CD A U.S. Environmental Protection Agency Industrial Environmental Research EPA-600/7-78-017
til f^ Office of Research and Development Laboratory •m*7O
Research Triangle Park, North Carolina 27711 February 197o
EFFECT OF HANDLING
PROCEDURES ON SAMPLE
QUALITY
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S.
Environmental Protection Agency, have been grouped into seven series.
These seven broad categories were established to facilitate further
development and application of environmental technology. Elimination
of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy—Environment Research-and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from
the effort funded under the 17-agency Federal Energy/Environment
Research and Development Program. These studies relate to EPA's
mission to protect the public health and welfare from adverse effects
of pollutants associated with energy systems. The goal of the Program
is to assure the rapid development of domestic energy supplies in an
environmentally—compatible manner by providing the necessary
environmental data and control technology. Investigations include
analyses of the transport of energy-related pollutants and their health
and ecological effects; assessments of, and development of, control
technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
I
This document is available to the public through the National Technical
Information Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-017
February 1978
EFFECT OF HANDLING PROCEDURES
ON SAMPLE QUALITY
by
J.W. Adams, T.E. Doerfler, and C.H. Summers
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02140
Contract No. 68-02-2150, T.D. 10501
Program Element NO.EHB529
EPA Project Officer: Larry D. Johnson
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
TABLE OF CONTENTS
Page
TABLE OF CONTENTS iii
LIST OF TABLES iv
LIST OF FIGURES vi
ACKNOWLEDGEMENT vii
SUMMARY viii
I. INTRODUCTION 1
II. EXPERIMENTAL CONSIDERATIONS 3
A. General Considerations 3
B. Experimental Design 3
C. Sample Identification £
D. Sample Formulation 6
E. Handling Cycle 8
F. Analytical Procedures 8
G. Data Analysis and Interpretation 9
III. EXPERIMENTAL RESULTS 11
IV. DISCUSSION OF RESULTS 22
A. Statistical Analysis of Experimental Data ... 22
B. Interpretation of Results 25
V. CONCLUSIONS AND RECOMMENDATIONS 37
REFERENCES 38
APPENDICES 40
I
APPENDIX A - Formulation Procedures for
Samples 41
APPENDIX B - Analytical Results of Metallic
Content Contained Within Green
Liquor, White Liquor and Green
Liquor Extract 45
APPENDIX C - Gas Chromatography Conditions
Employed During Analysis .... 48
iii
-------
LIST OF TABLES
Table No. Page
1 Admissible Experimental Conditions ... 5
2 Model Compound Chemical and Physical
Properties 7
3 Analytical Results from Stored Probe Wash
Samples (Set I) (Percent Deviation
from Expected) , „
4 Analytical Results from Stored Sorbent
Trap Condensate Extract Samples (Set II)
(Percent Deviation from Expected) ... 14
5 Analytical Results from Stored Sorbent
Resin Samples (Set III) (Percent
Deviation from Expected) 15
6 Statistical Summary of Standard Calibra-
tion Samples 16
7 Results of Extraction Step Experiment
(Percentage Deviations from Unspiked
Standard) 17
8 Analysis of Variance - Probe Wash Experi-
ment (Phenol) 23
9 Probe Wash Experiment (Set I) (Average
Percentage Deviation from Expected) . . 24
10 Probe Wash Experiment (Set I) Signifi-
cant Two-factor Interactions 26
11 Analysis of Variance - Condensate Extract
Experiment (Phenol) 28
12 Condensate Extract Experiment (Set II)
(Average Percentage Deviation from
Expected) 29
13 Condensate Extract Experiment (Set II)
Significant Two-factor Interactions . . 30
continued.
iv
-------
LIST OF TABLES (continued)
Table No. Page
14 Analysis of Variance, Sorbent Resin
Experiment (Phenol) . . . . 32
15 Sorbent Resin Experiment (Set III)
(Average Percentage Deviation from
Expected) 33
APPENDIX A Table No.
Al Formulation Procedure for Probe Wash
Samples (Set I) 42
A2 Formulation Procedure for Sorbent Trap
Condensate Extract Samples (Set II) . 43
A3 Formulation Procedure for Sorbent Resin
Samples (Set III) 44
APPENDIX B Table No.
Bl Metallic Content of Green and White
Liquors by ICPOES 46
B2 Metallic Content of White Liquor, Green
Liquor and Green Liquor Extract by
Emission Spectrometry 47
APPENDIX C Table No. (
Cl Gas Chromatographic Conditions .... 49
-------
LIST OF FIGURES
Figure No. Page
1 GC/MS Chromatogram of Set II Sample
P-NN-H-RL-S 19
2 GC/MS Chromatogram of Set II Sample
P-NN-R-RD-S 20
3 GC/MS Chromatogram of Set III Sample
AG-AA-C-CD 21
4 Probe Wash Experiment (Set I) Significant
Two-Factor Interactions 27
5 Condensate Extract Equipment (Set II),
Significant Two-Factor Interactions . . 31
6 Sorbent Resin Experiment (Set III)
Significant Two-Factor Interaction . . 34
vi
-------
ACKNOWLEDGEMENTS
The authors wish to express their sincerest gratitude
to Drs. Judith C. Harris and Philip L. Levins, as well as to
Messrs. Stephen Spellenberg and James Stauffer of ADL for
their contributions during the design, performance and analy-
sis of the experimental scheme. Appreciation is also extended
to Drs. Larry Johnson, Robert Statnick and Raymond Merrill of
the Process Measurements Branch, Industrial Environmental
Research Laboratory of the Environmental Protection Agency at
Research Triangle Park for their suggestions and comments ,on
experimental planning and review.
vii
-------
SUMMARY
The results of a study designed to evaluate the effects of
typical shipping and storage handling procedures on organic mate-
rials collected in Level 1 studies are presented. Specific parame-
ters reviewed included sample container composition (amber glass
and high density linear polyethylene), head space composition (air
or nitrogen), temperature (38°C, 21°C, 5°C), lighting condition
(dark and diffuse sunlight) and catalytic species content. Three
unique sample sets, representative of fractions obtained during a
Level 1 environmental assessment, were formulated containing six
model organic compounds. A three-week simulated shipping and
storage cycle was used as representative of elapsed time between
sample collection and analysis. All three experiments were de-
signed in accordance with statistical principles appropriate for
conducting factorial experiments. Experimental results were ana-
lyzed by the Analysis of Variance technique in order to assess the
relative effect of each shipping/storage condition studied.
The major findings of these experiments are as follows:
• Generally, any reasonable combination of shipment temperature,
head space composition, and storage condition can be utilized
for the short term storage of Level 1 survey analysis samples.
• These same conditions can be employed for the short term
storage of samples undergoing quantitative chemical analysis
where the accepted range of variation is i 10 percent.
• Amber glass containers are preferred for sample storage over
polyethylene bottles, as it has been found that spurious con-
taminants may be extracted from the container walls by contact
with organic solvents.
• Diffuse sunlight can penetrate polyethylene containers and
cause degradation in species that are photoreactive. Airiber
glass on the other hand will absorb the harmful wavelengths
prohibiting photodegradation.
viii
-------
I. INTRODUCTION
Sampling is conducted as part of a Level 1 environmental assess-
ment study which provides data on the overall composition of process
and effluent streams. The collected samples constitute a small por-
tion of the stream and are assumed to be representative of the entire
stream. Therefore, the immediate goal of any process involving
sample collection is to insure that the results obtained from that
portion are truly representative of the whole. As such, all sam-
pling and analysis schemes are developed using best state of the
art procedures in an attempt to minimize the occurrence of un-
characteristic changes. If interferences are encountered, ap-
propriate steps are implemented to either quantify their contri-
bution or to eliminate them altogether. Failing this, the methods
are rejected and alternative procedures explored.
Despite the rigorous method development practices, there still
are areas where changes can occur and not be detected. One such
instance occurs between the time a sample is first placed within a
container, and when it is removed for analysis. During this time,
the sample may be exposed to a number of adverse conditions, such
as heat or light, that can result in change or degradation.
Normally, it is preferable to minimize the length of time that
any sample is retained prior to analysis, as this reduces the avail-
able time for changes to occur. However, in many instances, this
cannot be done, and alternative precautionary preservation proce-
dures must be devised to control sample quality.
Three factors are of concern in the development of any preser-
vation technique: contamination, loss and chemical reactivity.
Any one or combination of these factors can happen during the ship-
ment and storage of samples, compromising the value of their analysis.
It is important, therefore, that practices be designed to alleviate,
if not eliminate, their occurrence.
i
Many handling procedures are documented within the literature,
all of which are designed to fulfill a specific requirement. For
example, plastics of any sort (excluding Teflon) have been shown to
adsorb sample constituents or introduce extraneous contaminants (1, 2).
Specific classes of contaminants introduced include plasticizers,
antioxidants, colorants and stabilizers (3). Thus, many researchers
indicate a preference for glass containers, equipped with Teflon or
aluminum foil cap inserts (1, 2, 4). However, even this technique
is not foolproof as both Teflon and aluminum may be contaminated
with up to 400 parts per billion of di-2-ethylhexyl phthalate (5).
-------
Elaborate container and apparatus cleaning procedures are out-
lined in a number of references (6, 7, 8). Coupled with extensive
pre-utilization container storage specifications, these procedures
minimize spurious contamination but do not completely eliminate
their possible occurrence.
During actual shipment and storage operations, many researchers
(1,7,9,10) have noted preferences for either sub-zero temperature
conditions or freeze drying of sample media, but even these pre-
cautions do not eliminate possible compositional variations over
extended periods of time.
To further complicate matters, many desirable handling practices
cannot be implemented and rigorously followed under adverse field
collection conditions. Assuming that the analyst is able to collect
samples and package them according to defined practices, their desti-
nations and environments are not easily regulated unless they are
always under the immediate control of the investigator during ship-
ment and storage.
Recently, the Process Measurements Branch of the Industrial
Environmental Research Laboratory (IERL) published a manual detail-
ing specific procedures to be used for the collection and analysis
of materials during a Level 1 environmental assessment. However,
since little quantitative data was available describing the impact
of handling procedures, few recommendations were presented for
proper shipment and storage practices. To supplement the manual,
the Process Measurements Branch of the IERL commissioned this study
to review normally encountered procedures and provide data describ-
ing the impact of chosen conditions upon the sample quality of
organic species in the samples.
Although a complete review of methods is impractical, it is
hoped that sufficient data will be produced to highlight practices
that may be practically implemented and result in minimum sample
degradation and contamination.
-------
II. EXPERIMENTAL CONSIDERATIONS
A. General Considerations
Three main factors contribute to the degradation of sample
quality: contamination, sample loss, and chemical reactivity or
instability. Given appropriate conditions, any one or combination
of these factors may occur and become a source of interference
within the retained sample, detracting from its representative
nature. A number of alternative shipment and storage procedures
have been proposed and implemented, many of which include provi-
sions for the regulation of at least one of the following parameters.
• shipping and storage time
• sample container composition and quality
• shipment and storage temperature
• container head space composition
• storage lighting condition
• presence of catalytic species
Little quantitative information is available detailing the poten-
tial impact of these three factors upon mixed samples. Within this
study, various levels of each of these parameters were experimentally
controlled to a assess their contribution to degradation within
samples.
It was decided at the outset that test samples would be developed
which were representative of the type of sample collected during a
Level 1 environmental assessment and which contained a variety of
classes of compounds. These representative samples would be pack-
aged, stored under simulated shipping conditions and returned to the
laboratory where they would await analysis. The simulated shipment
time would span one week and the laboratory storage time an addi-
tional two weeks. Analyses would be performed within one week of
the termination of the storage period.
B. Experimental Design
Using criteria presented in the Level 1 Procedures Manual,
three experiments were designed to study the effects of the follow-
ing parameters, or factors, on sample integrity:
• sample container composition (amber glass or polyethylene),
• head space (air, nitrogen, or a combination),
• shipment and storage temperatures (38°C, 21°C, 5°C),
-------
• storage lighting conditions (dark or diffuse sun), and
• catalytic species content (acid extractable stainless steel
components).
Sets of samples were devised to simulate representative probe
wash (Set I), sorbent trap condensate extract (Set II), and sorbent
resin samples (Set III) which are obtained during a Level 1 source
assessment. Once formulated and packaged, each sample was handled
according to practices normally encountered during transport and
storage cycles.
Some combinations were eliminated as non-applicable, such as
the conversion back to an air head space over samples after prelimi-
nary packaging under nitrogen. Furthermore, a variation in cataly-
tic species content was considered only in the Set II (condensate
extract) experiment. The sorbent resin experiment (Set III) was
conducted with additional limitations in the factor/level combina-
tions tested. However, some of the factor/level combinations were
repeated in order to achieve greater precision in estimating experi-
mental error. The final array of factors and levels chosen for the
three experimental sets is given in Table 1.
As the experimental program progressed it was established that
certain economies in experimentation were possible by employing
statistical principles of experimental design. Therefore, only
twenty-seven unique combinations were actually tested in the con-
densate extract experiment (Set II). However, the set of twenty-
seven input conditions were selected according to the theory of
fractional factorial experimentation (11, 12), so that meaningful
inferences could be made about each factor independently, as well
as certain predetermined two-factor interactions. These concepts
are described in greater detail in Section IV-A.
C. Sample Identification
A coding system was developed that would allow for the immediate
identification of any sample without reference to a master log. The
code was derived from the first letter of each important shipping
or storage condition; an example of which is given below:
Catalytic
Head Space Shipment Storage Species
Container Composition Temperature Condition Content
AG- NN- H- CD- S
+ t t .f .+
Packaged Shipped in Na Shipped under Stored in Contains acid
in an Stored in N£ elevated the cold extractable
amber temperature and in the stainless
glass (hot) dark steel consti-
bottle tuents
A complete list of sample coding abbreviations is presented in Table 1.
-------
FACTORS
TABLE 1
Admissible Experimental Conditions
LEVELS
(1)
Container
(2)
Shipment/Storage
Head Space
Composition
(3)
Shipment
Temperature
(4)
Storage Condition
Temperature/Lighting
(5)
Catalytic Species
Code
AG
P
AA
AN
NN
H
R
C
RD
KL
CD
N
S
Possible Combination^
Probe Wash
Experiment
(Set I)
Amber Glass
Polyethylene
Air/Air
Air /Nitrogen
Nitrogen/Nitrogen
Hot (37°C)
Room (21 °C)
Cold (5°C)
Room/ Dark
Room/Light
Co Id /Dark
Not analyzed
33 x 2 = 54
Condensate Extract
Experiment
(Set II)
Amber Glass
Polyethylene
Air/Air
Air/Nitrogen
Nitrogen/Nitrogen
Hot
Room
Cold
Room/Dark
Room/Light
Co Id/ Dark
None
Stainless Steel Com-
ponents
33 x 22 = 108
Sorbent Resin
Experiment
(Set III)
Amber Glass
(Not analyzed)
Air/Air
(Not analyzed)
Nitrogen/Nitrogen
Hot
Room
Cold
Room/ Dark
(Not analyzed)
Cold/Dark
Not analyzed
3 x 22 = 12
-------
D. Sample Formulation
Model compounds selected for review within this study were
chosen on the basis of the following criteria:
• Representative of a variety of different compound classes.
• Representative of a boiling point range where losses due to
volatility may be possible, but do not pose a major
influence.
• Included in concurrent Level 1 analysis studies.
• Applicability to ready analysis by gas chromatography.
Based on these selection criteria, the following six compounds
were subsequently used:
Phenol
N-Methylaniline
4-Chlorobenzaldehyde
Acenaphthene
n-Hexadecane
4,4'-Dichlorobiphenyl
Specific structural and physical parameters of the chosen species
are detailed in Table 2.
Using information detailed within the "Level 1 Procedures Manual"
(6), three sets of simulated field samples were formulated, namely
probe wash (Set I), sorbent trap condensate extract (Set II), and
sorbent resin (Set III). Each set was compounded such that typically
expected field concentrations of selected model species were contained
within the individual sample. To minimize the variability of compound
concentration within any given sample, master batches of each sample
type were premixed and aliquots removed for packaging.
As compounded, the probe wash and condensate extract series con-
tained between 30 and 50 milligrams of each model compound, diluted
to a final volume of 300 milliliters with carrier solvent. As out-
lined in the Procedures Manual, the probe wash (Set I) solvent was
1:1 methylene chloride: methanol, and the condensate extract (Set II)
solvent was methylene chloride. The sorbent resin samples contained
approximately 3 to 5 milligrams of each model compound on 15 +1 grans
of XAD-2. In this instance a master batch of spiking solution was
made up in pentane and 5 mis pipetted•onto the resin.
During formulation of the condensate extract set, an attempt was
made to simulate conditions encountered during field utilization of
a. SASS train. It was reported that sorbent trap condensate samples
obtained from sources having hydrochloric acid mist and sulfur dioxide
within the effluent stream, contained quantities of hydrochloric and
-------
TABLE 2
Model Compound Chemical and Physical Properties
Species
Phenol
Class
Oxygenated
compound
(phenol)
N-Methylaniline Nitrogen
compound
Structure
Molecular
Weight
94
-NHCH3 107
Boiling
Point °C
182
196
4-Chloroberiz-
aldehyde
Acenaphthene
n-Hexadecane
Oxygenated
ygenated r ^
compound Cl -(' Y>-CHO
(Aldehyde) V^=L/
141
Polynuclear
Aromatic
Hydrocarbon
Straight Chain
Hydrocarbon CH3(CH2)ii+CH3
154
226
214
277
287
4,4'Dichloro-
biphenyl
Halogenated
Compound Cl
( Polychlorinated
biphenyl)
-// \\-Cl
223
315
-------
sulfuric acid and exhibited a characteristic pale green color (13).
As the sorbent trap is fabricated of stainless steel, the color
possibly results from the etching of the metal by the acid and is
indicative of trace metallic content within the condensate liquor.
To approximate this condition, and assess its potential impact upon
storage stability parameters, simulated green and white liquor solu-
tions were formulated and extracted with methylene chloride for
Set II samples. Specific details of this process and all other
sample formulation operations are provided within the appendices of
this report. (Tables A1-A3)
E. Handling Cycle
After formulation, samples were placed within the desired con-
tainer type, with either a nitrogen or air head space, and sealed.
They then were subjected to a three-week handling cycle. During
the first week, or simulated shipment phase, selected samples were
stored within closed boxes and placed in controlled temperature
zones to simulate conditions expected to arise during truck trans-
port from a remote collection point to a central laboratory. Three
such zones were used: one of 38 °.C (100°F), one of 21°C (70°F), and
one of 5°C (40°F). All samples were then unpacked and rerouted for
two additional weeks to a storage location. Samples requiring head
space composition change were opened, purged with nitrogen and then
resealed before being stored. During the storage phase, three
basic combinations existed:
21°C (70°F) and diffuse sunlight (northern window)
21°C (70°F) and dark (closed cabinet)
5°C (40°F) and dark (closed refrigerator)
F. Analytical Procedures
Gas chromatography (GC) was utilized as the major quantitative
analysis tool during this evaluation. Specific experimental details
are presented in the appendices of this report (Table C-l). As the
probe wash and condensate extract samples were already dilute solu-
tions, no sample preparation was required prior to analysis. For
sorbent resin samples, a continuous 24-hour Soxhlet extraction was
required to remove the spiked compounds from the XAD-2 resin. Since
the sample recovery of this step was unknown, several non-stored
spikes were similarly prepared and analyzed along with the stored
samples.
This was accomplished by placing 15 il gm of XAD-2 resin into
an amber glass bottle, and spiking it with 5 mis of a freshly pre-
pared standard solution. The contents of the bottle were then imme-
diately transferred into a Soxhlet thimble and extracted for 24-hours
with roughly 200 milliliters of methylene chloride. The resulting
solution was brought to a final volume of 250 milliliters with
methylene chloride, from which a 5 microliter aliquot was removed
-------
and directly analyzed by GC. Additional details of this procedure
are discussed in Section III of this report.
In all cases, a fresh standard solution was prepared immediately
before the analysis of individual sample sets. This standard was
required for several reasons. First, this sample was useful for
monitoring the instrumental parameters. If integrated peak area
responses were found to change between repetitive injections of the
standard solution, steps could be taken to retune the GC to original
conditions.
Similarly, as it was impractical to analyze all stored samples-
on the same day, the fresh standard provided a mechanism through
which the data from non-contiguous analysis could be compared. At
the beginning of each day, the standard solution was injected
several times to insure that the instrumental response was equiva-
lent to prior repetitions. If not, all parameters were rechecked
and adjusted, prior to proceeding with the analysis of the stored
samples.
Finally, this sample served as a calibtation standard. Formu-
lated to duplicate the composition of the stored samples, when they
were originally packaged and placed in the simulated shipping se-
quence, this standard solution represented the stored samples in an
unaltered state. By directly comparing the observed peak area re-
sponses of the stored sample to the fresh standard, and correcting
for minor initial compounding variations, it becomes possible to
compute a percentage deviation attributable to the imposed handling
procedures.
The metallic content of the green and white liquors and the
methylene chloride extract of the green liquor was determined by
spark emission spectroscopy (ES). Attempts to obtain more accurate
quantitative data for the methylene chloride extract by the induc-
tively coupled plasma optical emission spectroscopy (ICPOES) tech-
nique was frustrated by the lack of compatibility between this tech-
nique and methylene chloride. However, quantitative results were
obtained for both of the aqueous phase solutions (green and white
liquor) by ICPOES. All results of these analyses are given in
Appendix B, Tables B-l and B-2.
G. Data Analysis and Interpretation
The extent of sample degradation is computed as the ratio of the
stored samples' area response to that of the freshly prepared standards,
For example, in the hypothetical situation where a sample initially
containing 100 pg/ml of phenol yielded an integrated area of 500
microvolts-seconds and a standard containing 105 ug/ml phenol exhibi-
an area response of 515 microvolts-seconds, the equation would be
-------
microvolts-seconds
sample
microvolts-seconds
standard
100 yg/ml
A value of 1.00, representing the expected concentration, is
subtracted from this value and the resulting decimal expressed as
a percentage, i.e.,
1.019 - 1.00 = 0.019 - 1.9%
This figure is the percentage deviation of the stored samples' final,
post handling, concentration from the initially compounded amount as
represented by the fresh standard solution. As computed, positive
values imply that an increase in concentration has occurred within
the returned sample, while negative values indicate sample loss.
Values approximating zero infer no change.
As day to day fluctuations are known to influence chromato-
graphic responses for duplicate sample injections, data analysis was
limited to experiments performed on the same day under similar con-
ditions. The average response factors of bracketing standard solu-
tion injections was determined, and individual samples compared
directly with this value. This procedure was useful for normalizing
the effects of instrumentation variability.
Experimental results were analyzed according to a standard sta-
tistical technique known as the Analysis of Variance (ANOVA). The
basic assumption implicit in utilizing this technique is that the
experimental observations (i.e., the deviations of the stored sample
from the standard) are random variables, with well-defined distribu-
tional properties, that can be expressed as an additive linear func-
tion of the parameters that have been controlled in the experiment.
If the appropriate assumptions are satisfied, it is then possible to
test hypotheses regarding the effect of the individual parameters,
or factors, that have been varied over two or more levels. If there
is reason to believe that the necessary assumptions do not apply for
the data, alternative analytical methods could be utilized; e.g.,
transforming the data, using non-parametric test methods, and/or
performing an Analysis of Co-variance in which the standard calibra-
tion value is the co-variate and the observed peak area response the
dependent variate. These techniques are described in detail in most
statistical methods texts (11).
10
-------
III. EXPERIMENTAL RESULTS
Results obtained from this three-set stability review are pre-
sented in Tables 3 through 5. The tabulated values reflect percent-
age variation of the stored samples from the expected sample content,
based upon the initial formulation criteria. As such, values near
zero imply best handling practice implementation, while negative
values infer species loss and positive values imply an increase in
concentration or, more likely, contamination. Deviation on either
side of ideal is viewed as equally bad, as corrective practices
would need to be implemented in either case.
Generally, the variability attributed to the analysis itself
has been found to be small. Table 6 reflects the average peak area
response, standard deviation and coefficient of variation for each
compound within the standard preparation, as determined by repeti-
tive injections throughout the individual analysis sequences. In
addition, the 95 percent confidence levels surrounding the calcula-
ted means are presented.
As mentioned, the samples of Set III (sorbent resin set) con-
sisted of the model compounds supported on XAD-2 solid sorbent resin;
a combination which is unsuitable for direct GC analysis. In order
to analyze these samples, it was necessary to extract them within a
continuous Soxhlet extraction apparatus. The resulting solutions
were then adjusted to the appropriate final volume and directly
analyzed by GC.
Since the extraction process could possibly have interfered with
the fina.l results obtained, six non-stored samples were prepared and
analyzed simultaneously with the stored samples to estimate the effect
of this step. The results obtained are given in Table 7.
Observed average sample recoveries from the extraction process
are found to range from 91.4 to 98.4 p'ercent. However, as several
of the listed data groups exhibit large variability around the com-
puted average, no attempt has been made to adjust the results of
Set III samples to account for the extraction process effects.
Instead, tabulated values include all factors involved in sample
storage, handling, and extraction.
Generally the data show that most compounds are recovered within
±10% of their original concentration for all three sets of conditions.
The results are reviewed and analyzed in detail in Section IV.
Interpretation of the results obtained during Sets II and III
analyses of n-Hexadecane concentration was difficult, due to the in-
ordinately high recoveries which can not be assigned to any single
parameter. Within Set II analyses, these variances, ranging from 10
to 65 percent above the formulated concentrations, were initially
felt to arise from contamination. Observations made during analysis
11
-------
10
TABLE 3'
Analytical Results from Stored Probe Wash Samples (Set I)
(Percentage Deviation from Expected)
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Sample
Identification
© © (D © *
AG-AA-H-RD
-AA-R-RD
-AA-C-RD
AG-AA-H-RL'
-AA-R-RL
-AA-C-RL
AG-AA-H-CD
-AA-R-CD
-AA-C-CD
AG-AN-H-RD
-AN-R-RD
-AN-C-RD
AG-NN-H-RD
-NN-R-RD
-NN-C-RD
AG-AN-H-RL
-AN-R-RL
-AN-C-RL
AG-NN-H-RL
-NN-R-RL
-NN-C-RL
AG-AN-H-CD
-AN-R-CD
-AN-C-CD
AG-NN-H-CD
-NN-R-CD
-NN-C-GD
Phenol
14.3
4.9
0.0
-8.2
-0.5
-1.6
-6.0
-1.9
-2.3
9.6
-15.5
-3.2
-0.4
-2.1
-4.2
-8.2
-7.6
-13.7
0.0
3.2
2.6
-4.1
-17.6
-16.1
2.3
2.0
-26.3
N-Methyl.
.aniline
24.1
0.5
6.4
-0.3
5.7
3.6
5.5
13.7
2.4
4.0
-0.7
1.9
5.2
1.0
0.8
-0.8
-0.5
-3.8
3.1
7.6
8.3
0.7
-8.8
-10.0
-0.3
4.9
7.7
Acenaphthene
3.6
-1.6
0-0
-2.3
3.8
1.2
-1.4
-0.4
-2.1
2.9
-6.0
-2.7
1.8
-1.4
-1.2
-2.3
-1
-3
9
5.4
2.7
5.8
5.8
.2
1.2
0.8
3.1
2.8
3.6
a-Hexadecane
10.2
5.3
5.0
0.4
6.3
3.7
3.8
5.5
2.7
8.0'
-2.2
1.4
6.0
3.4
1.7
-0.4
-2
9
7.7
5.3
8.6
9.8
1.8
5.6
5.6
7.7
8.1
8.3
4,4'-Dichloro-
biphenyl
3.6
-6.3
1.5
-8.5
7.2
3.6
-2.7
1.0
0.6
2.8
-6.7
11.0
0»0
-0.5
-1.5
-8.5
-7
-1
6
8.7
6.7
8.2
9.2
,4
2.1
2.1
2.0
1.3
3.2
* Denotes factor combination observed during experiment (see Table 1)
continued....
-------
Table 3 (continued)
Analytical Results from Stored Probe Wash Samples (Set I)
(Percent Deviation from Expected)
Run
No.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
Sample
Identification
© (D(D ©*
P-AA-H-RD
-AA-R-KD
-AA-C-RD
P-AA-H-RL
-AA-R-RL
-AA-C-RL
P-AA-H-CD
-AA-R-CD
-AA-C-CD
P-AN-H-RD
-AN-R-RD
-AN-C-RD
P-NN-H-RD
-NN-R-RD
-NN-C-RD
P-AN-H-RL
-AN-R-RL
-AN-C-RL
P-NN-H-RL
-NN-R-RL
-NN-C-RL
P-AN-H-CD
-AN-R-CD
-AN-C-CD
P-NN-H-CD
-NN-R-CD
-NN-C-CD
Phenol
2.2
1.8
-17.6
4.7
9.8
-16.2
-14.7
31.6
2.9
7.6 .
17.2
0.4
1.7
15.5
-7.9
12.7
6.4
8.7
12.7
11.7
6. '8
-2.0
1.2
-8.3
-16 . 8
-15.5
-9.1
N-Methyl-
aniline
6.6
4.0
6.1
3.4
-4.5
-10.7
11.4
18.9
7.4
10.5
27.8
6.8
6.4
24.4
5.1
1.1
2.5
-4.5
-2.6
-4.2
-5.6
1.7
16.2
8.0
-8.3
2.0
7.0
Acenaphthene
4.7
2.6
-2.3
5.7
5.1
4.3
0.3
3.2
4.4
6.7
9.5
2.5
4.6
7.7
0.3
4.0
8.3
4.0
6.0
4.0
3.6
-1.3
-4.3
3.6
4.7
-10.2
-0.7
n-Hexadecane
-4.2
-5.0
-7.2
-2.7
0.0
-0.2
-4.3
-1.4
2.5
-1.2
3.5
-2.2
-3.6
-0.7
-2.4
-0.7
3.7
0.7
-2.2
-0.7
1.5
-6.3
-9.2
-0.5
-2.5
-12.5
-2.5
4,4'-Dichloro-
biphenyl
6.1
0.0
-4.6
1.7
-1.6
1.1
-1.0
-0.6
5.6
7.8
11.4
3.9
1.6
3.4
-2.9
-2.0
-15.6
7.5
4.4
-1.6
10.2
-2.0
-5.6
1.1
-1.
-15.
-7.0
Denotes factor combination observed during experiment (see Table 1)
-------
Table 4
Analytical Results from Stored Sorbent Trap Condensate Extract Samples (Set II)
(Percent Deviation from Expected)
Phenol
0.8
7.6
8.4
2.9
3.0
-6.2
-0.8
-3.4
-2.3
-0.3
0.0
5.0
6.0
-6.8
0.0
-0.3
0.0
-6.3
1.9
-3.3
O.8
3.6
2.6
-1.3
-1.0
2.9
-3.1
+ Equivalent sample identification codes indicate separate preparations of duplicate samples.
Presented in randomized order as performed in experiment.
* Denotes factor combination observed during experiment (see Table 1).
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Sample
Identifies ti
9-N?Sc?S
P-AN-C-RL-S
P-AN-C-RL-S
AG-AN-H-RD-N
P-AA-H-CD-S
AG-AA-R-RL-N
AG-NN-R.-RD-S
AG-AA-H-CD-S
AG-NN-H-RL-S
P-AA-C-RD-S
P-AN-R-CD-S
P-NN-H-RL-S
P-NN-H-RL-S
AG-AA-C-RD-S
AG-NN-C-CD-N
P-NN-H-RD- S
P-AN-H-RU-N
AG-AN-R-CD-S
P-NN-R-RD-S
P-NN-C-CD-N
P-AN-R-CD-S
P-AA-C-RD-S
P-AA-H-CD-S
P-AA-R-RL-N
P-AA-R-RL-N
P-AN-H-RD-N
AG-AN-C-RL-S
H-Methyl-
aniline
2.6
-6.0
-16.1
21.6
12.8
7.1
15.3
12.5 '
15.2
8.6
10.2
-10.9
-12.2
6.8'
4.5
12.9
15.2
6.9
13.5
10.6
5.4
8.5
12.5
-2.3
3.1
17.5
9.1
4-Chloro-
benzaldehvde
3.0
7.1
8.9
0.7
9.0
-0.7
.10.1
2,1
3.1
5.2
5.2
13.0
« 2.5
0.0
0.7
6.0
«.8
1.5
6.0
2.3
5.4
1.4
7.6
3.3
-1.5
W.3
5.3
Acenaph-
thene
-0.7
2.3
7.4
2.5
0.7
-4.4
3.6
-1.6
0.9
0.0
0.3
3.4
0.7
-8.2
-8.0 *
0.3
0.3
-6.7
1.3
-4.4
-2.0
-0.7
0.7
-3.3
-4.7
3.8
-4.0
n-Hexa-
decane
0.9
1.9
35.6
0.9
53.7
19.6
25.4
22.8
23.3
42.0
43.9
*5.2
64.2
14.9
.14.2
18.6
18.9
16.6
24.2
14.1
42.8
46.3
20.1
15.2
14.4
21.5
20.3
4,4'-Dichloro-
biphenvl
2.9
3.9
11.6
-3.4
•2.4
4.2
4.2
12.0
-5.0
-5.8
7.4
10.5
-0.5
-1.0
0.0
8.3
13.2
8.7
2.9
4.9
4.8
5.8
11.1
7.3
-0.5
•6.8
-1.9
-------
TABLE 5
Analytical Results from Stored Sorbent Resin Samples (Set III)
(Percent deviation from expected)
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Sample
Identification"*"
© vD ©*
AG-NN-R-RD
AG-AA-C-CD
AG-AA-H-CD
AG-NN-H-RD
AG-NN-C-CD
AG-AA-R-CD
AG-NN-R-CD
AG-NN-H-CD
AG-AA-R-RD
AG-NN-C-RD
AG--AA-H-RD
AG-AA-C-RD
AG-AA-H-CD
AG-NN-R-CD
AG-NN-C-RD
AG-AA-C-CD
AG-AA-R-RD
AG-NN-H-RD
Phenol
0.0
3.1
-0.5
1.1
4.3
-9.9
1.8
1.8
0.0
6.0
-0.6
1.2
0.6
3.5
-4.1
-2.9
3.9
4.9
N-Methyl-
aniline
-3.4
1.3
2.9
-9.0
2.9
-2.1
8.1
1.3
1.3
2.7
5.0
6.8
3.6
-6.0
1.6
6.3
2.8
6.3
4-Chloro
benzaldehyde
-4.0
-2.1
-1.2
-6.1
-4.0
-6.1
-0.7
-3.0
-1.5
-3.0
-0.8
-0.8
-3.1
-0.8
-4.9
-4.2
-8.8
-11.2
Acenaph-
thene
-2.4
-1.0
0.0
-6.9
-3.7
-5.6
-1.2
0.8
0.8
-1.3
0.0
-0.5
-3.2
-1.1
-2.2
-0.9
-0.7
-1.4
n-Hexa-
decane
5.2
24.2
19.7
11.7
15.1
15.3
15.5
19.0
18.6
17.1
18.4
18,8
15.9
18.4
17.0
17.9
19.7
17.5
4,4'-Dichloro'
biphenyl
-1.4
-0.3
1.2
-8.2
-5,1
1.0
0,3
-1.4
-2,1
3.1
-1.1
-2.2
-4.9
-2.6
-3,7
-1.3
-1.0
-2.0
+ Equivalent sample identification codes indicate separate preparation of duplicate samples.
* Denotes factor combination observed during experiment (see Table 1).
-------
Table 6
Statistical Susmary of Standard Calibration Samples
Set I
Probe Wash
Series
50 Observations
Set II
Condensate
Extract
Series
-
12 Observations
Set III
Sorbent Resin
Series
16 Observations
Mean
st'd dev.
coef. of var.
95*
confidence limits
Mean
st'd dev.
coef. of var.
95*
confidence limits
Mean
st'd dev.
coef. of var.
95*
confidence Units
Phenol
274
36.9
13. 5Z
263.5 - 284.5
376
9.4
• * 2.5*
370.1 - 381.9
15.6
1.1
7.1*
15.0 - 16.2
N-Methyl-
anillne
356
27.3
7.7*
348.2 - 363.8
366
8.3
2.3*
*
360.8 - 371.2
18.5
1.4
7.6*
17.8 - 19.2
4-Chloro -
benzeldehyde
not
analyzed
261
9.8
3.8*
1
254.8 - 267.2
16.0
0.9
5.6*
15.5 - 16.5
Acenaphthene
265
9.7
3.7*
262.2 - 267.8
295
10.9
3.7*
288.2 - 301.8
42.0
2.6
6.2*
40.6 - 43.4
n-Hexadecane
504
20.1
4.6*
498.3 -509.7
456
15.1
.3.3*
446.5 - 465.5
21.9
1.7
7.82
21.0 - 22.8
4,4'-Dlcbloro-
biphenyl
200
21.8
10.9*
193.8 - 206.2
236
19.4
8.2*
-
223.8 - 248.2
29.6
1.8
6.1*
28.6 - 30.6
This species-was not present in Set I samples.
-------
TABLE 7
Results of Extraction Step Experiment
(Percentage Deviations from Unspiked Standard)
Standard
Spikes
Run No. 1
2
: 3
4
5
6
Average
95% Confi-
dence Interya]
Limits
Phenol
-5.6
-1.9
-8.6
-14.1 '-
-9.0
-3.0
-7.03
-2.3
to
-11.7
N-Methyl
Aniline
-0.5
+3.4
+3.9
-12.5
-3.7
-5.3
-2.45
-8.9
to
+4.0
4-Chloro-
Benzaldehyde
-8.5
-3.6
-4.2
-7.8
-13.7
-13.7
-8.58
-4.0
to
-13.6
Acen aph thene
-1.4
-2.0
-2.3
-2.1
-6.4
-0.7
-2.48
-0.4
to
-4.6
n-Hexa-
decane
-1.4
-1.3
-1.3
-2.3
-2.3
-0.9
-1.58
-1.0
to
-2.2
4,4'-Dichloro-
biphenyl
-1.7
-4.1
-4.5
-2.0
-1.3
-0.7
-2.38
-0.7
to
-4.0
-------
tended to support this theory, as a shoulder was noted on the chro-
ma togram part way through the hexadecane peak. In addition, where
a relatively smooth baseline had once existed, several smaller but
reproducible peaks were in evidence. To determine the contaminants'
composition, two representative samples of the condensate extract
set were studied by capillary gas chromatography-mass spectrometry
(GC-MS). The chromatograms obtained from these analyses are presented
in Figures 1 and 2.
Referring to Figure 1, it is clear that a homologous series of
olefins is present within the retained sample (P-NN-H-RL-S), and
that the species identified as a CIG olefin is the probable cause of
much of the observed extra hexadecane content (65%). As olefins are
major constituents of both mold release compounds and polyethylene
bottles, it is likely that the contaminant series results from the
extraction of container walls by the methylene chloride.
However, the excessive n-hexadecane content observed in most
of the samples of Set II and Set III cannot be singly attributed to
the extraction of contaminants from the polyethylene sample bottles.
With reference to Figure 2, a chromatogram of a second sample
(P-NN-R-RD-S) from Set II, where a 25 percent excess of n-hexadecane
had been reported, indicates that even though the same homologous
series of contaminants is present, their apparent contribution to the
reported excess is not large (even on a proportional basis). Further-
more, reference to the results of Set I analyses, indicates that not
all samples stored in polyethylene bottles of the same lot and under
similar conditions will become contaminated.
Comparably, review of a typical sample from the serbent resin
set (see Figure 3) shows that no olefin species are contained within
a sample exhibiting a net gain of 24 percent n-hexadecane. As all
the tested resin combinations in this set, and many samples in Set II
as well, were stored in amber glass bottles, the extraction of olefins
from the container walls is not considered likely.
Thus, the origin and nature of all factors influencing the
n-hexadecane concentration within Set II and Set III samples is not
completely clear. While it is obvious that some of the samples have
become contaminated with components extracted from the walls of the
polyethylene sample bottles, this alone cannot account for all of the
encounted hexadecane excesses. Other factors, including possibly in-
accurate sample formulation, must be considered as having influenced
some of the reported results, making the interpretation of Set II and
III ji-hexadecane results impossible.
18
-------
VO
300
350
•"""I
400 450
Figure 1. GC/MS Chromatogram of Set II Sample P-NN-H-RL-S
-------
100
ii
i
•s
*
I
M-l
a)
H
O
to
r-4
O
M
s
el
50
100
200
250
300
350
400
450
Figure 2. GC/MS Chromatogram of Set II Sample P-NN-R-RD-S
-------
00 |
r-
•r
9
iH
£
1
o
0
01
£
50 100
I
f
4) V
X <
1
s
.0
O
l-c
O
6
^
1'
150 200 250 300
i
a
n
a
»1
B
i T"j ij i i i i . i i
350
-------
IV. DISCUSSION OF RESULTS
A. Statistical Analysis of Experimental Data
The changes in composition (percent deviation) of each compound,
given in Tables 3-5, are observed to vary over the set of experi-
mental input conditions considered. The statistical technique ANOVA
provided a mechanism to allocate this variability in the data to
identifiable sources; namely, those conditions which were controlled
in the design of the experiment (i.e., type of container, head space
composition, shipment temperature, and storage condition). Further-
more, the ANOVA technique provides an estimate of the normal or
chance variation in the data that has been an inherent part of
performing the experiment. By comparing the magnitude of the dif-
ference attributable to varying each factor over the levels tested
to the normal variation in the data, those factors having a measur-
able (i.e., statistically significant) effect can be identified.
To illustrate this concept, an ANOVA table is given in Table 8
for the set of 54 phenol observations reported in Table 3. Although
the basic experiment has not been replicated, it is possible to
estimate chance variation (i.e., experimental error) by assuming
all three and four factor interactions are negligible. For the
phenol data, the experimental error variance is estimated to be
82.81: thus the standard error is V82.81 = i 9.1 percent on a per
observation basis.
The ANOVA table indicates those effects and two-factor inter-
actions that appear to contribute significantly to the variability
of the data. To illustrate this concept, the average values of
each level of each factor are given in Table 9. For example, the
•significant contribution attributable to varying types of container
on the phenol data can be explained by the fact that aniber glass, on
the average, yielded a loss of 3.7 percent, while the 27 experiments
using polyethylene yielded a positive deviation of 1.8 percent phenol.
This difference of 5.5 percent is statistically significant, since
the ANOVA table provides the data to calculate a 95 percent confi-
dence interval about this difference:
Estimated "container effect" = (-3.7 - 1.8) ± (2.048)\
- -5.5% ± 5.1%
Therefore, this experiment indicates the likelihood that the use of
polyethylene results in phenol measurements of 0.4 percent to 10.6
percent higher than would be observed using amber glass, a positive
(non-zero) contribution.
22
-------
TABLE 8..
Analysis of Variance
Probe Wash Experiment (Phenol)
Source of Degrees of
Variation Freedom
Main Effects
Container 1
Head Space Composition 2
Shipment Temperature 2
Storage Condition 2
Interactions
Container x Head Space
Composition 2
Container x Shipment
Temperature 2
Container x Storage
Condition 2
Head Space Composition x
Shipment Temperature 4
Head Space Composition x
Storage Condition 4
Shipment Temperature x
Storage Condition 4 i
Error 28
Total 53
Sum of
Squares
406.3
38.6
675.2
574.3
426.4
363.5
159.6
327.9
689.1
175.7
2318.8
6155.5
Mean
Square
406.3
19.3
337.6
287.1
213.2
181.7
79.8
82.0
172.3
43.9
82.8
F
4.91*
< 1
4.08*
*
3.47
2.57
2.19
< 1
< 1
2.08
< 1
Standard Error per sample = \82.81 = ' + 9.
* Indicates significance at 5% test level
23
-------
TABLE 9
Probe Wash Experiment (Set I)
(Average Percentage Deviation from Expected)
Factor
Container
Head Space
Composition
Shipment
Temperature
Storage Con-
dition (Temper-t.
ature/Light)
Level
Amber Glass
Polyethylene
Air /Air
Air/Nitrogen
Nitrogen/Nitro-
gen
Hot
Boom
Cold
Room/ Dark
Room/light
Cold/ Dark
Average
Standard Error Per Sample
No. of Ob-
servations
27
27
18
18
18
18
18
18
18
18
18
54
Significant Interactions:
Phenol
-3.7 j
f *
1.8 \
0.2
-1.8
-1.3
0.4 x
- 2.5>*
-5.8^
1.3,
1.3 >*
-5.6'
-1.0
9.1
None
N-Methyl
aniline
3.0
5.1
5.8
2.9
3.5
4.0
6.1
2.0
7.8i
-O.lj>**
4.4'
4.1
6.3
Cont.x Hd.)
Sp. Comp. -7
Cont. x K
5to.r.Cond.J
Acenaph-
thene
°-7},
3.0 )
1.6
1.5
2.4
2.2
1.6
1.7
1.8x
3.5 > *
0.2^
1.8
3.4
None
n-41exa-
decane
4-7L*
-2.2\
1.1
0.7
1.8
0.8
0.9
2.0
0.9
2.1
0.7
1.2
3.5
None
4,4'-
Dichloro-
biphenyl
1.2
0.1
0.4
0.5
1.1
0.5\
-1.5 i*
3.0'
1.7
1.3
-1.0
0.6
4.9
Hd. Sp.
Comp.x :i
Stor.
Cond.
to
* Indicates significance at 5% test level.
** Indicates significance at 1% test level.
-------
Similarly, all other significant differences are identified in
Table 9. The Set I experiment was structured in such a way that
it was possible to investigate not only the influence of each factor
independently, but also the interactive or combined effect of all
factor pairs. Therefore, significant two-factor interactions are
also given in Table 9. The nature and interpretation of the two-
factor interactions observed to be significant are demonstrated
by the two-way tables and graphs given in Table 10 and Figure 4.
For example, it is observed that n-methylaniline readings were ap-
proximately 2 percentage points higher with glass than with poly-
ethylene when tested with air/air or nitrogen/nitrogen head space
compositions. However, this situation was very different when air/
nitrogen was used, in that glass yielded nearly 10 percent lower
readings than polyethylene.
The condensate extract (Set II) experimental design was dif-
ferent from Set I in that all possible factor/level combinations
were not tested. It had been possible as a result of analzying the
data from Set I to eliminate some of the small effect combinations,
The plan used, known as a Fractional Factorial design, required only
twenty-seven unique test conditions in order to estimate each factor
independently. Furthermore, only the two-factor interactions involv-
ing the container effect were estimable. This analysis required the
assumption that all other two-factor and higher order interactions
had a negligible effect on the data.
An illustrative example of the components of variation are given
for phenol data in Table 11. It is observed that only the container
effect is statistically significant; i.e., an increase between 1.4
and 7.4 percent can be expected in phenol measurements when poly-
ethylene is used instead of amber glass.
Table 12 summarizes the results of all ANOVA's computed for Set
II. Significant interactions are illustrated in tabular (Table 13)
and graphic form (Figure 5) as described for Set I data. Similar
analytical results are given for the Sorbent Resin (Set III) experi-
ment in Tables 14 and 15.
B. Interpretation of Results
The overall purpose of this set of experiments has been to detect
measureable differences in the recoveries of select model compounds
that are directly attributable to the variation of specific shipment
and storage procedures. Subsequent to detection, the extent of dif-
ference found would be directly compared with the accuracy require-
ments of a Level 1 survey analysis to determine if, in fact, the
measured variability could interfere with the desired goals of such
a survey.
As such, the laboratory experiments were designed and conducted
in such a way that differences were likely to be detected if, indeed,
25
-------
TABLE 10
Probe Wash Experiment (Set I)
Significant Two-factor Interactions
COMPOUND
N-Methylaniline:
Container
Glass
Poly
Head Space Composition
AA
AN
NN
6.8
4.7
-2.0
7.8
4.3
2.7
Container
Glass
Poly
Storage Condition
RL
CD
4.8
10.9
2.5
-2.8
1.8
7.1
4,4*-Dichlorobiphenyl:
Storage Condition
Head Space
Composition
AA
AN
NN
RD
RL
CD
0-0
5.0
0 0
0.6
-2.9
6.2
0.5
-0.6
-3.0
* Indicates significance at 5% test level.
26
-------
Compound: N-Methylaniline
Container X Head Space Composition
12
.1 8
-------
TABLE 11
Analysis of Variance
Condensate Extract Experiment (Phenol)
Source of Degrees of
Variation Freedom
Main Effects
Container
Head Space,
Composition
Shipment
Temperature
Storage
Conditions
Catalytic
Species
Interactions
Container x Head
Space Composition
Container x Shipment
Temperature
Container x Storage
Condition
Container x Cata-
lytic Species
Error
Total
1
2
2
2
1
2
2
2
1
11
26
Sum of
Squares
113.8
31.0
29.6
5.9
4.7
34.0
12.9
12.6
38.3
122.1
405.0
Mean
Square
113.8
15.5
14.8
2.9
4.7
17,0
6.4
6.3
38.3
11.1
F
**
10.16
1.40
1,33
<1
<1
1.53
<1
<1
3.45
Standard Error per sample = -y/11.10 = 3.33%
** Indicates significance at the 1% test level.
28
-------
TABLE 12
Condensate Extract Experiment (Set II)
(Average Percentage Deviation from Expected)
Factor
Container
Head Space
Composition
Shipment
Temperature
Storage Condi-
tion (Tempera-
ture/Light)
Catalytic
Species
Level
Amber Glass
Polyethylene
Air/Air
Air/Nitrogen
Nitrogen/Nitro-
gen
Hot
Room
Cold
Room/Dark
Room/ Light
Cold /Dark
Stainless Steel
None
Average
Standard Error Per Sample
No. of
Observa-
vations
9
18
9
9
9
9
9
9
9
9
9
18
9
27
Significant Interaction
Phenol
-2.9)
1.5J
-1.1
1.5
-0.3
0.7
-1.5
0.8
0.3
0.3
-0.6
0.3
-0.6
0.01
3.33
None
N-Methyl-
aniline
11.0)
4.8 j *
7.7
7.1
5.7
9.4 )
8.0} *
3.2)
13.3 j
- -1.4 /**
8.7)
5.8
8.9
6.8
3.78
Cont. x|
Stor. ?**
Cond. /
4-Chloro-
benzalde-
hyde
2.5\ A
5.9 j
2.9
6.1
5.2
6.6
3.9
3.8
5.6
4.6
4.1
5.5
3.2
4.8
3.62
None
Acenaph-
thene
-2.9)
0.3)
-2.4
0.4
-0.3
1.3)
-1.7 >*
-1.8J
0.3
-0.2
-2.4
-0.1
-2.1
-0.8
2.49
Cont. x)
Ship. /*
Temp . /
n.-Hexa-
decane
17.6
30.2
27.7
22.5
27.8
32.3
24.5
21.1
23.6
28.9
25.5
32. 3^
13. 3/
26.0
16.9
None
4,4'-
Dichloro-
biphenyl
2.0
5.4
3.9
5.7
3.1
5.2
5.3
2.3
3.4
3.3
6.0
4.4
3.9
4.3
5.4
None
VO
* Indicates significance at 5% test level.
** Indicates significance at 1% test level.
-------
TABLE 13
Condensate Extract Experiment (Set II)
Significant Two-factor Interactions"*"
COMPOUND
N-Methylaniline:
Storage Condition
Container
Glass
Poly
RD
RL
CD
14 = 6
12.7
10.5
-7.4
8.0
9.0
**
Acenaphthene:
Shipment Temperature
Container
Glass
Poly
Hot
Room Cold
0.6
1.6
-2.5
-1.3
-6.7
0.6
Only those interactions involving the factor "Container" are
estimable from this design.
^Indicates significance at 5% test level.
Indicates significance at 1% test level.
30
-------
o
CD
Q
Compound: N—Methylaniline
Container X Storage Condition
15
10
-10
RD RL
O - Amber Glass
A— Polyethylene
CD
Compound: Acenaphthene
Container X Shipment Temperature
c
tg
'
-
<0 9
'> *•
tt)
O
-6
Hot Room Cold
O— Amber Glass
^— Polyethylene
FIGURE 5. CONDENSATE EXTRACT EQUIPMENT (SET II)
SIGNIFICANT TWO-FACTOR INTERACTIONS
31
-------
TABLE 14
Analysis of Variance
Sorbent Resin Experiment (Phenol)
Source of Degrees of Sum of Mean
Variation Freedom Squares Square F
Main Effects
Head Space 1 33.1 33.1 1.64 (non-sig.)
Shipment Temperature 2 7.4 3.7 <1
Storage Condition 1 6.2 6.2 <1
Interactions
Head Space Composition x
Shipment Temperature 2 3.5 1.8 <1
Head Space Composition x
Storage Condition 1 17.9 17.9 <1
Shipment Temperature
Storage Condition 2 8.2 4.1 <1
Error 8 161.4 20.2
Total 17 237.7
i^^^^^^^^
Standard Error per sample = \20.2 = +4.5%
32
-------
TABLE 15.
Sorbent Resin Experiment (Set III)
(Average Percentage Deviation from Expected)
Factor
Bead Space
Composition
Shipment
Temperature
Storage Condi-
tion
(Temp/Light)
Level
Air /Air
Nitrogen/Nitro-
gen
Hot
Room
Cold
Room/Dark
Cold /Dark
Average
Standard Error Per Sample
No. of
Observa-
tions
9
9
6
6
6
9
9
18
Significant Interactions
Phenol
-0.6
2.1
1.2
-0.1
1.3
1.4
0.2
0.8
4.5
None
N-Methyl-
aniline
3.1
0.5
1.7
0.1
3.6
1.6
2.0
1.8
5.6
None
4-Ctjloro-
benzalde-
hyde
-3.2
-4.2
-4.2
-3.6
-3.2
-4.6
-2.8
-3.7
2.1
Hd. Sp.
Comp . x ^
Ship.
, Temp. ,
Acenaph-
thene
-1.2
-2.2
-1.8
-1.7
-1.6
-1.6
-1.8
-1.7
2.2
None
n-Hexa-
decane
18.7
15.2
17.0
15.4
18.4
16.0
17.9
16.9
4.3
None
4,4'-
Dichloro-
biphenyl
-1.2
-2.3
-2.7
-1.0
-1.6
-2.1
-1.5
-1.8
3.3
None
oo
* Indicates significance at the 5% test level.
-------
Compound: 4—Chlorobenzaldehyde
Shipment Temperature
Hot Room Cold
Head Space
Composition
Air
Nitrogen
-1.7
-6.8
-5.5
-1.8
-2.4
-4.0
Head Space Composition X Shipment Temperature
o
+3 o
.s
Q
35
^
-6
-8
Hot
Room
Cold
O -Air/Air
A -Nitrogen/Nitrogen
Indicates significance at the 5% test level.
FIGURE 6. SORBENT RESIN EXPERIMENT (SET III)
SIGNIFICANT TWO-FACTOR INTERACTION
34
-------
they did exist (that is, the design was "powerful" in the statistical
sense). Furthermore, it should be recognized that these experiments
were intended to yield estimates of the magnitude of the true, but
unknown, average effect caused by varying a parameter over two or
more levels, and that these estimates were to be measured with
reasonable analytical precision. Thus, as an example, it was found
in Set I experiments that a sample stored in polyethylene can be
expected to yield a larger phenol concentration than comparable
samples stored within amber glass containers. Furthermore, the con-
centration of phenol in polyethylene can be expected to be from 0.4
to 10.6 percent higher on the average.
Although this suggests that such a difference is "real" in the
sense that repetitions of this experiment would yield similar find-
ings, the added contribution of even 10 percent is considered to be
well within normally accepted analytical limits for Level 1 analyses.
With this in mind, it becomes clear that none of the tested com-
binations produce differences large enough to impact upon the results
of a Level 1 survey analysis. Thus, any reasonable combination of
shipment temperature, storage condition, container, head space com-
position and catalytic species content can be used for a short
period of time. Furthermore, inasmuch as a ±10 percent range is
normally acceptable for quantitative chemical analysis of samples,
it appears that most of these same conditions can be directly ap-
plied to quantitative analysis routines. However, the tabulated
data also suggest that some caution should be exercised before these
conditions are unilaterally applied.
One occasion implying caution, occurs during the analysis of the
n-Hexadecane results in Set II experiments. As was previously men-
tioned, a homologous series of olefins apparently was extracted
from several, but not all, of the polyethylene containers used
during this experimental sequence. The Cig member of this series
contributed directly to the observed excessive n-hexadecane concen-
trations, as it was integrated as part of the model compound GC
response, causing an average increase of 26 percent to be found
within the stored samples. However, the distribution of the olefin
series was not uniform throughout all similarly handled samples
implying that some of this observed variability was produced by other
factors which can not be identified by the experimental conditions
used. As the origin of all factors contributing to the excessive
hexadecane content are unclear, judgment of the handling condition
effect upon species similar to n-hexadecane must be reserved.
What is important from this experiment, however, is the fact
that it emphasizes that polyethylene containers should not be used
for the storage of samples in organic solvents. While other results
of this study imply that plastic containers do not affect the model
compounds recoveries (due to loss or concentration), this sequence
of experiments shows that undesirable contaminants can arise during
confinement within polyethylene and interfere with eventual analysis.
35
-------
As there is no way to insure that spurious contaminants will not be
extracted from the polyethylene, only those materials unlikely to
react with the container should be retained in them.
A review of N-Methylamiline data from both Set I and Set II also
is of interest. As listed, only those samples stored in diffuse sun-
light and at room temperature exhibit a net decrease in the N-Methyl-
aniline content. From Tables 9 and 12 it is noted that decreases of
0.1 and 1.4 percent are encountered while other tested storage condi-
tions (room or cold temperature and in the dark) produce increases
ranging from 4.0 to 14.0 percent of the intially formulated concen-
trations. The noted decreases are even more pronounced when a review
of the storage condition x container composition interaction is made.
As seen from Tables 10 and 13, those samples stored in polyethylene
and exposed to diffuse sunlight lose between 2.5 and 7.5 percent
of the initially compounded N-Methylaniline content, while all other
tested conditions (including both polyethylene and amber glass con-
tainers stored in the dark, and amber glass stored in diffuse sun-
light) show increased levels ranging from approximately 1.5 to 15.0
percent.
As N-Mathylaniline is known to be photoreactive, it is presumed
that the noted species loss is attributable to some amount of photo-
oxidation. While the amber glass bottle appears to have adequately
screened out the reactive ultraviolet light, the polyethylene did
not. This is expected because polyethylene is transparent in the
ultraviolet region and will allow some light to penetrate and degrade
the aniline. The amber glass, on the other hand, absorbs the ultra-
violet light prohibiting photodegradation of the amine. Once again
this implies that the utilization of polyethylene should be restric-
ted to those instances where it is known that stored samples will
not degrade solid samples.
36
-------
V. CONCLUSIONS AND RECOMMENDATIONS
Excluding the storage of certain materials within polyethylene
containers or in the presence of diffuse sunlight, any reasonable
combination of handling procedures appears to be suitable for the
short term storage of Level 1 survey analysis samples. As Level 1
procedures only require the identification of sample constituents
within a range of ±2 or 3 times of what actually exists, the rela-
tively small perturbations encountered during this experimental
review will not detract from the value of the accumulated data.
In fact, based on these results, it appears that several of the
tested conditions, including shipment temperatures (below 38°C) ,
storage conditions (storage at either room or reduced temperature
and in the dark) head space compositions (air or nitrogen) and
amber glass containers, are appropriate for the storage of quanti-
tative samples where variation must be held to ilO percent.
The utilization of polyethylene sample bottles for storage of
organic materials is not appropriate, as it has been demonstrated
during Set II experiments that spurious contaminants can be extrac-
ted even from rigorously precleaned containers. Furthermore, poly-
ethylene is not suitable for the storage of photo-reactive sub-
stances. As noted from Sets I and II, N -Methylaniline results,
samples stored for short periods of time in sunlight (even in-
direct) can be lost through photo-degradation. "While amber glass
bottles absorb the harmful ultraviolet wavelengths, polyethylene,
which is transparent, allows the light to penetrate. Therefore, it
is recommended that these containers only be used for the storage
of samples that are known not to degrade.
Furthermore, it is recommended that attempts be made to minimize
the contact of organics species with catalytic agents. Although the
presence of acid extractable stainless steel constituents has not
been proven to be significant during this analysis , the effect of
these and other catalytic species has not been fully explored.
37
-------
REFERENCES
1. Bruce, H. E. and Cram, S. P. "Sampling of Marine Organisms and Sedi-
ments for High Precision Gas Chromatographic Analysis of Aromatic
Hydrocarbons" in Marine Pollution Monitoring, Procedings of Workshop
held at NBS, Gaithersburg, MD, NBS Special Publication 409, pp. 181-
182, 1974.
2. Federal Working Group on Pest Management," Guidelines on Sampling and
Statistical Methodologies for Ambient Pest Monitoring" Washington,
D.C. (1974).
3. Lawrence, W. H., Mitchell, J. L., Guess, W. L. and Avtian, J.,
"Toxicity of Plastics Used in Medical Practice" Journal of Pharma-
ceutical Sciences, 52_, 958-963, 1963.
4. Goerlitz, D. F. and Brewn, E. "Methods for the Analysis of Organic
Substances in Water," Techniques of Water Resources Investigations,
USGS Cat. #19:15/5, BK 5, Chapter A3 (1972).
5. Giam, C. S. amd Chan, H. S., "Control of Blanks in the Analysis of
Phthalates in Air and Ocean Biota Samples," Accuracy in Trace Anal-
ysis: Sampling, Sample Handling, Analysis, NBS Special Publication
No. 422, Vol. 2, pp. 701-708, 1976,
6. Hamersma, J. W., Reynolds, S. L., Maddalone, R. F., IERL-RTP Proced-
ures Manual: Level 1 Environmental Assessment, EPA Publication,
EPA-600/2-76-160a, 1976.
7. Hertz, H. S., et. al., "Methods for Trace Organic Analysis in Sedi-
ments and Marine Organisms," Marine Pollution Monitoring, NBS Special
Publication 409, pp. 197-199, 1974.
8. Hamilton, P. B. and Myoda, T. T., "Contamination of Distilled Water,
HCl, and NH.OH with Amino Acids, Proteins and Bacteria," Clinical
Chemistry 20, pp. 687-691, (1974).
9. Straughan, D.,"Field Sampling Methods and Techniques for Marine
Organisms and Sediments." Marine Pollution Monitoring, NBS Special
Publication 409, pp. 183-187, 1974.
10. Bristol, D. ''Effects of Storage Conditions on Residues of 2,4-D and
2,4-DCP on Potatoes," Accnracy in Trace Analysis: Sampling, Sample
Handling, Analysis, NBS Special Publication, Vol. 2, pp. 737-746, 1976.
11. Snedecon, G.W., and Cochran, W.G., Statistical Methods,
Iowa State University Press, Ames, Iowa, 1967.
38
-------
References (continued)
12. Addelman, S., and Kempthorne, 0., Orthogenal Main Effect Plan,
Aeronautical Research Laboratory Report No. 79, Nov. 1961.
13. Private Communication with Dr. Robert Statnick of IERL
Laboratory, Research Triangle Park, North Carolina.
39
-------
APPENDICES
40
-------
APPENDIX A - Formulation Procedures for Samples
TABLE Al - Formulation Procedure for Probe Wash
Samples (Set I)
TABLE A2 - Formulation Procedure for Sorbent Trap
Condensate Extract Samples (Set II)
TABLE A3 - Formulation Procedure for Sorbent Resin
Samples (Set III)
41
-------
TABLE Al
Formulation Procedure for Probe Wash Samples (Set I)
Amber Glass Samples - 27 samples of 300 mis each
9 liters total volume of a 1:1 mix of methylene chloride:methanol
cone Mg/ml
Phenol 168
N-Methylaniline 167
Acenaph th ene 9 5
n-Hexadecane 166
4-4'-Dichlorobiphenyl 103
Polyethylene Sample - 27 samples of 300 mis each.
9 liters total volume of a 1:1 mix of methylene chloride:methanol
cone yg/ml
Phenol 155
N-Methylaniline 162
Ac enaph then e 10 3
n-Hexadecane 164
4,4'-Mchlorobiphenyl 94
42
-------
TABLE A2
Formulation Procedure for Sorbent Trap
Condensate Extract Samples (Set II)
White liquor extract - unspiked - 9 samples of 300 mis each.
White liquot - 3 liters of an 0.05 M HC1 and 0.005 M H2SOit
in distilled water prepared (metallic content analyzed by
ICPOES) and then extracted with an equal volume of tnethylene
chloride. Methylene chloride extract retained as white
liquor extract. To this, the following concentrations of
model compounds was added.
Phenol
N-Methylaniline
4-Chlorobenzaldehyde
Acenaphthene
n-Hexadecane
4,4'-Dichlorobiphenyl
Green liquor extract - spiked - 18 samples of 300 mis each.
Green liquor - A piece of 316 stainless steel placed in 100 mis
of 1.0 M HC1 and stirred for 3 hours, producing green acidic
solution.
A 6-liter volume of 0.05 M HC1, 0.005 M H2S04 and distilled water
was prepared using the concentrate as a base (metallic content
analyzed by ICPOES). The total volume was then extracted with
an equal amount of methylene chloride. Methylene chloride (green
liquor extract) retained and metallic content analyzed. (See Ap-
pendix B). Aqueous phase discarded.
ug/ml
Phenol 165
N-Methylaniline. 169
4-Chlorobenzaldehyde 167
Acenaphthene 101
n-Hexadecane 168
4,4'-Dichlorobiphenyl 101
43
-------
TABLE A3
Formulation Procedure for Sorbent Resin Samples (Set III)
Amber Glass bottles - 24 samples
Stock solution - seven model compounds added to 200 mis pentane
to produce following concentrations:
Phenol
N-Methylaniline
4,-Chlorobenzaldehyde
Acenaph thene
n-Hexadecane
4,4'-Dichlorobiphenyl
Five mis of stock solution pipetted onto 15^ 1 gm of XAD-2
resin contained in amber glass bottles. After handling period,
samples were removed and extracted (Soxhlet) with methylene
chloride for 24 hours. Extract volume brought to 250 mis and
analyzed.
44
-------
APPENDIX B - Analytical Results of Metallic Content
Contained Within Green Liquor, White Liquor
and Green Liquor Extract
TABLE Bl - Metallic Content of Green and White Liquors
As Determined by Inductively Coupled Plasma
Optical Emission Spectrometry (ICPOES)
TABLE B2 - Metallic Content of White Liquor, Green
Liquor and Green Liquor Extract by Emission
Spectrometry As Determined by Spark Emission
Spectrometry
45
-------
TABLE Bl
METALLIC CONTENT OF GREEN AND WHITE
LIQUORS BY ICPOES
(Values in yg/ml)
"^^vQompound
Element — ..
Li
Be
B
Na
Mg
Al
Si
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
Sr
Mo
Ag
Cd
Sn
Ba
Pb
White
Liquor
ND
ND
ND
ND
ND
ND
0.01
ND
ND
ND
ND
ND
0.004
ND
0.13
ND
0.08
ND
ND
0.02
ND
ND
ND
ND
ND
Green
Liquor
ND
ND
ND
0.9
0.01
0.11
0.5
ND
0.15
0.003
0.009
2.3
0.22
8.3
0.005
1.6
0.14
0.13
0.008
0.31
ND
ND
ND
0.003
ND
Detection
Limits
(in H20)
0.3
0.001
0.05
0.3
0.006
0.05
0.01
1
0.001
0.002
0.003
0.005
0.001
0.003
0.005
0.01
0.01
0.008
0.002
0.01
0.01
0.001
0.01
0.001
0.01
ND = not detected
46
-------
TABLE B2
METALLIC CONTENT OF WHITE LIQUOR, GREEN LIQUOR
AND GREEN LIQUOR EXTRACT BY
EMISSION SPECTROMETRY
(Values in yg/ml)
\Cpmpound
Element ^\.
Li
Be
B
Na
Mg
Al
Si
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
Sr
Mo
Ag
Cd
Sn
Ba
Pb
White
Liquor
ND
ND
ND
0.5
0.005
ND
0.01
ND
0.01
ND
ND
ND
ND
0.01
ND
ND
0.05
ND
ND
ND
0.001
ND
ND
ND
ND
Green
Liquor
ND
ND
ND
3
0.01
0.05
0.5
ND
0.01
ND
ND
4
0.8
10
ND
3
0.10
ND
0.1
0.1
0.005
ND
ND
ND
ND
Green
Liquor
Extract
ND
ND
ND
0.05
0.001
ND
0.005
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.001
ND
ND
ND
ND
Detection
Limits
0.1
0.001
0.005
0.05
0.001
0.01
0.01
10
0.005
0.01
0.01
0.01
0.005
0.01
0.05
0.05
0.001
0.05
0.01
0.01
0.001
0.01
0.05
0.1
0.01
ND - Not detected
47
-------
APPENDIX C - Gas Chromatography Conditions Employed
During Analysis
48
-------
TABLE Cl
Gas chromatographic conditions:
Instrument: Varian Assoc. Model 2700 with FID
Column: 6' x 1/8" SS containing 10% 0V - 101 on 100/120 Supelcoport
Conditions:
He 30 ml/min
H_ ^30 ml/min flows adjusted slightly to optimize performance
air ^300 ml/min
injection port 260°-270°C
detector 300°-310°C
Program: 50° (5 min) 10°/min 215° (4 min)
Calculations: Spectra-Physics System I Computing Integrator
49
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-78-017
2.
3. RECIPIENT'S ACCESSION- NO.
4. TITLE AND SUBTITLE
Effect of Handling Procedures on Sample Quality
5. REPORT DATE
February 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J.W. Adams, T.E. Doerfler, andC.H. Summers
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02140
10. PROGRAM ELEMENT NO.
EHB529
11. CONTRACT/GRANT NO.
68-02-2150, T.D. 10501
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final: 12/76-11/77
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES IERL-RTP project officer is Larry D. Johnson, Mail Drop 62,
919/541-2557.
16. ABSTRACT
JThe report gives results of an evaluation of the effects of typical shipping
and storageTiandling procedures on organic materials collected in Level 1 environ-
mental assessment (EA) studiesJ Parameters reviewed included: sample container
composition (amber glass and high-density linear polyethylene), head space compo-
sition (air or nitrogen), temperature (38 C, 21 C, and 5 C), lighting (dark and diffuse
sunlight), and catalytic species content. Three sample sets, representing fractions
obtained during a Level 1 EA and containing six model organic compounds, were used.
A simulated 3-week shipping and storage cycle represented elapsed time between
sample collection and analysis. All three experiments were in accordance with
statistical principles appropriate for conducting factorial experiments. Experimental
results were analyzed using analysis of variance, to assess the relative effect of each
shipping/storage condition studied.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Pollution
Organic Compounds
Sampling
Quality
Materials Handling
argo Transportation
Storage
Shipping Contai-
ners
Containers
Analysis of
Variance
Pollution Control
Stationary Sources
Environmental Assess-
ment
13B
07C
14B 13D
15E,13H
12A
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
58
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
50
------- |