United Stales
 Environmental
 Agency
Targeted National Sewage Sludge Survey


Sampling and Analysis Technical Report
                               January 2009

-------

-------
U.S. Environmental Protection Agency
       Office of Water (430IT)
   1200 Pennsylvania Avenue, NW
       Washington, DC 20460
         EPA-822-R-08-016

-------
                      Acknowledgments and Disclaimer
       The Targeted National Sewage Sludge Survey was made possible by the assistance and
cooperation of numerous staff working at each of the sewage treatment facilities involved. The
staffs of the facilities contacted during the course of this survey were, without exception,
knowledgeable, friendly, helpful, and deservedly proud of their efforts to protect the
environment and serve their local constituencies.

       This technical report has been reviewed and approved for publication by the Office of
Science and Technology.  This report was prepared with the support of Battelle Memorial
Institute, Computer Sciences Corporation, and their subcontractors, under the direction and
review of the Office of Science and Technology. EPA would like to thank the Alexandria
Sanitation Authority, Alexandria, VA, who opened their doors and assisted EPA and contractor
staff during hands-on training in sampling techniques in July 2006.

       Neither the United States government, nor any of its employees, contractors,
subcontractors, or other employees makes any warranty, expressed or implied, or assumes any
legal liability or responsibility for any third party's use of, or the results of such use of, any
information, apparatus, product, or process discussed in this report, or represents that its use by
such a third party would not infringe on privately owned rights.
TNSSS Sampling and Analysis Technical Report              i                                   January 2009

-------
TNSSS Sampling and Analysis Technical Report                  ii                                               January 2009

-------
                                 Table of Contents
                                                                                   Page
Acknowledgments and Disclaimer	i
Executive Summary   	v
Section 1  Background and Organization	1
          1.1  Regulatory and Surveys History	1
          1.2  Targeted National Sewage Sludge Survey	1
          1.3  Content  of this Report	3
Section 2  Survey Objective and Design	4
          2.1  Survey Objective	4
          2.2  Target Population	4
          2.3  Stratification	4
          2.4  Facillity Selections	5
Section 3  Sample Collection	7
          3.1  Training	7
          3.2  Sample Collection	7
          3.3  Site-specific Deviations	88
          3.4  Representative Samples	9
          3.5  Field Duplicates	9
          3.6  Sample Labeling and Tracking	100
          3.7  Packing  and Shipping Samples to the Repository	10
          3.8  Storage and Shipments to Laboratories	11
          3.9  Shipping Issues	12
          3.10 Sampling Summary	13
          3.11 Equipment Blanks	13
Section 4  Sample Analyses	15
          4.1  Analytes of Interest	15
          4.2  Analytical Techniques	18
          4.3  Laboratories	18
          4.4  Method Modifications	19
               4.4.1  Ensuring Consistent Method Sensitivity	19
               4.4.2  Anions in Sewage Sludge	201
               4.4.3  Modified GC/MS Procedures for Semivolatile Organics	21
               4.4.4  PBDE Analyses	23
               4.4.5  Pharmaceutical Analyses	24
               4.4.6  Steroid and Hormone Analyses	26
               4.4.7  Focusing Quality Control on the Survey-specific Analytes of Interest....28
Section 5  Data Review Procedures	29
          5.1  General  Review Procedures	29
          5.2  QC Acceptance Criteria	30
          5.3  Data Qualifiers and Database	30
          5.4  Data Review Findings	31
               5.4.1  Reporting Limits	31
               5.4.2  Calculation Errors	32
               5.4.3  Blank Contamination	32
               5.4.4  Extract Dilution	33
TNSSS Sampling and Analysis Technical Report
January 2009

-------
               5.4.5  Matrix Spikes	34
               5.4.6  Interferences	34
               5.4.7  Recovery Issues	37
          5.5  Revised Results for BDE-209	38
Section 6  Survey Results	39
          6.1  Summary Results	39
          6.2  Investigation of Results for Metals	42
          6.3  Comparison of Metals Results to Current Standards	43
          6.4  Analytical Completeness	43
          6.5  Analytical Sensitivity	44
          6.6  Equipment Blank Evaluation	47
               6.6.1  Semivolatile Organics and PAHs	47
               6.6.2  Metals	48
               6.6.3  PBDEs	49
               6.6.4  Anions	49
          6.7  Field Duplicate Results	51
               6.7.1  Anions	52
               6.7.2  Metals	52
               6.7.3  Semivolatile Organics and PAHs	53
               6.7.4  PBDEs	54
               6.7.5  Pharmaceuticals	54
               6.7.6  Steroids and Hormones	56
               6.7.7  Results in Liquid Samples	58
          6.8  Matrix Spike and Duplicate Results	59
               6.8.1  Matrix Spike Results for Anions	60
               6.8.2  Duplicate Results for Anions	62
               6.8.3  Matrix Spike Results for Metals	62
               6.8.4  Duplicate Results for Metals	65
               6.8.5  MS/MSD Results for Organics	67
               6.8.6  MS/MSD Results for PBDEs	69
               6.8.7  Qualification of Sample Results based on MS/MSD Results	70
               6.8.8  Labeled Compound Recoveries for Isotope Dilution Methods	71
Section 7  References  	72

Appendix A  Solids Leaching Procedure for Anions	73
Appendix B  Method Modifications for the PBDE Analyses	74
Appendix C  QC Acceptance Criteria	75
TNSSS Sampling and Analysis Technical Report             iv                                   January 2009

-------
                               Executive Summary
       This Sampling and Analysis Technical Report ("Technical Report") describes the
sampling and analysis activities conducted by EPA in support of the Targeted National Sewage
Sludge Survey (TNSSS). The TNSSS was designed to: 1) obtain updated occurrence
information on nine analytes of potential concern, and 2) obtain occurrence information on a
number of contaminants of emerging interest identified by EPA and the National Research
Council (NRC). The objective of the survey was to obtain national estimates of the
concentrations of these pollutants in sewage sludge for use in assessing if exposures may be
occurring and whether those levels may be of concern.

       Final sewage sludge is defined as the liquid, solid, or semi-solid residue generated during
the treatment of domestic sewage, receiving secondary treatment or better, in a treatment works,
which may include sewage sludge processed to meet land application standards.  The publicly
owned treatment works (POTWs) included in the survey were selected without consideration of
their sewage sludge use or disposal practices.

       For this survey, EPA focused its efforts on POTWs that treat more than one million
gallons of wastewater per day (MOD). This group of facilities collectively generates
approximately 94 percent of the wastewater flow in the nation. To be eligible for the survey,
EPA also required that a POTW be located in the contiguous United States and employ
secondary treatment or better. From the 3,337 POTWs that met the criteria, EPA statistically
selected 74 facilities in 35 states for the survey and collected biosolids samples from those
facilities. Whether the facility recycles the sewage sludge to land or disposes of it via
incineration or surface disposal was not a consideration for selecting a facility for inclusion in
the survey. By using statistical methods, the concentration measurements can be extrapolated to
the entire population of 3,337 POTWs.

       EPA collected samples between August 2006 and March 2007. EPA collected 84
samples of sewage sludge from 74 facilities, one from each of 64 POTWs, as well as two
samples at the remaining ten facilities (either because the facility had more than one treatment
system and produced two types of final sewage sludge, or for quality assurance purposes.  EPA
conducted analysis of sewage sludge samples for 145 analytes, including four anions
(nitrite/nitrate, fluoride, water-extractable phosphorus), 28 metals, four polycyclic aromatic
hydrocarbons, two semi-volatiles, 11 flame retardants, 72 pharmaceuticals, and 25 steroids and
hormones.

       The survey used both well-established multi-laboratory validated EPA procedures as well
as three analytical methods that were developed or updated for the survey. The two new
methods are single-lab validated methods for pharmaceuticals (EPA Method 1694),  and steroids
and hormones (EPA Method 1698). The updated multi-lab validated method is for flame
retardants (EPA Method 1614).

       EPA took steps to ensure that the results were comparable across all of the facilities
sampled.  The percent solids in the various sewage sludge samples range from 0.14 to 94.9.  To
ensure comparability of results, all  sample results are reported on a dry-weight basis.
TNSSS Sampling and Analysis Technical Report             v                                   January 2009

-------
       EPA subjected all of the analytical results generated by the laboratories to data review
procedures.  These procedures used review protocols to ensure that the results met EPA's
objectives for data quality.

       This Technical Report includes the number of samples in which each analyte was
reported, along with minimum and maximum measurements.  Reported concentrations and
frequency of detects are limited by the sensitivity of the analytical methods used. Some analytes
were found in all 84 samples, while others were found in none or only a few of the sewage
sludge samples. The minimum concentration is the lowest value reported as present in any
sample. EPA did  not report a minimum or maximum value for those analytes that were not
detected (i.e., a situation that  occurred for some of the pharmaceuticals, steroids and hormones).
For these situations, EPA used "ND" to indicate that the minimum and maximum values were
"not detected." The maximum concentration is the highest value reported as present in any
sample.

       Briefly, the survey found:

    •   The four anions were  found in every sample.

    •   27 metals were found  in virtually every sample, with one metal (antimony) found in no
       less than 72 samples.

    •   Of the six  semivolatile organics and poly cyclic  aromatic hydrocarbons, four were found
       in at least  72 samples, one was found in 63 samples, and one was found in 39 samples.

    •   Of the 72 pharmaceuticals, three (i.e., cyprofloxacin, diphenhydramine, and triclocarban)
       were found in all 84 samples and nine were found in at least 80 of the samples.
       However,  15 pharmaceuticals were not found in any sample and 29 were found in fewer
       than three samples.

    •   Of the 25 steroids and hormones, three steroids (i.e., campesterol, cholestanol, and
       coprostanol) were found in all 84 samples and six steroids were found in at least 80 of
       the samples. One hormone (i.e., 17a-ethynyl estradiol) was not found in any sample and
       five hormones were found in fewer than six samples.

    •   All of the flame retardants except one (BDE-138) were found every sample or all but one
       sample.

       It is not appropriate to speculate on the significance  of the results until a proper
evaluation has been  completed and reviewed.  EPA plans to evaluate the pollutants identified by
the survey as being present in sewage sludge.  As its  first priority, using the survey information,
EPA has begun assessing the  nine pollutants identified from the 2003 biennial review as needing
updated concentration information and molybdenum to determine whether additional action may
be necessary. Later this year, EPA expects to initiate evaluations of other pollutants in the
survey that may warrant further consideration.  The evaluations will depend on the availability of
data needed to conduct the evaluations.
TNSSS Sampling and Analysis Technical Report             vi                                   January 2009

-------
                                      Section 1
                         Background and Organization

1.1    Regulatory and Surveys History

       Sewage sludge is the solid, semisolid, or liquid organic material that results from the
treatment of domestic wastewater by municipal wastewater treatment plants, also known as
publicly owned treatment works (POTWs).  The U.S. Environmental Protection Agency (EPA)
uses the terms sewage  sludge and biosolids interchangeably, but others often refer to biosolids as
sewage sludge that has had additional processing for land application.

       Section 405(d)  of the Clean Water Act (CWA) requires that the EPA establish
requirements for the use or disposal of sewage sludge. The Standards for the Use or Disposal of
Sewage Sludge are found at Part 503 of Section 40 of the Code of Federal Regulations (40 CFR
503, hereafter simply "Part 503").

       These regulations establish numeric limits, management practices, and operational
standards to protect public health and the environment. Sewage sludge is typically used by land
applying to fertilize crops or reclaim mined lands, or disposed either by landfilling  or surface
disposing, or by incinerating. States may adopt additional or more stringent regulations for the
use or disposal of biosolids.

       Additionally, Section 405(d) of the CWA requires EPA to review existing sewage  sludge
regulations at least every two years (i.e., biennial review).  The purpose of such reviews is to
identify additional toxic pollutants, and promulgate regulations, if needed, for those pollutants
consistent with the requirements set forth in the CWA.

1.2    Targeted National Sewage Sludge Survey

       The Agency periodically conducts surveys to determine what may be present in sewage
sludge.  EPA has conducted three previous sewage sludge surveys:  1) a 40-city survey in 1982 to
develop information on the fate and effects of priority pollutants in wastewater treatment plants
and estimates of pollutant concentrations in sewage sludge; 2) a National Sewage Sludge Survey
in 1988-1989 to gather information on sewage sludge use or disposal practices and to obtain
updated information on the concentration of over 400 pollutants in the Nation's sewage sludge;
and 3) a National Sewage Sludge  Survey in 2001 to obtain updated national estimates of dioxins
and dioxin-like compounds in sewage sludge managed by  land application.

       In conducting the 2003 biennial review (68 FR 75531), EPA identified 15 analytes that
needed further evaluation. EPA subsequently reduced the list of analytes to nine based on a
biosolids exposure and hazard assessment. EPA also determined it needed updated
concentration data for  more refined risk assessment and risk characterization of these nine
pollutants (see Table 1).
TNSSS Sampling and Analysis Technical Report             1                                   January 2009

-------
                    Table 1.  2003 Biennial Review Analytes Requiring
                                 Additional Information
Analyte
Nitrate
Nitrite
Barium
Beryllium
Manganese
Silver
Fluoranthene
Pyrene
4-Chloroaniline
Type
Anion
Anion
Metal
Metal
Metal
Metal
PAH
PAH
Semivolatile organic
       Inclusion of pollutants in the TNSSS does not reflect a determination that their presence
in sewage sludge adversely affects human health or the environment.  Rather, EPA decided that
updated or new concentration data were needed to assess exposure and help in evaluating
whether levels of these pollutants in sewage sludge present environmental or human health
concerns.

       Given the national scope of the survey, EPA expanded the list of analytes to reflect the
Agency's interest in collecting concentration data for other pollutants.  The expanded list
included 24 additional metals that could be analyzed at little extra cost at the same time as the
four metals (barium, beryllium, manganese, and silver) included in the list of nine pollutants
above; molybdenum because of the Agency's interest in determining the need for a revised
numeric standard for it in land-applied biosolids; and other analytes because of their widespread
incidence and use, as well as emerging concern.  The latter category included:

    •   benzo(a)pyrene (found in coal tar, automobile exhaust fumes, tobacco and wood smoke,
       charbroiled food, and burnt toast);
    •   2-methylnaphthalene (found in nonstructural caulking compounds and sealants, synthetic
       resins, rubber adhesives, and wall coverings);
    •   bis (2-ethylhexyl) phthalate (widely used as a plasticizer in manufacturing of items such
       as cosmetics, toys, tools, and laboratory equipment);
    •   fluoride (used in topical and systemic therapy for preventing tooth decay, as well as many
       other uses);
    •   water-extractable phosphorus (correlated with phosphorus concentration in runoff from
       soils amended with manure and biosolids and an effective indicator of loss that may
       contribute to algae buildup in surface waters);
    •   11 polybrominated diphenyl ethers (PBDEs). Four of the PBDEs were of most interest
       because of available human health information that may be useful for future risk
       evaluation efforts. PBDEs are used as flame retardants in a wide array of products,
       including building materials, electronics, furnishings, motor vehicles, plastics,
       polyurethane foams, and textiles; and
   •   97 pharmaceuticals, steroids, and hormones because of broader emerging interest in these
       analytes.
TNSSS Sampling and Analysis Technical Report
January 2009

-------
       EPA began sampling in August 2006, using the procedures described in Section 3, and
completed sampling in March 2007. Analyses of survey samples were conducted as described in
Section 4.

1.3    Content of this Report

       This report describes the design, sampling, and analysis activities for the TNSSS. The
report addresses the  following topics:

   •   Survey Objective and Design
   •   Sample Collection
   •   Sample Analyses
   •   Data Review Procedures
   •   Survey Results
   •   References
   •   Appendices
TNSSS Sampling and Analysis Technical Report             3                                   January 2009

-------
                                     Section 2
                         Survey Objective and Design
2.1    Survey Objective

       The TNSSS was designed to: 1) obtain updated occurrence information on nine analytes
of potential concern, and 2) obtain occurrence information on a number of contaminants of
emerging interest identified by EPA and the National Research Council (NRC) that may be
present in sewage sludge generated by POTWs.

2.2    Target Population

       For this survey, EPA focused its efforts on POTWs that treat more than one million
gallons of wastewater per day (MOD). This group of facilities collectively generates
approximately 94 percent of the wastewater flow in the nation.  To be eligible for the survey,
EPA also required that a POTW be located in the contiguous United States and employ
secondary treatment or better.  EPA selected POTWs meeting the criteria from information in the
2004 Clean Water Needs Survey and the 2002 version of the Permit Compliance System. From
the 3,337 POTWs that met the criteria in either data source, EPA used a stratified random
sampling design to statistically select 74 facilities in 35 states for the survey and collected
biosolids samples from those facilities. Whether the facility land applies the sewage sludge or
disposes of it via incineration or surface disposal was not a consideration for selecting a facility
for inclusion in the survey. By using statistical methods, the concentration measurements can be
extrapolated to the entire population of 3,337 POTWs.

2.3    Stratification

       EPA selected POTWs for inclusion in the survey using a random sampling  design
stratified for flow. EPA divided the 3,337 facilities in the sample population into three
categories, based on their design flow:

   •   Flow rate of 1 to 10 MOD;
   •   Flow rate of 10 to 100 MOD; and
   •   Flow rate of greater than 100 MGD

EPA then selected a proportionate number of POTWs from each of the above stratum at random.

       POTWs with flow rates less than 1 MGD were not included in the survey. However, the
combined flows of all such facilities represent less than 6% of the total flow of all POTWs
nationwide.
TNSSS Sampling and Analysis Technical Report              4                                   January 2009

-------
2.4    Facility Selections
       EPA invited 80 POTWs to voluntarily participate in the survey. The initial written
invitation was followed by a telephone call.  These communications outlined the nature of the
survey, the analytes of interest, and the timeframe for completion. EPA also assured each
facility that samples sent to the laboratories for analysis would be submitted as "blind" samples,
such that the results from any given sample could not be associated with a particular facility.

       These communications identified facilities that did not meet the criteria for POTW
selection in the peer-reviewed survey design and outlined above.  Some POTWs provided only
partial treatment of their wastewater, while others employed wastewater lagoons which do not
typically produce sewage sludge, as defined, on a routine basis. Ultimately, EPA eliminated
eleven POTWs and found five replacement POTWs.  As a result, EPA selected 74 POTWs that
met the stated criteria. The rationale for each replacement POTW is included in Table 2.

 Table 2.   POTWs  Selected for Sampling
Facility Name and Flow Group
Sugar Creek VWVTP
Aldridge Creek VWVTP
Phoenix VWVTP
Valley Sanitary District STP
San Francisco
El Estero VWVTP
Santa Rosa
Stockton Water Quality Plant
Los Angeles County Sanitation District
Boulder WWTP
South Windsor
Three Oaks WWTF
Orange County Northwest WRF
Tampa
Albany
Americus-Mill Creek
Boone STP
Calumet Water Reclamation Plant
Plainfield WWTP
Lake County DPW, New Century STP
Dupage County-Knollwood STP
Blucher Poole WWTP
William Ross Edwin WWTP
Parsons
Topeka
Mayfield WWTP
Eunice
Jefferson Parish East Bank WWTP
Nantucket
Salisbury
Mechanic Falls Treatment Plant
Benton Harbor-St. Joseph WWTP
Wixom WTP
Festus Crystal City STP
Elizabeth City WWTP
Hillsborough WWTP
Flow Stratum
1 100MGD
1 100MGD
10 100MGD
1 
-------
 Table 2.   POTWs Selected for Sampling
Facility Name and Flow Group
Beatrice
Wildwood Lower WTF
Middlesex County Utility Authority WRC
Verona TWP DPW
Buffalo
Canajoharie VWVTP
Geneva A-C Marsh Creek STP
NYC DEP - Jamaica WPCP
North Tonawanda STP
Clermont County Commissioners
Bedford
Metropolitan Sewer District Little Miami
VWVTP
Northeast Ohio Regional Sewerage
District Southerly VWVTP
Delaware County Alum Creek VWVTP
Mingo Junction STP
Duncan Public Utilities Authority
City of Klamath Falls WWTF
Western Westmoreland Municipal
Authority
Allegheny County Sanitary Authority
Greater Pottsville Area Sewer Authority
Punxsutawney
South Kingstown WWTF
Plum Island WWTP
Lawson Fork WTP
Elizabethton
Amarillo
Dallas Southside WWTP
Trinity River Authority of Texas
Fredericksburg
Odo J. Riedel Regional VWVTP
Wagner Creek WWTP
Tyler Southside WTP
Spanish Fork City Corporation
Buena Vista
Everett City SVC Center MVD
Beaver Dam
Elkins VWVTP
Huntington
Flow Stratum
1  100MGD
1 100MGD
1 100MGD
1 100MGD
1 
-------
                                      Section 3
                                Sample Collection
       EPA collected samples of the final treated sewage sludge at each of the 74 POTWs that
ultimately participated in the TNSSS. EPA developed a sampling and analysis plan that was
peer-reviewed and describes the sample collection procedures in detail. EPA revised the plan
periodically during the survey to address the changing list of facilities and to add updated contact
information for the laboratories that performed the analyses. As noted in Section 1.2, EPA
sampled between August 2006 and March 2007.

3.1    Training

       Prior to the start of sampling, the biosolids samplers were trained by the contractor.  The
contractor, with assistance from the Alexandria Sanitation Authority, Alexandria, VA, provided
the samplers with instructions on sampling techniques, sample point selection, required
paperwork, sample packing, and shipping techniques. The  samplers also toured the Alexandria
Sanitation Authority to become familiar with typical sewage treatment processes.  The tour
included demonstrations and hands-on training in the collection of sewage sludge.
Demonstrations included how to collect a range of samples, from liquid to dewatered sewage
sludge.

3.2    Sample Collection

       EPA began the sample collection process by identifying the number and nature of the
types of sewage sludge produced at each facility.  This effort took place during telephone
conversations with the plant staff well in advance of sampling. Details were confirmed with
plant staff upon gaining access to the final treated sewage sludge. Access to the treated sewage
sludge was generally not difficult.  However, in several instances, the samplers worked with
plant staff to obtain samples from difficult locations where  there might be safety concerns.  Two
facilities required that their personnel collect the actual samples. These instances are described
in Section 3.3.

       Grab samples were collected using sampling equipment appropriate to the type of sewage
sludge (liquid or solid) and the analytes of interest. To avoid or minimize contamination from
sampling equipment, plastic equipment was used to collect  samples for analyses of metals and
anions, and stainless steel equipment was used to collect  samples for analyses of all the organics.

       Liquid samples were collected as free-flowing materials from storage tanks, transfer
lines, taps, and hoses. After purging any lines used to collect samples, liquid samples were
collected directly into the final sample containers shown in  Table 3.  Where possible, plant staff
turned on mixing equipment in any storage tanks prior to sampling so that the collected liquids
were representative of the bulk sewage sludge.

       Solid samples included dewatered sewage sludge. These samples were collected from a
belt press, filter press, drying bed, centrifuge, compost pile, or other source on site.  The sampler
collected small grab samples from multiple areas of any large piles, or multiple grabs from any
continuous processes (e.g., belt press).  Small grabs were composited in a large pre-cleaned

TNSSS Sampling and Analysis Technical Report             7                                   January 2009

-------
container of appropriate construction, mixed well, and the mixed sample was transferred into the
final sample containers (see Table 3). Several kilograms of material were collected for each type
of treated sewage sludge and mixed. Any mixed material that remained after all the sample
containers were filled was returned to the sewage sludge process for disposal.

       Grabs of solid samples for anions and metals analyses were collected with a large pre-
cleaned plastic serving spoon and mixed in a pre-cleaned plastic wastebasket. Grab solid
samples for organics analyses were collected using a pre-cleaned stainless steel  scoop and mixed
in a pre-cleaned stainless  steel bowl (8 to 12 quarts).  Separate sampling equipment was used for
each facility and all equipment was cleaned prior to shipment to the facility.

       Sample containers were purchased from commercial suppliers who provided certificates
of analysis for common contaminants of interest (e.g., metals, semivolatile organics, pesticides,
PCBs). The cleaning procedures applied by the vendors were presumed to be sufficient for the
other analytes in the survey for which routine testing by the vendor was not performed. At least
two containers of each type were used to prepare equipment blanks as an overall check on
possible contamination (see the discussion of equipment blanks later in Section  3.11). Both the
high density polyethylene (HDPE) and glass containers were wide-mouth designs, sealed with
screw caps containing a polytetrafluorothylene (PTFE) lid liner.

 Table 3. Sample Containers for Solid and Liquid Sewage Sludge, by Analysis Fraction
Analysis Fraction
Metals
Polycyclic Aromatic Hydrocarbons (PAHs)
and Semivolatiles (as one analytical
fraction)
Inorganic Anions
Polybrominated Diphenyl Ether
Congeners
Antibiotics and Drugs
Steroids and Hormones
Archive Samples - for use in the event of
breakage, lab accident, or for future EPA
studies
Total Containers per Sampling Point
Solid Sample Container
500-mL wide-mouth HDPE
500-mL wide-mouth glass
500-mL wide-mouth HDPE
500-mL wide-mouth glass
500-mL wide-mouth glass
500-mL wide-mouth glass
2 500-mL wide-mouth HDPE and
4 500-mL wide-mouth glass
12
Liquid Sample Container
500-mL wide-mouth HDPE
1000-mL wide-mouth glass
500-mL wide-mouth HDPE
1000-mL wide-mouth glass
1000-mL wide-mouth glass
1000-mL wide-mouth glass
2 500-mL wide-mouth HDPE and
4 1000-mL wide-mouth glass
12
3.3    Site-specific Deviations

       Site-specific conditions at two facilities required modifications to the equipment
protocols. At one facility, the sewage sludge was discharged from a two-story tower with a belt
press directly into a dump truck parked below. Access to the bed of the truck was not practical,
even if the discharge was stopped temporarily. Therefore, the staff at this facility routinely
collected samples in a polyethylene container mounted on the end of a length of PVC pipe. In
the presence of the EPA-contractor sampler, the facility staff inserted the device into the
discharge from an opening in the tower, collected a small grab sample, and pulled the device
back through the opening. Successive grab samples were collected through that opening and
quickly placed into either the stainless steel or plastic compositing containers held by the EPA-
contactor sampler. The sampler carried the containers down from the tower, mixed the samples
in the compositing containers as described above, and then placed the samples into the
appropriate final containers. The total contact time of each grab sample with the polyethylene
sampling device was on the order of 30  seconds.
TNSSS Sampling and Analysis Technical Report
January 2009

-------
       At a second facility, sewage sludge was discharged into a roll-off dumpster body in a
narrow building which allowed limited access. In the presence of the EPA sampler, one of the
facility staff collected the sample in a 5-gallon plastic bucket lined with a trash bag. A stainless
steel scoop was used to remove aliquots of sewage sludge from the center of the bucket and
transfer the material to a stainless steel bowl.  Once that portion was removed, a plastic spoon
was used to transfer the sewage sludge to a plastic bowl.

3.4    Representative Samples

       The TNSSS was designed to collect sewage sludge samples that were representative of
various types of sewage sludge.  For bulk sewage sludge, collecting representative samples
presented a challenge at some facilities.  For example, at one facility that composted its final
sewage sludge, samples were collected from one of the long piles of sewage sludge mixed with
woods chips.  The piles were upwards of 50 feet long and over 6 feet high, with  sides sloping up
at roughly a 45 degree angle. Samples were collected from the oldest sections of the rows at the
facility to represent the length of the typical composting period at the facility, which ranges from
one to six months, depending on the season.

       Samples of biosolids materials were taken by digging into the side of the compost pile at
roughly six points along its length, on both sides of the pile, a foot or more off the ground to
avoid materials in contact with the concrete substrate. Materials removed from the pile often
contained large chunks of wood or small branches. Because these materials would not fit into
the sample containers, they were removed from the compositing containers before mixing the
bulk sample.  Once the bulk sample was well mixed, the samples were transferred to the final
sample containers.  This procedure was repeated twice: 1) for samples for the organic
parameters, using stainless steel equipment and glass containers, and 2) for the metals and
anions, using plastic equipment and containers.

       At another facility which produced liquid sewage sludge, samples were collected from a
catwalk atop a 1-million gallon storage tank. Sewage sludge was introduced into the tank by
water cannon with a 4-inch diameter discharge nozzle. Plant personnel turned on the water
cannon and throttled back the flow to a relative trickle and the sampler held each sample
container in the edge of the stream until it was full. The containers were capped once they were
full and wiped down before packing. Neither of these situations means that the samples were not
representative or that the Agency can not rely on the results obtained.  It simply points  out the
complexities and challenges with sampling sewage sludge generated by the variety of treatment
processes and management options available nationally.

3.5    Field Duplicates

       The sampling plan called for collection of field duplicate samples at 10% of the facilities.
A field duplicate sample is a second sample collected at the facility using similar procedures and
equipment as the original sample for quality control purposes.  The results of the field duplicate
sample can be compared to the results of the original sample as a means of assessing the overall
precision of the sampling and analysis processes.
TNSSS Sampling and Analysis Technical Report              9                                   January 2009

-------
Note:  The 10% frequency for field duplicates is a common, but arbitrary, choice designed to
       balance the cost of the additional samples against the desire to assess the precision of the
       sampling and analysis processes.

       Eight facilities were originally selected for collection of field duplicates. This number
was ultimately reduced to six because two of the facilities at which field duplicates were to be
collected were dropped from the survey and not replaced (for reasons described in Section 2).

       Separate EPA sample numbers were assigned to the field duplicates so they were not
identifiable by the laboratories as duplicates.  The results of the field duplicate analyses are
discussed in Section 6.

3.6    Sample Labeling and Tracking
       ^,   ™4               ,,,,,,       •       EPA Sample No. 68408
       The EPA contract samplers  labeled each  container
                                                            POTW Sewage sludge
                                                            Date collected
with a preprinted EPA Sample Number.  An example label is
shown at the right.  The EPA sample number was specific to
each sewage sludge sample at the facility.
                                                            Sampler Initials
      In addition to labeling each container, the samplers
prepared an EPA Traffic Report that documented the origin of the samples.  The sampler
recorded the name of the facility,  date of sampling, sampler's name, and shipping airbill number
on the traffic report. The numbers of sample containers of each type (e.g., four plastic and eight
glass) that were collected were recorded.  The traffic report prepared at the site allowed EPA to
track shipments of samples to the EPA Sample Repository at Microbac Laboratories in
Baltimore, MD.

3.7   Packing and Shipping Samples to the Repository

      The sample containers were packed for shipping using procedures described in the peer-
reviewed sampling and analysis plan.  Each sample container was either encased in bubblewrap
bag or layers of bubblewrap sheeting to prevent its movement during shipping. Samples were
packed into sturdy plastic ice chests. All of the samples from a single site could be packed, with
ice and bubblewrap, in one 48-quart ice chest, or two 28-quart ice chests, depending on
availability.

      The samplers purchased ice near each facility, or the POTW provided ice, and packaged
it in one-gallon self-sealing plastic bags.  Approximately one pound of ice was used for each
sample container (e.g., four bags containing two pounds of ice each were used to cool eight
samples in a 28-quart ice chest). To prevent leakage during shipping,  each ice chest was lined
with bubble wrap and two trash bags.  Samples and ice were packed into the inner bag, and then
tied or sealed shut with tape.  Additional packing materials were placed around the inner bag, if
needed,  and the outer bag sealed shut.  The completed traffic report was  placed in a plastic bag
and affixed to the underside of the lid of the ice chest with either tape or a plastic airbill pouch.

      Each cooler was shut with a layer of duct tape placed horizontally across the seam
between the ice chest and its lid. Packaging tape or filament tape was applied to the cooler
vertically, one  band near each end of the ice chest, to secure the lid.

TNSSS Sampling and Analysis Technical Report             10                                  January 2009

-------
       Ice chests were shipped overnight from full-service FedEx offices to the EPA Sample
Repository.  Samples collected from multiple facilities in a given day were shipped at the end of
the day, or depending on the logistics, sent separately from different FedEx locations.  Each
sample shipment was tracked through the carrier's web site and EPA confirmed receipt at the
repository.

       Over the course of the survey, samples from two facilities were hand carried to the
repository in Baltimore, because the facilities were located nearby.  These samples were packed
in a similar fashion as those sent by FedEx, except that no airbills were prepared.

3.8    Storage and Shipments to Laboratories

       When samples arrived at the sample repository, the staff inspected the ice chests for
external damage or leakage (none occurred) and placed them in one of two walk-in freezers
dedicated to EPA samples and maintained at -11°C. Freezing at - 11°C reduces microbiological
activity and the rates of any chemical reactions that might lead to changes in the sample.

       To streamline the  shipping logistics and manage both shipping and analytical costs, EPA
shipped batches of 15 to 20 samples from the repository to the contract laboratories for analyses.
EPA prepared new traffic reports that listed the samples in each shipment.  Additional shipments
were sent to the laboratories as more facilities were sampled. In all, six shipments were sent to
the laboratory performing the  analyses of metals, anions, and organics, with the last shipment
being the two samples collected at the last facility. For the PBDEs, pharmaceuticals, and
steroids and hormones analyses, more samples had been collected and stored at the repository by
the time those analyses began.  Ultimately, three shipments were made to the laboratory
performing the PBDE analyses and three shipments were made to the laboratory performing the
Pharmaceuticals, steroids, and hormones analyses.

       During the packing process prior to shipment, EPA examined the samples  for signs of
breakage, including cracked glass jars and cracked lids. Although some cracking  and breakage
did occur in jars or lids, as described below, it was observed that neither cross-contamination
occurred due to frozen conditions nor were any samples lost or not available for analysis.

       Only three incidents of breakage were observed and these containers were  not shipped to
laboratories, but were packaged in separate plastic bags and returned to the freezer.  The jars may
have cracked during freezing,  as the cracked jars were all of the 1000-mL  size used for liquid
samples.  While the samplers took care not to fill any of the jars to more than 90% of their
capacity, the high water content of the liquid sewage sludge samples could lead to greater
expansion during freezing.  This may have resulted in a greater risk of breaking the 1000-mL
glass jars than their 500-mL counterparts used for solid samples.

       All samples sent from the repository were shipped frozen, with large quantities of ice
added to  all coolers. The laboratories inspected the samples on receipt and reported that all the
coolers still contained ice. The laboratories also reported that a small number of samples  (<15)
were received with cracked lids or jars, which may have cracked during shipping.  In each case,
the laboratory transferred the sample to a suitable clean container or otherwise protected the
contents, such that no samples were lost as a result of shipping. The laboratories stored the

TNSSS Sampling and Analysis Technical Report              11                                   January 2009

-------
samples frozen until analysis.  Each laboratory returned a copy of the traffic report, with the date
and time of sample receipt documented on the form.

       Methods for the analysis of pharmaceuticals, steroids, and hormones in sewage sludge
were not available when sample collection began in August 2006.  Therefore, samples were
stored frozen (-11°C) until methods for these analytes were available.  Except for the two
samples collected from one plant in March 2007, all of the samples analyzed for
Pharmaceuticals, steroids, and hormones were stored frozen for 11 to 15 months.

       For the purpose of this survey, EPA assumed that all the analytes of interest were stable
in sewage sludge samples stored at -11°C. Because the analyses for metals involved the "total
recoverable" concentrations (e.g., all of the metal that can be recovered during digestion with a
strong acid), holding frozen samples for extended periods is not a concern because the metals
will be recovered even if there were any residual microbiological or chemical activity in the
frozen samples. For the anions and semivolatile organics, EPA also did not anticipate any
substantive changes in the concentrations of these analytes during storage due to the relatively
low storage temperature.

       The samples analyzed for pharmaceuticals,  steroids, and hormones were held in storage
longer than any of the other samples.  The stabilities of these analytes have not been studied by
EPA.  However, EPA began a holding time study of pharmaceuticals in aqueous samples and
sewage sludge in late 2008 that may provide some data to assess the freezer storage stability of
these analytes in the near future.

3.9    Shipping Issues

       EPA tracked each sample shipment through the carrier's web site, or confirmed receipt
by contacting the laboratory directly.  Only two shipping problems occurred, a surprisingly small
number, given the number of shipments involved. In one instance, the laboratory performing the
metals and anions analyses was sent the incorrect number of sample containers, which was
resolved.  In a second instance, the airbills applied to two similar ice chests containing samples
for different laboratories were switched and the wrong types of containers were sent to each
laboratory. The error was discovered when the samples were received at each laboratory the
next morning.  Both laboratories kept the samples frozen until they were able to ship them to the
intended recipient.
TNSSS Sampling and Analysis Technical Report             12                                   January 2009

-------
3.10   Sampling Summary

       The overall scope of the sample collection effort is summarized in Table 4 below.

 Table 4. Summary Sampling Statistics
People
A total of 12 samplers visited 74
plants and collected 84 samples.
This included 6 field duplicates
and additional samples at 4 plants
that produced 2 types of final
sewage sludge.
Each sampler visited between 2
and 14 plants and collected 12
jars of sewage sludge per plant,
for a total of 1,002 jars.
Travel
46 one-way airline
flights
12 samplers spent
107 days on the
road and drove over
19,000 miles
Shipping
More than 1 50 containers of sampling supplies and
equipment were shipped to field locations during the
survey. Over 3,100 pounds samples and ice packed in
108 coolers were shipped to the EPA Sample
Repository in Baltimore, MD, via FedEx. An additional
4 coolers were hand delivered to the repository.
EPA shipped 427 jars to commercial labs for the
analyses of metals, anions, organics, PBDEs,
Pharmaceuticals, steroids, and hormones. Samples
from each site remain in an archive at the repository
for possible future analyses.
3.11   Equipment Blanks

       All of the equipment that came into contact with samples during the collection process
was made of stainless steel (for organics) or plastic (for metals and anions) and was used for only
one facility to avoid potential cross-contamination between sites.  Prior to sending equipment to
the field, all of the stainless steel and plastic scoops, spoons, and bowls were washed thoroughly
with a non-phosphate detergent, rinsed three times with tap water, rinsed once with reagent
water, inverted and air dried. Once dry, stainless steel equipment was wrapped in aluminum foil
and plastic equipment was sealed in plastic bags.  No field cleaning of equipment was performed
during the survey.  Liquid samples were placed directly into appropriate containers, while solids
samples were placed in appropriate containers using scoops and spoons.

       There is no relevant clean solid "reference matrix" for sewage sludge that could be easily
used to prepare equipment blanks for the variety of analytes in this survey. Therefore,
equipment blanks were created as follows.  Two sets of the relevant equipment were sent to each
laboratory performing analyses of anions, metals,  semivolatiles organics, and PBDEs (e.g., two
stainless steel compositing bowls, two stainless steel scoops, and two glass jars for the organics).
For the semivolatile organics and PBDEs, there were two styles of bowls, based on availability,
and two sizes of glass jars (500-mL and 1000-mL wide-mouths).  Because the glass jars were
purchased with certificates of analysis for common organics, EPA did not anticipate any effect
due to the jar size.  Therefore, EPA assembled one blank using each style of stainless steel bowl
and included one jar size with one style bowl  and  the other jar size with the other bowl. For the
plastic equipment, EPA used only one style of plastic compositing bowl and one size high
density polyethylene resin (FtDPE) jar, so four identical sets of equipment were sent to the
laboratory performing the metals and anions analyses (two sets for each class of analytes).

       Each laboratory used the equipment to prepare the equipment blank by rinsing each piece
with the solvents or solutions used to prepare  field samples. For organics, including PBDEs, the
laboratories poured the same volumes of extraction solvents used for samples over the scoop into
the bowl. The analyst carefully swirled the solvent in the bowl to contact the  majority of the
inner surface, and then poured it into the glass jar.  For metals, the laboratory  used the acidic
sample digestion solution to contact all of the plastic materials. For the anions, the laboratory
used reagent water to "extract" the samples in a similar fashion. The laboratory treated the
TNSSS Sampling and Analysis Technical Report
13
January 2009

-------
solvent or solution as if it came from a nominal size field sample and the reported the results
accordingly, on a dry-weight basis.

       The results for these equipment blanks, or rinsates, represent worst-case estimates of the
potential contributions of the equipment to the final sample results. The worst-case nature of the
estimate reflects the fact that the samples were solids, and even with vigorous mixing during the
compositing steps, not all of the surface area of the sample contacted the entire surface of the
bowl.  The samples themselves did not contain solvents, or acidic solutions, so the potential
transfer of contaminants would be expected  to be much less than using the sample preparation
solutions. Finally, any contaminants that were transferred from the equipment would be
associated with the bulk composite and the final sample sent to the laboratory was only a portion
of the total material in the bowl. In practice, 6 to 8 liters of solid sewage sludge were
composited to provide material to fill eight 500-mL glass jars, or partially fill eight 1000-mL
jars, for the organics analyses.  For the metals and anions, only four plastic 500-mL jars were
filled.  The remainder of the composited material was returned to the sewage sludge disposal
process.

       The results of the equipment blank analyses for the anions, metals, and semivolatile
organics are discussed in detail in Section 6.6.  Based on EPA's experience with the semivolatile
organic analytes of interest in the survey, and given the fact that EPA had exhausted the supply
of one of the types of glass containers used to collect samples, EPA opted not to submit
equipment blanks to the laboratory analyzing the pharmaceuticals, steroids, and hormones.
There was no issue with organic compounds in general regarding contamination of equipment
(e.g., through ambient air, equipment supplier, or laboratory contamination). Thus, the Agency
does not believe that contamination with pharmaceuticals, steroids and hormones would have
been an issue.
TNSSS Sampling and Analysis Technical Report              14                                   January 2009

-------
                                      Section 4
                                 Sample Analyses
4.1    Analytes of Interest
       The TNSSS was designed to do two things: 1) obtain updated occurrence information on
nine pollutants of potential concern, and 2) obtain occurrence information on a number of
contaminants of emerging interest identified by EPA and the National Research Council (NRC).

       As discussed in Section 1, EPA identified nine pollutants (shown in bold in Table 5) for
further evaluation of occurrence in sewage sludge.  This evaluation was based on an assessment
of chemical pollutants for which EPA had adequate data (e.g., human health benchmark values,
and information on fate and transport in the environment).

       Given the national scope of the survey, EPA expanded the list of analytes to reflect the
Agency's interest in collecting concentration data for other chemicals (see Tables 5 and 6).  The
expanded list included 24 additional metals that could be analyzed at little extra cost at the same
time as the four metals (barium, beryllium, manganese, and silver) included in the list of nine
pollutants above; molybdenum because of the Agency's interest in determining the need for a
revised numeric standard for it in  land-applied biosolids;  and  other analytes because of their
widespread incidence and use  and emerging concern.  The latter category included:

      •   benzo(a)pyrene (found in coal tar, automobile exhaust fumes, tobacco and wood
          smoke, charbroiled food, and burnt toast);
      •   2-methylnaphthalene (found in nonstructural caulking compounds and sealants,
          synthetic resins, rubber adhesives, and wall coverings);
      •   bis (2-ethylhexyl) phthalate (widely used as a plasticizer in manufacturing of items
          such as cosmetics,  toys, tools, and laboratory equipment);
      •   fluoride (used in topical and systemic therapy for preventing tooth decay, as well as
          many other uses);
      •   water-extractable phosphorus (correlated with phosphorus concentration in runoff
          from soils amended with manure and biosolids and an indicator of loss that may
          contribute to algae buildup in surface waters);
      •   11 polybrominated diphenyl ethers (PBDEs).  Four of the PBDEs were of most
          interest because of available human health information that may be useful for future
          risk  evaluation efforts. PBDEs are used as flame retardants in a wide array of
          products, including building materials, electronics, furnishings, motor vehicles,
          plastics, polyurethane  foams, and textiles; and
      •   97 pharmaceuticals, steroids, and hormones because of broader emerging interest in
          these analytes.
TNSSS Sampling and Analysis Technical Report             15                                  January 2009

-------
 Table 5. Primary Target Analytes for the TNSSS, by Analyte Class
Analyte Class
Metals
Polycyclic aromatic hydrocarbons
(PAHs)
Other semivolatile organics
Inorganic anions
Polybrominated diphenyl ethers
(PBDEs), including the Tetra, Hexa,
Penta, and Deca congeners
Analyte
Aluminum
Antimony
Arsenic*
Barium
Beryllium
Boron
Cadmium*
Calcium
Chromium*
Cobalt
Copper*
Iron
Lead*
Magnesium
Benzo(a)pyrene
Fluoranthene
bis (2-Ethylhexyl) phthalate
Fluoride
Nitrate
2,4,4'-TrBDE (BDE-28)
2,2',4,4'-TeBDE (BDE-47)
2,3',4,4'-TeBDE (BDE-66)
2,2',3,4,4'-PeBDE (BDE-85)
2,2',4,4',5-PeBDE (BDE-99)
2,2',4,4',6-PeBDE (BDE-100)
Manganese
Mercury*
Molybdenum*
Nickel*
Phosphorus
Selenium*
Silver
Sodium
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc*
2-Methylnaphthalene
Pyrene
4-Chloroaniline
Water-extractable phosphorus
Nitrite
2,2',3,4,4',5'-HxBDE (BDE-138)
2,2',4,4',5,5'-HxBDE (BDE-153)
2,2',4,4',5',6-HxBDE (BDE-154)
2,2',3,4,4',5',6-HpBDE (BDE-183)
2,2',3,3',4,4',5,5',6,6'-DeBDE
(BDE-209)

 The 9 pollutants in bold are those selected in the December 2003 Biennial Review
 * Metals currently regulated at 40 CFR 503
       Among the other "new and emerging contaminants" of concern in the NRC report were
various pharmaceuticals, steroids, and hormones for which several EPA organizations were
developing methods at the time that the TNSSS was being planned. EPA included certain
pharmaceuticals, steroids and hormones in the TNSSS for which analytical methods were
developed. Given the time required to develop and test new methods, EPA proceeded with the
sample collection effort for the TNSSS as described in Section 3, and stored  samples for the
analyses of these analytes of interest until such time as the new methods for these classes of
compounds were more fully developed.  The drugs, antibiotics, steroids, and  hormones added to
the TNSSS are shown in Table 6.
TNSSS Sampling and Analysis Technical Report
16
January 2009

-------
 Table 6.  Pharmaceuticals, Steroids, and Hormones Included in the TNSSS
Analyte Class
Antibiotics and their degradation
products, disinfectants, and other
antimicrobials
Other drugs
Steroids
Hormones
Analyte
Anhydrochlortetracycline
Anhydrotetracycline
Azithromycin
Carbadox
Cefotaxime
Chlortetracycline
Ciprofloxacin
Clarithromycin
Clinafloxacin
Cloxacillin
Demeclocycline
Doxycycline
Enrofloxacin
4-Epianhydrochlortetracycline
4-Epianhydrotetracycline
4-Epichlortetracycline
4-Epioxytetracycline
4-Epitetracycline
Erythromycin
Flumequine
Isochlortetracycline
Lincomycin
Lomefloxacin
Minocycline
Norfloxacin
1 ,7-Dimethylxanthine
Acetaminophen
Albuterol
Caffeine
Carbamazepine
Cimetidine
Codeine
Cotinine
Dehydronifedipine
Digoxigenin
Digoxin
Diltiazem
Campesterol
Cholestanol
Cholesterol
Coprostanol
Desmosterol
Androstenedione
Androsterone
17a-Dihydroequilin
Equilenin
Equilin
17a-Estradiol
17(3-Estradiol
(3-Estradiol-3-benzoate
Ofloxacin
Ormetoprim
Oxacillin
Oxolinic acid
Oxytetracycline
Penicillin G
Penicillin V
Roxithromycin
Sarafloxacin
Sulfachloropyridazine
Sulfadiazine
Sulfadimethoxine
Sulfamerazine
Sulfamethazine
Sulfamethizole
Sulfamethoxazole
Sulfanilamide
Sulfathiazole
Tetracycline
Triclocarban
Triclosan
Trimethoprim
Tylosin
Virginiamycin

Diphenhydramine
Fluoxetine
Gemfibrozil
Ibuprofen
Metformin
Miconazole
Naproxen
Norgestimate
Ranitidine
Thiabendazole
Warfarin

Epi-coprostanol
Ergosterol
(3-Sitosterol
(3-Stigmastanol
Stigmasterol
Estriol
Estrone
17a-Ethynyl estradiol
Norethindrone
Norgestrel
Progesterone
Testosterone

TNSSS Sampling and Analysis Technical Report
17
January 2009

-------
4.2    Analytical Techniques

       Table 7 presents the analytical techniques applied to samples in the TNSSS.  The target
reporting limits in the table are based on a consensus of what might be achievable in a sewage
sludge sample. The actual reporting limits achieved are discussed in Section 6.5.

 Table 7. Analytical Methods or Techniques
Analyte Class
28 Metals, including mercury
4 Polycyclic aromatic hydrocarbons
(PAHs) and 2 semivolatiles (as one
analytical fraction)
4 Inorganic anions, including water-
extractable phosphorus (WEP)
11 PBDE Congeners*
72 Pharmaceuticals
25 Steroids and hormones
Method or Technique
ICP/AES, ICP/MS, and CVAA
(EPA Methods 200.7, 200.8, and 245.1)
GC/MS, with selected ion monitoring (SIM), after
solvent extraction and gel permeation
chromatography (GPC) cleanup
(EPA SW-846 Method 8270C)
EPA Methods 340.2, 353.2, and 365.3, after
leaching of the solid sample with reagent water
with a study-specific protocol
High resolution GC/MS, draft EPA Method 1614
High performance liquid chromatography (HPLC)
with tandem MS/MS detection, using an early draft
of EPA Method 1694f
High resolution GC/MS, using an early draft of
EPA Method 1698f
Target Reporting
Limit (dry weight)
3 to 4 mg/kg
100 to 300 ug/kg
2 to 8 mg/kg
5 to 200 ng/kg
Not specified
Not specified
 "The list of target PBDE analytes was limited to the following 11 PBDE congeners: 28, 47, 66, 85, 99, 100, 138, 153, 154, 183,
 and 209, which include those identified in the method as being of potential environmental or public health significance.  There are
 some differences between the specifics of the method and the procedures used for the TNSSS (see Section 4.4.4 of this report).
 t The laboratory solicitation and contract were issued prior to the December 2007 formal release of EPA Methods 1694 and 1698
 and as a result, there are some differences between the specifics of those methods and the procedures used for the TNSSS (see
 Sections 4.4.5 and 4.4.6 of this report).
 mg/kg = milligrams per kilogram  ug/kg = micrograms per kilogram   ng/kg = nanograms per kilogram

       As indicated, the survey used both well-established multi-laboratory validated EPA
procedures as well as three analytical methods that were developed or updated for the survey.
The two new methods are single-lab validated methods for pharmaceuticals (EPA Method 1694),
and steroids and hormones (EPA Method 1698). The updated multi-lab validated method is for
flame retardants (EPA Method 1614).  These three methods have  not yet been promulgated  at 40
CFR Part 136 for compliance monitoring in CWA programs,  including the analysis of sewage
sludge.

4.3    Laboratories

       EPA awarded competitive-bid analytical contracts to the following commercial
laboratories:

    •  Columbia Analytical Services for the analyses of the metals, anions, and PAHs and
       semivolatile organics
    •  Severn  Trent Laboratories  (now part of Test America) for the analyses of the PBDEs
    •  Axys Analytical Services for the analyses of pharmaceuticals,  steroids, and hormones.
TNSSS Sampling and Analysis Technical Report
18
January 2009

-------
4.4    Method Modifications

       From an analytical standpoint, sewage sludge is a challenging matrix. The concentrations
of pollutants present in samples vary depending on the nature of the inputs to the treatment plant.
In addition to the pollutants of interest, sewage sludge contains a number of other components
that are potential interferences in the analyses of the pollutants of interest. These components
include lipids and other naturally occurring materials, as well as materials that may be added to
the sewage during processing (e.g., surfactants, ferric chloride, polymeric colloids, or lime).
These components can manifest themselves as interferences at all stages of the analytical
process, from sample preparation through the determinative analysis.

       Another analytical challenge with a national survey of sewage sludge is that the various
treatment process and disposal practices used nationwide lead to differences in the moisture
content of the final sewage sludge sent for use or disposal. Some of the facilities from which
samples were obtained in the TNSSS produce liquid final sewage sludge, while others produce
solid sewage sludge. Among the sewage sludge that were pourable liquids, the percent solids
(hereafter percent solids) content ranged from less than 1% to about 4%, across treatment plants.
For the solids, the percent solids content ranged from 5% to 99%.  These differences in the form
(liquid vs. solid) of the sewage sludge and the range of moisture or solid contents have direct
effects on the analyses of the samples.  The differences also affect how the data for the survey
can be interpreted.

       Recognizing these challenges, EPA structured the laboratory subcontracts for the various
analyses to achieve the most uniform results across facilities as practical.  These modifications
are described below.

4.4.1   Ensuring Consistent Method Sensitivity

       Many analytical methods applicable to biosolids instruct the laboratory to prepare a
specific known weight of a solid material for analysis (e.g., some methods for organics specify
using 30 g of sample).  However, that sample aliquot may contain significant amounts of
moisture.  Those same methods may treat samples that are pourable liquids as if they contain
little or no solids, and specify using a known volume,  such as 1 L, for the analysis, although that
volume may contain measurable solids as well.  These differences in how liquid and solid
samples are prepared and analyzed, as well as the differences in the amount of solids or moisture
in the two types of samples, mean  that any measure of method sensitivity (e.g., a reporting limit
or a detection limit) will depend on the initial mass or volume chosen for analysis and its
moisture content.

       EPA considered these effects on sensitivity and comparability when it planned the
TNSSS.  EPA minimized the potential sensitivity differences by instructing the laboratories to
determine the percentage of solids (percent solids) of each sample first, and then use that
information to select a portion of the sample for the analysis  that contains the method-specified
sample weight or volume on a dry-weight basis.  In addition, even when the laboratories
prepared liquid samples using procedures designed for aqueous samples (e.g., liquid-liquid
extraction with an organic solvent), they were instructed to report the  results in weight/weight
units (e.g., ng/kg, ng/kg, or mg/kg) appropriate for the class of analyte, adjusted for the moisture
content of the sample (e.g., 100% dry sample).

TNSSS Sampling and Analysis Technical Report             19                                   January 2009

-------
       The laboratory that performed the analyses of the metals, anions, and PAH and
semivolatile organics was instructed to determine the percent solids separately for each class of
analytes. EPA examined the percent solids data from each class for the first 35 samples and
determined that there was no statistically significant difference between the three measurements
on each EPA sample (see  Section 6.7.7). Nevertheless, in order to ensure the most consistent
sensitivity across the survey, EPA instructed the laboratory to continue to determine percent
solids on each sample for  each class for the remainder of the project.

4.4.2   Anions in Sewage Sludge

       Methods for determining anions (e.g., nutrients) in sewage sludge samples require that
the anions be dissolved in water and separated from the solid material. In planning the TNSSS,
EPA considered several approaches to preparing the  sewage sludge samples for the analysis of
anions, including:

   •   EPA Method 1685, Nitrate/Nitrite-N in Water and Biosolids by Automated Photometry,
       Draft January 2001;
   •   EPA Method 1688, Total Kjeldahl Nitrogen in Water and Biosolids by Automated
       Colorimetry with Preliminary Semi-automatic Digestion, Draft January 2001;
   •   A water extraction (or leaching) procedure developed at the Pennsylvania State
       University (Vadas, P. A. and Kleinman, P. J.  A., 2006).

       The two draft EPA methods cited above have only been validated in a single lab.  Neither
of these methods has been promulgated at 40 CFR Part 136 for compliance monitoring in CWA
programs, nor for the analysis of sewage sludge samples. As a result, few, if any, laboratories
routinely run samples using  those draft methods.

       The method developed at Pennsylvania State University was published in the literature
and has been used by several mid-western states that regulate the application of manure and
biosolids to agricultural lands.  However, the authors only used that procedure for the analysis of
phosphorus, and not the other anions.

       After reviewing the Vadas and Kleinman leaching procedure, EPA concluded that the
leachate was amenable to  analyses of the anions of interest in the TNSSS by existing commonly
used EPA methods that are approved at 40 CFR 136. Therefore, EPA decided to use the
leaching procedure to prepare all of the sewage sludge samples and have the anion analyses
performed using the following EPA methods:

   •   Method 340.2, Fluoride, Potentiometric, Ion Selective Electrode, March 1983
   •   Method 353.2, Nitrate-Nitrite, Colorimetric, Manual  Cadmium Reduction, March 1983
   •   Method 365.3, Phosphorus, All Forms, Colorimetric, Ascorbic Acid, Two Reagent,
       March 1983
TNSSS Sampling and Analysis Technical Report             20                                   January 2009

-------
       Appendix A of this report provides the TNSSS-specific directions for the leaching
procedure.  Briefly, the procedure involves:

   •   Determining the percent solids of the original sample using standard procedures
   •   Weighing a sample aliquot equal to 0.5 g (dry weight) into a plastic bottle
   •   Adding 100 mL of reagent water
   •   Shaking the bottle on a shaker table for 60 minutes at 70 RPM
   •   Centrifuging the mixture for 10 minutes at 2000 RPM
   •   Filtering the sample by gravity through a Whatman #2 filter
   •   Preserving the aqueous leachate sample by adding H2SO4 to pH <2
   •   Analyzing the leachate sample within 48 hours (the holding time for nitrate/nitrite)
   •   Reporting all results in mg/kg, based on the original 0.5-g sewage sludge sample weight.

       The leaching procedure includes steps that will preserve the relative proportions of nitrate
and nitrite after the sewage sludge sample is leached (e.g., acid preservation and analysis within
48 hours). However, it is not possible to determine whether or  not the leaching process itself
resulted in oxidation of nitrite to nitrate, or vice versa.  Therefore, EPA decided that nitrate and
nitrite would be analyzed as the combined parameter nitrate/nitrite, using Method 353.2. This is
approach is acceptable for determining better exposure scenarios for nitrate/nitrite during land
applicable of biosolids.

       In addition, the analytical approach for the survey included determining the element
phosphorus (P) as part of the suite of metals. The water-extractable phosphorus (WEP)
determined using the leaching procedure above is a useful predictor of the concentrations of
phosphorus that might be available for runoff from land to which sewage sludge has been
applied. The ratio of the two forms of phosphorus (WEP/P) is an indication of the proportion of
the total phosphorus applied that may contribute to runoff. That ratio may be of interest to those
states that regulate land application of sewage sludge.

Note:  The reader is cautioned about making comparisons between the anion results from this
       survey and data from other sources that may have used different procedures to leach the
       anions from the sample. EPA's use of dry-weight reporting units may facilitate such
       comparisons, but other differences among leaching procedures may still influence the
       results.

4.4.3   Modified GC/MS Procedures for Semivolatile Organics

       Gas chromatography, coupled with mass spectrometric detection (GC/MS), is the
backbone of the analytical methods for many organic pollutants, including the PAHs and
semivolatile organics of interest in the TNSSS.  The most common form of GC/MS analysis is
known as "full scan" GC/MS and involves examining all of the mass fragments in a wide mass
range that exit the GC column and reach the MS detector. A typical GC/MS method will scan a
mass range from 35 to 450 atomic mass units (amu) once every second.  The masses of any
materials that exit the GC column in that range will be recorded and used to identify the
pollutants of interest.
TNSSS Sampling and Analysis Technical Report             21                                   January 2009

-------
       Some EPA GC/MS methods include several hundred target analytes, and while full-scan
GC/MS is a powerful technique, it involves tradeoffs in sensitivity and selectivity in order to be
applied to large number of analytes simultaneously. Full-scan GC/MS can also be subject to
interferences from other materials in the sample that are not of interest.

       Some of these "co-extracted" interferences include biolipids that can be removed from
the sample extracts using established cleanup procedures. One such procedure is gel permeation
chromatography (GPC), which segregates the relatively small pollutant molecules from the
larger lipids and other interferences on the basis of molecular size. Based on our experience in
two previous sewage sludge surveys, EPA required every sample extract analyzed for the PAHs
and semivolatile organics be subjected to GPC cleanup before analysis.

       When a GC/MS instrument is operated in full-scan mode, it scans the entire mass range
very quickly and there is relatively little time to observe the results at any given mass within that
range.  This places some practical limits on the sensitivity of the procedure.  However, those
limitations can be overcome by using the technique known as selected ion monitoring (SIM). In
SIM, the MS instrument only looks for a small subset of masses (ions) in the overall mass range.
These masses are  ones associated with the list of target analytes, and any other masses that exit
the GC column are simply  ignored. The ions that are ignored may be those associated with
interferences, including biolipids, or those from analytes in a full-scan method that are not of
interest for a given project. Because the MS can spend more time looking for fewer masses, the
sensitivity of the instrument for those pollutants with those masses can increase 10-fold or more
over that of a full-scan procedure.

       EPA examined the results for the first 50 samples that were analyzed by full-scan GC/MS
and found that for many of those samples, the reporting limits achieved by the laboratory were
much higher than  anticipated.  In other words, analytes were reported as "not detected"  at
concentrations greater than the reporting limits desired for the survey.  Many of the increased
reporting limits were due to large amounts of interferences that remained in the sample  extract
even after GPC cleanup, which required dilution of the extract. The laboratory also diluted a
smaller number of samples in order to get one or two target analytes (usually bis [2-Ethylhexyl]
phthalate) within the instrument calibration range, resulting in a loss of sensitivity for other
analytes present at much lower concentrations.

       In response to these early findings, EPA required that the laboratory reanalyze the
extracts for about  35 samples using a SIM procedure.  That procedure included masses that
represented the four PAHs and two semivolatile organics that are target analytes in the survey, as
well as the surrogate compounds and internal standards used in the full-scan method. The
masses used in the SIM procedure are shown in Table 8. Based on successful analyses  of these
extracts, the SIM procedure was the only GC/MS procedure employed for the analysis of the
remaining samples in the survey.

 Table 8. Selected Ion Monitoring Parameters for Organic Analytes
Type
Target
Analytes
Analyte
4-Chloroaniline
2-Methylnaphthalene
Fluoranthene
Pyrene
bis (2-Ethylhexyl) phthalate
Quantitation Mass
127
142
202
202
149
Approximate Retention Time (min)*
9.71
10.49
15.04
15.32
16.61
TNSSS Sampling and Analysis Technical Report             22                                   January 2009

-------
 Table 8. Selected Ion Monitoring Parameters for Organic Analytes
Type

Surrogates
Internal
Standards
Analyte
Benzo(a)pyrene
Nitrobenzene-ds
2-Fluorobiphenyl
p-Terphenyl-di4
1 ,4-Dichlorobenzene-d4
Naphthalene-ds
Acenaphthene-d-io
Phenanthrene-dio
Chyrsene-di2
Perylene-di2
Quantitation Mass
252
82
172
244
152
136
164
188
240
264
Approximate Retention Time (min)*
18.12
8.67
10.94
15.49
7.90
9.59
11.77
13.57
16.63
18.18
 •Retention times are specific to the GC column and operating conditions used for these samples.  Retentions times on other
 columns or instruments will differ. These data are presented solely to illustrate the relationships among the target analytes,
 surrogates, and internal standards.

4.4.4  PBDE Analyses

       The original plan for the survey was to analyze for all 209 of the PBDE congeners using
the latest draft of EPA Method 1614, a high resolution GC/MS isotope dilution procedure.
Method 1614 is a highly sensitive procedure that can determine PBDEs at the part per trillion
(ng/kg) levels in solid samples.  It employs at least five cleanup techniques, including GPC, to
remove interferences from sample extracts.  The method employs isotope dilution quantitation,
in which PBDE congeners synthesized using only carbon 13 (13C), a stable (nonradioactive)
isotope of carbon, are added to the sample prior to extract. These isotopically labeled congeners
do not occur in nature and are used as internal standards to quantify the unlabeled PBDE
congeners. Because the labeled congeners are carried through the entire sample preparation,
cleanup, and analysis process, they can be used to correct for any loss of analytes during the
overall analysis, providing a more accurate result for each unlabeled target analyte.  For this
reason, EPA uses isotope dilution in many of the 1600-series methods, including those for
dioxins, furans, PCBs, and PBDEs.

       Shortly after work began on the first batch of survey samples, the laboratory and EPA
realized that additional efforts would be required to adequately determine PBDEs in sewage
sludge. Ultimately, EPA agreed to permit the use of a range of modifications to the original plan
to overcome as many analytical difficulties as practical in the time permitted for the TNSSS.
Appendix B contains a detailed list of those modifications. The most significant modifications
were:

   •  Due to the high levels of some congeners and some interferences, the initial sample size
       was reduced from 10 g to 2 g for extraction, and ultimately to 0.2 g (dry weight).
   •  Samples were not spiked with labeled congeners prior to extraction.
   •  The sample  extracts were concentrated to a final volume of 10 mL after extraction.
   •  A 1-mL aliquot of the extract was removed for spiking and cleanup.
   •  Labeled congeners were spiked into the 1-mL extract, which was then carried through the
       following cleanup steps described in the method: silica gel, GPC, and alumina.  The
       remaining 9 mL of extract were retained in case dilutions or additional cleanups were
       required.
   •  After cleanup, extracts were analyzed by HRGC/HRMS.
TNSSS Sampling and Analysis Technical Report
23
January 2009

-------
   •   The list of target analytes was limited to the following 11 PBDE congeners: 28, 47, 66,
       85, 99, 100, 138, 153, 154, 183, and 209, which includes those identified in the method as
       being of potential environmental or public health significance.
   •   Data were examined to determine if a larger sample size (greater than 0.2 g) was required
       to achieve the desired sensitivity, or if the extract required dilution to keep analytes
       within the calibration range.  The latter was much more common.
   •   If congeners were present above the upper limit of the calibration range, the laboratory
       evaluated whether or not the peaks had saturated the detector. If saturation did not occur,
       the results were reported, but flagged in the database as exceeding the calibration range.
       This approach reduced the dilutions that had to be analyzed for each sample to a practical
       number.
   •   Because the labeled congeners are not spiked into the samples before extraction, matrix
       spike and matrix spike duplicate samples were prepared and analyzed periodically to
       provide an estimate of extraction efficiency.
   •   Results for the 11 congeners were corrected for any losses of analytes that occur during
       the many cleanup steps, but this is not equivalent to the true isotope dilution quantitation
       procedure in the original method because it does not correct for the efficiency of the
       sample extraction procedures.

       These steps were successful in overcoming challenges inherent in sewage sludge
analyses, and produced useful data for the purposes of the TNSSS.  However, these steps
involved some trade-offs, most notably the loss of true recovery correction of isotope dilution
quantitation and the degree to which the very small sample size represents the bulk material.

4.4.5   Pharmaceutical Analyses

       The analysis of the  pharmaceuticals was performed using an early draft of EPA Method
1694. EPA Method 1694 employs high performance liquid chromatography (HPLC) to separate
the analytes of interest and tandem mass spectrometry (MS/MS) to detect them. HPLC (or
simply LC) is a technique that allows the analysis of polar compounds in polar solvents.  It has
advantages over gas chromatographic (GC) methods for the pharmaceuticals because GC
methods involve introducing the analytes into the instrumentation in a gaseous form and many of
the pharmaceuticals are not easily volatilized.  Some have boiling points that are above the
operating temperatures of a GC system and others will break down when heated (i.e., they are
"labile").

       Tandem mass spectrometry involves the use of two quadrapole mass  spectrometers in
series, with a collision cell  between them, such that selected ions produced in the first MS unit
are directed into the collision cell and further fragmented before being sent to the second MS for
detection.  The only ions passed through the collision cell are those selected by the instrument as
representing the analytes of interest.  These fragments, or "product" ions, are characteristic of the
"precursor" compound and are used to positively identify the analyte in the presence of other
analytes and potential interferences.  The MS/MS detector  can be operated in an ionization mode
that produces positive ions from the analytes of interest, or in a mode that produces negative
ions.  The method for pharmaceuticals involves four analytical fractions, three of which operate
in the positive ionization mode and one of which uses the negative ionization mode. As
employed in this method, the tandem MS unit is operated at unit mass resolution (e.g., it can
distinguish between masses that differ by one atomic mass unit).

TNSSS Sampling and Analysis Technical Report              24                                   January 2009

-------
       The Pharmaceuticals method employs solvent extraction procedures to isolate the
analytes of interest from the sewage sludge samples. Many of the analytes are weak acids or
weak bases that ionize in aqueous solutions, losing or gaining a proton from a water molecule.
The extraction procedures in the pharmaceutical method involve adjusting the pH of the sample
to provide more favorable conditions for isolating the analyte from the sample matrix.

       An aliquot of each sewage sludge sample containing a consistent dry weight of solids is
mixed with a phosphate buffer solution with a pH of 2.  This pH adjustment causes the ionized
acid forms of the analytes of interest to gain protons and become neutrally charged molecules
that are less  soluble in water than their ionized forms and more soluble in a  polar organic  solvent.

       A suite of stable isotopically labeled standards (forms of the analytes that do not occur
naturally) is spiked into the sample and the sample is further mixed. The analyst adds
acetonitrile to the buffered sample, ultrasonically agitates the mixture  for 30 minutes, and
centrifuges the solution to separate the solvent extract from the solids. After decanting and
collecting the acetonitrile, the analyst performs a second extraction of the solids with  fresh
aqueous buffered acetonitrile, and a third extraction using acetonitrile  alone. All three acid
extracts are combined for cleanup.

       The base extraction is conducted in a similar fashion, using a second aliquot of the
original sample, but the pH of the sample is adjusted to  10 with an ammonium hydroxide
solution. Three extractions are performed, two with aqueous buffered acetonitrile and a third
extraction with acetonitrile alone.  All three base extracts are combined for cleanup.

       The combined acid extract is concentrated to remove the acetonitrile and prepared for
cleanup by adding disodium EDTA and diluting the solution to 200 mL with reagent water.  The
aqueous solution is processed through a solid-phase extraction (SPE) cartridge, which traps the
analytes of interest.  Potential interferences are removed by eluting the cartridge with reagent
water and discarding that eluant. The analytes of interest are eluted from the cartridge with
methanol, followed by  1:1 acetone:methanol.  The eluant is evaporated to near dryness,
reconstituted in methanol, spiked with the method-specified internal standards, and brought to a
final volume of 4 mL with a 0.1% formic acid buffer solution.

       The base extract is subjected to a similar SPE cleanup procedure, but the cartridge is
eluted with methanol, followed by 2% formic acid in methanol.  The combined eluant is
evaporated to near dryness, reconstituted in methanol, spiked with the method-specified internal
standards, and brought  to a final volume of 4 mL with a 0.1% formic acid buffer solution.

       Four separate LC/MS/MS analyses are performed for the pharmaceuticals: three on the
acid extract and one on the base extract. Separate chromatographic conditions are associated
with each of the four fractions.  The LC/MS/MS instrument is operated in the multiple reaction
monitoring (MRM) mode, which monitors a series of precursor/product ion transitions that are
characteristic of each target analyte.

       The labeled analytes spiked into the sample prior to extraction  are used to perform
isotope dilution quantitation for all of the target analytes that have labeled analogs. For those
target analytes for which  labeled analogs are not readily available, quantitation is performed
using the labeled analog of a similar compound in that fraction. As a result, all of the target

TNSSS Sampling and Analysis Technical Report              25                                   January 2009

-------
analyte concentrations are corrected for the recovery of the labeled analogs, thus accounting for
extraction efficiencies and losses during cleanup.

       The approach to ensuring more consistent method sensitivity described in Section 4.4.1
of this report was applied to the pharmaceutical analyses as well.

4.4.6  Steroid and Hormone Analyses

       The analysis of the steroids and hormones was performed using an early draft of EPA
Method 1698.  EPA Method 1698 employs GC to separate the analytes of interest and high
resolution mass spectrometry (HRMS) to detect them.  As employed in this method, the mass
spectrometer achieves a resolution of at least 5,000.  The target analytes in the method have
molecular weights that range from about 100 to 500, such that the MS can distinguish between
analytes with molecular weights that differ by 0.02 to 0.1 atomic mass units, and identify them in
the presence of potential interferences.

       The steroids and hormones method also employs solvent extraction procedures to isolate
the analytes from the sewage sludge samples.  An aliquot of each solid sample containing a
consistent dry weight of solids is spiked with a suite of stable isotopically labeled standards. The
sewage sludge samples that are firm solids are extracted in a Soxhlet extractor with 60:40
acetone:hexane.  The extract is split into two portions, with l/25th of the extract used for steroids
analysis and the other 24/25th used for the hormones analysis. This split ratio is used to
compensate for the fact that the steroids and their interferences are present at higher levels in the
samples than the hormones.

       The sewage sludge samples that are pourable liquids are extracted using a liquid-liquid
solvent extraction with methylene chloride.  The extracts of these samples also are split into two
portions, with a split ratio of 1:99, for the steroids and hormones respectively.

       For both forms of sewage sludge (liquid and solid), the separate extract portions are
subjected to cleanup using a layered alumina-Florisil (LAP) column.  Following cleanup, each
sample extract is concentrated to approximately 0.1 mL and the extract solvent is exchanged to
pyridine.

       In order to be amenable to GC analysis, the steroids and hormones are  converted to
compounds that are more volatile than their native forms. Both the steroids and hormones are
derivatized to their trimethyl-silyl ethers using N,O-bis(trimethylsilyl) trifluoroacetamide with
trimethylchlorosilane (BSTFA:TMCS). The derivatized extracts are concentrated and spiked
with the method-specified internal standards and the final extract volume  is adjusted to 500 uL.

       Separate GC/HRMS analyses  are performed for the steroids and for the hormones, using
the same chromatographic conditions for each analytical fraction, but running  the two fractions
at different dilutions. The GC/HRMS instrument is operated at a mass resolution of at least
5,000, and at least two exact masses are monitored for each target analyte.

       As with the pharmaceuticals method, the labeled analytes spiked into the sample prior to
extraction are used to perform isotope dilution quantitation for all  of the target analytes that have
labeled analogs.  For those target analytes for which labeled analogs are not readily available,

TA/SSS Sampling and Analysis Technical Report             26                                   January 2009

-------
quantitation is performed using the labeled analog of a similar compound in that fraction.  As a
result, all of the target analyte concentrations are corrected for the recovery of the labeled
analogs, thus accounting for extraction efficiencies and losses during cleanup.

       EPA worked with the laboratory to adjust the procedures for the steroids and hormones to
address problems encountered early in the survey. In particular, the laboratory modified one of
the cleanup procedures.  As written, the steroids and hormones method involves solvent
extraction of the sample and processes that extract through a layered alumina-Florisil (LAP)
cleanup column before splitting the extract into two unequal portions (1/25 and 24/25) for
analyses of steroids and hormones, respectively.  However, some of the steroids and hormones
are present in both  extract portions and during the early part of the survey, the high levels of
steroids were creating analytical challenges during the analysis of the hormones fraction.

       The laboratory proposed a  solution to the problem that involved modifying the order in
which the extract was subjected to cleanup and split into two portions.  Instead of running the
entire extract through the LAP column and then splitting it into two unequal portions, the raw
extracts were split first and then run through separate LAP cleanups, one for steroids and one for
hormones.  The LAP cleanup for the steroids was run as described in the original procedure.

       The laboratory modified the LAP procedure for the hormones to remove most of the
steroids from the hormone fraction. The steroids were eluted from the LAP column with
methylene chloride and that eluant was discarded. The hormones then were eluted from the
column with methanol, as described in the original procedure.

       The laboratory tested the method modifications prior to starting analyses of the survey
samples by extracting large samples of sewage sludge from another source (e.g., not a TNSSS
sample), spiking the extracts with the hormones of interest to ensure that they were present, and
splitting each extract into two portions.  One of the portions was run through the LAP column as
described in the method. The other portion was subjected to the modified LAP procedure.  Both
extract portions were analyzed for hormones by GC/HRMS and the results were compared.  The
change to the LAP  procedure dramatically reduced the levels of steroids that remained in the
hormone fraction, and improved the results for the hormones.

       As noted in Section 4.2, the target analyte list for the TNSSS was based on a variety of
factors, but was not driven by lists of analytes in any individual methods.  EPA had included
Mestranol as a target analyte in early versions of the survey plan for the TNSSS.  However,
Desogestrel was never listed as a target analyte for the survey. Therefore, this  modification to
the LAP cleanup procedure sacrificed one hormone in favor of improvements for all the other
target hormones. Given the difficulties inherent in sewage sludge analyses, EPA judged this to
be a reasonable compromise.
TNSSS Sampling and Analysis Technical Report             27                                   January 2009

-------
4.4.7  Focusing Quality Control on the Survey-specific Analytes of Interest

       Because the analytes of interest for the TNSSS often are a very limited subset of the total
number of analytes listed in some of the relevant methods, EPA did not require the laboratories
to consider those other analytes in either preparing or evaluating the quality control operations
associated with these samples. For example, EPA only required the laboratory to spike the
analytes of interest for this  survey into such QC samples as matrix spike/matrix spike duplicates
and ongoing precision and recovery (OPR) samples, not all the analytes that may be listed in a
given method.  Alternatively, if the laboratory chose to use spiking solutions that contained
additional analytes beyond  those in Tables 5 and 6, EPA did not require them to assess the
results for those non-survey target analytes or take corrective actions if those non-survey
analytes failed to meet the acceptance criteria.

       This approach allowed the laboratories to focus their efforts on the survey-specific
analytes of interest in the face of the analytical challenges presented by sewage sludge matrices.
It also allowed EPA to control the costs for the analyses to some degree by eliminating
reanalyses related to potential QC failures associated with non-target analytes.  It also reduced
the costs for EPA's data review efforts  described in Section 5.
TNSSS Sampling and Analysis Technical Report             28                                    January 2009

-------
                                      Section 5
                             Data Review Procedures
5.1    General Review Procedures

       EPA assessed the results for all of the samples analyzed during the survey using well-
established procedures described in this section. The analysis involved in the TNSSS was
complex and a number of analytical challenges were faced. Biosolids is one of the most
challenging environmental matrices known due to the high solids content and matrix
interferences present. When conducting analyses in sewage sludge matrices it is expected that
some results will have to be qualified to accurately reflect the uncertainty of the values.

       EPA subjected all laboratory results to a comprehensive review for completeness and
compliance with project and method specifications to ensure that the data met the objectives of
the survey. A multi-stage review process was used and designed to identify and correct data
deficiencies as early as possible and maximize the amount of usable data generated.

       Trained staff reviewed the data using established review process designed to identify and
correct data deficiencies as early as possible and maximize the amount of usable data generated
during the TNSSS. EPA encoded the data quality information gathered during the review in the
final results database using a series of qualifiers and reasons.  EPA did not exclude data unless a
result was flawed such that no reasonable use could be made  of it.  The four stages of the review
process are described below.

       In the first stage of the data review process, EPA performed a data completeness check.
Specifically, EPA evaluated elements in the laboratory submission to verify that results for
specified samples were provided, that data were reported in the correct format, and that relevant
information, such as preparation and analysis logs, was included in the data package.  EPA
initiated corrective action procedures to resolve any deficiencies identified.

       The second stage of the data review process focused on an instrument performance
check.  EPA verified that calibrations, calibration verifications, standards, and calibration blanks
were analyzed at the appropriate frequency and met method or survey performance
specifications. Corrective action procedures were initiated to resolve any deficiencies identified.

       The third stage of the data review process focused on  a laboratory performance check.
EPA verified that the laboratory correctly performed the required analytical procedures and was
able to demonstrate a high level of precision and accuracy. During this stage, EPA evaluated
quality control (QC) elements such as the ongoing precision and recovery (OPR) tests, method
blanks, and other QC operations. Again,  corrective action procedures were initiated to resolve
any deficiencies identified.
TNSSS Sampling and Analysis Technical Report             29                                   January 2009

-------
       In the fourth stage of the data review process, EPA examined method/matrix performance
data to discern whether any QC failures resulted from laboratory performance or difficulties with
the method or sample matrix.  EPA evaluated labeled compound and surrogate spike results and
other performance data.  The reviewers also verified that proper sample dilutions were performed
and that necessary sample cleanup steps were taken.  As with previous steps, corrective action
procedures were initiated with the laboratory to resolve any deficiencies identified.

       The objective of the data review process was to document the quality of all of the data in
the TNSSS and identify any limitations that might affect their end use. The EPA database on the
mainframe contains data qualifiers applied to results from the  TNSSS, individually, and by
analyte class.

5.2    QC Acceptance Criteria

       As noted in Section 4, the analytes of interest for this survey were subject to quality
control. Appendix C presents a summary of the QC acceptance criteria for all of the analyte
classes.

       The laboratory performing the analyses for the anions, metals, PAHs and semivolatiles
used its own in-house limits routinely applied to soil  samples (as opposed to aqueous samples),
except as noted in Appendix C.  The laboratory prepared liquid sewage sludge samples using
procedures applicable to aqueous samples (e.g., a liquid-liquid solvent extraction procedure for
the organics) and the laboratory ran an aqueous laboratory control sample (LCS) and applied
acceptance limits appropriate to that set of procedures.  The laboratory derived their acceptance
limits for solid samples from historical data for sewage sludge samples. Given the difficulties
evident in the analyses of the sewage sludge samples, one might expect that statistically derived
acceptance criteria for sewage sludge samples would differ from those for soil samples.
Therefore, recoveries falling slightly outside of the limits for soil are not necessarily fatal flaws
in these analyses. However, for the sake of transparency, EPA noted all recovery problems in
the database.

       The laboratory performing the analysis of the PBDEs using draft EPA Method 1614
employed the default acceptance limits in the draft method.

       The methods used for the pharmaceutical, steroid, and  hormone analyses include QC
acceptance criteria.  The acceptance criteria include statistically derived limits that are based on
data from a single laboratory.  As a result, the acceptance limits for some QC operations are
fairly wide (e.g., 5 - 200% for labeled  compound recoveries).  The laboratory that performed the
pharmaceutical, steroid, and hormone analyses for the TNSSS was also the laboratory that
helped EPA develop these methods. Therefore, EPA believes that the use of the single-
laboratory specifications is a reasonable approach.

5.3    Data Qualifiers and Database

       After reviewing each data package, the reviewers and the database development staff
created an analytical database in MS Access that contained all field sample results from the
survey. Appendix D of this report contains a table with the data applied to the results during the
review process for TNSSS pollutants and entered into the database.

TNSSS Sampling and Analysis Technical Report              30                                   January 2009

-------
       At intervals during development and upon completion of the database, EPA performed
various checks to verify the accuracy of the database, including checks for consistent analyte
names, CAS numbers, and data qualifier flags.  After completing the review of the results for
each analytical fraction, EPA uploaded the data from an Access database to the EPA mainframe
using standardized procedures. The final database for the survey includes all of the qualifiers
that were applied to the individual sample results.

Note:  Except for the "exclude" qualifier in the database, the presence of data qualifiers is not
       intended to suggest that data are not useable.  Rather, the qualifiers are designed to
       caution the user about an aspect of the data that does not meet the acceptance criteria
       originally established for the project.

5.4    Data Review Findings

       The data review process was crucial in identifying the sensitivity issues associated with
the full-scan GC/MS analyses discussed in Section 4. By examining the first sets of sample
results shortly after they were delivered by the laboratory, EPA was able to take corrective action
and instituted the use of selected ion monitoring for all PAH analyses in the survey to achieve
the needed analytical sensitivity.

       The data review process uncovered only a few other issues with data quality. Typical
issues included:

   •   Reporting unit issues
   •   Calculation errors for specific samples
   •   Blank contamination
   •   Extract dilution issues
   •   Matrix spike issues
   •   Interference issues
   •   Recovery issues

Specific examples  of these issues are presented in Section 5.4.1 to 5.4.7.

5.4.1   Reporting Limits

       As expected, some samples with low percent solids  were prepared and analyzed as
aqueous samples. However, when reporting the results for  the earliest such samples, the
laboratory performing the metals analyses provided results  in the weight/volume units typically
applied to aqueous samples.  When contacted by the data reviewers regarding this error, the
laboratory revised these results and reported them on the basis of the dry weight of solids in the
samples, as originally requested.
TNSSS Sampling and Analysis Technical Report              31                                    January 2009

-------
       The laboratory analyzing the pharmaceuticals, steroids, and hormones reported all the
sample results on a dry-weight basis, as required, including those samples that were pourable
liquids.  However, the laboratory used reporting units of nanograms per gram (ng/g), instead of
the method-specified units of nanograms per kilogram (ng/kg).  This error was discovered during
the early stages of the review of these data.  However, because the units of ng/g are equivalent to
jig/kg, after consulting the laboratory and considering the concentrations of the analytes in the
samples, EPA decided to accept the numerical results in ng/g (e.g., EPA kept the numbers
reported by the lab), but changed the units in the EPA database to |ig/kg (i.e., 100 ng/g is  equal
to 100 jig/kg).

5.4.2   Calculation Errors

       EPA noted that the fluoride results for two sewage sludge samples appeared to be  much
higher than any other results for this analyte. EPA contacted the laboratory and asked them to
check these two results. The laboratory reported that the raw data for these two results were off
by two orders of magnitude, due to a transcription error. The laboratory corrected the error  and
resubmitted the results for those two samples. The corrected results were included in the  survey
database.

5.4.3   Blank Contamination

       EPA examined the results for each analyte in every sample and compared the results to
the concentrations of analytes found in each of the method blanks associated with each batch of
samples. For all of the analyte classes, the concentrations of the analytes found in the method
blanks were generally well below the concentrations in the field samples.

       In some cases, there were low levels of analytes in some method blanks.  For all of the
analytes in the TNSSS, EPA used a common approach to evaluating blank contamination known
simply as "the 5x and lOx rules." Under routine circumstances, EPA qualified a field sample
result if the concentration of an analyte was not at least 10 times the amount found in the blank.
The rationale for the lOx rule is that under the worst of circumstances, in which the material
found in the blank is coming from a source within the laboratory, the amount in the blank would
only represent 10% of the amount in the field sample, and that small contribution is likely within
the overall measurement error.  If the amount in the sample is between 5x and  lOx the amount in
the blank, EPA normally will qualify the sample result as a maximum value because the potential
contribution from the laboratory could be as high as 20%.  Below 5x the amount in the blank,
EPA considered the field sample result to be a non-detect at the nominal reporting limit.

       The method blanks for the steroids and hormones presented greater challenges.  Three of
the steroid and hormone analytes, Cholesterol, Stigmasterol, and p-Sitosterol, were frequently
found in the method blanks, and a fourth analyte, Progesterone, was occasionally found in those
blanks.  During the course of the survey, the laboratory determined that the steroids were
adhering to the glassware used to prepare samples, perhaps originating from an earlier sewage
sludge sample prepared in the same glassware. Although the laboratory took steps to improve its
glassware cleaning procedures to prevent such "carryover," they were not able to completely
eliminate the occurrence of some target analytes in the method blanks.
TNSSS Sampling and Analysis Technical Report             32                                   January 2009

-------
       Therefore, EPA flagged the steroids and hormones results, where applicable. Normally,
EPA evaluates results for samples that have been diluted for analysis by multiplying the blank
result (which is usually not diluted) by the dilution factor for the sample. For example, if a
sample extract is diluted by a factor of 50 to bring one or more target analytes within the
calibration range of the instrument, EPA will multiply the concentration of an analyte found in
the blank by that dilution factor of 50 before comparing the blank and the field sample results.
That approach represents a worst-case assumption that the contaminant may be present in the
solvent used to extract the samples, thus using more of that solvent to dilute the extract would
add more of the contaminant.

       In the case of the steroids and hormones, the laboratory checked its reagent water and
extraction solvents and found no  such contamination.  As noted above, the laboratory surmised
that the material in the blanks was being transferred from the surface of the glassware used to
prepare the blank, but it originated from an earlier sample processed with the same glassware.
Thus, the amount, or concentration, found in the blank may represent the material that the  sample
extract might pick up from similar glassware, but there would be no likely contribution from the
solvents themselves, and thus no  need to consider the dilution factors of the samples.  The
dilution factors for samples in thus survey often were quite high.

      Even after instituting this minor change to the data review procedures, there were still 24
instances in which the levels in the blank led EPA to consider the results for one or more of the
steroids to be a non-detect. These 24 instances represented 14 field samples, with 1 to 3 analytes
affected in each of those samples. In many cases, the original result for the field sample was not
only less than 5x the blank results it was actually 2x to 3x lower than the amount found in  the
blank. Of the 14 affected samples, 5 were in the first batch of 12 samples analyzed and reported.
The 6 subsequent batches of samples had lower levels of these analytes in the blanks, with 2
batches having no samples set to  non-detects  and 2 more batches with only 1 difficult analyte.
These improved blank results are evidence of the effectiveness of the revised glassware cleaning
procedures instituted by the laboratory.

5.4.4 Extract Dilution

       Surrogate compounds were added to each sample for the semivolatile and PAH analysis
as a measure of sample extraction efficiency.  However, if the sample contained  large amounts of
the target analytes, the laboratory often had to dilute the sample extract to bring the results for
that analyte within the instrument's calibration range.  If the dilution factor was high enough, the
surrogates may have been present in the diluted extract at such low levels that they could not be
detected. Without measurable  results for the  surrogate, EPA could not evaluate the efficiency of
the extraction procedures for that sample.

       An important part of the data review process was to ensure that results for a given sample
included those where the  surrogates could be measured as well as results where all of the
analytes were within the calibration range.  This often requires that the laboratory provide  results
from multiple analyses of the sample at different dilution factors.  EPA worked with the
laboratory analyzing the semivolatiles and PAHs to obtain data from as many analyses as
practical and minimize situations in which only the most dilute analysis (without observable
surrogates) was provided.  However, the exceptionally high levels of bis (2-Ethylhexyl)
TNSSS Sampling and Analysis Technical Report              33                                   January 2009

-------
phthalate in some samples made the situation more difficult.  EPA flagged in the database any
samples in which the surrogates were diluted out.

5.4.5  Matrix Spikes

       Due to the often high and variable levels of many analytes in the samples, some of the
MS/MSD samples were spiked at levels too low to provide useful recovery data. This issue
occurred frequently for the anions, metals, semivolatile organics and PAHs. It also was an issue
for the PBDE analyses when EPA added matrix spike analyses to those analyses to make up for
the method modifications that eliminated the use of true isotope dilution quantitation (see
Section 4.4.4). Matrix spike samples were not required for the pharmaceutical, steroid, and
hormone analyses because those methods use isotope dilution quantitation and therefore provide
a sample-specific recovery correction of every analyte.

       The difficulty in generating  useful matrix spike results was exacerbated by the different
approaches to calculating MS/MSD recovery provided in different methods. EPA initially
qualified the results for samples associated with "under spiked" MS/MSD samples as estimates.
However, EPA recalculated the recoveries of the spiked analytes  using an alternative equation
designed to address this issue (see Section 6.7) and the bias and precision of the recalculated
results were further evaluated.

5.4.6  Interferences

       All of the methods used in the TNSSS included  some form of identification criteria for
the analytes of interest. These criteria included absorption or emission wavelengths,  retention
times, and mass spectrometric criteria. The GC/HRMS method for the steroids and hormones
involves monitoring two characteristic ions for each analyte.  As with other many EPA
GC/HRMS methods (e.g., Method 1613B, 1614, and 1668A), the responses for both  of those
ions  are used to quantify the concentration of the analyte in the sample, using an approach called
dual-ion quantitation.

       The method for the steroids  and hormones stipulates that the ratio of the abundances of
the two ions must be within a certain percentage of the theoretical ratio for the analyte. The
theoretical ratio is determined from natural  abundances of the exact masses of all of the
component elements that make up both of the ions monitored for the analyte. The acceptance
criterion in the method for the steroids and hormones is a ± 30% window around the theoretical
ion abundance ratio for the analyte.

       The purpose of this criterion is to ensure that the analytes of interest are adequately
separated from one another and from  any potential interfering compounds.  As described in the
method, if the ion abundance ratio is not met, the analyte cannot be positively identified. The
method instructs the analyst to note any instances where both ions from the analyte are present,
but the ion abundance ratio acceptance criterion is not met. This  approach to reporting results
allow the data users to decide how best to use the data.
TNSSS Sampling and Analysis Technical Report              34                                   January 2009

-------
       The most likely reason that the ion abundance ratio does not meet the acceptance
criterion is that there is a positive interference for one of the two ions monitored. That positive
interference increases the response for that ion, affecting the ion abundance ratio (IAR).  Because
the areas (abundances) for the peaks representing the two ions are added together to calculate the
concentration of the analyte, the positive interference translates into a higher reported
concentration for the analyte of interest.  Thus, the reported result is an estimated maximum
possible concentration (EMPC) for the analyte.

       There were 129 instances of EMPCs in the steroid and hormone results out of the total of
2100 results, or about 6% of all steroid and hormone results.  In all, 76  of the 84 survey samples
had one or more analytes qualified as an EMPC, and 17 of the 25 analytes were affected.  The
frequencies at which EMPCs were reported are shown in Table 9.  In some instances, the
observed ion abundance ratio was only marginally outside of the method acceptance limits, and
in other instances, the differences were much greater.

       Table 9.  Frequency of Estimated Maximum Possible Concentrations (EMPCs)
Analyte
Campesterol
Estrone
Testosterone
Stigmasterol
Ergosterol
Desmosterol
(3-Sitosterol
Norethindrone
Norgestrel
# of EMPCs Reported
50
21
11
10
6
5
4
4
4
Analyte
Androstenedione
(3-Stigmastanol
Equilin
17 a-Dihydroequilin
Androsterone
(3-Estradiol 3-benzoate
Equilenin
Progesterone

# of EMPCs Reported
3
3
3
1
1
1
1
1

       The affected analytes include both steroids and hormones, and therefore represent
separate analytical runs in the method.

       The degree to which the reported results for an EMPC exceed the actual concentration
cannot be determined exactly because a number of factors are involved, including the effect of
the interferences and the fact that the results are quantified by isotope dilution. In order to place
some bounds on the likely effects, EPA examined the survey results for Testosterone in detail.
The  11 EMPCs for Testosterone have reported concentrations that range from 34.8 to 2040
ug/kg.  The reported ion abundance ratios for these 11 Testosterone results ranged from 0.01 to
2.34, with the laboratory's QC acceptance limits for the ratio being 2.38 to 4.42.  The lowest ion
abundance ratio (0.01) was  associated with the highest reported result (2040 jig/kg), but there is
no apparent relationship between the other ratios and results.

       Although the actual  areas of the two peaks monitored for each analyte are presented in
the hard copy raw data provided by the laboratory, these peak areas are not readily accessible in
the electronic data. Rather than retrieving several hundred peak areas (2 peaks for each of 129
EMPCs) from the raw data, entering them into a spreadsheet, and checking for data entry errors,
EPA took an alternative approach.

       Although the actual  peak areas were not in an  accessible electronic format, EPA had the
ion abundance ratios themselves in the database.  In order to investigate the likely effects of the
interferences, EPA arbitrarily assigned the area of the first peak a value of 10,000 area counts.
TNSSS Sampling and Analysis Technical Report
35
January 2009

-------
Using the reported ion abundance ratios, we calculated the corresponding area of the second
peak. As described in the method, the concentration of an analyte is calculated from the sum of
the areas of both peaks, so we summed the two "dummy" areas for each sample result.  EPA also
calculated the area of the second peak using the theoretical ion abundance ratio for the analyte
and summed those two areas.  Since the sample concentration is proportional to the sum of the
areas, we compared the two calculated sums to determine a factor that could be used to convert
the reported result to the result that might have been calculated without the positive interferences
that affected the ion abundance ratio.  The results for the 11 EMPCs for Testosterone are shown
in Table  10.

 Table 10. Potential Effects of Ion Abundance Ratio on Reported Concentrations of
          Testosterone
Sample
1
2
3
4
5
6
7
8
9
10
11
Dummy
Abundance
Massl
10,000
10,000
10,000
10,000
10,000
10,000
10,000
10,000
10,000
10,000
10,000
Reported
IAR
0.01
0.27
0.54
0.55
0.68
1.2
1.22
1.4
1.99
2.04
2.34
Dummy
Abundance
Mass 2
1 ,000,000
37,037
18,519
18,182
14,706
8,333
8,197
7,143
5,025
4,902
4,274
Sum of
Dummy
Masses
1,010,000
47,037
28,519
28,182
24,706
18,333
18,197
17,143
15,025
14,902
14,274
Ratio of Sum from
Reported IAR to that
from Theoretical IAR
0.012813046
0.275127374
0.453781513
0.459203036
0.523809524
0.705882353
0.711181770
0.754901961
0.861302380
0.868421053
0.906657274
Reported
Cone.
2040
97.9
238
701
46.9
65.7
42.3
122
238
34.8
67.3
Adjusted
Cone.
26.1
26.9
108
322
24.6
46.4
30.1
92.1
205
30.2
61.0
       The results in Table 10 are sorted by the reported ion abundance ratios. For ease of
discussion, the samples addressed in this table were numbered 1 to 11, in order of reported ion
abundance ratio. The sample numbers have no other significance. As can be seen for Sample 1,
the potential implications of the ion abundance ratio could be dramatic.  However, for many of
the other samples, the effects are far less apparent.  The reported concentrations already include
the recovery correction inherent in isotope dilution, so the adjusted concentrations include the
same degree of correction.

       EPA performed similar calculations for the 50 reported EMPCs for Campesterol.
Because none of the observed ion abundance ratios were as far from the theoretical ratio for this
analyte, the overall effects on the reported results were smaller.  The applicable conversion
factors for Campesterol range from 0.173 to 0.907, meaning that the results for this analyte might
be reduced by factors from 1.1 to 6.  Given the time required, EPA did not perform these
calculations for all the other analytes in which EMPC results were reported by the laboratory.

Note:  The adjusted concentrations shown in Table 10 illustrate the potential change to the
       reported results.  However, the calculations shown assume that the ion abundance ratio
       problem is the result of a simple positive interference in one of the two ions for the
       analyte. The  actual cause may be more complicated.  Therefore, EPA included the
       original results reported by the laboratory in the survey database, not the adjusted
       concentrations described above. EPA flagged each result in the database that did not
       meet the method-specified ion abundance ratio as  an "EMPC."
TNSSS Sampling and Analysis Technical Report
36
January 2009

-------
5.4.7  Recovery Issues

       Three other analytical issues were noted that had significant potential effects on the
survey data for the pharmaceuticals, steroids, and hormones.  EPA identified each such instance
in the EPA database with the qualifier EXCLUDE to prevent those results from being used to
determine the national estimates of the concentrations of pollutants in biosolids.

       Two of those issues involved recoveries of the isotopically labeled analytes used to
perform isotope dilution quantitation. The third issue involved the "internal standards" added to
the sample extracts immediately prior to instrumental analyses and used to measure the
recoveries of the labeled analytes.

       There were 27 instances where EPA excluded pharmaceutical results from the database,
involving 15 analytes. For 13 of those analytes, EPA excluded the results for 3 samples because
the labeled compounds for those analytes were not recovered from the samples. These were not
simply instances of lower than expected recoveries of the labeled compounds, but rather, no
recovery at all (zero).  The inability to recover the labeled compound spiked into the sample
suggests a significant analytical problem beyond the routine analytical challenges presented by
sewage sludge samples.  All three samples were extracted on the same date and analyzed
together several days later.

       There were two instances where the results for the analyte Fluoxetine were excluded
because the labeled compound for this analyte was not recovered from those samples.

       The remaining 12 results were excluded for the analyte Minocycline. All 12 instances
came from the same extraction batch and were excluded because the  laboratory did not recover
the native (unlabeled) analyte in the laboratory control sample (LCS, also known as an ongoing
precision and recovery sample, or OPR) that was prepared at the same time as this batch of 12
samples. The laboratory did not report numerical results for Minocycline in these 12 samples
and flagged all of the Minocycline results with an "NQ" flag in the database, indicating that the
laboratory was not able to quantify the analyte.

       There were 65 instances of steroid and hormone results being excluded from the
database, involving 14 analytes. In 42 of those instances, involving 7 analytes, the data were
excluded because the laboratory found only trace levels (e.g., extremely low) of the internal
standard, Pyrene-dio, added to the sample extract immediately before analysis and used to
measure the recovery of the labeled compounds added to the sample before extraction. Trace
levels of the internal standard occurred in 6 samples analyzed in the same batch, suggesting a
possible problem with the addition of the internal standard (e.g., the automated injection of the
internal standard may have failed or been incomplete).

       The remaining 23 excluded results involved 7 other analytes where there was no recovery
of the labeled compound in the same 6 samples. In these cases, the issue was not simply that the
recovery could not be measured because of the very  low level of the internal standard present,
but rather that there was no measurable signal for the labeled compound itself.
TNSSS Sampling and Analysis Technical Report             37                                   January 2009

-------
5.5    Revised Results for BDE-209

       In March 2008, the laboratory that performed the PBDE analyses in 2007 prepared a new
set of calibration standards for the PBDEs, compared the responses of the new standards to data
from the standards used for the TNSSS analyses, and found that the results for one of the target
congeners (BDE-209) were markedly different. The laboratory also compared the results to a
standard from a second source and found similar differences, ultimately tracing the difference to
a vial of the original BDE-209 standard that was mislabeled by the manufacturer.

       After discussions with the laboratory, EPA decided that the BDE-209 results could be
recalculated based on the responses in the single-point calibration verification standard analyzed
with each batch of samples and that this recalculation would mathematically adjust the results to
a more accurate value. This verification standard was prepared from a separate source from the
initial calibration standards and was not affected by  labeling error. The laboratory recalculated
the BDE-209 results for all 84 field samples and the 2 equipment blanks and resubmitted the
corrected data in late May 2008. The resubmitted BDE-209 results were reviewed and uploaded
to the survey database on the EPA mainframe.
TNSSS Sampling and Analysis Technical Report             38                                   January 2009

-------
                                     Section 6
                                  Survey Results
6.1    Summary Results
       Table 11 provides a summary of the results for all 84 samples in the first phase (i.e., all
analytes excluding pharmaceuticals, steroids, and hormones) of the survey, listing the number of
samples in which each analyte was reported, along with the minimum and maximum
concentrations.  Tables 12 and 13 provide the results for the pharmaceuticals and steroids and
hormones, respectively. All sample results are reported on a dry-weight basis, based on the
percent solids in the original sample. The percent solids in the various sewage sludge samples
range from 0.14 to 94.9.  The units for pollutants vary with the class of analyte, as shown.  This
summary includes the results for the six field duplicate samples and the four POTWs that
generate more than one type of sewage sludge.  The minimum concentration is the lowest value
reported as present in any sample. EPA did not report a minimum or maximum value for those
analytes that were not detected. That situation only occurred for some of the pharmaceuticals,
steroids and hormones, and EPA used "NA" to indicate that the minimum and maximum values
were "not applicable."

 Table 11. Summary of Results for Metals, Anions, Organics, and PBDEs
Class
Solids
Anions
Metals
Analyte
Percent Solids
Fluoride
Nitrate/Nitrite
Water-extractable phosphorus
WEP ratio
Aluminum
Antimony
Arsenic*
Barium
Beryllium
Boron
Cadmium*
Calcium
Chromium*
Cobalt
Copper*
Iron
Lead*
Magnesium
Manganese
Mercury*
Molybdenum*
Nickel
Phosphorus
Selenium*
Silver
Sodium
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc*
Units
%
mg/kg
unitless
mg/kg
# Detects
84
84
84
84
84
84
72
84
84
83
80
84
84
84
84
84
84
84
84
84
84
84
84
84
84
84
84
80
78
83
84
84
84
Observed Dry-weic
Minimum
0.43
7.6
1.6
11.0
0.00065
1400
0.45
1.18
75.1
0.04
5.70
0.21
9,480
6.74
0.87
115
1,575
5.81
696
34.8
0.17
2.51
7.44
2,620
1.10
1.94
154
0.02
7.50
18.50
2.04
0.70
216
jht Concentration
Maximum
93.5
234
6,120
9,550
0.33920
57,300
26.6
49.2
3,460
2.3
204.0
11.8
311,000
1160
290
2,580
299,000
450
18,400
14,900
8.3
132
526
118,000
24.7
856
26,600
1.7
522
7,020
617
26.3
8,550
TNSSS Sampling and Analysis Technical Report
39
January 2009

-------
 Table 11.  Summary of Results for Metals, Anions, Organics, and PBDEs
Class
Organics
(PAHs and
Semi-
volatiles)
PBDEs
Analyte
4-Chloroaniline
2-Methylnaphthalene
Fluoranthene
Pyrene
bis (2-Ethylhexyl) phthalate
Benzo(a)pyrene
BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209
Units
i-ig/kg
ng/kg
# Detects
63
39
77
72
84
64
84
84
84
84
84
84
56
84
84
84
83
Observed Dry-weic
Minimum
51
10
45
44
657
63
2,200
73,000
1,800
3,200
64,000
13,000
1,900
9,100
7,700
2,100
150,000
jht Concentration
Maximum
5,900
4,600
12,000
14,000
310,000
4,500
160,000
5,000,000
110,000
150,000
4,000,000
1,100,000
40,000
410,000
440,000
120,000
17,000,000
 * Metals currently regulated at 40 CFR 503
 Table 12.  Summary of Results for Pharmaceuticals
Analyte
Percent Solids
Acetaminophen
Albuterol
Anhydrochlortetracycline
Anhydrotetracycline
Azithromycin
Caffeine
Carbadox
Carbamazepine
Cefotaxime
Chlortetracycline
Cimetidine
Ciprofloxacin
Clarithromycin
Clinafloxacin
Cloxacillin
Codeine
Cotinine
Dehydronifedipine
Demeclocycline
Digoxigenin
Digoxin
1,7-Dimethylxanthine
Diltiazem
Diphenhydramine
Doxycycline
Enrofloxacin
4-Epianhydrochlortetracycline
4-Epianhydrotetracycline
4-Epichlortetracycline
4-Epioxytetracycline
4-Epitetracycline
Erythromycin-total
Flumequine
Fluoxetine
Gemfibrozil
Units
%
pg/kg
Mg/kg
# Detects
84
2
1
1
52
80
39
0
80
0
1
74
84
45
0
0
20
39
19
3
0
0
4
69
84
76
14
0
31
1
8
80
77
0
79
76
Observed Dry-weight Concentration
Minimum
0.14
1,120
23.2
125
94.3
10.2
65.1
NA
8.74
NA
1,010
7.59
74.5
8.68
NA
NA
9.59
11.4
3.48
96
NA
NA
1,130
1.39
36.7
50.8
12.1
NA
126
974
35.7
47.2
3.1
NA
12.4
12.1
Maximum
94.9
1,300
23.2
125
1,960
6,530
1,110
NA
6,030
NA
1,010
9,780
47,500
617
NA
NA
328
690
24.6
200
NA
NA
9,580
225
5,730
5,090
66
NA
2,160
974
54.9
4,380
180
NA
3,130
2,650
TNSSS Sampling and Analysis Technical Report
40
January 2009

-------
 Table 12. Summary of Results for Pharmaceuticals
Analyte
Ibuprofen
Isochlortetracycline
Lincomycin
Lomefloxacin
Metformin
Miconazole
Minocycline
Naproxen
Norfloxacin
Norgestimate
Ofloxacin
Ormetoprim
Oxacillin
Oxolinic Acid
Oxytetracycline
Penicillin G
Penicillin V
Ranitidine
Roxithromycin
Sarafloxacin
Sulfachloropyridazine
Sulfadiazine
Sulfadimethoxine
Sulfamerazine
Sulfamethazine
Sulfamethizole
Sulfamethoxazole
Sulfanilamide
Sulfathiazole
Tetracycline
Thiabendazole
Triclocarban
Triclosan
Trimethoprim
Tylosin
Virginiamycin
Warfarin
Units
# Detects
54
1
3
2
6
80
32
44
29
0
83
1
0
1
29
0
0
46
3
2
2
3
5
1
2
0
30
8
1
81
58
84
79
24
0
15
0
Observed Dry-weight Concentration
Minimum
99.5
3,140
13.9
33.3
550
14.2
351
20.9
99.3
NA
73.9
5.91
NA
39.4
18.6
NA
NA
3.83
14.3
179
35.9
22.9
3.58
5.61
21.5
NA
3.91
191
21
38.3
8.42
187
430
12.4
NA
43.5
NA
Maximum
11,900
3,140
33.4
39.8
1,160
9,210
8,650
1,020
1,290
NA
58,100
5.91
NA
39.4
467
NA
NA
2,250
22.8
1,980
58.7
140
62.2
5.61
23.2
NA
651
15,600
21
5,270
239
441,000
133,000
204
NA
469
NA
 NA = Not applicable, because the analyte was not reported in any sample
 Table 13. Summary of Results for Steroids and Hormones
Analyte
Percent Solids
Androstenedione
Androsterone
Campesterol
Cholestanol
Cholesterol
Coprostanol
Desmosterol
17 a-Dihydroequilin
Epicoprostanol
Equilenin
Equilin
Ergosterol
17 a-Estradiol
Units
%
pg/kg
pg/kg
# Detects
84
32
50
84
84
81
84
58
1
83
1
15
53
5
Observed Dry-weight Concentration
Minimum
0.14
108
21.3
2,840
3,860
18,700
7,720
2,730
98.4
868
60.6
22.3
4,530
16.1
Maximum
94.9
1,520
1,030
524,000
4,590,000
5,390,000
43,700,000
94,400
98.4
6,030,000
60.6
107
91,900
48.8
TNSSS Sampling and Analysis Technical Report
41
January 2009

-------
 Table 13. Summary of Results for Steroids and Hormones
Analyte
17(3-Estradiol
(3-Estradiol 3-benzoate
17 a-Ethinyl-estradiol
Estriol
Estrone
Norethindrone
Norgestrel
Progesterone
(3-Sitosterol
(3-Stigmastanol
Stigmasterol
Testosterone
Units
# Detects
11
18
0
18
60
5
4
19
73
83
76
17
Observed Dry-weight Concentration
Minimum
22
30.2
NA
7.56
26.7
21
43.8
143
24,400
3,440
11,000
30.8
Maximum
355
1850
NA
232
965
1,360
1,300
1,290
1,640,000
1,330,000
806,000
2,040
 NA = Not applicable, because the analyte was not reported in any sample

6.2    Investigation of Results for Metals

       After compiling the results from the first phase of the survey, EPA investigated the
potential causes for the maximum results for calcium, iron, phosphorus, and silver. By
reviewing the sampler's field notes and ultimately contacting the POTWs by telephone, EPA
found that:

   •   The maximum result for calcium (311,000 mg/kg) was for a sample of Class A sewage
       sludge produced by a process known as advanced alkaline stabilization with subsequent
       drying. The alkaline stabilization process involves addition of large amounts of lime
       (calcium carbonate) to the material. The final sewage sludge is sold as a soil amendment.

   •   The maximum concentrations of iron (299,000 mg/kg) and elemental phosphorus
       (118,000 mg/kg) occurred in the same  sample. The facility from which this sample was
       collected adds ferric chloride during its wastewater treatment process to reduce the level
       of phosphorus in its effluent discharge. This treatment step results in high levels of iron
       and phosphorus in the sewage sludge from this plant.

   •   The maximum result for silver (856 mg/kg) occurred in a sewage sludge sample from a
       POTW that employs a "complete mix activated sewage sludge process" and disposes of
       its sewage sludge by incineration.  There are two major industrial dischargers to this
       POTW, but neither employs silver. The plant is not aware of any other instances of high
       silver results in their sewage sludge. There were no apparent calculation or transcription
       errors. The laboratory noted that this result was determined by ICP/MS, and when the
       laboratory re-examined its result for the ICP/AES analysis of the same sample, silver was
       present in that analysis at about 900 mg/kg, thus seemingly confirming the ICP/MS
       results.
TNSSS Sampling and Analysis Technical Report
42
January 2009

-------
6.3    Comparison of Metals Results to Current Standards

       As noted in Section 1 of this report, the sewage sludge regulations at Part 503 include
standards for land application of nine metals.  These standards are based on the dry-weight
concentrations. Table 14 illustrates the maximum results from this survey for the nine metals.
Table 14. Comparison of Survey Maximums to Existing Regulatory Limits
Pollutant
Arsenic
Cadmium
Copper
Lead
Mercury
Molybdenum
Nickel
Selenium
Zinc
Dry-Weight Concentration in mg/kg
Land application ceiling
75
85
4,300
840
57
75
420
100
7,500
Survey Maximum
49.2
11.8
2,580
450
8.3
132
526
24.7
8,550
Number of TNSSS Results
Over Ceiling
0
0
0
0
0
2
3
0
1
   Maximum results that exceed the land application ceiling are shown in bold.

Note:  It is critical to note that the selection of facilities to be sampled in this survey was not
       based on whether they managed their  sewage sludge by land application, nor was a goal
       of this survey to assess compliance. In fact, a number of the facilities disposed of their
       sewage sludge by incineration or placement in a landfill, and those facilities need not
       meet the ceiling concentrations shown in Table 14.

       As shown in bold in Table 14, three metals had observed concentrations in this survey
that exceeded the land application ceiling concentrations (molybdenum, nickel, and zinc). The
maximum observed concentrations for all other regulated metals were below the land application
ceiling concentrations.

       Five samples in this survey, collected from four facilities, contained metals that exceeded
the land application limits  (two of those samples were a pair of field duplicates collected from
one facility). One sample  exceeded the limits for both molybdenum and nickel. Of the four
facilities involved, one incinerates its sewage sludge on site and the other three send their sewage
sludge to landfills.  None of the facilities that actually dispose of their sewage sludge by land
application exceeded the limits.

6.4    Analytical Completeness

       "Completeness" is  a quality assurance measure of the number of samples collected
compared to the number of useable results produced.  Although the laboratories experienced a
number of difficulties with the samples from this survey and not all of the results met all of the
acceptance criteria in the applicable analytical methods, laboratories made acceptable efforts to
overcome these challenges and adequately document any QC issues encountered. In all cases
laboratories provided acceptable documentation for every sample in the survey.
TNSSS Sampling and Analysis Technical Report
43
January 2009

-------
6.5    Analytical Sensitivity

       As noted earlier in this report, the two previous national sewage sludge surveys
experienced analytical sensitivity challenges for some analytes.  This was largely due to the co-
extracted interferences present in the challenging matrix and the wide variation in the solids
content of treated sewage sludge disposed of nationwide. EPA designed the TNSSS for practical
analytical sensitivity, taking the steps outlined in Section 4 to ensure that the results were directly
comparable across facilities in the survey.

       Table 7 lists the target reporting limits for all classes of analytes in the survey. Those
targets were based on EPA's decision of what was practical in sewage sludge. The anions were
found in every sample and all but six metals were found in the 100 percent of the survey samples
(see Table 11).  Thus, sensitivity is not a concern. Therefore, the focus of analytical  sensitivity is
on the actual reporting limits for the other analyte classes and on the metals that were not found
in every sample (i.e., antimony, beryllium, boron, thallium, tin, and titanium). Table 15 provides
a comparison of those target limits with the actual reporting limits.

 Table 15. Analytical Sensitivity
Analyte Class
Metals
PAHs and
semivolatiles
PBDEs
Anions
Target Reporting Limit (dry weight)
3 to 4 mg/kg
100 to 300 |jg/kg
5 to 200 ng/kg
2 to 8 mg/kg
Actual Reporting Limit for Non Detects
Antimony
Beryllium
Boron
Thallium
Tin
Titanium
4-Chloroaniline
2-Methylnaphthalene
Fluoranthene
Pyrene
Benzo(a)pyrene
bis (2-Ethylhexyl) phthalate
BDE138
All anions detected in all samples
0.05 mg/kg
0.02 mg/kg
5 mg/kg
0.02 mg/kg
5 mg/kg
2 mg/kg
1 Dug/kg
(by SIM)
NA
5000 ng/kg
NA
 NA = Not applicable, analyte reported in all samples in the survey

       As can be seen in Table 15, the laboratories not only met, but far exceeded, the majority
of the sensitivity targets. The reporting limits for boron and tin were only slightly higher than
the target range. Four of the six samples in which tin was not reported include situations where
the results for tin in the samples and their associated method blank differed by less than a factor
of five.  As part of the data review process described in Section 5, the tin results were reset in
those four samples to non-detects at the nominal reporting level.

       Bis (2-Ethylhexyl) phthalate is a common laboratory contaminant and EPA reviewed the
results for all of the method blanks to ensure that the laboratory was not the source of the bis (2-
Ethylhexyl) phthalate in the samples. Although the laboratory reported bis (2-Ethylhexyl)
phthalate in many of the method blanks for organics, the levels in the samples are two to four
orders of magnitude higher than the levels in the blanks, indicating that sensitivity was not an
issue.

       The reporting limits for the other five organics were well below the target range using the
selected ion monitoring modifications described in Section 4.  The full-scan GC/MS results
TNSSS Sampling and Analysis Technical Report
44
January 2009

-------
included in the database are for samples in which the majority of the analytes were present above
the original target reporting limits in Table 11, thus sensitivity was not an issue for those
samples.

        All of the PBDE congeners except BDE-138 and BDE-209 were detected in all of the
survey samples, often at high levels relative to the calibration range of the method.  There was
only one non-detect for BDE-209.  BDE-138 was not detected in 30 of the survey samples with
the reporting limit of 5,000 ng/kg.  That reporting limit is significantly higher than the original
target, but reflects the need to adjust the sample size and extract dilution to accommodate the
very high levels of the other congeners that the laboratory reported in the  survey samples, which
were often 10 to 20 times higher than the BDE-138 concentrations. Therefore, sensitivity was
not a significant issue for the PBDE analyses.

       Because the methods for the pharmaceuticals, steroids, and hormones were under
development at the time the TNSSS began, EPA did not set any goals for  sensitivity for these
analytes. The reporting limits are based on the "Minimum Levels" in the  methods, which were
not optimized for the analysis of sewage sludge samples, but apply to all solid matrices. The
Minimum Level is the concentration in the sample that is equivalent to the concentration of the
lowest calibration standard analyzed by the laboratory.  When used as a reporting limit, that
value is adjusted for the nominal sample size and the moisture content of the sample (i.e., it is a
dry-weight concentration). The advantage of the approach used for the TNSSS to ensure
consistent sensitivity is that the reporting limit for each analyte was the essentially the same for
all the samples, regardless of the moisture content of the original sample.

       The Minimum Levels for the 72 pharmaceuticals vary by analyte,  and range from 2 |ig/kg
for analytes such as Albuterol, Digoxin, and Erythromycin, to 1,000 |ig/kg for 1,7-Dimethyl-
xanthine.  The Minimum Levels for the steroids and hormones ranged from 21  |ig/kg for
Testosterone, to 2,500 |ig/kg for Cholesterol.  Table 16 presents the Minimum Levels for the
Pharmaceuticals and Table 17 presents the Minimum Levels for the steroids and hormones.
TNSSS Sampling and Analysis Technical Report             45                                  January 2009

-------
Table 16.  Minimum Levels for the Pharmaceuticals
Minimum Level |jg/kg (dry-weight)
2
Albuterol
Digoxin
Diltiazem
Erythromycin-total
Roxithromycin
Sulfadimethoxine










4
Cimetidine
Dehydronifedipine
Diphenhydramine
Ormetoprim
Oxolinic Acid
Ranitidine
Sulfamerazine
Sulfamethazine
Sulfamethizole
Sulfamethoxazole






10
Azithromycin
Carbadox
Carbamazepine
Clarithromycin
Cotinine
Flumequine
Fluoxetine
Gemfibrozil
Miconazole
Ofloxacin
Sulfachloropyridazine
Sulfadiazine
Sulfathiazole
Thiabendazole
Trimethoprim
Warfarin
20
Cloxacillin
Codeine
Enrofloxacin
Lincomycin
Lomefloxacin
Naproxen
Norgestimate
Oxacillin
Penicillin G
Triclocarban






Minimum Level ug/kg (dry-weight)
40
Ciprofloxacin*
4-Epioxytetracycline
4-Epitetracycline
Cefotaxime
Chlortetracycline
Clinafloxacin
Doxycycline
Isochlortetracycline
Oxytetracycline
Penicillin V
Tetracycline
Tylosin
100
Sarafloxacinf
4-Epianhydrotetracycline
4-Epichlortetracycline
Anhydrochlortetracycline
Anhydrotetracycline
Caffeine
Demeclocycline
Digoxigenin
Ibuprofen
Norfloxacin
Sulfanilamide

400
MetforminA
4-Epianhydrochlortetracycline
Acetaminophen
Minocycline
Triclosan







1,000
1,7-Dimethylxanthine











There were 3 analytes with unique MLs. They were grouped with other analytes in this table for simplicity. The actual MLs are
shown below.
* Actual ML = 35 ug/kg
t Actual ML = 91 ug/kg
A Actual ML = 200 ug/kg
Table 17.  Minimum Levels for the Steroids and Hormones
Minimum Level ug/kg (dry-weight)
21
17 a-Dihydroequilin
17 a-Estradiol
17 a-Ethinyl-estradiol
17(3-Estradiol
Androsterone
Equilenin
Equilin
Estriol
Estrone
Norethindrone
Testosterone
p-Estradiol 3-benzoate
42
Norgestrel











104
Androstenedione
Progesterone










Minimum Level ug/kg (dry-weight)
500
Campesterol
Cholestanol
Coprostanol
Epicoprostanol
Stigmasterol
1,500
(3-Sitosterol
p-Stigmastanol



2,500
Cholesterol
Desmosterol
Ergosterol


TNSSS Sampling and Analysis Technical Report
46
January 2009

-------
       There were 15 pharmaceuticals and 1 hormone that were not detected in any sample.
EPA examined the supporting QC data, such as the OPR samples and labeled compound
recoveries, looking for indications of methodological issues.

       There were several pharmaceutical analytes that exhibited occasional low recoveries in
an OPR aliquot, no recoveries in an OPR aliquot, or low labeled compound recoveries in a
sample. For example, 1,7-Dimethylxanthine was only found in 4 samples, and there were 12
samples (prepared in one batch) that were associated with an OPR aliquot with low recovery of
this analyte.  The other 72 samples were analyzed in 6 other batches associated with acceptable
OPR recoveries for 1,7-Dimethylxanthine.

       There were 13 results for Warfarin where the sample exhibited low labeled compound
recovery and EPA qualified the non-detect results for the analyte.  There were no issues with
labeled compound recoveries or OPR recoveries for Warfarin in the other 71 samples in the
survey. Therefore, EPA does not believe that the fact that Warfarin was not detected in any
sample is an indication of a methodological challenge (i.e., that the method did not work).
Rather, it may be related to method sensitivity or breakdown of this analyte during sewage
treatment.

       For the steroids and hormones,  17 a-Ethinyl-Estradiol was the one analyte that was not
reported in any sample. EPA noted four instances where the labeled compound associated with
this analyte was not recovered at all, leading EPA to exclude those four non-detect results from
the survey database.  In three other instances, EPA noted that the recovery of the labeled
compound was low, but there was some recovery. However, EPA did not note OPR issues for
this analyte, which suggests that in those specific samples there may be issues associated with
the specific biosolids samples being analyzed rather than pervasive analytical problems.

6.6    Equipment Blank Evaluation

       Equipment blanks were prepared as described in Section 3.11 in each of the laboratories
that analyzed sewage sludge samples in the first phase of the TNSSS.  Equipment blanks were
prepared for anions, metals, semivolatile organics and PAHs, and PBDEs.  Equipment blanks
were not prepared for the pharmaceuticals, steroids,  and hormones in the second phase of the
TNSSS.

       EPA evaluated the results for the equipment blanks by comparing them to their
associated method blanks and to the sample results.  This was done to determine if any analytes
of concern were present at levels that might affect EPA's use of the data.

6.6.1   Semivolatile Organics and PAHs

       For semivolatile organics and PAHs, the laboratory reported only one target  analyte in
either equipment blank, BEHP, at a concentration of 10 |J,g/kg, based on a nominal 10-g sample
weight. However, BEHP is a common laboratory contaminant and the method blank associated
with both equipment blanks was reported to contain 11 ng/kg. In addition, all of the field sample
results for BEHP were  orders of magnitude higher than either the method blank or equipment
blank results, ranging from 657 to 310,000 ng/kg. Therefore, there is no evidence that the
compositing equipment contributed BEHP to any of the field samples.

TNSSS Sampling and Analysis Technical Report             47                                  January 2009

-------
6.6.2   Metals
       The laboratory reported low concentrations of eight elements in the two equipment
blanks. Table 18 presents the results for the two equipment blanks. The reported concentrations
are based on the nominal 1-g dry-weight aliquot used for the field samples. The laboratory
reported all other metals as non-detects in the two equipment blanks.

       The concentrations of some of the metals in the equipment blanks exceeded the
concentrations in the associated method blank,  although in some instances only marginally.  For
example, the laboratory reported nickel in the method blank at 0.05 mg/kg, and at 0.1 mg/kg in
one equipment blank, but nickel was not detected in the other equipment blank. The laboratory
reported lead in one equipment blank at 0.03 mg/kg, which is only marginally above the
laboratory's reported detection limit of 0.02 mg/kg for lead, and lead was not detected in the
other equipment blank. Therefore, the equipment blank results are not unexpected, or of
significant concern.

                 Table 18.  Equipment Blank Metal Results
Analyte
Barium
Calcium
Copper
Lead
Magnesium
Manganese
Nickel
Zinc
Concentration (mg/kg)
Equipment Blank 1
27.2
26.9
1.33
0.03
2.8
0.07
0.1
3.3
Equipment Blank 2
27.9
22.3
1.56
ND
2.1
ND
ND
6.7
                 ND = not detected

       The concentrations in the field samples generally were several orders of magnitude
higher than those in the equipment blanks (see Table 11). However, EPA took a conservative
approach to evaluating the potential impact of the equipment blanks. EPA compared the results
for each analyte in each of the solid samples against the results for both equipment blanks and
flagged in the database any solid sample result that was not at least five times higher than the
result in either of the two equipment blanks.

       Barium was the only metal analyte affected, with the results in four solid sewage sludge
samples less than five times the results in the equipment blanks (27.2 and 27.9 mg/kg,
respectively).  Those results are shown in Table  19. The value of 5 is a multiplier that is often
used for the evaluation of blanks.  It indicates that the amount of contaminant in the equipment
blank might account for 20% or more of a given field sample result.

       Table 19. Comparison of Sample Results and Equipment Blank Results for Barium
Sample
68320
68343
68344
68393
Result (mg/kg
128
75.1
78.6
134
Ratio of Sample Result to Equipment Blank Result
Sample/Equipment Blank 1
4.7
2.8
2.9
4.9
Sample/Equipment Blank 2
4.6
2.7
2.8
4.8
TNSSS Sampling and Analysis Technical Report
48
January 2009

-------
       The fact that the results for these four samples are less than five times the equipment
blank results is not clear evidence that any portion of the barium in the samples was actually
derived from the sampling equipment. As noted in Section 3, equipment blanks for solids are
conceptually different than those for aqueous samples and sampling equipment. In the 2001
NSSS, equipment blanks were prepared by placing wet sand in the compositing bowls used for
collecting samples. Aliquots of the  sand had been analyzed for dioxins, furans, and PCBs and
the sand was found to be free of these analytes, so it could be mixed with reagent water to
simulate wet sewage sludge.  Equipment blanks prepared in that manner in the 2001 survey
demonstrated that the stainless steel equipment used to collect samples did not contribute any
dioxins, furans, or PCBs.

       In this survey, sand could not be used for equipment blanks because it contains metals.
No readily available solid reference material is free of metals and resembles sewage sludge, so it
was not practical to prepare equipment blanks for solids that mimic the way that the sample
comes in contact with the sampling  equipment.

6.6.3   PBDEs

       The laboratory did not detect any of the PBDE congeners in the two equipment blanks
above their nominal reporting limits. The reporting limits for these blanks ranged from about
100 ng/kg to 20,000 ng/kg, for the various congeners, based on a nominal sample size of 10 g.
However, because the sample size for the field samples was reduced to 0.2 g, EPA examined the
equipment blank in greater detail.  In addition to using reporting limits that were based on the
low-point of the instrument calibration, the laboratory estimated the signal-to-noise based
detection limits for each analyte in each sample. Those estimated  detection limits were markedly
lower than the nominal reporting limits (e.g., 1 ng/kg versus 100 ng/kg).  The low levels of
PBDEs detected in the equipment blanks were similar to those reported in the method blanks,
indicating that the sampling equipment did not contribute additional PBDEs to the samples.

6.6.4   Anions

       The equipment blanks for the anions analysis contained low levels of fluoride and
nitrate/nitrite. The results for both analytes in both equipment blanks were below the
laboratory's reporting limits,  but above their detection limits, and the  laboratory reported the
results as "estimated." Neither analyte was detected  in the method blanks associated with the
equipment blanks or the field samples. Table  20 presents the results for the anions in the two
equipment blanks.

                 Table 20.  Equipment  Blank Anion Results
Analyte
Nitrate/Nitrite
Fluoride
Concentration (mg/kg)
Equipment Blank 1
5.2
11
Equipment Blank 2
1.9
18
TNSSS Sampling and Analysis Technical Report             49                                   January 2009

-------
       The equipment blank results for water-extractable nitrate/nitrite are two to five times
lower than the laboratory's nominal reporting limit. The concentration of water-extractable
nitrate/nitrite in the field samples ranges from 1.6 mg/kg to more than 6,100 mg/kg. Of 41
sewage sludge samples that contained less than 10 mg/kg of nitrate/nitrite, 29 of those are solid
samples that came in contact with the plastic sampling equipment.  However, the other 12
samples with less than 10 mg/kg nitrate/nitrite are liquid sewage sludge that was never in contact
with the equipment. Because of this, EPA concluded that nitrate/nitrite levels observed in blanks
were acceptable for the purposes of this report.

       Fluoride is added to drinking water by many municipal water systems for its dental decay
prevention benefits. It also is present in toothpastes and mouthwashes that are rinsed down the
drain after use.  Fluoride is soluble in water, which is, in fact, the basis of the analytical results
for this survey, as the samples were leached with reagent water as described in Appendix A, and
the leachate was analyzed for all of the anions. Figure 2 is a plot of the fluoride concentrations
versus the percent solids in each survey sample.

       Water is removed from sewage sludge by a variety of means, including several types of
presses, centrifugation, and air drying. Given the solubility of fluoride, one would expect that
water removed by mechanical means such as presses and centrifugation would take some portion
of fluoride with it, leaving the solids with lower concentrations then at the start of the water
removal process.
     I  5°
       40
                r:»     %  *  *  • *.
                  t  f        * * **
                *» *     t  •;  *•*
                        »      « *
                                       Fluoridt (mflftfl, dry weight)
                Figure 2. Plot of Fluoride Concentration versus Percent Solids
TNSSS Sampling and Analysis Technical Report
50
January 2009

-------
As can be seen in Figure 2, there is an inverse relationship between solids and fluoride in these
samples, with most of the highest fluoride concentrations in the samples with the lowest solids
contents.

Note:  All of the fluoride results are reported on a dry -weight basis in this survey, so the results
       are directly comparable across samples with different percent solids.

       The presence of fluoride in the two equipment blanks may indicate that the reagent water
rinse did not remove some of the fluoride in the tap water used to wash the equipment. As with
the nitrate/nitrite results, the amounts of fluoride reported in the two equipment blanks are 2 to 4
times lower than the laboratory's reporting limits. EPA treated the fluoride results in the field
samples in the same fashion described above for nitrate/nitrite, flagging any results in solid
samples less than five times the higher equipment blank result as an  estimate, but retained the
data in the database.  A total of 62 fluoride results for solid samples were flagged in the database.
However, there also were 12 liquid samples  with fluoride concentrations in that same range of up
to five times the higher equipment blank result.  These samples never contacted the sampling
equipment, so the fluoride in them cannot be attributed to the equipment.

       As noted in Section 3, the sampler mixed a large quantity of sewage sludge in each
plastic bowl, and only placed a portion of that material in jars for analysis. The moisture in the
original sewage sludge  samples was not free flowing for the solid samples, thus it would not
contact the surfaces of the equipment in the  same way that the reagent water used to prepare the
anions equipment blank did.  EPA does not believe that the qualifiers applied to these results
significantly compromise data usability.

6.7    Field Duplicate Results

       As part of the quality assurance  effort, EPA collected field duplicate samples to assess
the overall precision of the sampling and analysis approach for the survey. Of the 80 facilities
originally selected for sampling, 8 were chosen at random for collection of a field duplicate. As
described in Section 3, two of those facilities were not sampled and not replaced, and one other
field duplicate was collected at a different plant than originally planned.

       The tables in this section present the  relative percent differences (RPDs) between the two
results in each of the six pairs of field duplicate samples,  for each analytical class. The RPD is
used as the measure of precision because both results from the pair are measured concentrations
and there is no "true" concentration to be used in the comparison. The formula for RPD is
shown below where Result 1 and Result 2 represent the concentrations reported in the two
samples in each pair. The vertical bars in the numerator indicate it is the absolute value of the
difference, and the factor of 100 converts the value to a percentage.
                                    (Result 1 + Result 2)
                                            2

       In cases where both of the results in a field duplicate pair were non-detects, there was
little point in comparing the sample-specific reporting limits.  EPA indicated these instances with
ND, for "not detected."  When only one of the results in a field duplicate pair was a non-detect,
EPA did not calculate the RPD and indicated these instances with NC, for "not calculated."

TNSSS Sampling and Analysis Technical Report              51                                   January 2009

-------
       EPA did not establish a formal acceptance limit for the RPD of field duplicates in this
project. However, 50% is often used as a default limit that reflects the sum of the anticipated
analytical variability and the variability in the sample collection process.

6.7.1   Anions

       Table 21 summarizes the field duplicate results for the anions. The majority of the RPD
values for the anions are less than 20%, and are well within the expected variability of
laboratory duplicate analyses.  Using the default limit of 50% for field duplicates, the 16 of 24
observed RPD values in Table 21 that are below 20% indicate better than anticipated results for
these QC samples.

    Table 21. Comparison of Field Duplicate Results for Anions
Analyte
Fluoride
Nitrate/Nitrite
Water-Extractable Phosphorus
Total Solids
Relative Percent Difference (%)
Pair 1
11.9
19.0
84.3
8.1
Pair 2
18.3
17.1
26.0
0.2
Pair3
8.0
16.1
33.4
0.8
Pair 4
35.1
15.7
39.5
58.7
PairS
8.1
2.2
9.0
0.9
Pair 6
3.6
36.2
20.5
1.0
       The RPD values above 20% occur primarily for the water-extractable phosphorus (WEP)
and most of the analytes in Pair 4. The WEP results are more variable than those for fluoride or
nitrate/nitrite.  All three analytes are extracted from the sewage sludge at the same time using the
same leaching procedure. The WEP differences between some of the duplicate pairs may simply
be due to variability.

       Field duplicate Pair 4 exhibited large RPD values for the anions and all other analytes
(see Tables 21 - 26). EPA examined the sample collection information and found that this was a
liquid sewage sludge collected from a large storage tank.  Because of safety concerns at the
facility, the sampler observed as one of the facility staff opened a series of valves, flushed liquid
sewage sludge through the piping, and collected each aliquot of the first sample, then collected
the second sample.  The percent solids results for these two samples are 1.85% and 1.01% in the
aliquots use for the anions analyses.  However, the percent solids in the aliquots used for the
metals analyses are 4.27% and 1.01%, while the aliquots for the organics are 0.61% and 3.2%
respectively.  These data for total solids suggest that liquid sewage  sludge was not particularly
homogeneous and the sampling procedures used for this facility did not result in true duplicate
samples.

6.7.2  Metals

       Table 22 presents the field duplicate results for the metals. The majority of the RPD
values for the metals (125 of 174) are less than 20%, and 135 RPD values are less than 30%.  As
with the anions, results for field duplicate Pair 4 are markedly higher than the other five pairs.
TNSSS Sampling and Analysis Technical Report
52
January 2009

-------
     Table 22. Comparison of Field Duplicate Results for Metals
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Phosphorus
Selenium
Silver
Sodium
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
Total Solids
Relative Percent Difference (%)
Pairl
0.0
ND
6.2
4.1
0.0
13.7
7.1
4.3
7.5
2.7
3.4
4.7
7.1
10.5
4.8
25.6
9.1
5.1
4.7
2.7
5.0
1.6
6.1
6.6
9.9
8.3
8.1
4.6
0.4
Pair 2
1.3
3.6
0.0
59.8
3.6
2.9
1.7
1.9
6.8
0.2
1.7
3.6
0.6
0.7
5.4
95.3
5.0
3.3
0.6
1.1
1.8
41.8
3.1
0.5
3.9
4.1
1.6
1.6
1.2
Pair3
4.2
3.0
4.5
4.6
0.0
1.6
3.7
2.5
27.9
4.6
5.4
2.5
3.9
4.9
3.1
2.7
0.5
4.3
4.3
12.5
10.2
4.2
0.0
ND
27.2
3.6
4.3
5.4
0.4
Pair 4
100.5
97.5
109.2
100.7
83.3
115.1
106.1
112.2
120.3
110.4
100.7
106.5
105.4
112.5
105.8
89.5
112.6
107.9
106.1
118.7
95.4
121.6
112.5
94.1
96.8
100.7
107.5
106.5
123.5
PairS
7.7
19.3
30.8
10.8
25.0
12.6
31.5
11.2
29.4
22.4
10.3
9.6
30.5
8.5
12.1
11.8
31.2
27.7
11.0
31.3
8.8
5.2
41.4
9.3
4.6
28.6
28.0
10.7
2.7
Pair 6
0.7
2.3
7.3
1.5
7.4
0.1
7.3
1.5
0.9
9.1
2.9
0.6
8.5
1.2
1.0
7.1
8.4
11.4
0.6
9.8
16.9
3.2
0.0
23.5
16.2
8.7
8.2
1.5
1.0
     ND = not detected in either sample in the duplicate pair

6.7.3  Semivolatile Or games and PAHs

       Table 23 presents the field duplicate results for the semivolatile organics and PAHs.  A
total of 29 of 42 RPDs are less than 20% and 31 of 42 RPDs are less than 50%. The issues with
Pair 4 are apparent in the organics data as well (see Section 6.7.7).

     Table 23. Comparison of Field Duplicate Results for Semivolatile Organics and PAHs
Analyte
2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl) phthalate
Fluoranthene
Pyrene
Total Solids
Relative Percent Difference (%)
Pair 1
12.7
NC
2.4
16.7
16.7
0.0
1.2
Pair 2
36.6
7.4
3.1
3.6
0.0
6.9
4.3
PairS
ND
8.0
ND
0.0
0.0
ND
1.2
Pair 4
ND
41.8
143.9
157.8
144.3
128.9
136.0
PairS
ND
12.7
8.7
15.9
2.1
5.7
10.6
Pair 6
17.1
5.0
17.0
10.8
14.2
8.9
1.0
     ND = not detected in either sample in the duplicate pair
     NC = not detected in one sample in the duplicate pair and therefore RPD was not calculated
TNSSS Sampling and Analysis Technical Report
53
January 2009

-------
6.7.4  PBDEs

       Table 24 presents the field duplicate results for the PBDEs. A total of 36 of 72 RPD
values are less than 20% and 58 of 72 are less than 40%.  Field duplicate Pair 4 exhibits higher
RPD values than then other five pairs, but the differences are not as marked as for the other
analyte classes.  It is possible that the difference may be a reflection of the smaller sample size
used for the PBDE analyses compared to the other classes.  Due to the concentrations of PBDEs
in biosolids, which can interfere with analysis, the laboratory extracted just 0.2 g of the sewage
sludge, compared to samples of up to 10 g for other samples and analytes.  Extracting a smaller
sample eliminated the need for repeated dilution of the sample, resulted in fewer burdens to (i.e.,
didn't overwhelm) the laboratory equipment, and made quantitation possible.


     Table 24. Comparison of Field Duplicate Results for PBDEs
Analyte
BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209
Total Solids
Relative Percent Difference (%)
Pairl
19.6
16.7
28.0
45.3
17.4
14.7
NC
18.2
21.1
16.9
33.0
4.6
Pair 2
16.7
22.2
10.0
25.6
15.1
22.2
34.3
15.4
15.0
18.2
32.7
7.1
Pair3
11.2
9.1
6.5
2.3
19.0
9.5
ND
11.1
13.3
10.3
4.9
2.5
Pair 4
42.4
35.3
25.6
40.0
48.6
48.6
44.6
45.0
45.1
46.2
44.7
38.9
PairS
21.9
26.3
2.7
30.8
27.8
34.2
ND
27.9
28.6
28.6
41.9
28.2
Pair 6
19.4
9.5
14.6
20.2
1.0
14.6
8.0
8.7
10.5
19.4
25.0
0.0
     ND = not detected in either sample in the duplicate pair
     NC = not detected in one sample in the duplicate pair and therefore RPD was not calculated

6.7.5  Pharmaceuticals

       Table 25 presents the field duplicate results for the pharmaceuticals. As noted earlier, in
cases where both of the results in a field duplicate pair were non-detects, EPA indicated these
instances with ND, for "not detected." There are 270 instances of NDs for the pharmaceuticals.

       When only one of the results in a field duplicate pair was a non-detect, EPA did not
calculate the RPD and indicated these instances with NC, for "not calculated."  There were 18
instances of NCs for the pharmaceuticals. In Table 25, bold is used to indicate all of the RPD
values that exceeded 50%.

       Table 25. Comparison of Field Duplicate Results for Pharmaceuticals
Analyte
Acetaminophen
Albuterol
Anhydrochlortetracycline
Anhydrotetracycline
Azithromycin
Caffeine
Carbadox
Carbamazepine
Cefotaxime
Relative Percent Difference (%)
Pairl
ND
ND
ND
13.9
14.8
ND
ND
18.4
ND
Pair 2
ND
ND
ND
NC
9.4
ND
ND
6.0
ND
Pair3
ND
ND
ND
11.2
9.38
3.71
ND
9.32
ND
Pair 4
ND
NC
ND
31.8
50.9
72.4
ND
38.2
ND
PairS
ND
ND
ND
22.6
25.6
21.4
ND
11.1
ND
Pair 6
ND
ND
ND
15.3
18.9
ND
ND
19.7
ND
TNSSS Sampling and Analysis Technical Report
54
January 2009

-------
        Table 25. Comparison of Field Duplicate Results for Pharmaceuticals
Analyte
Chlortetracycline
Cimetidine
Ciprofloxacin
Clarithromycin
Clinafloxacin
Cloxacillin
Codeine
Cotinine
Dehydronifedipine
Demeclocycline
Digoxigenin
Digoxin
Diltiazem
1 ,7-Dimethylxanthine
Diphenhydramine
Doxycycline
Enrofloxacin
4-Epianhydrochlortetracycline
4-Epianhydrotetracycline
4-Epichlortetracycline
4-Epioxytetracycline
4-Epitetracycline
Erythromycin-total
Flumequine
Fluoxetine
Gemfibrozil
Ibuprofen
Isochlortetracycline
Lincomycin
Lomefloxacin
Metformin
Miconazole
Minocycline
Naproxen
Norfloxacin
Norgestimate
Ofloxacin
Ormetoprim
Oxacillin
Oxolinic acid
Oxytetracycline
Penicillin G
Penicillin V
Ranitidine
Roxithromycin
Sarafloxacin
Sulfachloropyridazine
Sulfadiazine
Sulfadimethoxine
Sulfamerazine
Sulfamethazine
Sulfamethizole
Sulfamethoxazole
Sulfanilamide
Relative Percent Difference (%)
Pair 1
ND
3.6
38.4
ND
ND
ND
ND
ND
ND
ND
ND
ND
NC
ND
22.0
48.1
ND
ND
ND
ND
ND
37.6
51.7
ND
25.5
3.4
ND
ND
ND
ND
ND
5.7
ND
ND
ND
ND
15.9
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Pair 2
ND
34.4
5.5
ND
ND
ND
ND
10
ND
ND
ND
ND
34.1
ND
0.8
27.4
ND
ND
NC
ND
ND
27.8
44.1
ND
10.1
13.1
2.2
ND
ND
ND
ND
1.80
ND
12.0
ND
ND
16.5
ND
ND
ND
NC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
PairS
ND
73.1
0.54
20.6
ND
ND
ND
10.5
27.8
ND
ND
ND
50.5
ND
11.9
15.7
ND
ND
2.5
ND
ND
24.4
2.25
ND
40.3
7.8
14.2
ND
ND
ND
ND
ND
9.2
11.8
18.2
ND
3.9
ND
ND
ND
ND
ND
ND
NC
4.03
ND
ND
ND
ND
ND
ND
ND
ND
ND
Pair 4
ND
116
50.7
19.0
ND
ND
39.4
NC
27.3
ND
ND
ND
32.9
ND
15.8
17.4
18.9
ND
36.3
ND
ND
55.2
32.5
ND
47.9
71.1
119.1
ND
ND
ND
ND
2.3
ND
68.5
59.2
ND
29.0
ND
ND
ND
ND
ND
ND
121.5
ND
ND
NC
ND
NC
ND
ND
ND
56.3
ND
PairS
ND
24.5
2.0
ND
ND
ND
NC
27.3
ND
ND
ND
ND
46.4
ND
29.5
15.8
7.2
ND
35.0
ND
ND
3.24
138
ND
7.5
4.7
3.0
ND
NC
ND
ND
187
5.1
15.5
1.8
ND
11.2
ND
ND
ND
NC
ND
ND
2.1
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Pair 6
ND
17.0
5.5
6.4
ND
ND
ND
19
ND
ND
ND
ND
2.96
ND
6.23
4.55
ND
ND
5.30
ND
NC
18.2
22.2
ND
18.5
2.6
0
ND
ND
ND
ND
33.8
6.2
NC
ND
ND
12.3
ND
ND
ND
9.04
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
TNSSS Sampling and Analysis Technical Report
55
January 2009

-------
        Table 25. Comparison of Field Duplicate Results for Pharmaceuticals
Analyte
Sulfathiazole
Tetracycline
Thiabendazole
Triclocarban
Triclosan
Trimethoprim
Tylosin
Virginiamycin
Warfarin
Total Solids
Relative Percent Difference (%)
Pair 1
ND
15.2
ND
6.1
5.7
ND
ND
ND
NC
1.7
Pair 2
ND
1.6
ND
10.2
12.4
ND
ND
22.1
ND
4.0
PairS
ND
5.09
20
0
37.3
30.8
ND
NC
ND
2.84
Pair 4
ND
2.3
0.8
13.2
38.3
ND
ND
ND
ND
88.0
PairS
ND
0.52
0.30
7.31
26.9
ND
ND
NC
ND
6.21
Pair 6
ND
13.1
0.49
5.90
41.0
NC
ND
ND
ND
3.54
        ND =  Not detected in both samples in the pair
        NC =  Not calculated because one of the results was a non-detect
        RPD values greater than 50% are shown in bold.

       EPA calculated 143 RPD  values for the pharmaceuticals, not including those for the total
solids in each sample. Of the 143 RPDs, 16 exceeded 50%, although some only marginally (e.g.,
50.5, 50.7, and 50.9%). Of the 16 values greater than 50%, 11 occurred in Field Duplicate Pair
4. As discussed above, EPA believes that the RPD values in Pair 4 reflect differences in the two
samples of liquid sewage sludge collected at that facility, with the RPD for the total solids at
88%.

       Many of the 270 instances where an analyte was not detected in either sample in the field
duplicate pair are a function of the low frequency at which some of the pharmaceuticals were
detected.  For example, as shown in Table 12, Acetaminophen was only detected in 2 of the 84
survey samples. Therefore, its frequency of occurrence was only 2.38%. The chance that a field
duplicate sample would be collected at any of the 74 POTWs in the survey was 8.1% (6 out of 74
plants). The likelihood of detecting Acetaminophen in both samples in a field duplicate pair is
on the order of 0.2% (e.g., 2.38% x 8.1%).  Therefore, the fact that Acetaminophen is listed as
ND in Table 25 for all 6 field duplicate pairs is not surprising.

       In contrast, Ciprofloxacin was reported in all 84 samples from the survey, including the 6
field duplicate samples.  Therefore, EPA was able to calculate an RPD value for each field
duplicate pair in Table 25. Except for field duplicate Pair 4, the RPDs for Ciprofloxacin indicate
good precision (four of the RPD values are less than 6%).

6.7.6   Steroids and Hormones

       Table 26 presents the field duplicate results for the steroids and hormones.
TNSSS Sampling and Analysis Technical Report
56
January 2009

-------
        Table 26. Comparison of Field Duplicate Results for Steroids and Hormones
Analyte
Androstenedione
Androsterone
Campesterol
Cholestanol
Cholesterol
Coprostanol
Desmosterol
17 a-Dihydroequilin
Epicoprostanol
Equilenin
Equilin
Ergosterol
17 a-Estradiol
17 (3-Estradiol
(3-Estradiol 3-benzoate
17 a-Ethinyl-estradiol
Estriol
Estrone
Norethindrone
Norgestrel
Progesterone
(3-Sitosterol
(3-Stigmastanol
Stigmasterol
Testosterone
Total Solids
Relative Percent Difference (%)
Pair 1
ND
ND
0
11.1
0.0
13.4
20.0
ND
17.3
ND
ND
38.4
ND
ND
ND
ND
ND
ND
ND
ND
ND
38.7
39.0
25.5
ND
1.7
Pair 2
9.7
ND
10.4
9.3
3.4
14.5
3.2
ND
38.4
ND
ND
ND
ND
ND
ND
ND
ND
8.3
ND
ND
20.5
8.8
3.8
5.1
ND
4.0
PairS

3.4
5.1
6.2
2.5
3.8
14.8
ND
13.4
ND
13.4
33.6
ND
ND
ND
ND
90.0
3.1
ND
ND
ND
16.2
14.3
34.5
ND
2.8
Pair 4
ND
ND
88.9
97.0
99.4
110.4
96.7
NC
109.7
ND
ND
93.8
ND
119.5
88.3
ND
NC

ND
ND
ND
87.4
87.0
83.6
ND
88
PairS
ND
NC
96.6
32.8
48.0
39.8
9.0
ND
32.8
ND
ND
24.4
NC
27.5
ND
ND
21.0
0.6
ND
ND
ND
44.4
33.1
61.0
ND
6.2
Pair 6
83.3
28.8
17.1
7.2
3.7
13.6
23.1
ND
5.6
ND
5.7
43.0
ND
ND
ND
ND
NC
4.5
ND
ND
65.9
3.2
30.6
2.5
NC
3.5
        ND =  Not detected in both samples in the pair
        NC =  Not calculated because one of the results was a non-detect
        RPD values greater than 50% are shown in bold.

       There were 6 instances of NCs for the steroids and hormones. For the steroids and
hormones, EPA calculated 74 RPD values, not including those for the total solids in each sample.
Of those 74 RPDs, 17 exceeded 50%. There were 70 instances where EPA did not calculate an
RPD value because both results for the analyte were non-detects (listed as ND in Table 26). As
discussed for the pharmaceuticals, the prevalence of the ND entries in Table 26 is largely a
function of the frequency of occurrence of the analytes across all samples. For example,
Norgestrel was only reported in 4 of 84 samples from the survey (4.76%), yielding a very low
likelihood it would be found in both samples from a field duplicate pair.  The occurrence of
Cholesterol in 81 of 84 samples is not surprising, given that it is excreted by humans.  That high
frequency of occurrence enabled us to calculate an RPD value for all 6 field duplicate pairs.

       Of the 17 RPD values  for steroids and hormones greater than 50%, 12 occurred in Field
Duplicate Pair 4. EPA believes that the RPD values reflect differences in the two samples of
liquid sewage sludge collected at that facility.  The differences are evident in the total solids
contents of each sample, where the RPD is 88%.
TNSSS Sampling and Analysis Technical Report
57
January 2009

-------
6.7.7  Results in Liquid Samples
       The results for all classes of analytes support EPA's conclusion that the sampling
procedures were appropriate and effective in collecting sewage sludge for the TNSSS.  The
exception was for the facility from which Field Duplicate Pair 4 was collected.  EPA believes
that the variability of the field duplicate results for Pair 4 is not typical of the variability observed
for the other liquid samples in the survey, but may reflect site-specific conditions.

       Given the variability of the results for Field Duplicate Pair 4 shown in the tables above,
EPA examined the results for those two samples in greater detail, comparing the two samples
from that one site to the liquid sewage sludge samples from other sites. EPA examined the
percent solids data from the anions, metals, and semivolatile organics analyses for all 19 liquid
sewage sludge samples collected during the survey.  Because the results for the PBDEs,
Pharmaceuticals, steroids, and hormones were delivered later in the TNSSS effort, EPA did not
include them in this more detailed analysis of liquid sample results.

       EPA examined the percent solids results generated during analyses of the anions, metals,
and semivolatile organics in all 19 liquid sewage sludge samples (19 samples x 3 classes = 57
measurements in all).  Based on the observed distribution of results, the data were transformed
by taking the natural log of the results and subjecting them to an F-test. The null hypothesis was
that the three aliquots from each of the two samples in Pair 4 did not have significantly different
variances from the variances of the three aliquots in each of the other liquid samples, indicating
that they came from the same population. Table 27 presents the log-transformed percent solids
data for all 19 liquid sewage sludge samples.

        Table 27. Log-transformed Percent Solids Data for Liquid Sewage Sludge
                 Samples
Liquid Sample
1
2
3
4
5
6
7-FD
8-FD
9
10
11
12
13
14
15
16
17
18
19
Natural Log of % Solids
Anions
1.284
3.795
0.560
4.126
1.543
0.482
0.615
0.010
2.434
2.425
1.677
-0.844
0.293
0.815
0.182
-0.673
0.270
2.072
0.788
Metals
1.747
3.773
0.560
4.126
0.270
0.482
1.452
0.010
2.407
2.434
1.128
-0.693
0.647
0.658
0.182
-0.844
0.207
1.991
-0.041
Organics
0.425
3.795
1.621
4.126
1.085
1.456
-0.494
1.163
2.477
2.370
0.920
-0.942
0.604
0.888
0.191
-0.892
0.182
1.652
0.775
Variance of Logs
0.450
0.000
0.376
0.000
0.416
0.316
0.953
0.443
0.001
0.001
0.153
0.016
0.037
0.014
0.000
0.013
0.002
0.049
0.226
         FD = Field duplicate pair
TNSSS Sampling and Analysis Technical Report
58
January 2009

-------
       Table 28 presents the results of the F-test. The pooled within-sample variance (the
standard deviation squared) is markedly greater for the two field duplicate samples, compared to
the remaining liquid sewage sludge samples in the survey.  The F-ratio in Table 28 is greater
than the critical value of F, leading to the conclusion that the variances in the percent solids
results for the field duplicate samples are greater than would be expected by chance from
samples in a single population.
Table 28. F-test Results for Field Duplicates and Other Liquid Samples
Pooled Variance (field duplicates)
Pooled Variance (non-field duplicates)
F-ratio
F-critical
p-value
0.698
0.122
5.731
2.650
0.001229
       Therefore, the variability exhibited by the field duplicate results for Pair 4 is not typical
of the variability that is apparent for the other liquid samples in the survey. The results for Pair 4
are not an indication that the sampling procedures used for the survey are inappropriate for liquid
samples. Rather, the results for Pair 4 may reflect site-specific conditions.  However, both sets
of results for all of the field duplicate pairs, including Pair 4, are included in the survey database.

6.8     Matrix Spike and Duplicate Results

       Matrix spike (MS) samples and matrix spike duplicate (MSD) samples, or matrix spike
samples and unspiked duplicate (DUP) samples, were prepared with batches of field samples
analyzed for the anions, metals, semivolatile organics and PAHs, and the PBDEs analyzed as
described in Section 4 without isotope dilution quantitation.  These QC samples served to
demonstrate the applicability of the methods to the matrices in question, e.g., liquid and solid
sewage sludge. The use of an MSD versus a DUP is generally called out in the method, with
MS/MSD pairs being the norm in methods for organics, and MS and DUP samples being the
norm in methods for metals and other inorganic such as anions. The results for those analyses
are discussed below by analyte class.  Appendix C presents the QC acceptance limits used by the
laboratories.

       Because the  methods used for the pharmaceuticals, steroids, and hormones use isotope
dilution quantitation, those methods do not require that separate MS/MSD samples be analyzed.
Rather, the recoveries of the isotopically labeled analytes spiked into every sample are monitored
and used to correct the results for the  target analytes.

       The MS/MSD results for the TNSSS are discussed in the subsections that follow,  by
analyte class.  The labeled compound recoveries for the isotope dilution methods are discussed in
Section 6.8.8.
TNSSS Sampling and Analysis Technical Report
59
January 2009

-------
6.8.1   Matrix Spike Results for Anions

       The laboratory prepared eight sets of matrix spike samples for the anion analyses.  Table
29 presents the results in terms of the percent recovery of each analyte spiked into the sample.

 Table 29. Matrix Spike Recoveries for Anions in Sewage Sludge

Analyte
Fluoride
Nitrate/Nitrite
Water-Extractable Phosphorus
Matrix Spike Recovery (%)
MS1
93
98
-24
MS2
74
101
88
MS3
66
103
75
MS4
82
95
102
MS5
93
100
90
MS6
60
90
95
MS7
68
101
80
MS8
77
96
95
       The laboratory's acceptance limits for matrix spike recoveries were 75 - 125%, and 18 of
24 recoveries in Table 29 met those limits. The laboratory reported the recovery of WEP in MS3
as 74.8%, before rounding. That value is only marginally out of the specification, and rounds to
75%, the lower limit of the laboratory's acceptance range.

       The laboratory reported the recovery of WEP in one matrix spike (MSI) as -24%.
Negative recoveries are not physically possible and the reported recovery is a function of the
manner in which it is calculated. Although most recent published EPA methods explicitly
include example calculations for QC parameters such as recovery, some older analytical methods
do not. EPA Methods 340.2, 353.2, and 365.3, used for the anion analyses, are among those
methods without example calculations. In the absence of project-specific requirements, the
laboratory performing the anions analyses relied on formulae from the SW-846 methods manual
from EPA's Office of Solid Waste. Chapter One of the manual includes  definitions of a number
of commonly used terms and provides the equation for the calculation of recovery shown below:
                                  %R =
                              100(xs-xu)
                                   K
where:
   %R =
   xs   =
   K   =
percent recovery
measured value for spiked sample
measured value for unspiked sample
known value for the spike in the sample
       This same basic formula appears in many individual EPA methods. The remainder of
this report refers to this equation as the "traditional approach" to calculating recovery. As
written, K is often interpreted as the amount of material spiked into the sample. That
interpretation ignores any "background" concentration of the analyte in the unspiked sample. In
practice, the equation produces recoveries that appear reasonable and generally meet
expectations in those samples where the background concentration of the analyte is either very
low, or where the amount spiked into the sample is much greater than the background amount.
However, the laboratory calculates negative recoveries any time the result in the spiked sample is
less than that in the unspiked sample, even if that is a function of inhomogeneity in the original
sample. Given the Law of Conservation Mass, whereby matter cannot be created or destroyed,
negative recoveries are physically impossible.
TNSSS Sampling and Analysis Technical Report
                                 60
January 2009

-------
       Laboratories commonly prepare MS samples without extensive knowledge of the
background levels of any target analytes in specific samples. Therefore, a laboratory may
blindly spike an amount that is similar to that already in the sample and the equation above will
perform poorly.

       Given the frequency at which negative and unrealistic recoveries are reported using the
equation above, EPA developed an alternative calculation that considers the result found in the
matrix spike sample in comparison to what was found in the unspiked sample plus the amount
spiked.  The alternative equation is as follows:

                               %R =	^	xlOO

where:

   %R = percent recovery
   Cs   = measured value for spiked sample
   Cu   = measured value for unspiked sample
   Cn   = nominal spike added to the sample

       Eliminating the subtraction operation in the numerator of the equations prevents the
occurrence of any negative values. Moving the concentration of the unspiked sample to the
denominator more effectively addresses the issue of the "background" concentration.

       In the case of the first MS sample in Table 29, the water-extractable phosphorus result in
the MS sample (271 mg/kg) was less than the result in the unspiked sample (317 mg/kg), despite
adding 189 mg/kg of phosphorus, leading  to a negative recovery (-24.3%) using the traditional
approach, as follows:

                       (100 x (271 - 317))/189 = -4600/189 = -24.3%

Using the alternative equation above and the same laboratory results, the recovery in the matrix
spike is calculated as:
                       100 x (2717(317 +  189) = 27100/506 = 53.6%

While the recovery of 53.6% is still below the acceptance limits used by the laboratory, it is a
more rational expression of the results in this sample.

       The alternative equation increased the recovery of WEP dramatically in this example.
However, it did not cause the recoveries that already appear reasonable to  exceed the acceptance
limits when they did not otherwise do so.  For example, Table 29 lists the recovery of
nitrate/nitrite in MS3 as 103%. This value is rounded down from 103.3652% determined via the
traditional calculation. Using the alternative equation, the recovery is 103.3466%, a trivial
difference that is removed when the results are rounded to the nearest whole percentage. Given
this trivial difference, EPA did not recalculate every recovery reported by the laboratory, but
focused this discussion on the negative recoveries alone.
TNSSS Sampling and Analysis Technical Report             61                                   January 2009

-------
6.8.2   Duplicate Results for Anions

       The laboratory prepared eight sets of duplicate sample analyses for anions. Table 30
presents the results shown in terms of the relative percent difference (RPD) for each analyte in
the sample. Although the laboratory used acceptance limits expressed to the nearest whole
percentage, Table 30 presents the RPDs to one decimal place to illustrate the small differences
between many of the duplicate pair results.

 Table 30. Duplicate Precision for Anions in Sewage Sludge
Analyte
Fluoride
Nitrate/Nitrite
Water-extractable Phosphorus
Relative Percent Difference (%)
Dup 1
0.9
3.6
6.1
Dup 2
2.7
0.0
5.2
Dup 3
4.5
23.0
18.8
Dup 4
14.6
17.1
4.6
Dup 5
5.4
2.2
9.4
Dup 6
12.2
8.3
6.8
Dup 7
0.4
6.9
4.1
Dup 8
14.6
10.7
11.1
       The laboratory's acceptance limit for precision (RPD) is 20% and all but one result in
Table 30 met that limit. The highest RPD value reported by the laboratory was 23.0% for
nitrate/nitrite in Duplicate 3. This RPD was only slightly outside of the 20% limit.

       Except as discussed above, the recovery data in Table 29 and the precision data in Table
30 demonstrate that the methods were generally accurate and precise when applied to sewage
sludge samples.

6.8.3   Matrix Spike Results for Metals

       The laboratory prepared 15 sets of matrix spike samples for the metals analyses. Table
31 presents the matrix spike recoveries in three parts (Tables 31 A, 3 IB, and 31C), rounded to
one-tenth of a percent, to illustrate some of the smaller differences.

       Because calcium and magnesium  are common components of soils and occur at levels
that vary widely, the laboratory did not spike these two metals into solid samples, and they do
not appear in Table 31. However, calcium and magnesium also are not major metals of concern
in this survey.

       The laboratory analyzed mercury  separately from any other metals.  Therefore, the results
for mercury appear at the bottom of the table because the laboratory prepared only 9 matrix spike
samples for mercury and did not necessarily use the same field samples for the mercury matrix
spike analyses as for the other metals.

       During the survey, as the laboratory gained experience with the sewage sludge samples,
they adjusted the amounts of some metals spiked into each matrix spike sample in an effort to
account for the background concentrations. For example, for aluminum, the spiking
concentrations ranged from about 400 mg/kg to about 3200 mg/kg, while copper spiking levels
ranged from about 40 mg/kg to 4000 mg/kg. These adjustments to the spiking levels were not
always successful in addressing recovery issues.
TNSSS Sampling and Analysis Technical Report
62
January 2009

-------
        Aluminum and iron presented problems with reported recoveries. The laboratory also
 reported negative recoveries for phosphorus, again a function of the assumption that the
 background concentrations are much lower than the spiking levels.  Therefore, Table 31 includes
 the recoveries reported by the laboratory using the traditional calculation from the methods (in
 the MS# columns) along with the alternative calculation described above (in the ALT# columns).
 Using the alternative calculation, all of the negative recoveries reported by the laboratory
 become positive values.  In addition, many of the very high recoveries (e.g., over 250%) are
 greatly reduced in magnitude.

Table 31A.  Matrix Spike Recoveries for Metals in Sewage Sludge, Calculated in the Traditional
           Fashion and with an Alternative Equation
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Molybdenum
Nickel
Phosphorus
Selenium
Silver
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
Mercury
Recovery (%)
MS1
177.7
44.3
102.0
99.2
108.7
105.1
110.7
93.9
101.1
83.2
-203.0
111.3
78.2
103.0
101.7
-101.5
102.3
68.0
104.8
96.6
47.3
101.4
102.2
50.8
106.0
ALT1
102.4
45.7
101.8
99.7
108.3
103.5
105.9
98.8
101.0
98.5
98.0
104.4
97.9
102.6
101.3
96.9
102.0
90.7
104.7
96.8
57.7
101.0
101.8
97.1
101.3
MS2
334.5
41.1
102.0
98.8
99.7
106.5
117.4
113.7
102.9
59.7
1543.8
108.3
135.9
108.7
105.5
-123.5
101.4
93.8
106.2
95.4
29.4
105.9
102.4
41.2
76.0
ALT2
104.2
43.7
101.7
99.5
99.7
105.9
108.6
105.0
102.7
96.7
111.3
103.0
108.1
105.7
103.8
91.2
101.3
98.2
106.2
95.9
43.2
104.3
101.9
96.6
97.0
MS3
68.8
75.7
101.5
101.9
102.3
110.6
112.9
110.5
107.6
79.5
-71.4
111.1
91.7
111.5
105.6
-203.9
102.8
103.3
107.4
95.6
16.3
107.3
106.0
99.9
110.8
ALT3
98.5
75.7
101.4
101.0
102.3
108.4
111.5
106.9
107.3
97.9
96.5
108.6
99.7
110.0
105.1
90.1
102.6
101.4
107.4
95.9
21.9
106.4
105.6
100.0
104.1
MS4
400
41.5
111.1
96.3
112.4
110.6
121.8
118.8
116.3
40.0
265.0
118.9
81.0
110.3
118.3
-46.0
112.9
83.1
108.9
91.3
55.5
117.7
119.6
72.0
106.6
ALT4
114.9
42.0
110.6
98.2
112.0
108.9
119.6
110.5
115.7
95.2
103.4
113.6
95.1
110.0
114.7
91.7
112.2
92.3
108.8
91.6
61.1
115.1
115.5
94.1
103.2
MS5
443.9
46.4
101.6
117.5
101.2
103.7
110.7
118.1
115.7
173.3
156.2
109.5
101.3
107.5
102.5
-41.8
103.5
101.8
108.4
81.1
10.8
101.4
103.7
125.3
62.5
ALTS
107.9
47.4
101.5
105.7
101.2
103.0
107.9
102.4
104.0
104.7
100.8
105.7
100.4
106.9
102.2
96.0
103.2
101.1
108.4
82.6
14.8
101.3
102.7
101.9
94.4
Table 31B.  Matrix Spike Recoveries for Metals in Sewage Sludge, Calculated in the Traditional
           Fashion and with an Alternative Equation
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Molybdenum
Nickel
Recovery (%)
MS6
236.2
72.4
103.0
103.7
93.0
105.7
110.3
93.5
102.4
111.3
94.7
107.2
98.8
109.6
100.3
ALT6
104.3
72.9
102.8
101.8
93.1
104.8
108.8
96.7
102.3
101.1
99.8
105.7
99.5
108.4
100.3
MS7
194.9
74.8
102.8
107.8
95.6
102.6
107.2
100.0
102.3
182.6
1167.5
110.1
99.3
108.2
99.8
ALT7
104.3
75.8
102.7
104.2
95.8
101.9
106.3
100.0
102.2
103.6
103.9
108.8
99.7
107.3
99.8
MS8
653.0
69.1
140.3
119.5
98.7
119.5
119.9
116.9
118.8
216.4
2126
329.8
206.4
101.9
150.0
ALTS
353.7
69.4
137.8
118.8
98.7
119.1
119.3
116.3
118.2
191.8
465.7
263.9
186.1
101.8
145.0
MS9
202.0
47.9
100.3
92.6
117.5
103.4
111.8
88.6
99.6
38.4
-1312.7
131.2
10.1
100.6
103.2
ALT9
103.2
49.1
100.2
96.9
116.7
102.3
109.0
96.6
99.6
93.6
92.5
113.6
93.7
100.5
102.4
MS10
485.0
74.3
119.1
103.0
100.9
127.0
107.7
184.4
110.3
304.4
250.4
166.5
102.7
107.1
184.9
ALT10
325.1
74.5
118.5
102.9
100.9
126.3
107.6
171.2
110.2
247.3
214.3
157.6
102.6
106.7
171.3
 TNSSS Sampling and Analysis Technical Report
63
January 2009

-------
Table 31B.  Matrix Spike Recoveries for Metals in Sewage Sludge, Calculated in the Traditional
           Fashion and with an Alternative Equation
Analyte
Phosphorus
Selenium
Silver
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
Mercury
Recovery (%)
MS6
-63.0
103.4
56.8
105.9
92.5
13.7
98.9
101.9
95.7
236.7
ALT6
96.5
103.1
85.1
105.9
93.1
20.7
99.0
101.7
99.5
145.5
MS7
243.4
101.7
33.4
105.7
85.4
22.5
115.8
99.7
109.4
116.2
ALT7
102.0
101.6
84.0
105.7
86.1
29.5
106.7
99.8
101.1
115.6
MS8
360.6
111.7
95.8
108.9
94.0
91.0
169.7
140.0
230.8
124.9
ALTS
263.5
111.3
96.0
108.9
94.0
91.0
162.1
137.4
201.3
109.0
MS9
-282.7
101.8
67.7
104.7
99.7
85.0
109.6
107.0
92.0
91.4
ALT9
91.1
101.7
87.3
104.7
99.7
90.8
105.8
105.6
92.2
94.4
MS10
320.6
114.4
114.2
103.9
83.6
79.0
139.8
100.1
160.4
-
ALT10
238.4
113.6
112.1
103.9
83.6
79.0
136.8
100.1
153.4
-
Table 31C.  Matrix Spike Recoveries for Metals in Sewage Sludge, Calculated in the Traditional
           Fashion and with an Alternative Equation
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Molybdenum
Nickel
Phosphorus
Selenium
Silver
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
Recovery (%)
MS11
229.9
77.7
101.2
107.7
101.4
119.5
104.4
105.3
103.5
155.8
505.6
100.0
100.3
105.3
105.0
384.0
100.3
90.9
99.4
90.2
90.5
100.9
100.0
461.3
ALT11
184.4
111
101.2
107.6
101.4
118.4
104.4
105.3
103.5
153.7
224.6
100.0
100.3
105.3
105.0
169.0
100.3
91.8
99.4
90.3
90.8
100.9
100.0
237.9
MS12
500.0
55.9
95.9
95.2
100.1
100.1
108.7
102.8
100.9
62.1
250
108.9
92.1
108.8
102.8
-80.2
93.1
94.5
104.5
88.9
35.7
109.1
107.8
86.1
ALT12
108.6
56.7
96.1
98.0
100.1
100.1
107.4
101.6
100.9
97.8
102.6
106.0
99.1
107.2
102.3
96.2
93.5
97.2
104.5
90.0
51.4
108.1
106.9
98.5
MS13
-395.9
75.2
69.7
47.7
94.5
88.7
101.9
77.4
95.8
-204.1
-1573.6
62.2
0.0
99.3
91.0
-142.0
91.8
41.0
101.0
91.8
23.9
92.6
108.2
-106.1
ALT13
75.8
75.6
70.5
82.3
94.5
90.7
101.7
87.6
96.0
74.7
77.4
82.6
89.0
99.3
92.3
76.9
92.2
79.4
101.0
92.2
26.8
93.7
106.2
111
MS14
301.9
54.7
-8.2
115.1
91.9
99.8
55.5
118.9
102.5
181.8
1209.8
120.5
178.0
-40.9
104.7
105.9
-98.2
79.2
95.4
84.0
27.0
102.5
89.1
181.8
ALT14
107.3
56.1
56.4
103.5
92.6
99.8
76.9
108.7
102.3
104.0
109.4
107.6
104.1
50.9
103.8
100.6
38.9
93.5
95.6
84.3
28.3
101.9
91.2
104.9
MS15
147.5
85.1
105.8
120.7
109.4
111.6
114.1
130.0
108.2
178
599.9
114.9
124.0
112.9
116.2
260.0
107.8
91.0
111.6
99.4
40.8
106.9
104.8
138.0
ALT15
101.9
85.4
105.6
108.8
109.2
106.8
111.9
104.7
107.8
105.9
103.2
109.8
105.9
110.5
109.5
102.3
107.3
97.4
111.6
99.4
47.7
105.0
104.4
103.8
       Using either calculation, many of the metals exhibited a small positive bias. A total of 12
 the 25 spiked metals have mean traditional recoveries between 101% and 115%, while 14 of 25
 metals have mean alternative recoveries in the same range.  Five metals exhibited a slight
 negative bias using the traditional calculation, with mean recoveries ranging from 80% to 97%.

       Antimony, phosphorus, and titanium exhibited mean traditional recoveries below 70%.
 The alternative calculation dramatically altered the mean recovery of phosphorus, raising it from
 39% to 120%, by virtue of eliminating the large number of negative values in the traditional
 calculation.
 TNSSS Sampling and Analysis Technical Report
64
January 2009

-------
       The mean recovery of titanium was only 43% in these 15 matrix spike samples, with only
four values in the acceptable range.  Similarly, the mean recovery of antimony was 62% in these
15 matrix spike samples, with only five values in the acceptable range. Neither titanium nor
antimony is among the nine pollutants initially selected for this survey (see Section 1), nor one of
the metals that currently has a regulatory standard in sewage sludge.

6.8.4   Duplicate Results for Metals

       Over the course of the survey, the laboratory analyzed 15 samples for metals in duplicate,
to assess analytical precision. Table 32 presents the results for those duplicate analyses.  The
laboratory prepared and analyzed only seven duplicates for mercury, and as with the matrix spike
results, not necessarily using the same field samples as for the other metals.  The exception is for
Duplicate 8, where the results for mercury are from the same sample as for all the other metals.
Table 32 is divided into several parts.  The vast majority of the RPD values are less than the
acceptance limit of 30%. The exceptions are almost exclusively in Duplicates 8, 10, and 11.  All
three of these samples were liquid sewage sludge.

Table 32A.  Duplicate Precision for Metals in Sewage Sludge
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorus
Selenium
Silver
Sodium
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
Mercury
Relative Percent Difference (%)
Dup1
14.4
8.3
2.5
2.3
6.3
13.8
1.3
1.2
0.7
0.4
3.2
0.7
3.9
4.3
2.3
1.2
1.2
1.6
3.6
12.4
1.0
0.0
1.0
51.1
1.1
0.0
3.2
17.5
Dup2
6.2
4.5
0.0
0.8
0.0
2.8
4.8
3.9
2.2
2.4
3.0
5.4
4.0
8.5
6.2
4.2
2.3
0.0
3.4
7.6
10.4
0.0
2.3
22.5
3.2
3.8
0.6
2.5
Dup3
0.7
ND
8.6
3.1
8.0
6.5
3.4
0.7
4.7
4.3
2.6
1.5
3.7
0.7
1.8
3.3
4.3
2.8
2.6
2.1
3.3
0.0
3.3
1.4
4.1
3.5
0.8
28.7
Dup4
13.6
21.9
6.8
1.4
18.2
13.9
5.2
17.1
13.6
10.1
2.1
7.3
0.5
27.3
3.2
14.1
2.9
2.4
3.3
1.7
2.0
22.2
9.1
18.2
2.9
6.1
0.5
14.1
Dup5
3.1
18.8
3.2
1.7
3.8
5.0
0.6
1.8
3.8
4.7
0.7
1.4
0.5
4.3
1.2
0.9
3.7
0.6
1.8
1.5
2.0
4.0
0.7
4.3
6.6
2.6
0.8
5.8
Dup6
8.1
10.0
10.0
5.8
4.4
6.1
3.1
6.1
8.0
5.3
5.5
7.5
1.6
6.6
5.6
16.2
9.3
5.9
9.1
8.0
4.4
2.9
2.1
1.2
3.1
3.7
5.9
29.9
Dup7
1.4
6.8
4.4
1.4
5.4
4.2
6.4
1.1
5.0
9.3
0.9
0.9
7.2
1.4
1.9
5.0
5.5
1.2
2.8
15.9
1.0
4.3
5.8
1.0
11.9
4.3
1.3
-
DupS
142.1
146.3
145.0
140.3
141.9
141.4
151.0
139.0
140.0
144.4
138.1
140.9
150.7
140.6
141.0
134.1
143.6
141.1
141.0
139.5
134.1
167.0
ND
142.5
148.9
149.4
141.1
136.1
TNSSS Sampling and Analysis Technical Report
65
January 2009

-------
Table 32B.  Duplicate Precision for Metals in Sewage Sludge
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Molybdenum
Nickel
Phosphorus
Selenium
Silver
Sodium
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
Mercury
Relative Percent Difference (%)
Dup9
5.0
20.5
1.8
3.2
6.2
8.7
2.5
3.1
2.9
2.2
1.9
3.0
2.3
3.4
2.3
1.9
0.6
2.4
0.0
2.7
1.9
2.2
6.2
12.6
7.3
2.4
1.7
-
DupIO
148.3
97.6
89.8
147.9
81.1
135.6
105.9
140.0
83.9
98.0
148.2
144.6
89.9
129.8
144.9
91.2
89.6
137.7
63.9
118.2
122.4
66.7
ND
ND
89.8
84.3
148.8
-
Dup11
94.0
30.9
41.7
43.6
ND
71.2
38.9
82.1
43.1
17.8
44.9
91.8
43.6
51.1
87.6
48.0
38.2
63.4
50.0
36.5
5.5
ND
48.4
68.5
53.1
46.9
94.5
-
Dup12
7.4
6.3
0.4
0.2
0.0
8.3
1.2
0.4
4.3
0.0
0.3
4.4
1.9
3.7
0.7
4.2
0.4
0.9
1.3
0.0
1.3
8.0
1.3
55.3
2.6
5.7
0.1
8.6
Dup13
20.4
21.9
7.7
21.8
15.2
11.5
17.5
24.6
23.9
27.0
23.1
19.8
28.9
19.1
13.5
6.1
32.5
11.3
2.2
26.8
21.3
6.5
21.0
11.1
20.4
2.4
19.9
18.9
Dup14
2.5
11.2
18.8
3.5
30.9
3.5
33.1
2.7
1.6
1.8
1.2
0.6
5.0
0.8
2.3
46.0
7.9
1.5
45.2
2.5
2.4
43.7
7.6
1.0
1.1
32.7
2.2
-
Dup15
0.1
10.7
7.3
2.3
0.0
0.7
11.6
2.4
1.9
9.5
2.1
1.3
8.8
1.6
2.3
6.0
11.2
2.0
8.6
4.6
1.3
0.0
9.1
4.4
10.5
8.2
1.6
-
       The differences apparent in the three liquid sewage sludge samples in Table 32
(duplicates 8, 10, and 11) cannot be attributed to differences between the percent solids results in
different containers, as was suggested for the field duplicate samples earlier.  Each of these
duplicates was prepared at the laboratory from the single 500-mL HDPE container of sewage
sludge from the particular POTW.

       The laboratory homogenized each sample before removing the two aliquots (the original
sample and the duplicate), but those procedures may not have been entirely adequate for samples
with very low solids contents, or the samples may have settled between collection of the two
aliquots. In addition, the laboratory only measured the percent solids of each sample once, for
the original aliquot. Therefore, if the duplicate aliquot used for the metals analysis had slightly
different solids content than that of the original sample aliquot, this would not be known or
reflected in the dry-weight results in this survey.

       EPA does not include laboratory duplicate analyses  in a survey database, since that would
provide two results for each sample chosen for this laboratory QC test.  Therefore, while the
RPD values in Table 32 may be useful in diagnosing laboratory issues, they do not influence the
survey data EPA may use.
TNSSS Sampling and Analysis Technical Report
66
January 2009

-------
6.8.5   MS/MSD Results for Organics
       Organic contaminants such as the PAHs and semivolatiles of interest in this survey are
generally less common in environmental samples. Therefore, most methods for organics specify
using MS and MSD samples as the means of assessing the applicability of the method to the
matrix of interest, rather than a single MS sample and a duplicate sample analysis.  The
advantage of spiking the analytes into both QC samples is that it avoids the difficulty of
comparing non-detect results to assess precision. If the laboratory  only analyzes an unspiked
duplicate sample and a compound is not found, there is no numerical result that can be compared
to the original sample result, which may also be a non-detect. While one can compare reporting
limits for non-detects, those limits may differ for legitimate reasons that do not reflect analytical
precision.

       The laboratory prepared and analyzed seven sets of MS/MSD samples for the organics in
this survey.  They prepared some of the MS/MSD samples in conjunction with the full-scan
GC/MS analyses and prepared others, later in the survey, with the selected ion monitoring (SIM)
analyses. Five of the seven MS/MSD pairs were solid samples  and two were liquid samples.
The laboratory employed separate acceptance limits for samples analyzed as liquids versus those
analyzed as solids. The results of all seven sets are summarized in Table 33.

 Table 33. MS and MSD Recovery and Precision for Organics
Analyte
2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl)
phthalate
Fluoranthene
Pyrene


2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl)
phthalate
Fluoranthene
Pyrene


2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl)
phthalate
Fluoranthene
Pyrene
MS/MSD 1 (solid)
MS
Rec
55
89
104
149
62
51
MSD
Rec
60
98
112
226
63
61
RPD
9
10
7
41
3
17
MS/MSD 4 (liquid)
MS
Rec
62
66
67
-72
43
57
MSD
Rec
67
80
80
16
44
66
RPD
9
19
18
314
2
15
MS/MSD 7 (solid)
MS
Rec
92
53
83
4622
78
77
MSD
Rec
97
57
95
4785
97
91
RPD
5
7
14
4
22
17
MS/MSD 2 (solid)
MS
Rec
19
-480
-641
-8024
-2101
-613
MSD
Rec
11
-503
-589
-9476
-1980
-754
MS/MSD 5 (liq
MS
Rec
99
97
169
-1362
290
278
MSD
Rec
108
95
94
-275
119
100
RPD
56
5
9
17
6
21
uid)
RPD
8
2
58
133
84
94
MS/MSD 3 (solid)
MS
Rec
120
671
131
-7160
93
105
MSD
Rec
143
685
115
-4296
105
146
RPD
18
2
14
50
12
32
MS/MSD 6 (solid)
MS
Rec
109
75
117
267
124
123
MSD
Rec
110
85
120
416
135
141
RPD
1
12
3
44
8
14

TNSSS Sampling and Analysis Technical Report
67
January 2009

-------
       The recoveries in Table 33 exhibit some of the same issues as for the anions and metals,
including large negative recoveries for some analytes, particularly bis(2-Ethylhexyl) phthalate.
The laboratory adjusted the amounts of the analytes spiked into samples over the course of the
survey.  However, the within-sample variability affected many of the recoveries, resulting in
some negative recovery values.  Therefore, EPA recalculated the MS/MSD recoveries and RPDs
using the alternative recovery equation described earlier.  Table 34 presents the recalculated
results, rounded to one-tenth of a percent, to illustrate some of the smaller differences.

 Table 34. Alternative MS and MSP Recovery and Precision for Organics
Analyte
2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl)
phthalate
Fluoranthene
Pyrene

2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl)
phthalate
Fluoranthene
Pyrene

2-Methylnaphthalene
4-Chloroaniline
Benzo(a)pyrene
bis (2-Ethylhexyl)
phthalate
Fluoranthene
Pyrene
MS/MSD 1 (solid)
MS
Rec
66.6
94.1
104.0
112.3
76.1
71.1
MSD
Rec
70.4
99.1
111.7
131.6
77.1
76.8
RPD
5.5
5.2
7.1
15.8
1.3
7.7
MS/MSD 4 (liquid)
MS
Rec
63.1
78.5
74.8
73.9
69.1
70.3
MSD
Rec
68.5
87.2
84.5
87.3
69.6
76.7
RPD
8.2
10.5
12.2
16.6
0.7
8.7
MS/MSD 7 (solid)
MS
Rec
92.5
91.2
90.0
526.1
88.3
88.0
MSD
Rec
97.3
91.9
97.2
541.5
98.6
95.3
RPD
5.1
0.8
7.7
2.9
11.0
8.0
MS/MSD 2 (solid)
MS
Rec
67.6
41.3
61.3
71.7
40.3
70.8
MSD
Rec
64.2
38.9
64.0
66.7
43.6
65.0
MS/MSD 5 (liq
MS
Rec
99.3
98.3
142.1
68.1
183.3
171
MSD
Rec
107.6
97.4
96.0
91.8
108.3
99.9
RPD
5.2
6.0
4.3
7.2
7.9
8.5
uid)
RPD
8.0
0.9
38.7
29.6
51.4
52.5
MS/MSD 3 (solid)
MS
Rec
109.7
670.6
113.0
82.2
98.6
101.1
MSD
Rec
120.8
685.0
106.0
89.2
101.0
110.0
RPD
9.6
2.1
6.4
8.2
2.4
8.4
MS/MSD 6 (solid)
MS
Rec
108.9
89.6
114.8
115.3
114.1
114.6
MSD
Rec
109.9
93.8
117.4
129.0
120.6
126.1
RPD
0.9
4.6
2.2
11.2
5.5
9.6

       As Table 34 illustrates, all of the negative recovery values are eliminated and the
exceptionally large negative and positive recoveries for bis (2-Ethylhexyl) phthalate were
reduced as well.  With the notable exceptions of 4-chloroaniline in MS/MSD 3 and bis (2-
Ethylhexyl) phthalate in MS/MSD 7, the recoveries range from 39% to 183%.

       A total of 74 out of 84 recoveries are less than 125% and 70 of 84 recoveries are in the
range of 70% to  130%. Only 4 recoveries are below 50%. A total of 38 of 42 RPD values are
less than 20%, with the other 4 RPDs between 30% and 53%. These alternative recovery data
demonstrate that the analytical methods employed for this survey exhibit precision and bias
within expected norms for the analysis of organics.
TNSSS Sampling and Analysis Technical Report
January 2009

-------
6.8.6  MS/MSD Results for PBDEs

       The laboratory performed the PBDE analyses using EPA Method 1614, which normally
employs isotope dilution quantitation of the analytes of interest. Because every sample is spiked
with the labeled compounds and their recoveries are measured in every sample, isotope dilution
methods do not require the analysis of MS/MSD aliquots to assess bias. However, as noted in
Section 4, the survey involved a number of modifications to the published method in order to
overcome analytical challenges presented by the sewage sludge samples. One of these
modifications was that the laboratory did not spike the labeled compounds into the sample before
extraction, but rather spiked the raw sample extracts.  Using this modification, the laboratory was
able  to successfully analyze the survey samples, but their data on labeled compound recovery
does not include an assessment of extraction efficiency. Therefore, in conjunction with that
modification, the laboratory agreed to prepare MS/MSD aliquots with each batch of field
samples analyzed by the modified procedure. Ultimately, four sets of MS/MSD analyses were
performed. Table 35 summarizes the MS/MSD results for the PBDEs.

            Table  35.  MS/MSD Recovery and Precision for PBDEs

Analyte
BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209

BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209
MS/MSD 1
MSRec
91
0
0
102
0
146
144
91
95
104
171
MSD Rec
128
3430
197
202
1750
547
180
256
224
123
2910
RPD
19
0
0
23
0
23
15
22
20
8.7
18
MS/MSD 3
MSRec
99
215
98
102
172
118
141
94
96
106
261
MSD Rec
100
121
98
105
116
98
94
102
101
98
143
RPD
1.8
5.4
1
3.2
3.6
5.1
20
4.4
3.5
3.8
8.7
MS/MSD 2
MSRec
61
0
26
44
0
0
104
0
0
54
0
MSD Rec
54
0
0
18
0
0
91
0
0
35
0
RPD
4.5
0
0
7.1
0
0
7.4
0
0
9.8
0
MS/MSD 4
MSRec
91
298
107
113
315
128
123
106
105
88
291
MSD Rec
60
0
69
61
0
2.9
84
36
40
58
179
RPD
35
0
35
41
0
51
34
45
44
34
28
            The laboratory reported recoveries as zero (0) any time the calculated recovery was
            negative, and reported the RPD as 0 when either of the recovery values was reported
            as 0.

       As noted in Table 35, the laboratory's reporting practices included substituting zero (0)
for any calculated negative recoveries, as well as for the RPD when either recovery value in the
MS/MSD pair is reported as 0. There are  16 negative recoveries reported as 0 for first 2
MS/MSD pairs and only 2 negative recoveries for the last 2 MS/MSD pairs. This is because the
laboratory increased their spiking levels for the later analyses.  There are also a number of
reported recoveries well over  100%, ranging as high as 3430% in one case.
TNSSS Sampling and Analysis Technical Report
January 2009

-------
       Given these recoveries and RPDs, EPA performed the alternative calculations described
in Sections 6.7.1 to 6.7.5 for the PBDE data.  Those alternative recoveries and RPDs are shown
in Table 36, rounded to one-tenth of a percent, to illustrate some of the smaller differences.

           Table 36. Alternative MS/MSP Recovery and Precision for PBDEs
Analyte
BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209

BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209
MS/MSD 1
MSRec
95.7
96.2
72.3
100.3
94.3
103.1
127.0
98.7
98.6
101.8
99.9
MSD Rec
116.1
126.0
124.7
126.4
121.2
129.9
148.4
123.8
120.5
111.5
120.4
RPD
19.3
26.8
53.2
23.0
24.9
23.0
15.5
22.6
20.0
9.1
18.6
MS/MSD 3
MSRec
99.5
107.2
98.5
101.3
104.1
104.2
136.7
96.8
99.3
103.8
111.9
MSD Rec
99.5
101.6
98.1
103.3
100.4
98.6
164.4
100.3
101.7
98.8
102.6
RPD
0.0
5.4
0.5
2.0
3.6
5.6
18.4
3.5
2.4
5.0
8.7
MS/MSD 2
MSRec
81.5
89.0
82.9
88.3
84.1
83.8
101.9
84.5
85.7
78.4
94.0
MSD Rec
51.7
77.3
76.5
82.1
78.0
71.7
94.6
76.8
77.6
71.4
79.5
RPD
44.8
14.1
8.1
7.2
7.6
15.5
7.5
9.6
9.9
9.4
16.7
MS/MSD 4
MSRec
91.8
120.8
106.2
109.0
121.3
110.0
121.1
103.0
102.2
89.3
172.4
MSD Rec
64.5
68.7
74.1
72.1
66.5
65.3
86.1
65.5
65.2
63.3
130.1
RPD
35.0
55.0
35.5
40.8
58.4
51.1
33.8
44.5
44.2
34.1
27.9
       Using the alternative calculations, all 18 of the negative values originally reported by the
laboratory as "0" were eliminated. The recalculated recoveries range from 52% to 172%.  The
laboratory employed acceptance limits of 50-150% for MS/MSD recoveries, and all but the one
recalculated recovery of 172.4% fall within that range.  The calculated RPD values range from
0% to 58.4%.  Only three recalculated RPD values were above the laboratory's acceptance limit
of 50%.

       As with the other analytical classes, these alternative recovery data demonstrate that the
analytical methods employed for this survey exhibit precision and bias within expected norms
for the analysis of PBDEs.

6.8.7  Qualification of Sample Results based on MS/MSD Results

       During review of the sample  results, EPA qualified any MS/MSD results reported by the
laboratory that fell outside of the relevant acceptance limits.  Those data qualifiers were carried
over into the results database.  However, the results of the alternative calculations shown in
Section 6.7 for all classes of analytes demonstrate that the shortcomings in MS/MSD recoveries
and precision are largely a function of the calculations in many EPA methods.
TNSSS Sampling and Analysis Technical Report
70
January 2009

-------
       Therefore, while retaining the data qualifiers that indicate that the recoveries and/or RPD
reported by the laboratories fell outside of the acceptance limits, EPA also added a qualifier to
affected samples to indicate that the alternative calculations suggest that method performance in
the sewage sludge matrix is not an immediate concern.  This new qualifier is "ACAP," for
"Alternative Calculation indicates Acceptable Performance," as stated in the database.

6.8.8   Labeled Compound Recoveries for Isotope Dilution Methods

       As noted elsewhere in this report, EPA's isotope dilution methods spike labeled analogs
of the target analytes into each sample and monitor the recoveries of those labeled compounds as
a measure of method performance. Therefore, EPA's isotope dilution methods do not require
that the laboratory prepare separate matrix spike samples. Rather than assessing extraction
efficiency and other aspects of method performance on  5% of the samples (i.e., 1 out of every 20
samples is used to create an MS/MSD pair), the isotope dilution methods generate extraction
efficiency data on 100% of the samples, and use those data to correct the final results for each
analyte for the recovery of its labeled compound.

       During the data review process, EPA checked the recovery of every labeled compound in
every field sample and QC samples against the method  acceptance criteria for that analyte.
When the labeled recovery fell outside of the acceptance criteria, EPA flagged in the database
the sample results for the associated unlabeled analytes.
TNSSS Sampling and Analysis Technical Report             71                                   January 2009

-------
                                     Section 7
                                    References
NRC, 2002. Biosolids Applied To Land: Advancing Standards And Practices, National Research
Council of the National Academies. July 2002.

USEPA, 2003. United States Federal Register. 68 FR 75531. Wednesday December 31, 2003,
pp.75531-75552.

Vadas, P. A. and Kleinman, P. J. A., 2006. Effect of Methodology in Estimating and Interpreting
Water-Extractable Phosphorus in Animal Manures, J. Environ. Qual., 35: 1151-159.
TA/SSS Sampling and Analysis Technical Report             72                                 January 2009

-------
                                     Appendix A
                    Solids Leaching Procedure for Anions
1. Using the percent solids determined by drying an aliquot of the sample of known weight
   overnight at 103 - 105°C and reweighing, weigh out a sample aliquot equal to 0.5 g dry
   weight in a 200- or 250-mL wide-mouth (screw-top) plastic bottle.

2. Add reagent water until the total mixture mass is 100.5 g.  The resultant solids:solution ratio
   is 0.5 g solids: 100-mL solution, or 1:200.

3. Seal bottles with screw top caps and place in a standard laboratory shaker set at 70
   revolutions per minute for 60 minutes.

4. Upon completion of agitation, centrifuge sealed bottles at 2000 rpm for 10 minutes.

5. Gravity filter the concentrate from step 4 using Whitman #2 filter paper.  (Suggest 150-mm
   circular filters folded and placed in simple plastic lab funnels.)

6. Adjust the pH of the filtrate to pH<2 with H2SO4 to preserve the nitrate/nitrite in the filtrate
   and store at 4°C until analysis.

7. The holding time for nitrate/nitrite is 48 hours, so samples must be analyzed for all three
   analytes within 48 hours  of preservation.

8. All sample results will be reported in mg/kg on the basis of the original 0.5-g sludge sample.
Adapted from Vadas, P. A. and Kleinman, P. J. A., 2006.
TNSSS Sampling and Analysis Technical Report             73                                   January 2009

-------
                                   Appendix B
               Method Modifications for the PBDE Analyses
      The following table summarizes the modifications made to the sample preparation and
analysis procedures described in EPA Method 1614 for the analysis of the PBDEs. EPA and the
laboratory instituted these modifications to address the interferences present in extracts of the
samples, and the relatively high levels of some PBDEs in the samples.

      The table provides a reference to the relevant section of the draft method, the original
method specification, and the revised approach employed for this survey.
Method Section
11.5.1
11.5.2
12.6
13
13.2.3
1.1.1 and 17
17.1
17.1
17.5
Topic
Sample size
Spiking labeled
standards
Macroconcentration
of extract
Cleanups
GPC cleanup
Target analytes
Isotope dilution
quantitation
Matrix spike
samples
Calibration range
Original Method Specification
10 g dry weight
Spike into sample before extraction
3-4 mL
May use GPC, silica gel, alumina,
or Florisil, if needed.
Process 5 mL of extract
All 209 possible BDE congeners
11 congeners determined by true
isotope dilution, and the remaining
congeners by internal standard.
The internal standards are the
labeled congeners for other PBDEs
that are added to the samples
before extraction.
Not used, due to isotope dilution
All results must be within the
calibration range, or must be
diluted to bring them within range
Revised Approach
As little as 0.2 g dry weight
Spike sample extract before
cleanup
10 mL
Must use silica gel, GPC, and
alumina, in that order
Process 1 mL of the 10 mL extract
Only 11 congeners:
BDE-28
BDE-47
BDE-66
BDE-85
BDE-99
BDE-100
BDE-138
BDE-153
BDE-154
BDE-183
BDE-209
11 congeners quantified using
labeled standards for 8 of those 1 1
congeners, all spiked into the
extract before cleanup. All results
are corrected for losses during the
cleanup steps, but not for initial
extraction efficiency.
Added periodic MS/MSD to
provide data on extraction
efficiency
Results for some congeners like
BDE-209 flagged "E" in the
database if above the calibration
range but not high enough to
saturate the detector system.
TNSSS Sampling and Analysis Technical Report
74
January 2009

-------
                                       Appendix C
                               QC Acceptance Criteria
QC Acceptance Criteria for the Targeted National Sewage Sludge Survey
Analytical Fraction
Anions
Metals
Organics
QC Parameter
LCS
MS Recovery
Duplicate Precision (RPD)
LCS for solid samples
LCS for liquid samples
MS/MSD Recovery
MS/MSD Precision (RPD)
LCS
MS/MSD Recovery for solid
samples
Analyte
Fluoride
Phosphorus
Nitrate/Nitrite
All analytes
All analytes
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Phosphorus
Selenium
Silver
Sodium
Thallium
Tin
Titanium
Vanadium
Yttrium
Zinc
All metals
All metals, except as noted
below
Mercury
Tin
Titanium
All metals
4-Chloroaniline
2-Methylnaphthalene
Fluoranthene
Pyrene
bis (2-Ethylhexyl) phthalate
Benzo(a)pyrene
4-Chloroaniline
2-Methylnaphthalene
Fluoranthene
Pyrene
bis (2-Ethylhexyl) phthalate
Benzo(a)pyrene
r
Acceptance Limits (%)
85-115
85-115
90-110
75-125
20
58-142
12-223
77-123
82-118
77-122
56-144
80-121
79-121
78-121
80-120
82-118
50-150
79-121
77-123
80-120
60-123
72-128
81-119
NA
76-124
61-139
56-145
76-124
NA
40-160
76-124
NA
79-120
85-115
70-130
60-128
50-150
50-150
30
39-95
12-159
47-139
52-129
39-174
49-144
10-104
39-116
49-130
58-110
53-172
50-126
  NA = Not applicable. The commercial reference material used for the LCS does not have certified values for these analytes.
TNSSS Sampling and Analysis Technical Report
75
January 2009

-------
  QC Acceptance Criteria for the 2006-2007 Targeted National Sewage Sludge Survey
Analytical Fraction
Organics
PBDEs
QC Parameter
MS/MSD Recovery for liquid
samples
MS/MSD Precision (RPD)
Surrogate Recovery
LCS
Labeled Compound
Recovery
MS/MSD Recovery
MS/MSD Precision (RPD)
Analyte
4-Chloroaniline
2-Methylnaphthalene
Fluoranthene
Pyrene
bis (2-Ethylhexyl) phthalate
Benzo(a)pyrene
All analytes, all matrix types
Nitrobenzene-ds
2-Fluorobiphenyl
p-Terphenyl-d-M
All analytes
All, except as shown below
IJC-BDE-209
All analytes
All analytes
Acceptance Limits (%)
10-62
10-109
10-150
10-136
10-150
10-152
40
35-128
43-133
49-137
25-150
25-150
20-200
50-150
50
                    QC Acceptance Criteria for Pharmaceuticals for the
                    Targeted National Sewage Sludge Survey
Analyte
VER (%)
OPR (%)
Acid-Extractable Fraction - Positive Electrospray lonization
Acetaminophen
Azithromycin
Caffeine
Carbadox
Carbamazepine
Cefotaxime
Ciprofloxacin
Clarithromycin
Clinafloxacin
Cloxacillin
Codeine
Cotinine
Dehydronifedipine
Digoxigenin
Digoxin
Diltiazem
1,7-Dimethylxanthine
Diphenhydramine
Enrofloxacin
Erythromycin
Flumequine
Fluoxetine
Lincomycin
Lomefloxacin
Miconazole
Norfloxacin
Norgestimate
Ofloxacin
Ormetoprim
Oxacillin
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
50-120
33-120
50-124
33-144
21 -137
8-186
50-120
8-154
5-200
5-200
34-129
50-124
42-120
8-183
5-148
11-120
50-138
48-120
50-125
50-158
36 - 200
49-125
5-120
17-120
27-120
50-135
36-120
50 - 200
50-120
5-200
TNSSS Sampling and Analysis Technical Report
76
January 2009

-------
                    QC Acceptance Criteria for Pharmaceuticals for the
                    Targeted National Sewage Sludge Survey
Analyte
Oxolinic Acid
Penicillin G
Penicillin V
Roxithromycin
Sarafloxacin
Sulfachloropyridazine
Sulfadiazine
Sulfadimethoxine
Sulfamerazine
Sulfamethazine
Sulfamethizole
Sulfamethoxazole
Sulfanilamide
Sulfathiazole
Thiabendazole
Trimethoprim
Tylosin
Virginiamycin
VER (%)
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
OPR (%)
42-124
5-200
5-200
38-120
17-200
50 - 200
5-200
50 - 120
50-148
50-142
50-120
50-120
5-189
41 -120
50-120
50-126
16-149
5-189
Tetracyclines
Anhydrochlortetracycline
Anhydrotetracycline
Chlortetracycline
Demeclocycline
Doxycycline
4-Epianhydrochlortetracycline
4-Epianhydrotetracycline
4-Epichlortetracycline
4-Epioxytetracycline
4-Epitetracycline
Isochlortetracycline
Minocycline
Oxytetracyclin
Tetracycline
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
50-135
7-141
45-172
5-200
22-166
18-120
5-200
40-150
50-142
50-173
5-200
5-176
50-183
50-155
Acid-Extractable Fraction - Negative Electrospray lonization
Gemfibrozil
Ibuprofen
Naproxen
Triclocarban
Triclosan
Warfarin
70-130
70-130
70-130
70-130
70-130
70-130
50-120
50-120
50-120
50-120
50-120
50-120
Base-Extractable Fraction - Positive Electrospray lonization
Albuterol
Cimetidine
Metformin
Ranitidine
70-130
70-130
70-130
70-130
50-133
5-120
50-149
24-160
TNSSS Sampling and Analysis Technical Report
11
January 2009

-------
                QC Acceptance Criteria for Pharmaceutical Labeled Compounds for
                the Targeted National Sewage Sludge Survey
Analyte
VER (%) OPR (%)
Recovery in Samples (%)
Acid-Extractable Fraction - Positive Electrospray lonization
13C2-15N-Acetaminophen
13C3-Caffeine
13C3-15N-Ciprofloxacin
Cotinine-ds
13C2-Erythromycin
Fluoxetine-ds
13C6-Sulfamethazine
13Ce-Sulfamethoxazole
Thiabendazole-d6A
13C3-Trimethoprim
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
5-200
5-200
5-200
5-120
50 - 120
50-126
5-157
50 - 146
50-146
50-177
19-200
31 - 200
37-181
5-145
23 - 120
40-148
12-120
40-129
32-140
50-172
Tetracyclines
Thiabendazole-deA
70-130
50-120
30-132
Acid-Extractable Fraction - Negative Electrospray lonization
Gemfibrozil-de
13C3-lbuprofen
13C-Naproxen-d3
13C6-Triclocarban
13Ci2-Triclosan
Warfarin-ds
70-130
70-130
70-130
70-130
70-130
70-130
38-122
28-122
34-131
5-172
5-168
50-177
21 -123
29 - 127
14-132
5-147
5-153
50 - 200
Base-Extractable Fraction - Positive Electrospray lonization
Albuterol-ds
Metformin-de
70-130
70-130
35-121
5-141
39-141
5-200
                 Thiabendazole-de is used as a labeled analog in both the tetracyclines and the acid
                  extractable-positive electrospray fractions, with separate acceptance criteria.
TNSSS Sampling and Analysis Technical Report
78
January 2009

-------
                   QC Acceptance Criteria for Steroids and Hormones for
                   the Targeted National Sewage Sludge Survey
Analyte
Androstenedione
Androsterone
Campesterol
Cholestanol
Cholesterol
Coprostanol
Desmosterol
17a-Dihydroequilin
Epicoprostanol
Equilenin
Equilin
Ergosterol
17a-Estradiol
17a-Ethinyl estradiol
17(3-Estradiol
(3-Estradiol-3-benzoate
Estriol
Estrone
Norethindrone
Norgestrel
Progesterone
(3-Sitosterol
(3-Stigmastanol
Stigmasterol
Testosterone
VER (%)
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
65-135
50-150
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
70-130
OPR (%)
5-200
50-121
40 - 200
50-164
5-200
34 - 200
5-200
45-151
50-197
5-200
5-200
5-200
50-120
50-123
50-176
5-189
5-193
50-173
45 - 200
46 - 200
5-200
5-200
29 - 200
50 - 200
50-136
            QC Acceptance Criteria for Steroid and Hormone Labeled Compounds
            for the Targeted National Sewage Sludge Survey
Analyte
Cholesterol-d/
17a-Ethinyl estradiol-d4
17(3-Estradiol-d4
Norethindrone-de
Norgestrel-de
Progesterone-dg
VER (%)
70-130
70-130
70-130
70-130
70-130
70-130
OPR (%)
50-120
50-120
50-120
37-120
36-120
5-200
Recovery in Samples (%)
50-120
50-120
29-132
12-120
7-120
5-200
TNSSS Sampling and Analysis Technical Report
79
January 2009

-------