorado, for their cooperation
in arranging the field test, and to Benton C. Clark, Judy Cook, and
Mike Thornton of Mar in Marietta Aerospace, Denver, Colorado, for
their help and information. Thanks also go to F. C. Garner, D. C.
Hillman, and D. E. Dobb at LEMSCo, Las Vegas, for their discussions,
recommendations and support; to Scott Wyma of Kevex Corporation for
the analyses; and to John R. Rhodes and Stan Piorek of Columbia
Scientific Industries Corporation for their consultation.
FOOTNOTES
An in situ measurement by a field—portable XRF system has no
homogenization technique as a part of sample preparation. The only
preparation necessary for an XRF in situ analysis is to clear a flat
surface on the soil. Therefore, the sample area of the in situ XRF
measurement cannot be considered homogeneous. Acceptance of an in
situ XRF measurement dictates the acceptance of a certain amount of
error in measurement’s accuracy. To validate the in situ XRF
measurement, the sample area of the in situ measurement is collected
in a volume of sample. The area of the in situ analysis represents
one sample and the volume of sample collected represents another.
The values of these two samples will closely approximate one another
but are technically not the same. However, the difference in the
two values should fall within the acceptance range of the overall
inaccuracy of the XRF in situ measurements. For the intents and
purposes of this report we will assume the in situ sample area and
the collected sample containing the same area to be one and the same
sample.
REFERENCES
Bernas, B., 1968. A New Method from Decomposition and
Comprehensive Analysis of Silicates by Atomic Absorption
Spectrometry. Anal. Chem., 40, 1682.
Buckley, D. E., and R. E. Cranston, 1971. Atomic Absorption
Analyses of 18 Elements from a Single Decomposition of
Alumnosilicate. Chem. Geol., 7, 273
Dolezal, J., J. Lenz, and Z. Suleck, 1969. Decomposition by
Pressure in Inorganic Analysis. Anal. Chim. Acta., 47, 517—527.
8— 37
-------
Furat, C. A., V. Tilliughast, and T. Spittler, 1985. Screening for
Metals at Hazardous Waste Sites: A Rapid Cost—Effective
Technique Using X—Ray Fluorescence. Proc. National Conference on
Management for Uncontrolled Hazardous Waste Sites. Hazardous
Materials Control Research Institute, Washington, DC, 93—96.
Kendal, D. S., J. H. Lowry, K. L. Bour, and T. J. Meszaros, 1984.
A Comparison of Trace Metal Determinations in Contaminated Soils
by XRF and ICAP Spectroscopies. In: Advances in X—ray Analysis,
Vol. 27, Ed. Cohen, Russ, and Leyden, Barrett and Predecki.
Plenum Publishing CorporatIon, pp. 467—473.
I4ernitz, S., and R. Olsen, 1985. Proc. National Conference on
Management of Uncontrolled Hazardous Waste Sites. Hazardous
Materials Control Research Institute, Washington, DC.
Rhodes, j. a., and C. B. Hunter, 1972. Particle Size Effects in
X—Ray Emission Analysis. X—Ray Spectrometry I, pp. 113—117.
U. S. Environflental Protection Agency, 1984. Chemical Analytical
Services for Inorganica, Exhibit D, Invitation for Bid,
(Solicitation Number WA84—70911J092), U. S. EPA, Washington, DC
8-38
-------
CANISTER-BASED SN 1PLERS FOR VOLATILE ORG ½NICS
William A. McClenny, Joachim Pleil, Monitoring Methods Section,
Environmental Monitoring Systems Laboratory, U.S. EPA, Research
Triangle Park, North Carolina
ABSTRACT
Canister—based samplers for volatile organics are a viable alternative
to samplers based on solid sorbent collection. Simple, low—cost units
are available that operate for sampling periods of from one minute to
several hours without the need for electricity. If electronic flow
controllers are used, constant sampling rates can be maintained over
24 hours. Sample integrity of compounds passing through the samplers,
and storage stability have been tested. Problems with sample
integrity due to contamination in sampler elements upstream of the
canisters are common, but have been prevented by proper cleaning
procedures. Storage stability over seven days at the 1.0 ppbv
concentration level has been established. Reactions among
co—collected compounds needs further study. The weatherized versions
of these units are suitable for monitoring near hazardous waste
landfills.
INTRODUCTION
Canister—based sampling for volatile organics has recently experienced
a revival in interest,., due to two main factors: (1) the
cha 1 a terization of Tenax ’ solid sorbent with respect to its use by
EPA , ; and (2) the demonstration that canisters can be used for
controlled saInplingfI i 5 long term (days) storage of volatile organics
of special concern. ‘ ‘ In particular, the use of distributed air
volume (DAV) sets for Tenax sampling, which allows the screening of
sample analyses to identify questionable results, shows that only
about 50% of the gata from DAV sets are acceptable based on the
screening criteria. This relatively low return in defensible data is
perceived as unacceptable as long as reasonable alternatives are in
prospect.
Canisters have been used for a number of years by scientists involved
in the study of the role of hydrocarbons in atmospheri. chemistry, and
of concentration trends in organic trace gases. However, a
systematic evaluation of canisters for the storage and retrieval of
“toxic” organics, mainly chloji ated and aromatic hydrocarbons has
only recently been provided. ‘ Based on this information, the
canisters have been used to sample for toxic organics at locations
remote from the analysis system and then shipped back through the mail
system for analysis. Commercial suppliers exist for the canisters,
for mailing cartons for the canisters and for the canister—based
samplers. The samplers are available in configurations suitable for
either indoor or outdoor sampling.
This discussion will touch on several issues that relate to the status
of canister—based sampling and the tradeoffs vis—a—vis solid sorbent
sampling. The reader should be aware that characterization of
8-39
-------
canister—based sampling is not complete and that continued field use
will provide the real test of the technique’s viability.
PREPARATION FOR SANPLING AND CERTIFICATION
Canisters are prepared by welding hemispheres or other shapes of
stainless steel together and attaching a suitable valve to the shell.
The interior surface is electro—polished. Procedures and recipes for
accomplishing canister preparation vary, although SUMMA’
electro—polishing is widely used. Cleanup of the canisters by the
user both initially and after each use consists of heating, evacuation
and repeated flushing with a “zero” gas. The sampler can takes
number of forms, several of which are shown in recent publications.
A generic design for the sampler is shown in Figure 1. The main
options are with or without a pump, and with or without weatherized
housing, including the temperature control elements shown in the
figure.
The canister is the last element in the sampling train. As such, it
is subject to contamination from all upstream elements. Hence, the
one over—riding consideration, regardless of sampler design, is that
these elements be clean. Furthermore, the compounds of interest must
not be lost or significantly delayed in transit to the canister. A
number of the early systems showed contamination. 8 However, by
selection of sampler components and proper cleaning , a number of
systems with minor or unobservable target compound contamination have
been demonstrated. Certain compounds, such as Freon 113 arid
tetrachloroethene, 8 appear to be conm n contaminants in the so—called
“K” type samplers, although samplers made by one experienced operator
do not appear to have contamination at ambient levels (the “R”—type
sampler).
Based on our experience, we have proposed a one—step certification
procedure that would identify any significant contamination in
candidate samplers. The system for generating samples for
certification sampling runs is shown in Figure 2. Concentrations of
target gases in the manifold system are generated by controlled
dilution of a reference concentration with humidified zero air. This
humidified standard gas is then introduced into a distribution
manifold and thus made available to the sampler inlet and for real
time analysis. The sampler’ s manifold pump is capped to reduced the
amount of dilution air needed for the certification. A series of
analytical runs are then performed by sampling directly from the gas
manifold and used as controls for a simultaneous, simulated sampling
run. The comparison agreement should be + 0.2 ppbv up to
concentrations of 4 ppbv and no greater than ± S for concentrations
greater than 4 ppbv. In our laboratory, reference standards
containing 40 volatile organics are analyzed in this manner. Prior
to, and during certification, the sampler is run at somewhat elevated
temperatures by disconnecting the fan that is used for internal
circulation of ambient air through the sampler box. This accentuates
any outgassing so as to facilitate identification of contamination and
also helps clean up a contaminated system. Generally, samplers that
are initially free of contamination remain so. This statement is
8-40
-------
supported by a year’s worth of data from analysis of duplicate
samplers located in Houston, TX at an EPA Toxic Air Monitoring System
(TANS) site. These samplers tracked each other very consistently fo
most compounds. Some example data is given in Figure 3 for
4-ethyltoluene and o--xylene.
SAIIPLINC-
Referring to Figure 1, when sampling w:ithout a pump and using an
electronic compensating flow controller, a constant flow rate can be
maintained from a canister vacuum up to a pressure of approximately
0.8 atmospheric. To extend the sampling duration to 24 hours using a
6—liter canister the flow rate must be regulated at about 3.5 cc/mm.
With a pump a maximum canister pressure in the range of 20—40 psig can
be attained, the exact value depending on type and condition of the
pump. When using a pump, provision should be made to flush the
sampling lines before diverting flow into the canister to minimize
contamination that could accumulate between runs. A similar procedure
can be implemented in the nonpumped system by adding a three—way valve
cmd a secondary air pump between the flow controller and canister.
The automated sampling systems can be started and stopped with an
electronic or mechanical timer. Before detaching the canister the
operator can check the canister pressure to insure that the correct
pressure valve has been reached. Such a check can also be made just
prior to the sampling run to assure that an evacuated canister has
been attached. In a pumped system the value of the canister pressure
relative to the expected value can indicate that there has been a
problem and help diagnose the location of a leak in the system. After
sampling, the canisters are shipped to the analysis system through the
mail (provided canister pressure does not exceed 40 psig).
ANALYSIS
Most current mass spectrometric systems for analysis of air samples
for trace level volatile organic compounds (VOCs) are configured for
thermal desorption of solid sorbent cartridge samples through which
5—80 liters of air have been drawn. After transfer of the
concentrated sample to a reduced temperature trap, the scanning (SCkN)
mode of operation is used to generate mass spectra which are then
systematically compared to a mass spectral library to identify unknown
compounds. To use a similar approach for can.ster samples, VOCs must
first be concentrated before analysis. Because of the time required
for this concentration step and the limited air volume available, this
approach has not proven to be practical. Instead, we have used the
selected ion monitoring (SIN) mode of our mass spectrometer. In this
mode the list of target compounds is defined and no spectral libraLy
searches are attempted. Identification ir. made by the dominant ion
fragment found in the source using retention time and two minor ion
fragments as qualifiers. There is an enhanced detection sensitivity
in the SIM mode compared to the SCAN mode, although unknown compounds
are not identified. This tradeoff has so far been acceptable, since
the list of VOC5 of current interest is still in double digits. In
the SIN mode, detection sensitivity is approximately equal to that of
8-41
-------
a flame ionization detector or, in our case, ± 0.2 ppb for
preconcentration of a 250 cc sample.
our current analytical procedure includes a preliminary analysis by
GC—FID/EcD to establish the level of concentrations and to provide
information on polar and non—targeted compounds. Reduced temperature
preconcentration of VOCs directly from the canister is followed by
separation of VOCs on a megabore capillary column. Based on this
initial run, the remaining sample volume is diluted, if necessary, to
fall within the calibration range of our GC/mass selective detector
(MSD). Provided unknowns are sufficiently high in concentration, a
SC1 N run on the GC/? ’ISD may be considered. . subsequent SIN run
requires sample conditioning through a dryer to prevent column
blockage by ice formation in the small bore capillary columns used on
the MSD system. Upon elution, water vapor also causes a problem in
the MSD’s El source by increasing pressure in the high vacuum region.
Sample introduction is accomplished using a positive pressure
canister, typically above - 8 psig. Because some canisters arrive at
the laboratory at lower pressures, pressurization with zero grade air
can be required. The dilution ratio is noted and applied to
subsequent analytical results. For analysis, a mass flow controller
is attached to the canister and set for approximately twice the flow
required by the analytical system. This flow is then vented past the
system inlet.
APPLICATIONS
The canister—based samplers have been deployed by EMSr.J, EPA in four
ambient air studies and two indoor air studies, and are being
implemented in EPA’s Toxic Air Monitoring System (TANS) at 10
locations in four cities. Applications near hazardous waste landfills
and waste water treatment facilities are planned for the fall of 1987.
DISCLAIMER
The research described in this article does not necessarily reflect
the views of the Agency and no official endorsement should be
inferred. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
REFERENCES
1. J.F. walling, “The utility of distributed air volume sets when
sampling ambient air using sold adsorbents,” Atmos. Environ. ,
18:855—859 (1984).
2. J.F. Walling, J.E. Bumgarner, J.D. Driscoll, C.M. Norris, A.E.
Riley and L.H. Wright, “Apparent reaction products desorbed from
Tenax used to sample ambient air,” AtInOS. Environ. , 20:15—57
(1986).
3. K.D. Oliver, J.D. Pleil and W.A. McClenny, “Sample integrity of
8-42
-------
trace, leve1 volatile organic compounds in ambient air stored in
SUMM ’—polished canisters.” Atmos. Environ. , 20(7) :1403 (1986).
4. MW. Hoidren and D.L. Smith, “Stability of volatile organic
compounds while stored in SUI1MA polished stainless steel
canisters.” Final report on EPA Contract 68—02—4127, WA—13,
Battelle Columbus Laboratory, Columbus, OH.
5. WA. McCleriny, J.D. Pleil, T.A. Luxnpkin and K.D. Oliver, “Toxics
monitoring with canister—based systems,” Paper 87—62.3, Preprint
for 80th APCP Annual Meeting, June 21—26, 1987, NY, NY.
6. Private communication from Dr. J. Walling, Pollutant Analysis
Branch, Environmental Monitoring Systems Laboratory, Research
Triangle Park, NC.
7. LA. Rasmussen and M.A.K. Kahlil, “Atmospheric halocarbons:
Measurements and analyses of selected trace gases,” Proceedings of
NA WASIonAtmospheric Ozone , 209—231 (1980).
8. W.A. NcClenny, J.D. Pleil, T.A. Luinpkin and K.D. Oliver, “Update
on canister—based samplers for VOCs,” Proceedin s of 1987 EP /APCA
Symposium on Measurement of Toxic and Related Air Pollutants , May
1987, Research Triangle Park, NC.
9. JD. Pleil, K.D. Oliver and W.A. McClenny, “Enhanced performance
of nafion dryers in removing water from air samples prior to gas
chromotographic analysis” JAPC 37:244—248 (1987).
8-43
-------
ToAC
ToAC
Figure 1. Updated sampler configuration.
8-44
-------
Calibration Gas
Cylinder
Zero Gas
Cylinder
MFC-Mass Flow
Controller
To exhaust
To analytical
system
directly
U,
Humidifier
Sampler
(See figure 1)
Figure 2. Apparatus used in canister certification.
-------
2.0-
1.8- (a)o-xylene
1.6-
t
0. 1.4-
c ,
0_P 1.)-
I
4 n
I.u—
Co
S
0
I
0.8- -
.— - I
LW
a 0.6-
E
45 — S
“P’ 0.4-
,1 •
0.2-
0.0—
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Sampler 5. concentration in parts per billion by volume
1.0- ‘I
0.9 - (b) 4-ethyltoluene -,
an
U.0
1-’
‘ 5
a. a-,
I).’—
0_P
•.i o u.u
45> 1 ’
I
C .o —
vc 0.5-
0=
0.4-
0.3-
0.2-
4$’
0.1-
I
F
I
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sampler 5, concentration in parts per billion by volume
Figure 3. Comparison of (a) o-xylene and (b) 4-ethyltoluene data from two
co-located samplers in an EPA field study. The dashed 45° line
indicates theoretically “perfect” agreement.
8-46
-------
CATALOG OF FIELD SCREENING NIETHODS
Andrew P. Szilagyi , Senior Environmental Scientist, CDM Federal
Programs Corporation, 13135 Lee—Jackson Highway, Fairfax, Virginia
22033; Claire M. Cesalman, Project Manager, Roy F. Weston, Inc., 955
L’Enf ant Plaza, Washington, D.C. 2OO2L ; and David A. Bennett, Chief,
Toxics Integration Branch, Office of Emergency and Remedial
Response, U.S. Environmental Protection Agency, Washington, D.C.
2OL 6O
ABSTRACT
The national Superfund program is decentralized in its conduct and
management. For this reason, knowledge and skills gained in one
Region or state, or even at one site, are not necessarily
transferred to others. To facilitate transfers of information about
methods for measuring and screening chemicals in the field and for
quick—turnaround analyses, the U.S. Environmental Protection Agency
(EPA) is developing a Catalog of Field Screening Methods. The
Catalog will be provided to users as both a pocket guide and on disk
in a dBase III system. The Catalog is provisional and will be
revised and updated as new information becomes available. This
paper describes the Catalog project initiation, sources of
information about the methods, and the development and contents of
the Catalog. The current status and plans for completion of the
Catalog and data base are detailed.. Finally, the paper indicates
both how to obtain copies of the Catalog and data base and how to
provide information to update the Catalog.
INTRODUCTION
Chemical constituents must be measured in various environmental
media during the course of many Superfund site investigations. For
example, measurements in surface and ground water, soil, sludge,
air, and containers (e.g., drums or tanks) may be needed at a given
site. Various types of monitoring and screening techniques are used
to characterize “hot spots,” define general site conditions, assist
in well placement and screen setting, compare off— and on—sita
conditions, estimate potential population exDosures, determine
removal efficiencies, and establish long—term monitoring.
The appropriate type of monitoring or screening at a given site
largely depends on the intended end use of the data and associated
data quality requirements. Specific criteria that should be
considered during selection of field methods include detection
limits of equipment versus expected concentrations of contaminants,
performance capabilities (accuracy and precision, for example), and
time and cost constraints. The iltimate goal of any data
8-47
-------
collection effort is to provide data of known quality to be used in
subsequent decisionmaking. Therefore, one or several laboratory or
field methods may be appropriate for any given data collection and
analysis activity.
Options for sample analysis include the Contract Laboratory Program,
other EPA, state, and private laboratories, mobile on—site
laboratories, and field screening techniques. Some differences
among these include:
• The Contract Laboratory Program (CLP) includes Routine
Analytical Services (RAS) and Special Analytical Services
(SAS) and provides a rigorous QA/QC and evidentiary
documentation program, resulting in data of known quality.
Cost is about lOOO per organic sample, and results are
provided to data reviewers in about 30 days.
• Other laboratories (EPA, state, or private) may follow
analytical methods that are similar to the CLP, but do not
necessarily provide the same degree of QA/QC and
evidentiary documentation. In some cases, costs may be
lover and turnaround time shorter. All these factors must
be weighed against data quality needs.
• Mobile laboratories and field analytical techniques often
operate under less well controlled conditions. However,
they can provide immediate results, which may be crucial,
for example, to determining further sampling on a given
site. Costs may be lower than CLP, but quantitative
results may not always be possible.
Experience gained in the first five years of the Superfund program
has shown that field analytical methods are suitable for use in many
applications. Quick turnaround of samples, gross characterization
of contaminants, and selection of sampling locations for more
rigorous analysis (for example, by the CLP) are some of the possible
applications. It also has become clear that methods are being
developed and used to meet these needs. Discussions among EPA and
contractor staff, occasional papers in scientific literature and at
conferences are beginning to bring these methods wider attention.
The primary factor that has limited the coordination and evaluation
by EPA of these developmental efforts is the decentralized structure
of the Superfund program. A mechanism to make field screening
methods available to all program participants is needed. To fill
this need, the Analytical Operations Branch (AOB) of EPA ’s Office of
Emergency and Remedial Response (OERR) initiated a project in FY 87
to increase coordination and facilitate distribution of this
information. This project has three objectives:
8-48
-------
• To produce a Catalog of field analytical methods, which
details demonstrated field screening and analytical
techniques and focuses on methods used in the 10 EPA
Regional Offices and the Office of Research and Development
(ORD) laboratories.
• To demonstrate and evaluate up to three methods in
conjunction with EPA Regions; and
• To create momentum in EPA and the private sector for
development, demonstration, and appropriate use of field
screening methods.
This paper describes the approach, progress to date, and future
direction of the Catalog development portion of the initiative.
APPROACH
The U.S. EPA Hazardous Site Evaluation Division began this project
in the Fall of 1986. As a first step, EPA contacted the 10 Regional
Offices and interviewed staff who understood the Regions’ overall.
use of field analytical techniques. Based on information obtained
during these interviews, EPA determined that contractor assistance
would be beneficial in the compilation and development of the Field
Screening Methods Catalog. Through the REM II program, EPA
contracted with CDM Federal Programs Corporation and Roy F. Weston,
Inc., to develop the Catalog and data base.
Information collection methods included additional telephone and
on—site interviews by contractor personnel to follow up on data
collected by EPA and review of EPA reports and papers published in
journals or presented at conferences. Contacts included the
Analytical Services Advisory Committee, the Regional Environmental
Services Divisions, Regional laboratory staff, EPA ORD laboratory
staff, and various contractors. ‘1ethods for further assessment were
selected from those identified by the various contacts. Selection
was based on their availability for use in field situations and
their ability to provide at least some quantitative data.
To date, approximately 30 methods that have been successfully used
in the field have been identified and are included in this Catalog.
The Catalog includes only those methods that were identified by
EPA’s initial interviews and for which sufficient data were
available based on contractor followup work. The methods presented
in this Catalog include analyses for metals, volatile and
semi—volatile organics, phenols, pesticides, PCEs, and polvcyclic
aromatic hydrocarbons. Several soil gas sampling techniques have
been included because of the expanding use of such techniques in
assessing contamination.
8-49
-------
Additional methods may exist within EPA, other governmental
agencies, and private industry. Future updates of the Catalog will
include these methods as appropriate.
After reviewing available documentation of each method, project
staff relied on individuals who had worked with each method to
provide additional details about their experiences using the
method. This information was used to fill gaps in information and
ensure accuracy. These and other field and laboratory personnel are
involved in the Catalog review process, which includes commenting on
drafts of the method descriptions and the data base.
The project provides both a pocket guide of the methods and a
computer data base, which are designed to serve the needs of field
staff and managers. For ease of use, a standard set of “fields” has
been developed to organize the information about each method. In
addition, the content of each field is standardized. The types of
information provided for each method include:
• Method name and number (number is specific to this Catalog);
• Suimnary and method description;
• Application, limitations, and instrumentation used;
• Performance specifications, including detection limit,
selectivity, accuracy, and precision/repeatability;
• Use of the method, including location, CERCLIS site number,
and matrix;
• Preparation, maintenance, and cleanup;
• Calibration;
• Analysis time;
• Capital costs; and
• Source of technical information.
In addition, the description of each method includes two “comment”
fields. At least one bibliographic reference is provided for each
method.
The pocket guide provides a concise description of each method,
which allows field staff to consider the range of analytical methods
that might be appropriate for the site while in the field. It
8-50
-------
provides only a brief overview of the method and refers to a source
of detailed instructions for use. The data base would be used in
the office for planning of field investigation and response
activities. It provides search capabilities based on four main
categories (chemical name, chemical class, CAS number, and method
type) and two subcategories (matrix and detection limit). For
example, the user might request information on analytical techniques
for PCBs. To limit the number of PC3 techniques displayed, the
computer system will ask which media are wanted (air, soil, water)
and whether the user wants to set detection limits (e.g., only show
the method if it detects less than 100 ug/kg). Similarly, the user
could ask for methods for volatile organics (chemical class) or for
all GC methods (method type). CAS number can be chosen to ensure
that the correct chemical is selected if confusion among chemicals
with similar names is possible.
The data base was developed using dBase III to run on IBM—compatible
microcomputers. It is available on two floppy disks in a compiled
version, and is menu—driven for ease of use. The system is designed
to prevent tampering or accidental changes or erasures. The data
base includes several options for printing requested information,
ranging from method titles only to the full method description.
STATUS AND PLANS FOR COMPLETION
To date, about 30 field sampling or screening methods have been
included in the Catalog. The methods in the Catalog include several
gas chromatography methods, two x—ray fluorescence methods,
ultraviolet fluorescence, fiber optic sensors, imznunoassay, mass
spectroscopy, and atomic absorption. The soil gas sampling methods
all utilize GC equipment for analysis. The Catalog also includes
variations of several analytical techniques based on sampling media,
such as bonded sorbents, or extraction chemicals (e.g., hexane
versus methanol). Figure 1 presents the titles currently included
in the Catalog.
The Catalog is under review by EPA Headquarters and Regional
personnel who are familiar with the development and use of the
methods. Based on this review, the descriptions may be revised.
The data base will also be reviewed. Both will be in final form,
ready for distribution, by early Fall of 1987. The review and
revision process will continue, however, with new methods being
added and current methods being revised as new data become
available. Several studies that are currently underway should
contribute significant information in the coming year.
8-51
-------
CONCLUSION
The limited amount of data available about precision and accuracy of
these methods in the field caused problems in completing the Catalog
entries. Field personnel using a method usually tested it to
demonstrate the method’s applicability to a particular site, but
often did not perform extensive testing or provide QA/QC
documentation. Several methods have been used at only a few sites.
While limited performance criteria or field experience caused some
problems in completing the Catalog entries, enough information is
available for preliminary evaluation of a method’s ability to meet
specific data requirements. It remains important in every case, no
matter how complete the Catalog entry or method protocol, for the
scientist performing the analysis to develop clear specifications
and operating procedures and to demonstrate acceptable performance
of the method in his or her hands for the site. It is necessary to
define data use and match the analytical method to this end use.
Often, the results of field screening methods should be verified by
CLP or another analytical procedure.
The Field Screening Methods Catalog, which contains approximately 30
sampling and analytical methods, will be available through EPA’s
Analytical Operations Branch. Call Carla Dempsey at (202) 382—7906
for information. This Catalog is a dynamic document; EPA will
update it to reflect the continuing rapid advances in field
screening methods. Suggestions for methods to be added or revisions
to methods included in the current version also may be provided to
EPA through Carla Dempsey.
8-52
-------
FIGURE 1——CURRENT CONTENTS OF FIELD SCREENING METHODS CATALOG
Headspace Technique Using an OVA (FID) for Volatile Organic
Compounds
Headspace Analysis Using HNu (PID) for Total Volatile Organics
Headspace Technique Using a Mobile CC for Volatile Organics
Imxnunoassays for Trace Organic Analysis
Total PNA Analysis Using an Ultraviolet Fluorescence
Spectrophotometer
Use of Fiber Optic Sensors in Environmental Monitoring
Air Monitoring for VOCs Using Programmed Thermal Desorber and CC
Pesticide Analysis Using a GC with ECD——Hexane/Methanol Extraction
Phenol Determination by Liquid—Liquid Extraction and CC Analysis
Use of Bonded Sorbents for Semi—Volatile Analysis
Use of Bonded Sorbents for Pesticide Analysis
VOC Analysis Using CC with Automated Headspace Sampler
PCB Analysis Using a GC in an On—site Laboratory——Hexane
Extraction
PCB Analysis Using a GC in an On—site Laboratory——Hexane/
Methanol/Water Extraction
PCB Analysis Using a GC in an On—site Laboratory——Hexane/Acetone
Extraction
PCB Analysis Using a GC in an On—site Laboratory——Hexane/Methanol
Extraction
Pesticide Analysis Using Isothermal CC with ECD——Hexane Extraction
PAH Analysis Using CC (FID) with Heated Column
X—Ray Fluorescence in Laboratory for Heavy Metals
X—Ray Fluorescence for Heavy Metals (On Site)
Field Atomic Absorption Analysis
Trace Atmospheric Gas Analyzer (TAGA)——MS/MS
Soil Gas Sampling Using a Perforated Tube
Soil Gas Sampling Using Mini—Barrel Sampler
Soil Gas Sampling Using Industrial Hygiene Samplers
Soil Gas Sampling for Downhole Profiling
Soil Gas Sampling Using Direct Injection——Stopper
Soil Gas Sampling Using Direct Injection——Auger
Soil Gas Sampling Using a One—Liter Syringe
Soil Gas Sampling Using Tenax Tubes
8-53
-------
A FIELD DEPLOYABLE ANALYTICAL INSTRUMENT FOR ANALYSIS OF
SEI4IVOLATILE ORGANIC COMPOUNDS OF SUPERFUND SITES
Edward B. Overton, Professor and Director, Institute for
Environmental Studies, Louisiana State University, Baton Rouge,
Louisiana; Steven J. Martin, President, Ruska Laboratories, Inc.,
Houston, Texas
ABSTRACT
This presentation discusses the single instrument thermal extraction
from solid samples of setnivolatile organic substances, such as the
base/neutral priority pollutants, and their subsequent analysis by
high resolution gas chromatography (GC) and/or gas chromatography
mass—spectrometry (GCMS), and the application of this type of
instrumentation to environmental analysis.
Site investigations and cleanup activities under Superfund often
require rapid analysis of samples for trace volatile and semi—
volatile organic compounds as well as heavy metals. The analytical
data are used by response officials for decision making in
characterizing and cleaning up sites. Traditional analytical
techniques require laboratory facilities to extract samples, cleanup
extracts, and instrumentally analyze the extracts. This process is
time consuming and expensive. Analytical data obtained from field
instruments can supplement data obtained from laboratory analysis of
samples. Field instrumentation that allows for the rapid analysis
of samples enables response officials to perform site investigations
and cleanup activities in a more efficient and cost—effective
manner. Field deployable analytical instrumentation for analysis of
volatile organic substances will be discussed elsewhere in this
symposium.
In this presentation we discuss the potential for the use of a
single field deployable instrument, the Pyran Level 2 Analyzer, for
the thermal extraction of solid samples with analysis of the
extracts by combined GCNS techniques. The instrumentation was
originally developed for use in petroleum exploration for analysis
of various source rock samples for trace levels of high molecular
weight petroleum hydrocarbons and NSO compounds. It consists of a
chemically inert all quartz thermal extractorpyrolyzer interfaced to
an all quartz cold trap—injector and cross—linked liquid phase
coated capillary GC column which elutes into either a flame
ionization detector or an ion trap detector (mass spectrometer).
The unit requires only compressed gas (helium and carbon dioxide)
and electrical power, and has been designed for use on oil rigs and
in seismic fans.
We believe that this instrument represents one of the few devices
that can be used for the analysis of environmental samples in the
8-55
-------
field, at reasonable cost per sample, and that has a wide range of
analytical capabilities. The Pyran instrument will be described in
detail and data from the analysis of several types of environmental
samples will be presented. Its applications and limitations in
field analysis will be reviewed.
INTRODUCTION
Effective chemical hazard assessment and mitigative actions for
hazardous waste site evaluations are dependent on a number of
considerations. With regard to a specific site, the following
information is needed:
1) identities and quantities of specific chemicals involved at
the site;
2) the toxic and reactive properties of specific chemicals at
the site;
3) knowledge concerning routes of, and probability of, exposure
to specific chemicals at or from the site;
4) environmental fates and transport mechanisms of the toxic
chemicals at the site.
Virtually all efforts associated with assessing hazards and
developing mitigative plans for chemical releases at Superfund sites
involve both identification and quantitative evaluation of specific
compounds at the sites. This implies a need for analytical
techniques that can provide useful chemical information, In a timely
fashion, from analysis of the various types of samples that are
encountered at hazardous waste sites. Ideally, the analytical
device that is the heart of this “problem solving” Instrument system
would have the following capabilities (Overton 1986):
1) portability or field deployability;
2) ability to identify and quantitate a wide variety of specific
chemical substances;
3) application for analyses of vapors, liquids and solid samples;
4) sensitivitIes to one—tenth the toxic concentrations;
5) rapid analytical response times compared to those available
by sending samples to a laboratory;
6) ruggedness and reliability;
8-56
-------
7) readily available, and simple to operate and interpret;
8) requires utilities that are readily available In field
situations.
Unfortunately, this “idealized” field analytical instrument does not
presently exist. Currently available analytical instruments are
simply unable to provide field deployable compound—specific
detection in a configuration that has all of the capabilities
described above. After due consideration for the information needs
associated with chemic a1 hazard assessments and mitigative actions,
and the analytical capabilities that are realistically achievable in
a timely fashion with field deployable instrumentation, we have
developed the following guidelines for a field deployable analytical
instrument. They include:
o identify primarily organic chemical contaminants;
o primary capability is associated with identifying and
quantitating toxic chemicals in a timely fashion with field
deployable equipment;
o analytical instruments should be broadly applicable without
on—site modifications or methods development for analysis of
vapor, liquid and solid samples;
o chemical data should be equivalent (or nearly so) to data
obtained from laboratory analysis of samples.
Based on these guidelines, we believe a compound—specific analytical
instrument system is needed for use in the field to provide
information on which to base environmental chemical hazard
assessments and mitigative actions. This analytical instrument
system should have the following capabilities:
1) sensitivities to one—tenth the toxic levels;
2) accuracies in compound identification equivalent to analyses
obtained from laboratory—bound instruments;
3) capability to be rapidly deployed at or near the hazardous
waste sites;
4) analytical response times in hours rather than days;
5. ability to be operated by specially trained technicians
rather than graduate chemists;
6. can be used to analyze waste materials and contaminants in
containers, soils and sludges, water and air.
8-57
-------
A great amount of research is being conducted to develop more
sophisticated, small, rugged analytical devices for a variety of
field applications. Much of this work is being conducted for
purposes related to U.S. Department of Defense activities. This
research seems to be divided into two basic efforts. One effort is
attempting to miniaturize existing analytical instruments such as
the gas chromatograph—mass spectrometer (Analytical Chemistry
1987). The other effort Is aimed at developing new solid state
sensors, and arrays of solid state sensors, whose data outputs are
examined by powerful multivariate statistical routines designed to
identify “patterns” in the data (Setter 1986). Virtually all of
these research efforts are aimed toward analysis of volatile
compounds In air samples.
A new field deployable analytical instrument has recently been
developed for the analysis of semivolatile, high molecular weight
marker compounds in source rock samples from petroleum exploration
activities. The Instrument uses thermal extraction of semivolatile
organic compounds into an inert, all quartz apparatus followed by
integrated gas chromatographic—mass spectrometric analysis of the
thermal extracts. This unit, called a Pyran Analyzer, was developed
by Rüska Instruments of Houston, Texas. We believe that this device
has great potential for application at hazardous waste sites to
analyze base—neutral and acid extractable types of aemivolatile
organic compounds. Additionally, using suitable trapping tech-
niques, the device should also be able to analyze volatile organic
compounds. The unit 18 rugged, potentially field deployable in a
van, and can have analytical turn—around times of approximately one
hour for moat types of analyses.
After a thorough review of field deployable analytical instruments
that have evolved over the past 8everal years, we have reached the
following conclusions:
1) Aside from transportable field laboratories, there is no
readily available, field deployable, analytical capability
for analyzing a wide variety of base—neutral and acid
extractable type organic compounds. Even the expensive Sciex
MS—MS analytical device has not proven useful as a general
analytical instrument for analysis of the variety of samples
commonly encountered at Superfund sites (Collins 1986). The
prohibitive coat and complexity of the Sciex further limit8
Its usefulness for most routine Superfund applications.
2) Virtually all conventional analyses of environmental samples,
for base—neutral and acid—extractable type organics
compounds, involve extraction with organic solvents, followed
by C MS analysis using conventional instrumental techniques.
All volatile analyses are based on purge and trap procedures
8-58
-------
except for those that are being developed for NOAA—EPA by our
Instrument Development Group (Overton 1986).
3) The Pyran Thermal Chromatographic System has great potential
for field deployable analysis of most of the sample types
that are encountered at Superfund sites.
4) The Pyran Instrument must be evaluated and, perhaps, modified
to meet the unique requirements for analytical application at
Superfund sites.
5) New procedures may be required in order to convert the Pyran
device into a field deployable, analytical instrument for
Superfund applications.
The Pyran Thermal Chromatograph is an instrument that was
specifically designed and developed to meet the analytical need of
petroleum exploration and development activities. It is a
self—contained extraction system and analyzer that is relatively
compact, rugged and designed for field applications. Figure 1 shows
the essential components of the Pyran Analyzer. It is constructed
of quartz to provide the chemical inertness and stability that is
needed for reproducible thermal extraction and pyrolysis of organic
compounds from source rock samples. Since quartz does not absorb
radiant energy, the quartz construction allows precise temperature
control at both subambient and elevated temperatures using a
computer controlled combination of cyrogenic cooling (liquid C0 2 )
and radiant heating. The Level I analyzer includes a thermal
extraction and pyrolyzer unit that is interfaced directly to a flame
ionization detector. The Level II unit includes a thermal extractor
and pyrolyzer module that is interfaced, with all quartz components,
to a specially designed capillary column gas chromatograph that has
no moving parts. Again, all quartz construction of the
chromatographic oven and column allows precise and reproducible
temperature control and programming from subambient to several
hundred degrees centigrade. The chromatographic effluents are
detected by an Ion Trap Mass Spectrometer and analyzed by
conventional data treatment software. Figures 2 and 3 show data
from analyses of a spiked synthetic sediment sample that was
thermally extracted and analyzed by both the Pyran Level I and II
units. These data confirm our belief that the Pyran Analyzer has
great potential for analysis of base/neutral and acid extractable
type compounds at Superfund sites.
The Level I analyzer thermally extracts organic components and
detects the substituents without any chromatographic separation
using flame ionization detection. It is designed to permit rapid
screening of samples and has analysis times of less than fifteen
minutes. The Level II analyzer has a thermal extraction module that
is interfaced to a GCMS analyzer. The GCNS unit has the capability
8-59
-------
to identify and quantitate specific substances that are thermally
extracted from the sample. The analyses times, including extrac-
tion, are generally on the order of one hour. The Level II analyzer
is designed to identify specific chemicals in a sample and to
measure their concentrations with analytical turn around times that
are significantly shorter than are available from conventional
laboratory GUMS analysis of environmental samples. The Pyran
Level I and II analyzers have great potential to provide analytical
data, in a timely fashion, for use at Superfund hazardous waste
sites. This potential, however, has not been totally proven or
substantiated.
The Pyran un.tt was originally designed for the automated analysis of
solid or liquid samples. In addition to these types of samples, we
believe the unit can be used for the analysis of gaseous samples as
well. Gaseous samples could be analyzed by adsorption onto solid
adsorbents and subsequent thermal extraction/GCMS analysis.
Milligram quantities of solid and liquid samples can be screened by
Level I analysis or analyzed directly by the Level II unit. Special
mass spectral techniques, such as the use of combined electron
impact-chemical ionization procedures, may be useful for use in the
analysis of the thermal extracts (Broadbelt 1987). Also, enhanced
“target compound” data treatment algorithms will undoubtedly be
necessary since the Pyran unit analyzes samples without any chemical
clean—ups.
The analytical potential of the Pyran Level I and II Thermal
iromatograph must be fully evaluated and characterized from a
stand—point of its application to Superfund analytical needs. This
process involves determining the precision, accuracy, recoveries,
detection limits and potential interferences for organic priority
pollutant type compounds in a variety of sample matrices (sludges,
sands, silts, clays, etc. with varying TOC content). This
validation procedure is necessary because the Pyran unit uses
methods of analysis that are significantly different from those used
in conventional laboratory analyses. Conventional laboratory
analyses, described in EPA methods 624 and 625 (Environmental
Protection Agency 1980), involve time consuming liquid—liquid
extractions, optional clean—up of the extracts with liquid—solid
chromatography (silica gel, alumina, fluorosil, etc.), and
instruments analysis with GCMS techniques. The clean—up procedures
are designed to remove interfering compounds, from naturally
occurring sources, before the extracts are submitted to analysis by
GCMS techniques. The Pyran unit, on the other hand, is a fully
automated unit that thermally extracts and analyzes samples in a
single step. It therefore baa much greater sample through—put
capacity than conventional EPA analytical procedures. However It
does not use any type of extract clean—up procedure. Consequently,
all organic compounds, not just the analytes of interests, will be
subjected to its capillary GEMS analysis. The Implication of this
8-60
-------
fact on the analytical results from Pyran Analysis will have to be
fully elucidated. It is obvious that total extraction and analysis
of all components in complex environmental samples will place a
substantial burden on the mass spectral procedures and data
treatment techniques that are needed to turn analytical data into
the identities and quantities of specific sample analytes.
Figure 4 shows data from the Pyran analysis of a “penta”
contaminated coal tar sludge that also contained low ppm levels of
octa and hepta chlorinated dioxins and furans. Examination of the
mass spectral data reveals the presence of numerous polynuclear
aromatic hydrocarbons but only barely detectable levels of
pentachiorophenol. Virtually the entire sample was composed of
organic material. These large quantities of polynuclear aromatic
compounds, and other organic components, made detection of trace
quantities of the chiorocarbons very difficult. Additional methods
must be developed before the Pyran unit can be used to completely
characterize these “difficult to analyze” types of samples.
In conclusion, we believe the Pyran Thermal Chromatographic Analyzer
has great potential for use in many Superfund analytical
applications. This potential should be evaluated more fully because
of the short analysis times and cost savings associated with its
field deployable analyses. Enhanced instrumental techniques, such
as chemical ionization GCNS analysis or daughter ion mass spectral
analysis, may be needed to fully exploit the analytical advantages
of the Pyran analyzer.
REFERENCES
Analytical Chemistry. “Explosives Sniffer Developed,” Anal. Chem .
1987, 59, 565 A.
Brodbelt, J.S.; Fouris, J.N.; Cook, R.G. “Chemical Ionization in an
Ion Trap Mass Spectrometer,” Anal. Chem . 1987, 59, 1278—1285.
Collins, R.V. Personal Communication 1986.
Environmental Protection Agency. 1980. Interim methods for the
sampling and analysis of priority pollutants in sediments and
fish tissue. EPA Environmental Monitoring and Support
Laboratory , Cincinnati, Ohio.
Overton, E. B.; Steele, C. F.; Naumann, S. B.; McKiriney, T. H.;
Kummerlowe, D. “Development of a Field Usable Analytical Device
for Hazardous Chemical Incidents,” Proceedings of the 1986
Hazardous Material Spill Conference , St. Louis, MO 211—216
(1986).
Stetter, J. R.; Jurs, P. C.; Rose, L. L. “Detection of Hazardous
Gases and Vapors: Pattern Recognition Analysis of Data from an
Electrochemical Sensor Array,” Anal. Chem . 1986, 58, 860—866.
8-61
-------
Figure 1 - Schematic diagram of the Pyran Levels I and II Thermal Extractor and
Analyzer. Level I analysis involves thermal extraction of milligram quantities from solid
samples that are placed in small quartz crucibles (8mm OD) in an all quartz pyrolysis
vessel. The extracted organic material is then detected by a flame ionization detector.
Level II analysis involves thermal extraction of solid samples in an all quartz pyrolysis
vessel, cold trapping of the extracted analytes, and flash evaporation of the analytes onto a
fused silica capillary gas chromatographic column. The capillary column is then
temperature programmed, using a combination of radiant heating and cryogenic cooling, to
produce the capillary gas chromatogram. Eluting analytes are detected by an Ion Trap mass
spectrometer.
8-62
-------
LABORATOUIES. INC.
CAPILLARY
GC COLUMN
a 0
‘-70 to 400 C,
PYROCELL
0 0
o to 600 C
SAMPLE LOADER
LEVEL II
TC Th /MS
QUARTZ ANALYZER
. -.. Liquid 002 Cooing
4- Sputter Purge
-*- Splitter Vent
- - Helium Gamer Gas
,IJ
)W DUNVALI. HOUSTON TTAS 5 106 1
?O 010115(110, ,OIJSTON, T(OAi 1(1 54
Level I — FID
Quartz Analyzer
cjo
SPLITTER
COLD TRAP J
—7Oto600C L
-------
Figure 2 - (top) Level I thermal extraction profile (FID) from the Pyran analysis of two
spiked synthetic sediment samples supplied by EPA; (middle) Total Ion Chromatogram,
from scan 700 to 1200, from the Level II Pyran Analyses of the EPA #1 spiked synthetic
sediment sample; (bottom) Total Ion Chromatogram, from scan 1200 to 2000, from the
Level II Pyran analysis of the EPA #1 spiked synthetic sediment sample.
8-64
-------
0
C
0
N
C
0
.
naphthalefle
phihatate
0
a
e
2
.C
I
.C
0
—
jI
E
[ . T I I II I I I I I j I ll J Il 1(111 II lI_ lj
100 127
788 808 908 1080 153 C
11:41 13:21 15:81 16:41 18:21 20:01
5C
i Li 1 1.1U05
02 45 o0 4 2 0Q
40
30
20
>
E
t J
U )
z
0
0
U)
U i
0
I-
0
U i
I—
Ui
0
I (I
a
phenol
IDlE (mm)
iGOx
101
trlchlo 0ph enol
0
•0
- _j
pyrono
MW 202
8-65
-------
Figure 3- (top) Total Ion Chroniatogram from the Level II Pyran analysis of EPA #2
spiked synthetic sediment sample; (middle) Total Ion Chromatogram from the Level II
Pyran Analysis, fmm scan 1500 to 1800, of EPA #2 spiked synthetic sediment samples;
(bottom) Extracted Ion Current Profiles (MJe 290 to 294,324 to 328, and 258 to 362)
from the Level II Pyran analysis of the same spiked sediment sample. These data show
detection of PCB’s in the sample.
8-66
-------
chrysane
641
13:21
dIn toto lUefle
12
324 A
I \ .._4 V\ f\j
31
358
to
362
—
hexach loro
r
TO’,
C EO)
PCB s
J uw
210
16 260C
26:41 33:21
PC .s
TOT
C}IRO)
1 etrac dOtO
1?5
29:11
_ _i I I I’ I I I I I
155 16 165 1788 1758
25:51 26:41 27:31 28:21 29:11
CR RO)
8-67
-------
Figure 4 - (top) Total Ion Chromatograph, from scans 1400 to 1900, from the Pyran
Level II analysis of a “penta” contaminated coal-tar sludge sample; (middle) mass
spectrums of the large peak at scan 1640 (note presence of ions at 268, 266 and 264 from
low levels of pentachiorophenol that coeluted with phenanthracene and anthracene;
(bottom) Extracted Ton Current Profiles (MIe 268, 266 and 264) from the Pyran Level II
GCMS analysis of the contaminants sludge samples shown in top and middle portion of
this figure.
8-68
-------
266
I 3 ’
14
15 Q
16
17O
18O
19
23:21
25: 1
26:41
28:21
3 :Q1
31:41
16
1 9
5163 8? 111 126
196
1øø
268
8-69
-------
DEVELOPMENT OF SAMPLING/MONITORING GUIDANCE
FOR THE RCRA HAZARDOUS WASTE REGULATORY PROGRAM
Gregory R. Swanson, Project Manager, Seth H. Schulberg, Project
Engineer, S—Cubed, San Diego, California; Ivan T. Show, Senior
Statistician, IWG Corporation, San Diego, California
ABSTRACT
Under the RCRA Hazardous Waste Regulations, it Is the responsibility
of waste generators to evaluate the wastes that they generate to
determine whether their wastes exhibit any of the hazardous waste
characteristics, and whether the concentration of toxic constituents
exceeds any regulatory thresholds established within the revised
hazardous waste listings. Wastes generated by industrial processes
are variable in their composition and volume, depending on the
nature of the product, the process, and various operational
factors. As a result, one sampling and analysis of a waste is not
sufficient to permanently identify a waste as hazardous or
non—hazardous. Rather, it is essential that a continuing sampling
and analysis, or monitoring, program be implemented by industrial
waste generators to identify whether specific wastes are hazardous
or non—hazardous over the lifetime of their generation.
This paper presents the results of an investigation into statistical
methods for determining the number and frequency of samplings needed
to adequately identify a waste as hazardous or non—hazardous over
the lifetime of its generation. These statistical methods were
supplemented with judgement factors and practical limitation to
develop a recommended approach for EPA/OSW to adopt as guidance
within the SW—846 Manual. The variability of industrial, processes
and the associated waste that they generate are considered as an
integral part of this recommended approach. Other aspects of
sampling/monitoring programs designed to fulfill regulatory
requirements, such as grab versus composite sampling, are also
discussed.
INTRODUCTION
Under the RCRA Hazardous Waste Identification Regulations (40 CFR
261), it is the responsibility of the waste generator to evaluate
the wastes that he generates to determine whether these wastes are
hazardous. Specifically, he must test each waste to determine
whether it exhibits any of the hazardous waste characteristics or
whether the concentration of toxic constituents in the waste exceeds
any concentration—based regulatory thresholds which may be
established within the hazardous waste listings.
8-71
-------
Wastes generated by industrial processes are variable ía their
composition and volume, depending on the nature of the product, the
process and various operational factors. Adlitionalif, the
me urement of waste characteristics or components is subject to the
unavoidable measurement variability inherent in the sampling and
analysis methods employed. As a result, one sampling and analysis
of a waste is not sufficient to permanently identify a waste as
non—hazardous. Rather, it is essential that a continuing sampling
and aflalysi8 program be implemented by industrial waste generators
to identify specific wastes as non—hazardous over the lifetime of
their generation.
The ultimate objective of this investigation was to provide guidance
to industrial waste generators to assist them in designing and
optimizing their monitoring programs. This objective was met by
developing an effective statistical method for estimating ho i
frequently a waste must be sampled to adequately characterize it as
non—hazardous. The statistical method was then supplemented with
judgement factors and practical limitations to form a preliminary
recommended approach for EPA/OSW to adopt as guidance within the
SW—846 Manual.
Refinement of the method is continuing based on input and comments
by EPA and others; the final guidance is scheduled for completion in
September, 1987.
REQUIREMENTS OF THE GUIDANCE AND GENERAL ASSUMPTIONS
To facilitate the selection of an appropriate statistical method for
estimating sampling frequency and the development of the associated
sampling/monitoring guidance, It was first essential to list the
general requirements and desirable elements of the guidance. These
were listed as follows:
o The guidance should be oriented towards defining how
frequently a generator must sample to confirm that their
waste Is non—hazardous. There is no minimum frequency for
Identifying a waste as hazardous.
o The sampling frequency should be designed to establish that a
waste is non—hazardous with a high level of con—
fldence.Therefore, greater waste variability should lead to a
requirement for more frequent sampling.
o ProxImity to regulatory thresholds is an important
consideration in the establishment of an appropriate sampling
frequency (i.e., wastes near a regulatory threshold need. to
be sampled more frequently).
o Volume of waste generated may be an element worthy of
consideration In estimating an appropriate sampling frequency
8-72
-------
(i.e., it may not be reasonable to request small—quantity
generators to sample as frequently as large quantity
generators).
o Statistical procedures for determining sampling frequency
should be simple in their presentation and may need to be
supplemented with judgment factors. The ability of the
regulated community (industry) to Implement the approach Is
an important cousideration.
In order to narrow the scope of this irlve8tigatiort and make the
development of the required guidance a manageable task, it was
essential that a series of general assumptions be adopted with
respect to the nature of industrially—generated wastes and the
hazardous waste regulatory framework. The general as aniptions that
were adopted include:
o A waste should be managed as hazardous only when the mean
waste value exceeds a regulatory threshold (a sequential or
weighted mean would be appropriate for a time series of
measurements).
o Industrially—generated wastes follow an approximate normal
distribution in the variability of their characteristics and
constituent concentrations under routine or continuous
operating conditions. Start—up or changeover conditions can
only be considered as separate events for the design of an
appropriate sampling/monitoring program.
o Each waste has a principal potentially hazardous
characteristic or constituent concentration for which a
regulatory threshold is established in the hazardous waste
regulat ions.
o It is not practical to separate measurement (sampling and
analysis) variability from process variability in the
statistical approach to estimating sampling freqtieucy. The
variability of any sampling/monitoring data must be
considered as combined process and measurement variability.
GENERAL ELEMENTS OF THE STATISTICAL METHOD
Based on the requirements and desirable characteristics of the
method previously developed and explained, the following additional
criteria were developed as guidelines for defining frequency of
sampling within the statistical method:
o For the statistical method to be useful, it must cause an
Increase in sampling frequency in response to (a) the mean
8-73
-------
waste concentration approaching or surpassing a regulatory
threshold or (b) large variability within the data set.
o Fre’quency of sampling should be based on trigger values that
signal an increase in sampling frequency (i.e., reduction in
the sampling interval). For instance, a trigger might be the
po tnt at which the probability is 90% that the sample mean is
greater than 75% of a regulatory threshold.
o A conclusion that a waste is hazardous should not be based on
a single concentration in excess of a regulatory threshold.
Such an event should, however, trigger more frequent
sampling, at least until sampling results return to a safe
zone.
o Cumulative averaging should be employed to focus on the
recent behav ioc of the waste generating process. This
procedure also reduces the probability of the pertinent
sample parameter exceeding a threshold by chance alone.
However, the autocorrelation among neighboring sample means
needs to be accounted for in any statistical tests.
o The method for determining sampling frequency should take
into account recent advances in understanding of sampling
variance components in the region of thresholds. This is a
different problem from approaching a method detection
threshold where measurement values are constrained or
truncated by the logical impossibility of values less than
zero.
DETAILED STATISTICAL METHOD
The above description of the desirable characteristics of the
statistical method could lead to any of a large number of
statistical techniques. The following techniques were considered
for determining sampling frequency:
o A modified Shewhart Chart (Control Chart) with the entries
normalized for sampling variance and based on at least two
consecutive means out of the safe zone to trigger increased
sampling beyond a calculated base level;
o PredictIon Intervals based on standard linear least squares
regression techniques;
o Time series based on Box and Jenkins methods, lag k—sample
autocorrelation analysis, semivariograms, or systems feedback
control models;
8-74
-------
o Non—parametric methods such as trend analysis, contingency
analysis, concordance, or runs tests.
The method that was selected is based on a modified Shewhart Control
1art. Methods used to construct the control chart are related to
linear least squares prediction intervals and incorporate certain
aspects of Box—Jenkins time series methods as applied to systems
feedback control. A complete mathematical description of the method
is presented in Appendix 1. A brief summary of the method is
provided below along with a description of the logic (see Figure 1).
The method is based on the calculation of a sequential mean of all
sample results that is weighted toward the most recent result. This
weighted mean, is calculated at each sampling tine and used in
direct comparison to the pertinent regulatory threshold, L, to
determine whether the waste is hazardous or non—hazardous (see
Figure 1).
The heart of the method incorporates two statistical t—tests (at
significance level ) to evaluate whether:
a) the weighted mean exceeds a defined percentage, P, of the
regulatory threshold (test of approach to the regulatory threshold),
or
b) the change in the weighted mean from the previous sampling, gw,t —
X is greater than a preset maximum, c (test of excessive
variance).
If either these two tests are not passed, then the sampling interval
is reduced by a constant factor. If both of these tests are passed
in consecutive time, then the sampling interval is increased by a
constant factor.
METHOD PARAMETER VALUES AND LIMITING CONDITIONS
Prior to applying the method, it is first necessary to establish the
value of various parameters of intended constant value that are
contained in the method and to set boundary conditions with respect
to minimum and maximum sampling intervals. The large number of
parameters in the method provides for a high degree of flexibility
in application to specific requirements but also requires the
application of well—founded judgement to establish appropriate
values for these parameters. Practical limitations dictate the
requirement for a minimum and maximum sampling interval to avoid the
potential extremes of a purely mathematical equation.
Suggested values for the parameters in the equations are provided
below. (See Appendix 1 for a description of the development of
8-75
-------
xt
Yes
.1
No
FiGURE 1. Logic Diagram for the Statistical Method
8-76
-------
these parameters and the associated equations). These suggested
values have been selected based on limited numerical simulation of
the method, real world practicality, and engineering judgement.
Suggested
Symbol Description Value
L Regulatory Threshold
P of Regulatory 0.68
Threshold (action
trigger level)
m Number of repeated 5
values which triggers
a decrease in sampling
frequency
Mean weighting factor 0.5
c Maximum acceptable 0.1L
deviation in sequential
sample means.
h Sampling frequency 0.25
regulating factor
The diversity of waste generators makes it impractical to propose
one set of maximum/minimum sampling intervaLs or significance
levels that would fairly apply to all generators. What is
reasonable for a high volume waste generator may be too stringent
for the low volume generator. Recognizing this fact, we propose
breaking down the world of waste generators into four categories
based on volume of waste generated. Each generator category would
have associated with it specific minimum and maximum sampling
intervals as well as a specific significance level to apply within
the two t—tests. Trie table below lists the authors 8uggested values
for the four recommended waste categories:
Waste Volume Sampling Interval Significance Level
Category Mm. Max. ( )
(Kg./Mo.)
100 2Yr. 2Yr. NA
100 — 1000 3 Mo. 2 Yr. 0.05
1000 — 10,000 1 Mo. 1 Yr. 0.10
10,000 — 100,000 1 Wk. 3.Mo. 0.2
Under this system the small volume waste generator C < 100 kg/mo.)
is exempted from the rigors of the statistical method. The low
8-77
-------
volume of waste represented by this category does not justify a
sophisticated monitoring program. The remaining categories and
their sampling intervals are self explanatory.
In considering the suggested 8lgflificanCe levels above, it is
Important to understand that the intent of the system is to reduce
the burden on the smaller generator. Thus, for example, it must be
demonstrated to a higher level of confidence that the smaller
generators weighted average waste value is in excess of the trigger
value. This higher confidence level requirement makes it less
likely that a given sample will force the smaller generator to
increase his sampling frequency. Conversely, the lower confidence
level (higher significance level) specified for the larger generator
makes it more likely that a given sample result will trigger an
Increase in sampling frequency.
Before any statistical. analyses may be performed, it is also
necessary to have some minimum quantity of data with which to work.
In order to allow use of this method as quickly as possible, the
authors suggest that at a minimum three samples be collected at the
minimum applicable sampling Interval. The mean of these three
samples can than be used to initiate the statistical method. If the
generator has valid historical data, It may be used. This
three—sample minimum is also applicable in the case of start up of a
monitoring program or for remonitoring a waste stream after a
significant process change.
METHOD PERFORMANCE
To test the method that was developed on a preliminary basis and to
evaluate tentative optimum values for the varIous parameters within
the method, a number of computer data simulations were run. The
results of these simulations, involving thousands of data points,
suggested the parameter values previously listed. The results of
two of these simulations are presented graphically In Figures 2—4
and 5—7, respectively to illustrate the functioning of the method
and to present the control chart concept for graphing the two test
statistics. Both examples were calculated based on a high volume
generator. In both simulations, it is assumed that the regulatory
threshold is 10 PPM and PL is 7.5 PPM. The graphs also depict only
every fourth data point to Improve graph readability.
The first example (Figures 2—4) is of a steady state process that
demonstrates a low degree of variance and a relatively stable mean
as shown in Figure 2. As can be seen, the mean is generally below
the PL of 7.5. As long as the mean is below PL, any data points in
excess of PL can be attributed to random variability and are not
statistically significant.
8-78
-------
Using this data to perform the two T tests described in the
statistical method produces the graph howa in Figure 3. This
figure graphs the T statistic (Ti) or the hypothesis that the mean
value of X is greater than PL and that the weighted average of the
previous data is significantly different from the weighted average
generated from the most recent data. Th confidence levels shown on
the graph correspond to the critical values of T, appropriate for a
high volume generator. The T—statistic graph indicates that in all
but three cases the mean does not exceed PL with any degree of
statistical significance. The three cases over the 95% confidence
limit signify that a generator who produced this data would have to
increase his sampling frequency.
The graph of the second T—statistic (T2) indicates that the
sequential sample means do not differ significantly iii most cases.
The graph shows 1 case where the difference in sequential sample
means may have been significant. This data point would trigger an
increase in sampling frequency based excess sequential sample mean
variability.
The third graph (Figure 4) shows the change In sampling Interval in
response to each new sample. From Figure 2 it can be seen that
initially most of the data Is below PL, thus no decrease in sampling
Interval is required, rather as sets of 5 samples below PL are
collected the sample interval increases to the maximum allowable
Interval. As the sample mean and the individual data points start
to exceed P1, the T—statistic triggers a corresponding decrease in
sample Interval.
The second example data set is shown in Figure 5. The first half of
this graph shows a data set with a gradually increasing mean and a
high degree of variability that eventually exceed PL. The second
half of the graph illustrates a data set with gradually decreasing
mean and variability.
Figure 6 provides the graph of the T—statls tics associated with this
data set. As can be seen from the graph there are 6 T—statistic
data points that Indicate with a high degree of probability that PL
has been exceeded. These 6 points correspond to the region iii
Figure 5 where the mean does rise above PL. In only one case does
Figure 6 indicate that the sequential means varied significantly.
From this graph one would expect the 6 points that exceeded PL and
the 1 point In(licating high variability In the sequential mean
should trigger a reduction in sampling interval. Indeed, as shown
in Figure 7, this is the case, as the mean exceeds PL, the sample
Interval does drop. As the sample mean and variability decrease the
sample lengthens until it stabilizes at the maximum interval.
8-79
-------
15
FIGURE 2. Example 1, x and Mean x Versus Time.
E
0.
0 .
10
5
0
Time (Arbitrary Units)
8-80
-------
t—Statistic. for Mean Xt>PL = 0
t—Statistic for
CL= Confidence Level
t—test a
2 Critical Values 0
CL=O.975
a
CL=O.950 a
. a
a a a
U
a a• a
U a
• a
U •
• • a • a a
a a a a a Us a a
• •• U a a
U
Time
FIGURE 3. Example 1, T-Statistic Versus Time.
r’ •‘
o-
-------
FIGURE 4. Example 1, Sampling Inte,val Versus Time.
365
I - .
244
>
—
C
122
0
Time
8-82
-------
15
• xt
0
Mean X
.
10
S 0
0 • 0 0
S 0
Time (Arbitrary Units)
FIGURE 5. Example 2, x and Mee’, x Versus Time.
-------
t—Statistic for Mean X >PL = 0
V -StatistIc for X SX ,t-i
CL= Confidence Level
t—te St
2. - Critical Values
—. .-CL= 0.975
a ‘ CL=0.95O
a
1-• * a S
a U U S U a
U U • U U
U U
a U •Ul S U U U
S • U U • S. U U
U • 0 U a • us a
S a
0 --e .o- e--- v -s
Time
FIGURE 6. Example 2, T-Statistic Versus Time.
8-84
-------
FIGURE 7. Example 2, Sampling Interial Versus Time.
I-
I
S
>
‘ I
V
S
C
a
C
a
a
a
E
S
C l )
365
244
122
0
Time
8-85
-------
APPENDIX I
Mathematical Description of the Statistical Method
In the method, we specify the regulatory threshold (L) and the percentage of the
threshold that may not be exceeded (P). We also select a parameter X (X:O
-------
r 1 = - ) [ x i - (5)
r = r 1 /r 0 (6)
4. Calculate the sequential variance.
s 2 ) = s 2 [ ) [ i + 2 ( r*1 ) r] (7)
5. Calculate the following test statistic.
( -PL )
T = “ i (8)
Js2 )i
If T 1 > the one-sided critical value for the student t-statistic at significance a, then it is
inferred that exceeds PL.
We now calculate a parameter (8) that will be used for a statistical test of excess
variability in the time series of x .
6. Calculate z = (r 0 + 2r 1 )/r 1 . (9)
7. 8 can be obtained by solving for the roots 8 of the equation
(1 — 8 — z(1 — 8) + z = 0 (10)
or
82 + (Z — 2) 8 + 1 = 0 (11)
o
-------
*This is done by setting a = 1, b = z —2, c = 1 and then choosing the root.
_ —baib 2 —4ac 2
8- 2a (1)
forwhich —1<8<1
8. Calculate the following variables based on deviations in between time step n
and n— i.
= — (13)
s 2 [ aj=—r 1 is (14)
where r 1 was calculated in equation 7.
9. Calculate the following test statistic.
T 2 = — c (15)
Js 2 [ at) ‘ ‘
Where c = a maximum acceptable deviation (li t — I). If 1> the two-sided
critical value for the student - t statistic at significance a1 then it is inferred that the
change from to is excessive.
10. exceeds PL (step 5) or a exceeds c (step 9) then
(M)r ••i = (1-h)(At) (16)
where (At) 1 , +1 is the time interval between the n and n +1 sampling periods and 0
-------
11. If neither test statistic (steps 5 and 9) exceeds its critical value for m time steps,
the time interval between sampling periods in lengthened by defining (M . ) =
(1 +h)( t) .
12. The process is iterative and is performed after each sampling period. Therefore,
first update the following variables:
+ {ftt) 1 —> tr •l
1—8 —> X
then return to step 1.
8-89
-------
RAPID FIELD ANALYSIS OF VOLATILE ORGANIC COMPOUNDS
IN ENVIRONMENTAL SAMPLES
UTILIZING A MICROCHIP GAS CI OMATOGRAPH
Edward B. Overton, Professor and Director, Torn H. McKinney, Charles
F. Steele, Edward S. Collard, Robert W. Sherman, Research
Associates, Institute for Environmental Studies, Louisiana State
University, Baton Rouge, Louisiana
ABS TRACT
A microchip gas chroinatograph wIth high resolution capillary columns
has been linked to an external personal computer allowing for rapid
analysis of environmental samples in the field. The instrument can
be used for the analysis of volatile organic compounds in air,
water, and soil samples during site Investigations and cleanup
activities. The time required for analysis is approximately two
minutes. A modified, temperature independent Kovats index system in
which the retention times of the reference compounds are calculated
at ambient temperatures rather than measured eliminates the need for
standard gases in the field and minimizes the need for temperature
control. Telemetry techniques allow the microchip gas chroniatograph
to be operated in the field by a technician while the
chromatographic data are analyzed by a chemist in a central
laboratory if necessary.
A portable sorption tube concentration device has been developed to
increase the overall sensitivity of the analytical procedure. The
concentrator may be used to analyze water samples for volatile
orgariics using a “purge and trap” technique.
The microchip gas chromatograph has been interfaced with an Ion Trap
Detector. This allows the speed and resolution of the microchip gas
chromatograph to be combined with the compound Identification
capability of a mass spectrometer.
INTRODUCTION
Cleanup activities under Superfund involving both emergency removals
and remedial actions require the analysis of samples for volatile
organic compounds. Analytical data obtained from field instruments
can supplement data obtained from laboratory analysis. Field
instrumentation that allows for rapid analysis of air and water
samples enables officials to perform site investigations and cleanup
activities in a more effective and cost—efficient manner. Field
instruments in common use today, such as the portable flame
ionization and photoionization detectors, are capable of vapor
detection but lack inherent analytical powers that can be found in
more sophisticated instruments.
8-91
-------
DIS JSSION
The Micromonitor 500 (MM500), manufactured by Microsensor Technology
Inc., Fremont, California, possesses the Inherent analytical
capability to greatly enhance the rapid acquisition of reliable
chemical data in field situations (Sadat and Terry 1984). The
analytical capabilities of the instrument are based on compound
separation with high resolution capillary columns and detection with
non—specific thermal conductivity detectors. The Micromonitor 500
contains several high resolution gas chromatographic modules with
different column lengths. Vapor samples are drawn into the
Instrument by an internal vacuum pump. The sample gas is then
routed to the appropriate gas chroinatographic modules and specific
volatile components are separated, detected, and then identified by
the instrument ’s internal computer system.
We have made modifications to the instrument to increase its
effectiveness for field environmental analysis and to overcome some
of its limitations as it comes from the factory. In addition, we
have developed a portable concentrator to increa8e the overall
analytical sensitivity. As it comes from the factory, the MM500 is
designed to be portable and has the capability to detect and
identify several dozen preselected chemical vapors (Wohltjen 1984).
The instrument is completely self—contained with an internal battery
and carrier gas supply. The compounds which each commercial version
of the Instrument can identify are selected by the user and preset
during construction at the factory. The instrument will detect up
to 10 of the preselected volatile chemicals during any one
analytical cycle. If one of these 10 gases is detected, the results
of the analysis will be displayed on a LCD readout in a manner such
as “BENZENE DETECTED 85 PPM.” The Instrument detects volatile
chemical by first predicting their retention times at the ambient
temperature and then integrating chromatographic peaks which elute
at the anticipated retention times.
Even considering the powerful analytical capabilities inherent in
the MM500’s design configuration utilizing microchip gas
chromatographic modules with high resolution capillary columns, the
instrument has several limitations that impact its ability to be
u8ed during responses to chemical releases. These limitations are
as follows. Commercial versions of the devise can only detect
compounds which have been preselected and can be identified by
parameters stored in its internal ROM library. Large quantities of
other vapors could be present but would go undetected by the MM500.
Further, the presence of common fuels which contain many Individual
components, such as gasoline, will confuse the MM500’s computerized
interpretation of data. Also, large quantities of a preselected
vapor will go undetected because when the capillary columns become
overloaded the retention times of the preselected compounds will be
outside of expected retention time windows. Finally, detection
8-92
-------
limits are too high by factors of ten to one hundred for commercial
versions of the device when they are ‘ised to detect toxic levels of
hazardous chemicals at or below TLV concentrations.
In order to circumvent these technical limitations and to take full
advantage of the considerable analytical capabilities resident in
the 45OO, we have developed methods to control the instrument and
retrieve raw chromatographic data with an external microcomputer.
The acquired data are processed by special software that allows for
rapid chemical analysis of vapor samples.
Our development efforts take advantage of the unique and
considerable computational powers of the Apple MacintoshTh
microcomputer linked, through a RS232 interface, to the MM500. The
software we have developed provides three basic modes of operation
f or the MM5OO—Macintosh 1 ensemble. Initial versions are designed
primarily for qualitative analysis while subsequent software will
provide capabilities for both qualitative and quantitative
analysis. All of these operational modes rely on the MM500 as the
analytical probe with data treatments and reductions done in the
MacintoshTh. The connection between the devices is bidirectional
bit serial giving the capability for the instrument control and data
transmission. The RS232 ports can be linked by a variety of methods
including twisted pair cables, modems using telephone lines,
radlofrequency modems, or any combination of these. Figure 1 is a
schematic diagram showing the various connections that have been
used to link the MM500 to a MaclutoshTh. These various connections
can be used to allow data to be collected in the field using a MM500
as the analytical probe and, shortly thereafter, to be interpreted
by a chemist at an analytical laboratory situatcd anywhere in the
country. This capability means that the analytical sensor can be
operated by a field technician with minimal training in chemical
analytical methods. The data can be interpreted as sites remote to
the analytical device by scientists with special training in the use
of chromatographic analytical methods.
The first of the three analytical modes of operation provides access
to all of MM500’s standard procedures using the external
microcomputer. A picture of the front panel, similar to that shown
in Figure 2, is displayed on the computer’s cRT. By moving the
cursor over buttons on the front panel’s display and clicking the
mouse, the operator can perform the same functions that could be
performed by actually rushing buttons on the front panel. However,
use of the Macinstosh external to the NM500 allows for the
analysis to be performed by a trained scientist situated at
locations remote to the analytical sensor. For example, the 1N5OO
can be placed in the field, convenient to sampling locatIons, and be
operated from a command post or analytical laboratory.
8-93
-------
Micflromonitor telemetry
Michromonitor Macintosh
2. Phone lines
modem
phone lines
RS-232 lines
modem
RF links
Macintosh
Michromonitor
Figure 1. Various ways to connect the MM500 with an
external microcomputer.
Conventional
R -232 line
Michromonitor
3. RF links
modem
Macintosh
Michromonitor
4. Phone lines and RF links
modem
modem
Macintosh
8-94
-------
Figure 2. CRT display of MM500s front panel.
8-95
-------
The second mode of operation for the MM500 package allows the devise
to analyze samples and dump raw digitized chromatographic data to
the Macintosh”. The data are converted into displays of the gas
chromotograms that were produced by the instrument ’s analytical
modules. These types of displays allow peak shapes, elution
profiles, peak areas and instrument drift to be evaluated. Figure 3
shows copies of the RT display for the chromatographic profiles
obtained from the analysis of the same sample by two different
analytical modules. The displays show the outputs from the high,
medium, and low gain amplifiers of the signal from the thermal
conductivity detector on each analytical module.
The third mode of operation is the most useful analytical package
but also the most complex. The MM500 analyzes a sample with a user
selected analytical module and stores data on the retention times
and peak areas of all peaks in the chromatogram. The ambient
temperature is also measured and stored. These data are transmitted
to the MacintoshTh. The software uses information on the retention
times of the reference peak (air peak) and ambient temperature to
calculate expected retention times for normal hydrocarbon standards
as if they had been analyzed concurrently with the actual sample.
Using these calculated retention times, for ethane (C 2 ) through
undecane (C 11 ), the software then calculates K.ovats retention
indices (Kovats, 1958) for all peaks in the analytical run. A
report, similar to that shown in Figures 4 and 5, is then displayed
on the MacintoshTh. The report tabulates retention times, retention
indices, peak areas, and the name of compounds in the retention
library that have the closest Index to the calculated Index for each
peak. A second part of the report shows a display of the compounds
in the library that have retention indices within 10% of the
retention index for each peak with compound names highlighted if the
retention index is withIn 1% of the peak’s index. A histogram,
resembling 9 gas chromatograms with very narrow peaks, is also
displayed as an aid for interpreting the analytical data. This mode
of operation provides data that can be used primarily for
qualitative identifications although peak areas are also given. It
is Important to emphasize that identifications based on gas
chromatographic retention time or index comparisons are limited by
the resolution of the chromatographic columns. Care must be
exercised when interpreting gas chroisatographic analytical data to
include as much information as possible concerning the origin of the
sample as well as any other facts that may aid accurate qualitative
interpretation of chromatographic data.
Each chromatographie module in the MM500 is designed to analyze
chemicals within a certain range of volatities. The most volatile
components are separated and detected with a module which uses a
four meter capillary column. The least volatile components are
analyzed on a module using a 0.5 meter capillary column. The range
of volatilities for compounds appropriate for analysis with the
8-96
-------
fl
tn trument
OE I
ir t columr I
R
E
S
p
0
N
S
I
FOe Instruments
GQsoHne columr% 2
_________________________ G soIine co!umn 2 _________________________
R
E
S
P
0
N
S
E
V. .... ...... . . -.
9 . -‘ ..< > 4) 4 - ,.
.. :- ‘ ..
Figure 3. CRT displays from the MM500’s analysis of a gasoline sample using
a 4 m by 0.1 trim ID DB 1701 column with 0.5 micron film thickness (top), and a
2 m by 0.1 mm ID DB 1701 column with 0.5 micron film thickness (bottom).
8-97
-------
ö File Instruments
Sample #1 t
k
A
R
E
A
0 5 10 15 20 25 30 35 40 45 50 55
Time(sec)
0983128:1
Somple #1
394.66 Sulfur Dioxide
400.00 Butane
409.48 1,3—Butadiene
462.11 Uater
472.70 Hydrogen Cyan I de
475.36 Ethylene Oxide
479.94 Ethyl Chloride
500.00 Pentane
504.18 Trichlorofluoromethane
555.39 1, 1—Dith1oroethyIer e
600.00 Hexane
——
Temperature 300.41?
Ref RI 5.77039
20 Peaks
RT INDEX AREA
5.7704 548692000.00
8.3913 399.01 697957.10
10.7315 461.42 275693.70
13.0248 498.60 1246971.00
14.4289 515. 19 73014.00
15.8330 531.15 445508.00
Figure 4. CRT display from MM500s analysis of a sample of
unknown content with bar plot shown in the foreground, using
a 4 m by 0.1 mm ID DB 1701 column with 0.5 micron film thickness.
8-98
-------
* File Instruments
Sbmple #L
4
Temp.ratur . 300.4 17
e1 RT * 5.77039
20 Peaks
Wc ter
Pen tant
Trlthlor’ofluoromethane
1, 1—Dich oroethy%ene
1, 1—Di MoroethyIene
Hexane
:ex ane
Trar —1,2—Dithloroeth Ie
Trans—1,2—D th1oroeth e
1, 1—DieMoroethone
1, 1, 1—Trichlor-aethane
k
A
AT INDEX AREA
5.7704 548592000.00
8.3913 399.01 67957.10 Butane
10.7315 461.42 5610123.00
_____ — 13.0248 498.60 1246971.00
14.4289 516 ,19 73014.00
15.8330 531.15 445508.00
20.2793 567.57 2981242.00
22.7130 583.00 1507473.00
26. 5444 603.85 440777.50
28.7974 613.81 29760.00
31.7927 626.23 113054.00
34.5073 635.30 1428984.00
400.00 48.0801 675.57 491712.00
462.11
472.70
475.36
479.94
500.00
504.18
555.39
600.00
.r)o n _________________ __________
Figure 5. CRT display from MM500s analysis of a sample of
of unknown content with analytical report window shown in the
foreground, using a 4 m by 0.1 mm ID DB 1701 column with 0.5
micron film thickness.
0 5
098:
Ethylene Oxide
Eth I Chloride
Pen tone
Tr I chlor ’ofluor ’omethane
1, 1—Dichloroeth tene
Hexane
.i ... t
Sulfur Dio
butane
1,3—Butadi
Mater
.._ _s_ —
8-99
-------
MM500 encompasses chemicals that have a room temperature vapor
pressure of approximately 1 mm of Hg and higher. This range
includes all of those compounds that are classified by the EPA as
volatile priority pollutants.
Table 1. contains the retention times and retention indices for
several representative vapors that were analyzed with a single
analytical module at three different temperatures. These
temperatures included room temperatures plus and minus 20°C.
Examination of this data reveals that retention times can change by
a factor of 200% for every 20°C change in operational temperature.
This means that retention times cannot be reliable used to provide
qualitative information on the identities of specific peaks. On the
other hand, retention indices change by less that 3% for every 20°C
change in operation temperature. These temperature independent
retention indices can be readily converted to meaningful qualitative
information about the volatile components in air samples.
Because the MM500 uses an extremely small sample loop (80
nanoliters) to collect vapors for analysis, its minimum detection
limit is approximately 10 ppm for most compounds. These detection
limits are too high for many environmental applications. However,
we have circumvented this limitation by designing a field deployable
sorption tube concentration device. This device is used to
concentrate volatile organic chemicals in large sample volumes
(200—400 ml) on ambient temperature sorption traps and desorb the
volatile organics at 250°C into much smaller volumes (2 ml). The
device Is outlined schematically in Figure 6.
This concentrator differs from others which are commercially
available in that the same very small Tenax traps are used
repeatedly, providing reproducible trapping characteristics. The
time required for sample concentration Is on the order of five
minutes. The recoveries of several volatile components are given in
FIgure 7 and Table II.
In addition to concentrating volatile components in air samples, we
have developed a technique to concentrate volatile components in
water and sediment samples (purge and trap technique) in the field
using the same portable concentrator. This technique and some
preliminary data are shown In Figures 8 and 9 and Table III.
The microchip gas chromatograph has been interfaced with the
Finnigan Model 700 Ion Trap Detector (Finnigan MAT, San Jose, CA)
allowing the speed and resolution of the microchip gas chromatograph
to be combined with the compound identification capability of a mass
spectrometer as shown in Figure 10. The two instruments were
interfaced by using a short length (2—5 meters) of capillary column
(OB 1701, 0.1 micron film thickness) to connect the gas
chromatograph module to the open split interface of the Ion Trap
8-100
-------
Table 1. RetentIon Indexes calculated at three temperatures
RT low RI low RI med RI med RI h RI h $
Temp(K)--> 280 300 320
HEXAt ’E 2.991 600.00 2.106 600.00 1.673 600.00
GAR9.O1’4TErRPO-LCR IDE 5.434 684.76 3.235 695.90 2.196 706.27
l-EPThNE 6.166 700.00 3.314 700.00 2.147 700.00
o-LcR3FtR 4 6.157 699.82 3.454 707.46 2.257 716.03
6.626 708.26 3.699 719.26 2.380 73L60
1 .2-DICHLOROETl-W JE 7.878 727.56 4.144 737.79 2.541 749.27
1T4J{c ROET1-MB iE 8.883 740.51 4.505 750.69 2.674 762.11
OCT.AI’E 15.898 800.00 6.396 800.00 3.166 800.00
1E1JB’E 17.471 811.08 7.460 820.81 3.725 829.15
NO’J At’E 38.124 900.00 14.079 900.00 5.944 900.00
-------
Volatile Organic Vapor Concentrator/Purge and Trap
Bubbler
Load 1
Load 2
Sample
out
Traps
Heater Controt
Temperature
Vent
Tenax GC
.1.
Sample
5 ml syringe
—a
0
500 ml alrbaQ
1 20 Volts AC
Fiu c (
-------
-n
-4
VOLATILES IN AIR
lppm v/v
BEFORE CONCENTRATION
I I
CONCENTRATION
(100:1)
5 6
1- CHCI 3
2- BENZENE
3- 1,2-DCP
4- BrCHCI 2
5- TOLUENE
6- C 2 C1 4
-l
( )
AFTER
2
1
3
L
6
seconds
8
-------
c
0
4 .
RECOVERY OF VOLATILE COMPONENTS
l5Oppb (v/v), 200 ml air sample
CCI4
CHCI3
Benzene
1,2-DCP
BrCHCI2
Toluene
CCI4 -
Br2CHCI
36.1
64.5
74.8
91.8
89.5
86.5
120.7
38.1
62,6
66.5
91.3
86.9
95
84.1
110.8
36.7
62.3
67.1
72.4
86.3
92.4
80.8
3. .9
63.1
69,5
85.2
87.6
91.3
82.4
115.7
,,1
1.2
4.6
11
1.7
4.3
2.3
7
Table 2
-------
FIGURE 8
Plastic Cartridge
Teflon tubing
Plastic Cartridge
1/2 to 2/3 fu
LOAD
BUBBLER
Glass Vial
plastic
tube
short
syringe
long syringe
teflon tubing
/
(
)
1/2 to 2/3 ‘ull
wood block
LOAD
teflon tubing
BUBBLER
8-1 05
-------
70 ng/rnI
Headspace
C i )
Purge and Trap
HCI 3
Benzene
/ 1 ,2 DCP
BrCHCI 2 /
0
Toluene
CICHBr2
5
10
Tftnr (Sc concis)
-------
60 ng/mI
% Recovery
Benzene
59.0
62.0
56.7
59.2
2.7
75.9
2.8
B rCHCI 2
88.4
96.0
89.1
91.2
4.2
Toluene
97.0
108.3
105.8
103.7
5.9
108.6
6.2
Lower
Purge and Trap
Detection Limits (ng/mI)
Chloroform
Benzene
1 ,2-Dichloropropane
Bromodichioromethafle
Toluene
Tetrachioroethylene
Chiorodibromomethafle
70
10
20
20
10
20
60
CHCI3
76.6
76.6
1 ,2-DCP
74.7
79.1
73.9
c 2 c 1 4
114.7
108.8
102.3
Br 2 CHC I
107.6
113.9
109.2
110.2
3.3
Table 3
-------
FIGURE 10
Concentrator/Purge and Trap - Michromonitor/ITD
10— 15 minutes
I
Purge and Trap
a
50 ml sIJrirKJe
500 ml airbag
- 20mg
T erature
Control
vent
TenexGC
4,
5.0 ml
sçjringe
direct
Michromonitor 500
R
Maci ntosh
Ion Trep Detector
IBM PC
8—108
-------
Detector (ITD). The interfacing column also serves as a separating
column. Figures 11, 12, and 13 schematically show the lTD and the
interfacing with the microchip gas chromatograph.
The ITO is a relatively small mass spectrometer (compared to
conventional mass spectrometers) capable of producing electron
impact spectra of compounds introduced into its source region. The
lTD is fundamentally different from conventional quadrapole
instruments In the manner in which the mass spectral information is
produced. Resulting mass spectra are similar to those obtained with
quadrapole instruments. However, the lTD is a mechanically simple
and rugged device when compared to quadrapole units which Is
important for its application to field analysis of environmental
samples.
Figure 14 shows a gas chromatograph/mass spectrometer analysis of a
mixture of seventeen different volatile priority pollutants. The
total run time for the analysis was slightly over two minutes.
CONCLUS ION
By modifying off—the—shelf chemical instrumentation and designing a
field deployable concentrator we have developed analytical tools
that allow for the rapid acquisition of reliable chemical data in
the field. The instrumentation can be used to characterize air,
water, and soil samples for volatile organic compounds during
hazardous waste site investigations and cleanup activities. Field
use of the instrumentation is currently being carried out by the
U.S. EPA, the U.S. Coast Guard, and Hazardous Materials Response
Branch of the National Oceanic and Atmospheric Administration.
ACKNOWLEDGEMENTS
This work was supported by the Hazardous Materials Response Branch
of the National Oceanic and Atmospheric Administration under
contract 85—ABC—00258.
REFERENCES
Kovats, E., Rely. Chem. , Acta 41, 1915., 1958.
Sadat, S., and Terry, S., “A High Speed Gas Analyzer,” American
Laboratory 16, 90—101, 1984.
WohltJen, H., “Chemical Microsensor and Microlnatrumentation,”
Anal. (lem . 56, 87A—lO3A, 1984.
8-109
-------
FIGURE 11
ION TRAP DETECTOR
LENS
ELECTRON
GATE \
TRANSFER LiNE i:1III::II1 TOP END CAP
ASSEMBLY
COLUMN.
___ RING ELECTRODE
BOTTOM END CAP
VOLTAGE
MULTIPLIER TO THE ELECTROMETER
____________ ELECTRON MULTIPLER
ELECTRON - ___________
CIRCUIT
_____ _______ ROUGHING
PUMP
TURBOMOLECULAR PUMP
8-110
-------
FIGURE 12
MANIFOLD
TO VACUUM
PUMP
2 METER
COLUMN
CAP
TO ION TRAP OPEN/SPLIT
INTERFACE
SAMPLE
INLET
I
1 METER
COLUMN
VALVE
8-111
-------
FIGURE 13
1o
III
ALUMINUM COLUMN SUPPORT
TO MS OPEN-
SPLIT INTERFACE
CAPILLARY
COLUMN
TEFLON ,4
BONDING
TUBING
____ MATERIAL
DETECTOR
INLET
0-RING
SEAL
8—112
-------
1 1, 1 -Dichloroethylene
2 tlethylene Chloride
3 Trans-i, 2-Dichioroethylene
4 1, 1-Dichioroethane
5 Chloroform
6 1, 2-Dichioroethane
7 Benzene
8 Carbon Tetrachioride
9 1, 2-Dichloropropan
10 Trichioroethylene
11 2-Chioroethyl Vinyl Ether
12 1, 1, 2—Trichioroethane
13 Toluene
14 Tetrachioroethylene
15 Ethyl Benzene
16 O-Xylene
17 1, 1, 2, 2, —Tetrachioroethane
co
—4
-J
tI1 I
IIII
•“ i
1O 2G 3 4
:26 :51 1:1? 1:43 2: 8
Figure 14
-------
SAMPLING TECHNIQUES FOR EVALUATION OF TARRY WASTE IMPOUNDMENTS
Roger A. Dhonau, Senior Environmental Engineer, C. John Ritzert,
Manager—Technical Services; Environmental Services Division, Lancy
International, Inc., Zelienople, PA
ABSTRACT
A 26 million gallon unlined industrial waste lagoon was investigated
with the intent of characterizing the waste’s general composition and
suitability for reclamation as an alternative fuel. Physical
characteristics of the waste varied considerably, providing unique
challenges to sample acquisition. The tar—like nature of the waste did
not lend itself to traditional sampling techniques and numerous
practical problems had to be overcome. The highly stratified nature of
the lagoon contents required development of a sampling approach that
would allow the identification of specific vertical and lateral areas.
The determination of an average composition for each strata was also
required. The sampling plan developed incorporated both statistically
random as well as intentionally subjective placement of sample
locations in order to accomplish these objectives. Results of this
investigation allowed the division of lagoon wastes into areas usable
as alternative fuels for cement kilns or incinerators, and those that
would not.
This paper discusses the logic of sample location selection and
parameters chosen to characterize this waste. Difficulties encountered
in sampling this type of waste will be discussed in conjunction with
techniques found to be successful in obtaining representative samples.
INTRODUCTION
An investigation was performed on an inactive lagoon which had received
wastes from a large coking facility and adjoining chemical plant over
many years. The intent of this investigation was to provide basic
characterization data for the preliminary economic evaluation of two
alternatives deemed to be most economically and environmentally
feasible for remediation of the site. These alternatives, reclamation
as a waste fuel and incineration at a commercial facility, were
selected for initial evaluation on the basis of the general
characteristics of the waste streams known to have been directed to the
lagoon over its years of operation. Physical characteristics of the
lagoon contents presented unique challenges to acquisition of the waste
samples needed for this investigation. A combination of several
techniques were required in order to both access the sampling sites and
to obtain the designated samples.
The lagoon was originally constructed in the early part of this century
when plant wastes were discarded in a swampy depression. As this
depression was filled, additional capacity was created by raising the
8—115
-------
elevation of the surrounding area via the placement of ash, coal, and
coke. Numerous expansions and contractions of the lagoon occurred over
the ensuing years, each raising the elevation of the surrounding area.
As was common practice at the time, no records of these activities were
maintained and consequently, depth of the lagoon was unkno n. At the
time of the investigation the lagoon surface was 87,000 ft
Although there were no records of quantities and types of wastes placed
in the lagoon over most of its active life, it was known that the vast
majority of wastes were various tars and sludges generated during the
coking operations. In addition, several other waste streams, such as
ammonia still lime sludges, scrap oven bricks and general plant wastes
were also placed into the lagoon. At the time field activities were
initiated the lagoon surface consisted of a two to six inch rubbery
crust overlain by one to six inches of water. This water layer was
found to be highly dependent upon recent weather conditions, as it
completely evaporating during dry periods. Beneath the crust was a
viscous tarry waste.
Obtaining representative samples of the material in the lagoon
presented several obstacles primarily related to the physical state of
the contents. The tar—like material would not support a normal drill
rig and the rubbery consistency of the crust make discrete samples
difficult to obtain. In that this was a preliminary investigation,
available techniques were also severely limited by economic
constraints.
SAMPLING LOCATIONS
The surface area of the lagoon w s divided into twelve grid areas
averaging approximately 7,250 ft each with one boring designated for
each area. This number of borings was determined to be sufficient to
properly discern variability in the lagoon contents while remaining
within the economic confines of the preliminary nature of this
investigation. Due to the irregular perimeter of the lagoon and
restrictions of the sampling technique, there was unavoidable
variation in the size of the grid areas.
The restrictions of the sampling technique were due primarily to the
mounding of coke, coal, and soil around portions of the lagoons banks.
Because of this physical obstacle, tow winches for the sampling
platform could not be properly placed to maneuver the sampling platform
into certain small regions of the lagoon. In order to access these
regions, it would have been necessary to move several thousand cubic
yards of material. Again, such an effort would not have been in
keeping with the preliminary nature of this investigation.
The purpose of the investigation dictated that spatial variations in
the wastes be established. As a means of vertically characterizing the
lagoon, random samples were identified through each core column. This
was accomplished with the use of random number tables utilizing a
sufficient number of elements to include all one—foot segments of an
individual core. In order to ensure that sample points would be
8-116
-------
distributed throughout the entire core length, random numbers were
selected from each five foot increment of the anticipated core length
(30 feet) plus additional numbers in the event tha... the lagoon was
found to be deeper, as was the case. This procedure also alleviated
the problem of obtaining an excessive number of samples in shallow
regions and too few samples in very deep locations. A fixed grid was
rejected as overly biased and risking the omission of significant waste
layers.
SAMPLE ACQUISITION
The chemical composition and tar like consistency of the waste made
sampling considerably more difficult than what is normally encountered.
Most of the waste was too viscous for equipment and procedures normally
utilized in liquid investigations and structurally too weak to be
retained by equipment used for sampling solid wastes. To complicate
the situation further, there were inclusions of very hard materials,
such as refractory bricks, dispersed throughout the lagoon.
In order to access the selected boring locations, a small floating
platform was constructed and equipped with a small skid mounted drill
rig. The platform was approximately ten feet by twelve feet with two
rows of sealed 55 gallon drums mounted beneath the platform to provide
buoyancy. Since the water layer was far too thin to provide any
significant buoyancy the platform was constructed to rest directly on
the rubbery crust rather than float above it. This situation provided
both advantages and disadvantages to the sampling efforts. By having
the platform in contact with the wastes, considerable stress was placed
upon the platform during moves between locations. As the platform was
pulled from one location to another, the lagoons rubbery crust would
often be stretched in the process. It was therefore necessary to
maintain tension on all tow and guide lines to the platform at all
times, in order to keep it from drifting as the crust returned to its
original position. On the other hand, direct contact with the waste
significantly increased the stability of the platform during the
sampling.
The rubber—like consistency of the upper layer of the lagoon contents
prevented use of a standard split spoon type sampler. It was,
therefore, necessary to use a piston type sampler where the waste could
be drawn into the sampler in a manner similar to a syringe. A split
spoon type sampler with a basket retainer was most successful in lower
zones of the lagoon contents where the waste had a more soil like
consistency. A split spoon with other inserts types was also utilized
in the sampling program with the type of insert depending upon the
consistency of waste encountered.
8 —117
-------
distributed throughout the entire core length, random numbers were
selected from each five foot increment of the anticipated core length
(30 feet) plus additional numbers in the event that the lagoon was
found to be deeper, as was the case. This procedure also alleviated
the problem of obtaining an excessive number of samples in shallow
regions and too few samples in very deep locations. A fixed grid was
rejected as overly biased and risking the omission of significant waste
layers.
SAMPLE ACQUISITION
The chemical composition and tar like consistency of the waste made
sampling considerably more difficult than what is normally encountered.
Most of the waste was too viscous for equipment and procedures normally
utilized in liquid investigations and structurally too weak to be
retained by equipment used for sampling solid wastes. To complicate
the situation further, there were inclusions of very hard materials,
such as refractory bricks, dispersed throughout the lagoon.
In order to access the selected boring locations, a small floating
platform was constructed and equipped with a small skid mounted drill
rig. The platform was approximately ten feet by twelve feet with two
rows of sealed 55 gallon drums mounted beneath the platform to provide
buoyancy. Since the water layer was far too thin to provide any
significant buoyancy the platform was constructed to rest directly on
the rubbery crust rather than float above it. This situation provided
both advantages and disadvantages to the sampling efforts. By having
the platform in contact-with the wastes, considerable stress was placed
upon the platform during moves between locations. As the platform was
pulled from one location to another, the lagoons rubbery crust would
often be stretched in the process. It was therefore necessary to
maintain tension on all tow and guide lines to the platform at all
times, in order to keep it from drifting as the crust returned to its
original position. On the other hand, direct contact with the waste
significantly increased the stability of the platform during the
sampling.
The rubber—like consistency of the upper layer of the lagoon contents
prevented use of a standard split spoon type sampler. It was,
therefore, necessary to use a piston type sampler where the i. aste could
be drawn into the sampler in a manner similar to a syringe. A split
spoon type sampler with a basket retainer was most successful in lower
zones of the lagoon contents where the waste had a more soil like
consistency. A split spoon with other inserts types was also utilized
in the sampling program with the type of insert depending upon the
consistency of waste encountered.
8-118
-------
The selection of sampler type was based upon the consistency of the
previous sample within a given column. In most instances the change in
waste consistency was gradual enough to allow proper selection of the
sampling device for a given depth. However, in severai instances the
waste consistency changed so quickly that they could not be adequately
anticipated. Unfortunately, this resulted in the collection of only
partial samples or complete loss of a given sample.
In order to obtain the samples the selected sampling device was driven
two feet into the waste, removed, and immediately followed with a
section of casing to the same depth. This casing section was then
cleaned out. The sampler was then passed down the casing to obtain the
next sample. This procedure was repeated with the sampler driven in
advance of the casing for the entire length of the sampling column.
The sampling column was extended to a depth at which natural soils were
encountered.
After each sampler was retrieved from the boring, the contents were
inspected and logged and a sample was taken for laboratory analysis
according to the randomly selected sample locations. Samples of
specific materials or interest were also identified and collected by
the field supervisor based on technical judgment. The samples were
placed in 500 ml wide mouth glass containers and refrigerated
immediately at the site. Samples were coded on—site as to boring
number, sample depth, date of collections, and field supervisor’s
initials. This information was also recorded, along with other notes
and observations, in a field log.
Samplers were steam cleaned after each use. No solvents or detergents
were used. In order to keep sampling and coring running smoothly, up
to four (4) samplers of each type were required: one for use on the
rig; one containing sample material being inspected by the field
supervisor; one in cleaning; and one in transportation to or from the
sampling platform to the inspection/cleaning location.
Each segment of every boring was inspected by the field supervisor
prior to either placement in a sample container or returning it to the
lagoon. Inspection included the following steps:
— Determination of percentage recovery
— Examination for color, texture, adhesion and inclusions. Colors
were referenced Munsell Color standards
— Inclusions were noted as to type, size, and approximate
percentage by volume
These observations were used as a basis for selection of supplementary
samples within a given boring and to assist in estimating the possible
origin of various types of wastes noted.
8-119
-------
FINDINGS
Physical characteristics of the wastes varied through a wide spectruni
of combinations. Color was generally absent, with most samples black
to grey and occasionally ranging to almost pure white (spent lime
slurry). The small percentages of samples that did show color
contained mostly dark hues of greens, browns, and reds. Texture varied
from honey like to that of a crumbly soil. Inclusion content varied
from one visible to 20 percent or more. These inclusions generally
appeared to be limestone, sand, coal, coke, and/or brick.
Subsequent chemical analysis determined a much wider range of
characteristics. As indicated earlier, the purpose of this
investigation was to obtain sufficient data to evaluate incineration
and waste fuel options. Under this guidance, analytical parameters
included the following:
Heat Content (BTU/lb) Sulfer (%)
Moisture (%) Viscosity (centipoise)
ASH (%) pH
These parameters are indicative of general composition and are of
primary importance for evaluation of incineration and waste fuel
options. Evaluation of the resultant analytical data allowed the
lagoon contents to be classified into three categories: materials
suitable for use as a waste fuel (cement kiln); materials suitable for
incineration; and materials not suitable for either of these options.
Typical analyses of waste samples for each category are as follows:
Possible Possible for
Waste Fuel Incineration Other
ph 2.4 6.9 11.7
Heat Value (BTU/lb) 9010 5,790 3,650
Sulfur (%) 5.6 3.1 1.2
Moisture (%) 15.2 22.0 15.4
Ash (%) 10.5 25.4 40.5
Viscosity (centipoise) 3,100 22,500 100,000
90°C
As can be seen by these values, a wide range of wastes were encountered
during this investigation. The most significant of these with respect
to sampling methodology was viscosity. At the upper range of
8-120
-------
viscosity, the waste had a Soil like consistency and could be sampled
in a similar manner. At the lower range a piston sampler or split
spoon with a flap value retainer was needed.
In summary, the chosen combination of commercially available soil
sampling equipment was found to be both efficient and cost effective in
the collection of tarry waste samples. The use of custom made or
complex and expensive equipment was not required. The devices selected
for this program not only allowed for acquisition of more than 80
percent of the targeted samples, but could be quickly and inexpensively
replaced when a device was damaged or lost through the course of the
program.
8-121
------- |