United States
Environmental Protection
Agency
Environmental Monitoring and
Support Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S4-86/043 March 1987
Project Summary
Development of Preservation
T— hniques and Establishment of
Times:
its of the
)ischarge
Elimination bybio,.i and Safe
Drinking Water Act
Huge S. Prentice and Daniel F. Bender
This project was designed to statis-
tically determine the length of time a
sample can be stored for analysis of ten
National Pollutant Discharge Elimina-
tion System (NPDES) compliance pa-
rameters and two Safe Drinking Water
Act (SDWA) parameters. Accepted
preservation and storage techniques
were used on spiked representative
real-world samples. Experimentally
determined maximum holding times
(MHT) were calculated for each pa-
rameter in three different water matrices
and under several different preservation
conditions for certain parameters.
Fifteen experiments were conducted
for the ten NPDES compliance parame-
ters and three experiments for the two
SDWA compliance parameters. The ten
NPDES parameters were phenols,
cyanide, mercury, ammonia, nitrate plus
nitrite, fluoride, total Kjeldahl nitrogen,
total phosphorus, total organic carbon,
and sulfide. Multiple experiments were
performed to evaluate MHTs for dif-
ferent preservation techniques for
phenols and mercury. The two SDWA
parameters were nitrate and fluoride,
with two experiments performed to
evaluate MHTs for different preserva-
tion techniques for nitrate.
This Project Summary was developed
by EPA's Environmental Monitoring and
Support Laboratory, Cincinnati, OH, to
announce key findings of the research
project that Is fully documented In a
separate report of the same title (see
Project Report ordering Information at
back).
Experimental Design and
Approach
The experiments performed during this
study are summarized in Tables 1 and 2.
For each of the 15 wastewater experi-
ments, three different wastewater sam-
ples were collected. For the three drinking
water experiments, three finished drink-
ing waters were collected from potable
water treatment plants treating different
raw waters using different methods. For
wastewater or drinking water parameters
with multiple experiments (e.g., phenols
and mercury in wastewater and nitrate in
drinking water), sample matrices were
collected at the same site, and mixing
and analyses were started on the same
day. The protocol to prepare samples for
a time-sequenced analysis of one pa-
rameter in a matrix was performed 54
times during the project (three matrices
-------
Table 1. Summary of Experimentally Determined MHTs for Wastewater Parameter Experiments
Experiment
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Parameter
Phenols
Phenols
Phenols
Cyanide
Mercury
Mercury
Mercury
Mercury
Ammonia
Fluoride
TKN
Nitrite
plus
nitrate
Organic
carbon
Total
phosphorus
Sutfide
Preservation
technique
Cool4°C
Cool4°C
1 g/L CuS04
H3P04pH<4
Cool4°C
Cool4°C
NaOHpH>12
HN03pH<2
0.05%K2Cr20J
HNO3pH<2
0.06% Kfr^Oj
HNO3pH<2
HNO3pH<2
Cool4°C
Conc.H,SO4pH<2
None
Cool4°C
Cool4°C
Cone. HzS04pH<2
Cool4°C
Cone. W«SO
-------
for 18 experiments). The protocol for pre-
paring samples for time-sequenced
analysis included:
1. Collecting grab samples and ship-
ping to the laboratory within 24
hours;
2. Compositing and homogenizing the
grab samples in a large-volume
mixing container;
3. Collecting three aliquots for analysis
from the composite prior to addition
of preservative;
4. Collecting three aliquots for analysis
from the composite containing the
preservative;
5. Adding the study analyte to the
preserved composite and mixing;
6. Collecting 100 aliquots (dispensed
to ensure a large enough population
for representative random sampling)
from the spiked, preserved compo-
site; and
7. Randomly selecting 18 fixed aliquots
of the 100 aliquots chosen for time-
sequences analysis.
The time sequence was either 0, 1, 2, 4,
8, and 16 days or 0, 2, 4, 8, 16, and 32
days.
The time-sequenced experimental data
were used to calculate experimentally
determined maximum holding times
(MHTs) according to a mathematically
acceptable definition. The experimentally
determined or calculated MHT was de-
fined as the 90 percent lower confidence
limit of a specified critical time (CT) (see
Figure 1). The CT was defined as the time
at which a 10 percent change in the
analyte concentration (from the intercept
at Day 0) occurred and when the precision
of the method allowed that 10 percent
change to be a statistically significant
difference at the 90 percent confidence
level. When the precision of the method
was not sufficient to discern a 10 percent
change in the analyte concentration, the
CT was that time at which the percent
change (greater than 10 percent) in the
analyte concentration represented a
statistically significant difference at the
90 percent confidence level. An upper
limit of a 15 percent change in the analyte
concentration was set for determining a
valid CT. The experimentally determined
MHTs were specified as either the 90
percent lower confidence limit (LCL) of
the CT or the maximum time of the
designed experiment when the LCL was
longer.
A statistical protocol was developed to
provide information to support the use of
a zero or first order kinetic model to
calculate a CT and its 90 percent LCL (or
1.1Cl
0.9CI
1O% Gain
Initial Concentration
Best-Fit Line of Experimental Data
MHT (LCL) CT
Time
Figure 1. Graphic representation of MHT definition.
MHT). Only the raw data zero-order model
(concentration versus time) and the
transformed on first-order model (natural
log of concentration versus time) were
evaluated since loss or gain of the con-
stituent was expected to follow zero-or-
first-order kinetics.
Results
A summary of the results for the
wastewater parameter experiments is
presented in Table 1. The experimentally
determined MHTs resulting from this
study were longer than proposed MHT
values published in the Federal Register
(44/244:75-28-75052, Tuesday, Decem-
ber 18, 1979) and recommended by EPA
in Methods for Chemical Analysis of
Water and Wastes, EPA-600/4-79-020,
for the following parameters in waste-
water:
Phenols Nitrate plus nitrite
Ammonia Total organic carbon
Fluoride Total phosphorus
Total Kjeldahl Sulfide
Nitrogen (TKN)
A complete set of mercury and cyanide
experimental results were unobtainable
because of difficulties with the analytical
procedures. The experimentally deter-
mined MHTs for fluoride and nitrate in
drinking water were longer than the EPA
required values and are summarized in
Table 2. Additional analyses for the nitrate
in drinking water experiments supported
experimentally determined MHTs of 32
days.
Conclusions
From the results obtained under the
set of conditions used for this study, the
following conclusions are drawn:
1. The experimentally determined
MHTs were longer than the proposed
holding times specified in the
Federal Register and the.fitcom-
mended values from EPA for the
analysis of the parameters in waste-
water samples listed below.
Phenols Nitrate plus nitrite
Ammonia Total organic carbon
Fluoride Total phosphorus
TKN Sulfide
For mercury and cyanide, the experi-
mentally determined MHTs were not
longer than the proposed or recom-
mended values.
2. The experimentally determined
MHTs were longer than the EPA
required holding time values for the
analysis of nitrate and fluoride in
samples from finished drinking
water.
3. Cyanide concentrations in waste-
water samples containing sulfur
-------
compounds, which are converted to
sulfide during the distillation proce-
dure, cannot accurately be deter-
mined using the currently approved
analytical method for cyanide.
4. The major reasons that some ex-
perimentally determined MHTs were
lower than the proposed and recom-
mended values were: the known
variability for mercury analyses,
especially at sample concentratons
less than 2.0 micrograms per liter
(jug/L), and the possible effective-
ness of 0.05 percent potassium
dichromate (K2Cr207) in samples
acidified with nitric acid (HN03) to
extract mercury from the walls of
new, HNOa-rinsed polyethylene
Hedwin Cubitainers®*. The slopes
of the best-fit regression lines were
not statistically similar for the two
experiments using polyethylene
Cubitainers®. Decreasing concentra-
tions with the time were observed
when samples were preserved with
HN03, except at the highest mercury
concentration where no statistical
change with time was observed. In-
creasing concentrations with time
were observed when 0.05 percent
K2Cr207 and HN03 were used as
preservative.
5. Different sample preservation tech-
niques were studied for nitrate,
phenols, and mercury analyses. The
two preservation techniques for
finished drinking water for nitrate
analysis yielded the same experi-
mentally determined MHTs. Of the
three preservation experiments for
phenols in wastewater, the two in
which H2SO4 was added to samples
stored in plastic or glass provided
the longest experimentally deter-
mined MHTs. Of the four mercury-
in-wastewater preservation experi-
ments, HNO3 and K2Cr207 added to
samples stored in glass provided the
best experimentally determined
MHTs.
6. The phenols experiments that used
H2S04 as the preservative consis-
tently exhibited higher coefficients
of variation for the unexplained
variability from regression diagnos-
tics compared to the copper sulfate/
phosphoric acid preservation experi-
ments. The highest CVs were from
Sample Matrix 3, which also had
the highest suspended solids con-
tent of the sample matrices collected
for the phenols experiments. Since
the higher variability is due to the
high suspended solids content, the
higher CVs were accepted as rea-
sonable for analytical variability,
therefore, allowing the calculated
MHT to be considered valid for that
matrix.
7. A statistically significant lack of
linear fit was encountered for the
experimental data due primarily to
larger day-to-day analytical vari-
ability compared to within-day
analytical variability. This phenom-
enon is not entirely unexpected, and
it has caused some concern from a
purely statistical standpoint. Specifi-
cally, the concern arises from de-
riving a precise procedure to evaluate
when the experimental data exhibits
a true lack of linear fit. A true lack of
linear fit for the experimental data
implies that an MHT cannot be
validly calculated. The procedure
used for this study represents a
reasonable approach for the deter-
mination of MHTs. Further holding
time studies should recognize the
problems encountered during this
study and possibly incorporate the
following two recommendations.
First, studies should provide a
mechanism to precisely define pro-
ject-specific analytical variability for
comparison to lack-of-fit variability.
Second, studies should use an
analysis of the residuals from the
linear regression line to detect out-
liers and determine the influence
from extreme observations.
Mention of trademarks or commercial pro-
ducts does not constitute endorsement or
recommendation for use.
H. S. Prentice. J. E. Singley. L J. Bilello, J. T. McClave. K. L Tuttle. E. M. Kellar. and
M. G. Schultz are with Environmental Science and Engineering, Inc.,
Gainesville, FL 32602.
Daniel F. Bender is the EPA Project Officer (see below).
The complete report, entitled "Development of Preservation Techniques and
Establishment of Maximum Holding Times: Inorganic Constituents of the
National Pollutant Discharge Elimination System and Safe Drinking Water
Act," (Order No. PB 87-132 833/AS; Cost: $18.95, subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No G-35
Official Business
Penalty for Private Use S300
EPA/600/S4-86/043
0000319 PS
------- |