EPA R2-73-227
MAY 1973 Environmental Protection Technology Series
Pyrographic Gross Characterization
of Water Contaminants
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington. D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five 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 five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
U. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, eguipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-R2-73-227
Kay 1973
PYROGRAPHIC GROSS CHARACTERIZATION
OF WATER CONTAMINANTS
by
Ihor Lysyj and P. R. Newton
Contract No. 14-12-802
Project No. 16040 EXD
Project Officer
H. Page Nicholson
Southeast Environmental Research Laboratory
Athens, Georgia 30601
prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
Washington, D.C. 20460
For sale by the Superintendent or Documents, U.S. Government Printing Office, Washington, D.C. 20402
Price $1.26 domestic postpaid or $1 QPO Bookstore
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the Environ-
mental Protection Agency, nor does mention of trade-
names or commercial products constitute endorsement
or recommendation for use•
-------
ABSTRACT
A method was developed for direct analysis of organic materials in
aqueous solutions based on thermal fragmentation followed by gas-
chromatographic separation and detection of the resulting derivative
compositions. Such thermal fragmentation processes were found to be
quantitatively linear, i.e., response is proportional to concentra-
tion, and independent for each organic compound in a mixture. In
addition, the pyrographic pattern produced by a mixture of organic
compounds is a simple arithmetic summation of contributing patterns
of each compound present.
A recorded pattern of pyrolytically produced fragments for a given
water sample reflects the total nature of its organic composition,
and can be interpreted and differentiated in a number of ways. Using
a priori established calibration patterns for individual components
to be found in a mixture, the pattern produced by a mixture can be
analyzed mathematically. The system can be calibrated in terms of
pure organic compounds, classes of organic materials, or any other
arbitrarily defined organic mixtures such as those found in industrial
waste effluents.
The practical hydrochemical methodology, which was developed as a
result of this effort, comprises analytical and data processing hard-
ware, mathematical logic, and computer procedures and programs for
data analysis and interpretation. The instrumentation was designed
and assembled in a unitized package, and required only electrical and
telephone lines for its operation. The Mark 1 prototype instrument
was tested in a trailer under field conditions of operation for 2 years,
An automated Mark 2 instrument was fabricated.
The practical utilization of this new hydrochemical tool was studied
in three areas: water pollution surveillance, waste treatment proc-
esses, and characterization of natural unpolluted water.
This report was submitted in fulfillment of Contract No. 14-12-802
under the sponsorship of the Environmental Protection Agency.
111
-------
CONTENTS
Abstract iii
Section I: Conclusions 1
Section II: Recommendations 3
Recommendation 1 3
Recommendation 2 5
Recommendation 3 6
Section III: Introduction 9
Section IV: Pyrographic Methodology 11
Principles of Pyrography 11
Instrumentation 16
Computer Procedures 26
Experimental Conditions 27
Section V: Application Research 37
The Concept of Multicomponent Pattern Recognition
and Differentiation 37
Section VI: Acknowledgements 55
Section VII: References 57
Section VIII: Publications 59
Section IX: Appendix 61
Computer Programs 61
Computations 61
DAY Program 61
CALIF Program 63
BAJA Program 65
BAJA-1 Program 73
BAJA-2 Program '. 80
Example of BAJA Computations 90
-------
ILLUSTRATIONS
1. Pyrogram of Oleic Acid 12
2. Mark 1 Pyrographic Analyzer 17
3. Cross-Sectional View of the Pyrolytic Furnace Unit . . 19
4. Mark 2 Pyrographic Analyzer Schematic 22
5. Rocketdyne Mark 2 Pyrographic Analyzer 23
6. Peak Intensity Versus Temperature 28
A-l. Flow Diagram of CALIF 64
A-2. Flow Diagram of BAJA 68
A-3. Flow Diagram of BAJA-1 75
A-4. Flow Diagram of BAJA-2 81
VI
-------
TABLES
1. Multiple Analyses of Mixtures of Glycerol, DL-Valine, and
Hexanoic Acid in Aqueous Solutions 14
2. CALIF for Albumin, Normalized Conditions 31
3. CALIF for Glucose, Normalized Conditions . 32
4. CALIF for Oleic Acid, Normalized Conditions 33
5. Comparison of Pyrograms for Glucose, Albumin, and Oleic
Acid 30
6. Analysis of Mixed Solutions 34
7. Typical Printout of BAJA Program 35
8. Pyrographic Fingerprints of Waste Effluents 39
9. Qualitative Analysis of Pyrograms for Various Industrial
Waste Effluents 41
10. Analysis of a Three-Component Mixture of Industrial Wastes . 43
11. Analysis of Mixed Solution of Industrial Waste Effluents . 45
12. Analysis of Mixed Industrial Wastes in Natural Water Matrix. 47
13. Reduction of Organic Load 49
14. Degradation of Organics in Poultry Waste 50
15. Organic Composition in Natural Waters 53
VII
-------
SECTION I
CONCLUSIONS
A 26-month field study has been conducted in Athens, Georgia, to (1)
determine the theoretical and practical feasibility of using the pyro-
graphic method of analysis as a tool in water pollution surveillance
and monitoring activities, and (2) evaluate performance of the Mark 1
pyrographic analyzer. As a result of this study, the principles of
quantitative pyrography were established and verified experimentally.
The application of the theory of multicomponent pattern recognition
and differentiation to the characterization of waste effluents was
demonstrated. Analytical hardware, mathematical logic, and computer
procedures, which are necessary for implementation of the concept,
were developed. All the system elements were integrated in the Mark 1
pyrographic analyzer and made operational. The advanced, automated
Mark 2 instrument was designed and fabricated.
The ability of this technique to perform direct analysis of mixed or-
ganic compounds at ppm levels in aqueous solutions and to determine
the identity and quantity of organic materials present was experi-
mentally demonstrated. A method was developed for gross definition
of industrial wastes. Using this method, the possibility of identi-
fication of sources of pollution, and the quantitative determination
of contributions of wastes from different sources to the total pollu-
tion load of a given water body, was demonstrated in laboratory exper-
iments. This method is based on the theory of multicomponent pattern
recognition and differentiation, and takes advantage of the premise
that each type of industrial waste has an overall chemical composition
characteristic to its operation. The overall characteristics of such
waste compositions expressed as pyrographic fingerprints permit iden-
tification of waste sources, and determination of their quantities in
a solution containing a number of various wastes.
Organic composition of various sources of unpolluted natural waters in
the southeast was characterized and a scheme for natural water classi-
fication was suggested. The possibility of using a pyrographic method
of analysis for control and optimization of treatment plants was
demonstrated.
The described concept and the mode of its practical use are new, and
further improvement of the method in terms of enlarging its analytical
capabilities and operational reliability is possible.
-------
SECTION II
RECOMMENDATIONS
Based on the results of this study, recommendatipns are made for future
research and practical utilization of pyrographic methodology. The
recommendations cover areas in which exploratory research was carried
out: water pollution monitoring, waste treatment processes, and classi-
fication of natural wastes. The following discussion defines each prob-
lem, describes the state of the art, outlines proposed solutions, and
states action needed.
Recommendation 1
Area of Application (Water pollution monitoring)
Need
The preservation of the quality of water resources has become a national
policy. As a first step in implementing this policy, water quality
standards were set by the states and approved by the federal government.
The setting of "standards" for waterways implies establishment of def-
inite rules, principles, or measures by the authority, an action which
is official or quasi-legal (Ref. 1). However, since legal actions can-
not be directed against a river or a waterway, the standards for natural
waters, to be effective in maintaining its quality, must be related
physically to the sources of water quality degradation (i.e., waste
effluents discharged by industries and municipalities). Since such dis-
charges are controlled by man, they can be legally regulated. However,
a solid scientific and equitable basis for such regulations must be
established.
The establishment and enforcement of effluent discharge standards is
probably the only realistic method for compliance with "water quality
standards" for the waterways. The objective and fair enforcement of
such standards requires practical and economical means for the monitor-
ing and surveillance of many pollution sources contributing to a typi-
cal water body. The monitoring system must deal with identities of
pollution sources and must be capable of determining quantitative con-
tributions from each industrial source.
State of the Art
The hydrochemical monitoring today is directed mainly toward the meas-
urements of water quality parameters, and generates information that
is not always correlatable in quantitative terms to specific sources
of pollution, especially in waterways with multiple waste discharges.
-------
The current situation is described by Ballinger (Ref. 2 ) as follows:
"It is worth noting that the available systems measure water quality
characteristics rather than specific pollutants. While these are im-
portant determinations, they are only measurements of pollution impact,
and do not identify the type or volume of waste causing the undesirable
condition. For truly effective pollution control, information on con-
centration of a polluting material, as well as the exact point of dis-
charge is required."
Consequently, while there are approximately 500 multiparameter water
quality monitoring instruments in use in the United States, they moni-
tor the impact of pollution on water quality rather than pollution
itself (Ref. 2). Although identification of specific organic and in-
organic pollutants can be undertaken with specialized procedures devel-
oped for particular situations, it is not practical in terms of time
and manpower to analyze for the wide variety of organic pollutants that
may exist in a waterway.
A method is needed for gross characterization of waste load contribu-
tions, from a given waste source to a river or a waterway for routine
monitoring for pollution control and enforcement.
Approach
This program addressed two fundamental questions dealing with water pol-
lution monitoring: (1) how to identify the source of pollution in
natural waters by positive physical means, and (2) how to determine
quantitatively the magnitude of contributions of various waste discharges
to a common waterway. The approach used in this study departed from con-
ventional working concepts of analytical chemistry (i.e., the resolution
of the mixtures into molecular species), followed by identification and
quantification. Instead, it was based on the theory of multicomponent
pattern recognition and differentiation, which permitted identification
of combined chemical compositions, rather than individual components.
Using a method of quantitative pyrographic analysis, it was possible to
perform direct analysis for organic composition in water samples. It
was possible to define particular industrial wastes as specific identi-
fiable entities in water by considering the pyrographic patterns they
produce, and to determine in laboratory experiments, identities and
quantities of a number of industrial-wastes present in a mixture.
Action Needed
Both the theory and practice of the described methodology are new, and
while the possibility for direct pollution monitoring of waterways is
indicated, additional research work will be required to reduce the pro-
posed concept to a practical tool of water pollution monitoring and
enforcement.
'4
-------
As far as application of this technique to pollution monitoring is
concerned, the following are needed:
1. Further development and modification of operating procedures
to increase the complexity of the pyrographic patterns. The
increased pattern complexity will produce more bits of infor-
mation from a given program and, hence, increase the resolu-
tion power of the technique.
2. Further improvement in the reliability of the system by mod-
ifications in the temperature control and sample injection
subsystems.
3. Evaluation of variations of chemical compositions in various
industrial waste discharges, and development of a statistical
model for defining a "mean" compositional standard or a "CALIF"
for a specific waste. The mean compositional pattern for a
given industry will be used in establishing the mathematical
equations for identification and quantification of waste from
such a discharge in a waterway.
4. Further study on identification and determination of concen-
tration of industrial wastes in model waterways.
Recommendation 2
Area of Application (Waste treatment processes)
Need
A principal means of industrial waste treatment in the southeast is
lagooning. Anaerobic and aerobic lagoons, and combinations of both,
are used extensively. With the establishment of water quality criteria,
a question was raised by Middleton and Bunch (Ref. 3) and Bersem and
Ryckman (Ref. 4) as to whether effluent quality produced by the oxida-
tion ponds was adequate. A recent EPA report (Ref. 5) states that:
"Operational data from oxidation ponds indicates that effluent quality
will not meet water quality criteria all the time," and "Performance
criteria cannot be predicted for facultative aerated lagoons prior to
actual operation."
To improve operations of waste treatment facilities, an analytical
methodology capable of monitoring compositional changes in waste water
streams is needed. The applicability of the pyrographic method of
analysis to this problem was investigated under one task of the program.
State of the Art
Application of the pyrographic method of analysis to the monitoring of
efficiencies of combined anaerobic-aerobic waste treatment was studied
under this program. The waste facility studied consisted of one small
-------
anaerobic and three large aerobic ponds. It was found, using pyrog-
raphy, that the anaerobic pond, which constituted a small part of
the physical plant of the facility, removed up to 80 percent of the
organic load. The second and third aerobic ponds, which constituted
more than 60 percent of the physical plant of the facility, removed
only 44 BOD equivalent units. It is apparent that such a type of
waste treatment facility is inefficient. In further studies, it was
found that in some cases, because of photosynthetic activity, the final
effluent had a higher organic load than that found in the intermediate
oxidation pond.
Approach
A pyrographic method of analysis was used for the study of efficiencies
of anaerobic and aerobic waste treatment processes.
Action Needed
It is indicated from this study that in combined anaerobic-aerobic sys-
tems, the anaerobic ponds function as the principal means for organic
load reduction. A study dealing with the optimization of existing lagoon
treatment systems and the proper engineering design of future plants
should be carried out.
Recommendation 5
Area of Application (Classification of natural waters
and eutrophication studies)
Need
The understanding of the eutrophication phenomenon, leading to a retar-
dation of aging processes in lakes, is essential to conservation of
natural fresh water resources. A method for type classification of
lakes, which is relatable to the eutrophication potential of a given
lake, is highly desirable. The residual organic matter in a lake may
be largely a manifestation of a lake's eutrophic state and may be
reflected in periodic algae blooms. The algae blooms fluctuate sig-
nificantly in time, but the composition of dissolved and particulate
organic matter is more stable (Ref. 6). For this reason, determination
of such compositions could provide simpler means for lake surveillance
than continuous monitoring for the appearance of periodic algae blooms.
The pyrographic method could provide effective and practical capability
for such a task.
State of the Art
The eutrophication-related measurements performed today include deter-
mination of phosphates and nitrates, organic carbon, and various other
-------
physical and chemical parameters. The more detailed analysis of organic
composition of water samples is not readily undertaken due to compli-
cated existing procedures.
Approach
It was shown by Birge and Juday (Ref. 6) that the organic background
of natural waters consists essentially of carbohydrates, proteins,
and lipids. The relative proportions of such materials form a pattern
which is relatable to total organic carbon concentration. Using this
model, a number of natural water sources in the southeast were analyzed
pyrographically. With the exception of the brown-colored, low-pH waters
of the coastal swamps of Georgia, a similarity was observed between the
organic composition reported by Birge and Juday and that found in the
southeast.
Action Needed
The study on organic composition of natural waters conducted under this
program was too limited in scope to draw any broad conclusions. A con-
siderable amount of field work will be needed to develop a practical
hydrochemical scheme for the classification of natural waters and to
relate such a classification scheme to the eutrophication phenomenon.
The enlargement of the model from three classes (proteins, carbohydrates,
and lipids) to a number of subclasses for each class will undoubtedly
increase the resolution of the model and lead to a more precise type
of classification.
-------
INTRODUCTION
This program is a continuation of a research and development effort
to establish means for rapid data acquisition to -control organic pol-
lution derived from industrial and municipal wastes. The initial 12-
month study was started on 28 June 1968 under FWPCA Contract No. 14-
12-424. It dealt with the design and assembly of working prototype
instrumentation, study of various modes of operation, and limited
preliminary field testing of the system. The essential components of
analytical hardware were assembled in a unitized package, and the
operating procedures were established. The field tests were carried
out using the Middle Oconee River Watershed as a hydrological model.
The preliminary data interpretation of the initial field experiments
indicated that application of pyrographic methodology in the area of
organic waste pollution control might be useful (Ref. 7).
Reported here are the results of a 26-month continuation study that
was initiated on 13 November 1969 under EPA Contract No. 14-12-802.
The overall objective of this exploratory research was to study and
define areas of potential applications of the pyrographic method of
analysis in water pollution control, especially in areas where ad-
vancement of the state of the art is needed. In addition, the pyro-
graphic methodology was further refined and principles of its opera-
tion were clearly defined. Specific tasks of this study included:
(1) computerization of an instrumental system for more efficient
data handling and computation of results, (2) development of proce-
dures for'qualitative and quantitative differentiation between vari-
ous industrial wastes present in admixtures, and (3) the application
of pyrography in the area of waste treatment process control. (The
term "qualitative" is used in this case for definition of identities
of waste sources, rather than identification of molecular species
present in each source.)
In the second phase of this program, design and fabrication of a sec-
ond, more advanced pyrographic instrument were accomplished. This in-
strument was needed to broaden the scope of research, and to permit
more detailed investigations of the utility of this technique in sev-
eral areas of potential applications. In addition, consideration was
given to the development of a general scheme for natural water class-
ification, and the establishment of a general index of water quality
based on the nature of organic content in water samples.
This effort was exploratory in nature and involved investigation of
potential uses of pyrography in water pollution control and identifi-
cation of specific areas where more detailed investigations are
needed.
-------
The object of this study was to determine the feasibility of the pyro-
graphic method of analysis for the qualitative and quantitative char-
acterization of organic pollutants in rivers and streams. For the
purpose of this report, pyrographic analysis is defined as the method
in which water samples containing organic materials are subjected to
high-temperature pyrolysis in the absence of oxygen whereby the or-
ganic species are fragmented to produce volatile gases. The gases
are then separated in a gas chromatographic column and detected with
a flame-ionization detector. The resulting "pyrogram" serves as a
fingerprint for the initial compounds, with the proportions of vola-
tile species formed being unique to the chemical structure of the
parent material.
Pyrography as a tool for identification of organic species is old in
the art. Greenwood, Knox, and Milne (Ref. 8) were able to charac-
terize various carbohydrates as early as 1961, while Winter and Al-
bro (Ref. 9) differentiated amino acids by pyrography in 1964.
Reiner (Ref. 10) identified bacterial strains using the method in
1965. Unfortunately, quantitative analysis has not been achieved
until now because of unsatisfactory reproducibility of the pyrograms.
It is believed this occurred because of the nonreproducible mode of
heat application. The reproducible mode of heat application on anal-
yzed materials is essential to reproducibility of pyrolytic processes
and, for this reason, an early task of this program was to develop a
means for maintaining controlled temperature of the pyrolysis and
providing for a reproducible mode of sample injection into the ther-
mal environment of the pyrolysis tube. The next task was to develop
the required computer programs, and then to test the method for ap-
plication to rivers and streams.
This report covers work performed on EPA Contract No. 14-12-802, for
the period November 1969 to February 1972. The experimental effort
was carried out at the Rocketdyne Field Laboratory, Athens, Georgia,
which was located on University of Georgia property, in the near vi-
cinity of the Southeast Environmental Research Laboratory of the En-
vironmental Protection Agency. For purposes of clarity, some of the
earlier work Rocketdyne performed for the EPA under Contract No. 14-
12-424 (Ref. 7) is included in this report. This earlier work led
to the present program.
10
-------
SECTION IV
PYROGRAPHIC METHODOLOGY
Principles of Pyrography
A pyrogram of a mixture of organic compounds is a composite pyrogram.
It is a simple arithmetic summation of pyrographic fingerprints of all
organic compounds present in the mixture. This is explained as fol-
lows. When a single organic compound is being pyrolyzed, a number of
derivative molecular fragments are produced. They are represented on
the potentiometric recorder chart as a series of peaks with varying
retention times and peak intensities (Fig- 1). Many organic compounds
produce similar products when pyrolyzed. These will manifest them-
selves on a pyrogram as peaks with identical retention times. However,
the intensities of such peaks will vary, producing a pattern specific
to that of the parent material.
If more than one organic compound is pyrolyzed simultaneously, the py-
rographic representation will be a summation of pyrograms produced by
individual compounds. In this case, the common peaks for each compo-
nent will be a summation of contributions from each compound. The
specific peaks will, of course, appear independently.
Whether 2 or 2000 compounds are present in the mixture, the common peaks
will be summations of the contributions from each contributing compound,
and specific peaks will appear with identities of their own.
Normally, it is possible to set a linear equation for each peak of a py-
rogram, and then solve a series of simultaneous equations for the un-
known concentrations of organic materials represented by the pyrogram.
The possibility of such a task was described by Lysyj, Nelson, and Webb
(Ref. 11) in 1970, in the following experiment.
Using an instrumental arrangement consisting of a gas chromatograph, a
pyrolysis tube, and a specially designed injection valve, pyrograms for
glycerol, DL-valine, and hexanoic acid were obtained. The intensities
of each peak for each material were determined and used in formulating
a series of linear equations. Mixed solutions of these materials then
were prepared and also analyzed pyrographically. The concentrations of
the individual components in a mixture were determined by relating the
intensities of all observed pyrographic peaks to the concentrations of
each component. The computations were performed by the least-squares
solution of linear equations, and showed good agreement with the known
composition of the samples.
This work demonstrated that the processes of pyrolytic fragmentation
when carried-out in the presence of large quantities of water were:
(1) quantitatively linear, (2) independent for each compound in a
11
-------
PEAK INTENSITY
•n
H-
eg
4
O
Op
H
O
I-"
-------
mixture, and (3) the pyrogram of a multicomponent mixture was a simple
arithmetic summation of the contributions from each component in the
analyzed mixture.
The maximum number of unknowns into which pyrographic data can be
solved is equal to the number of peaks observed on a pyrogram. The
pyrogram, of course, can be solved into a lesser number of composi-
tions, and a least-squares approach to the solutions can be used in
this case.
The organic compositions into which pyrographic information can be dif-
ferentiated are defined by calibrations, which provide the coefficients
for a series of simultaneous equations. The system can be calibrated
in terms of molecular entities, classes of organic materials, or any
other arbitrarily defined combination of organic molecules.
Table 1 lists the simplest form of mathematical differentiation and
compositional determination of mixed organic solutions. Here, three
compounds, glycerol, DL-valine, and hexanoic acids, were defined pyro-
graphically as the standards, and then the pyrogram of the mixture of
three is interpreted mathematically in terms of those standards. The
standards, however, do not have to be pure organic compounds. As a
matter of fact, any arbitrarily defined combination of organic species
present in a mixture can be classified as a standard. Such arbitrar-
ily defined groupings could be classes or organic compounds or multi-
component mixtures of organics as found in industrial and municipal
waste effluents. When a number of such defined organic compositions
are mixed in an aqueous solution, concentrations of each can be com-
puted. In this case, calibration constants for individual groupings
of organic matter are used as coefficients in linear equations.
A number of other ways can be used to differentiate composite pyrograms.
For example, it might be desirable to determine the concentration of a
specific chemical toxicant present in an admixture with a number of var-
ious industrial wastes. Such a system can be calibrated for a single
compound of interest and the groupings of compounds as found in indus-
trial wastes. Then the composite pyrogram could be differentiated in-
to a specific compound and industrial waste compositions.
Basically; the calibration of the system determines the type of infor-
mation that will be obtained from a composite pyrogram. The mathemati-
cal technique for treatment of pyrographic data in the described manner
was proposed at the "Symposium on Organics in Natural Waters," Univer-
sity of Alaska, 1968 (Ref. 12), and is based on the use of least-squares
solution to a number of linear equations. In a pyrogram from an organic
material, the area of a peak is directly proportional to the concentra-
tion of the material. The linear equation relating these quantities is
y. = a..c.
7i ij J
13
-------
TABLE 1. MULTIPLE ANALYSES OF MIXTURES OF GLYCEROL,
DL-VALINE, AND HEXANOIC ACID IN AQUEOUS SOLUTIONS
Compound
Glycerol
DL-Valine
Hexanoic Acid
Glycerol
DL-Valine
Hexanoic Acid
Glycerol
DL-Valine
Hexanoic Acid
Solution 1
Present,
ppm
33.3
33.3
33.3
57.1
28.6
14.3
25.0
50.0
25.0
Found ,
ppm
34.4
32.8
32.8
Solution
58.1
26.6
15.3
Solution
27.3
47.4
24.7
Difference,
ppm
+1.1
-0.5
-0.5
2
+1.0
-2.0
+1.0
3
+2.8
-2.6
-0.3
14
-------
where i identifies the peak, j identifies the material, y is the area
of peak i, a^.: is the area per unit concentration of the jtn material
at peak i, ana c is the concentration of the jth material.
•
In samples containing more than one organic material, the total area
of a peak in a pyrogram is the sum of the contributions from each ma-
terial. Therefore, Cj, C2> . . . , cm can represent the concentrations
of the materials in the sample and y\, V2> • • • > Xn are t^e observed
peak areas in the pyrogram. Then the total peak area (y^) of peak i
will be:
m
yl = 2 aijcj
Thus, the equation relating the area of the first peak to the unknown
concentrations would be:
With the proportionality constants (a^j) determined by calibration and
the peak areas (y^) observed for the sample, the pyrogram can be com-
pletely defined by a set of n equations in the m unknown concentrations,
The least-squares solution to this problem can be found in the follow-
ing set of derived equations:
Q w— l> Q Jr* 4- IN d o / /"* -*•••• ^ I > o o I /^
= i ailyi -Uj ail /Cl V^ ailai2/C2 V^ ailaim/Cm
n i n v /n o \ /n
v »v i^/v 2\^ iv ^
> a._y. = > a.0a.. ) c, + I > a._ I c_ + ••• + I > a.^a. )
^ 12-^1 \ A, i2 il / 1 \A, i2 / 2 \ ~, i2 im /
., . c
., . i2 im / m
1=1 i=l i=l
il) 61 + ( ai3ai2 )82 + '•• + (. ai3aim )
m
n /n \ / n \ /n9\
? a. y. = (Y a. a.. ) 8, + ( > a. a. 0 ) c0 + • • • + (j a. )
im7i VA im il/ 1 \^. im i2 / 2 v A. im /
This mathematical treatment of a pyrogram is capable of rapidly provid-
ing quantitative compositional information about a multicomponent or-
ganic mixture.
15
-------
Instrumentation
Two pyrographic instrumentation systems were fabricated during the
contract period. Design and fabrication of the Mark 1 pyrographic
analyzer began under EPA Contract No. 14-12-424 (June 1968 through
December 1969), while the data processing equipment was integrated
into the system in the early stages of this program. The experimen-
tal data reported here were obtained using the Mark 1 instrument.
The automated Mark 2 pyrographic analyzer was designed and built in
the last part of this program and only limited performance testing
of Mark 2 was carried out. Design of both systems will be discussed.
Mark 1 Pyrographic Analyzer
The analytical subsystem (Fig. 2) consists of the supply of carrier
gases (steam generator and cylinder of helium), sample injector; py-
rolysis unit and the flow-stabilization column, gas chromatographic
columns (detector and reference arms of the flow system), temperature
enclosure with temperature program capability for the columns, hydro-
gen flame-ionization detector, supplies of combustion gases (oxygen
and hydrogen), and associated flow and temperature controllers. The
data processing subsystem consists of an electrometer; a potentiometric
recorder, a digital integrator, a teletype (equipped with keyboard,
punch tape, and tape reader), an audio data set, and a telephone re-
ceiver of a commercial telephone service. The latter provides a link
with the remote time-sharing computer service.
The instrumentation system is contained in a functionally unitized
package, which could be used under field conditions with a minimum of
laboratory support. Its design was influenced by the fact that such
a unit would serve as a prototype for a generation of automatic, self-
contained monitors capable of characterizing organic water pollutants.
The principal design requirement of such a unit is integration of the
working function into a single instrumental unit with a high degree
of mechanical and electronic reliability. These functions include
sampling of a water stream, pyrolysis of injected constant-volume
water samples, gas chromatographic separation of the produced pyrolytic
fragments, detection of the eluted matter, recording of the detector
signal, and automatic reduction of the produced data.
As shown in Fig. 2, the upper-left compartment of the prototype instru-
ment contains the sample injection system, the pyrolysis chamber and
flow stabilization column, the gas chromatographic column in a ther-
mally controlled enclosure, the hydrogen flame-ionization detector,
and the flow and power controls. The lower-right compartment contains
the electrometer, while the digital integrator, audio-data set, and
telephone are installed in the upper-right compartment. The bottom
compartment is used for storage of oxygen and hydrogen gas supplies.
16
-------
SAMPLE
STEAM GENERATOR
1
FLOW CONTROLLERS !
H
COMPRESSED GAS
Hi nmiriuw'" "•"•'
./k Axk SMUT-OFF
'V ;f~f VALVES
]@ 00
~y~~[JREC. "EC
Hz
COMPRESSED GAS
Figure 2. Mark 1 Pyrographic Analyzer
17
-------
The principal feature of the design is minimization of the length and
volume of the connectors between the sampling valve and the pyrolysis
chamber. The components are arranged in a radial fashion, which
greatly minimizes the length of connecting tubing.
The connection between the pyrolysis chamber-flow stabilization unit
and the gas chromatographic column leading to the hydrogen flame-
ionization detector is also minimized. This design provides for rug-
ged connection of, and easy access to, the lines for the combustion
gases, hydrogen and oxygen, permitting easy access when checking for
gas leaks. The gas chromatographic columns, which are functionally
the heart of the system, might require a great deal of interchanging.
They are mounted on a solid-metal plate connected to, and supported
by, the body of the instrument housing. The mounting is such that
handling of the columns during connection and disconnection will not
stress the extremely fine, low-internal-volume connections between the
accumulator, sample loop, check valve arrangement, and pyrolyzing/flow-
stabilizing unit.
Sample Injection and Pyrolysis System. The operation of the sample in-
jector is based on the helium gas-pressurized transport of the water
sample from the metering loop of the injection valve into the pyroly-
sis chamber. A two-way valve is used in this arrangement. It is con-
nected to the helium accumulator and the pyrolysis chamber. The two
other ports are used for the sample loading. In the operation of the
injector, the helium accumulator is automatically pressurized from a
helium tank by proper positioning of the valve. The metering loop of
the valve (150-milliliter volume) is filled with a water sample, and
the valve is rotated, placing the loop containing the sample into
alignment with the helium accumulator and the pyrolysis chamber. The
helium accumulator pressure is higher than the pressure in the carrier
gas line, and forces the measured volume of sample, as a single slug,
into the pyrolysis chamber.
The pyrolysis chamber (Fig. 3), previously developed under a company-
sponsored program, is constructed from a heavy-walled tube of nickel.
Only pure nickel was found to be a satisfactory material for this ap-
plication. The stainless-steel and nickel alloy tubes both deterio-
rated after a few months of use. At the front end of the chamber, a
Swagelok male connector is welded to the nickel tube to provide a con-
nection with 0.159 cm (1/16-inch) tubing from the sample inj'ector. A
second Swagelok male connector at the rear of the chamber connects to
the flow-stabilization column. The pyrolysis chamber is filled with
loosely packed small metallic rods cut from nickel wire, which serves
to disperse the vaporized sample within the chamber.
Positioned axially about the pyrolysis chamber is a helical heating
coil of nichrome wire. The heating element is spaced between the py-
rolysis chamber and a ceramic primary insulation chamber. The ceramic
18
-------
(A) STATIONARY FURNACE HOUSING
(B) AND (C) END PLATES
(D) SHUTTER
(F) AND (K) HEATING ELEMENT TERMINALS
(G) PYROLYSIS TUBE
(H) INSULATION CHAMBER
(I) CLOSURE PLATE
(J) AND (N) THERMOCOUPLES
(L) HEAT SHIELD
(M) FLOW-STABILIZATION TUBE
(Q) CONNECTION TO INJECTOR
K
Figure 3. Cross-Sectional View of the Pyrolytic Furnace Unit
tube is closed at the back end with a circular plate of stainless
steel. The front end of this tube is positioned in an annular groove
of the front end plate.
The heat input to the pyrolysis chamber is controlled with a variable
transformer connected to the heating element. A secondary insulation
barrier is composed of a stainless-steel heat reflector, several lay-
ers of asbestos sheet insulation, and an outer heat reflector of
aluminum slip-fitted about the asbestos insulation. The combination
of metallic heat reflectors and the asbestos insulation provides ther-
mal stability by diminishing the heat flux from the ceramic insula-
tion chamber. The temperature in the outer zone, which contains the
flow stabilization column, is controlled by cylindrical shutters in
the furnace housing through a number of elongated windows. Concen-
trically positioned about the periphery of the housing is a rotatable
cylindrical aluminum shutter which is provided with windows of size,
shape, and number corresponding to those in the housing. By rotating
the shutter, heat loss from the outer zone can be increased or de-
creased. This design permits the maintenance of two independent
19
-------
temperatures, one in the pyrolysis chamber and the second around the
flow stabilization column, with only one heat source.
Gas Chromatographic Columns. The pyrographic instrument is designed
to work with water samples directly (i.e., no prior separation or
organic matter from the water matrix is involved). This means that
the water vapor from the sample matrix follows the same steps of sep-
aration and detection as organic fragments. Because water vapor is
present in much larger quantities than the organic constituents, the
effect of it on separation and detection operations must be minimized.
This requirement quite severely restricts the types of gas chromato-
graphic substrates and detectors that can be used for such applica-
tion. The following liquid substrates were tested: Carbowax 20M,
silicone oil 710, SE-30, m-phenylether (five ring), Apiezon-L, eth-
ylene glycol adipate, and diethylene glycol malonate; none performed
satisfactorily. The liquid substrates produced significant bleeds
of the stationary phase, making their use impractical. Only inert
materials, such as porous glass and Porapak, were found to be
satisfactory.
Porous glass is a totally inert, inorganic substrate, prepared by
leaching certain formulations of glass. The leaching process dis-
solves some of the glass components, leaving a honeycomb structure
of silicon oxide. The gas chromatographic separations reported for
porous .glass are similar to separations obtained with silica gel sub-
strates. However, because of large pore size, higher-boiling mater-
ials can be more readily eluted from this substrate than from silica
gel.
Previous studies at Rocketdyne demonstrated that hydrocarbons from
methane through hexadecane (Ref. 13) can be effectively separated on
porous glass. Like silica gel, porous glass is a polar substrate and
has the tendency to retain polar substances (e.g., amines, alcohols,
and carboxylic acids) with considerably greater affinity than nonpolar
substances (e.g., hydrocarbons). However, it was found that this dif-
ficulty can be overcome when aqueous samples are injected. It is be-
lieved that water vapor depolarizes polar sites on the solid matrix,
and improves the sharpness and resolution of pyrographic peaks.
The starting material for the column preparation was porous glass (Code
7039 made by Corning Glass Co., New York). The plates of glass were
ground and sieved. Material of 40 to 50 mesh size was used in this
study. Its physical properties were as follows: surface area, 173.2
m2/gram; pore volume, 0.109 ml/gram; average pore size, 25.2 A. 6.1 m
(20 feet) of 0.477-cm (3/16-inch), outer-diameter tubing was packed
with this material and installed in the pyrograph.
20
-------
.ie initial experiments with helium carrier gas produced pyrograms
>f very low intensity and poor resolution. However, as additional
vater samples were injected, the intensity of the peaks and the degree
of resolution improved. Finally, after approximately 15 injections,
a broad peak of residual organics (from previously injected samples)
was eluted, and the quality of the programs improved sharply. There-
after, water samples injected at 15- to 20-minute intervals produced
well-resolved, highly reproducible pyrograms. Once the porous glass
column was conditioned, it produced pyrograms of good quality.
In normal operation of the pyrographic system, the porous glass col-
umn is pretreated by injecting distilled water samples, before start-
ing analyses.
Detector. In the initial design of instrument provisions for dual
columns, dual-detection modes of operation were made. A dual hydro-
gen flame—ionization detector made by Hewlett-Packard Co. was selec-
ted and installed into the pyrographic apparatus. However, in the
actual operation of the instrument, a very stable baseline was ob-
tained with only one working porous glass column; all experiments des-
cribed herein were obtained in such a mode of operation.
The electrometer, a Hewlett-Packard Model 5750B, is a solid-state,
line-operated unit, which is of the differentiating type (i.e., it
will differentiate the output of two hydrogen flame-ionization de-
tectors). All components of the pyrographic system, which were com-
mercially available, were selected from suitable sources. Those com-
ponents not commercially available were designed and fabricated in-
house. The commercially procured components included:
1. A two-way, variable-volume valve, Model No. 2013, Carle
Instruments, Inc.
2. Dual-flame ionization detector, Hewlett-Packard Corp.
3. Electrometer, Model 5750 B, Hewlett-Packard Corp.
4. Digital integrator, Model 7600A, Hewlett-Packard Corp.
5. Teletype, Model ASR-33, Teletype Corp.
6. Audio-data set, Specialized Communications, Inc.
Mark 2 Pyrographic Analyzer
The main features of the Mark 2 pyrographic analyzer are similar to
that of the Mark 1. System components, however, have been rearranged
to provide for easier maintenance and repair. The schematic of the
system is shown in Fig. 4; Fig. 5 is a photograph of the console.
System modifications include an automated sample injection system,
versatile control of a temperature program, redesigned pyrolysis cham-
ber, and custom-built steam generator.
21
-------
Figure 5. Rocketdyne Mark 2 Pyrographic Analyzer
23
-------
Sample Injection. The instrument incorporates a manual and automatic
injection system. A sample input fitting on the front panel is nor-
mally used for manual operation. The reservoir valve directly above
the fitting is closed during manual operation. In automatic mode, a
syringe is left in place on the panel fitting and the reservoir valve
is opened. A large syringe with the plunger 'removed may be attached
to the fitting at the top front of the cabinet to serve as a reservoir.
The valve may be used to throttle the flow to the desired rate by ob-
serving the discharge at the lower cabinet recess.
When the control switch is moved to AUTOMATIC, power is applied to the
timer motor and the timer clutch if the actuator is in CHARGE position.
At the end of a time period (preset at any period up to 60 minutes),
the timer makes a circuit operating the INJECT solenoid valve. The
valve opens, admitting nitrogen to the actuator. When the injection
valves reach the INJECT position, the nitrogen charge in the accumula-
tor pushes the sample in the metering loop into the carrier gas flow
at the entrance to the pyrolysis chamber. In this position, the timer
clutch circuit is broken, allowing the timer to reset. Breaking the
clutch circuit also closes a set of contacts in the timer, which stops
the integrator analysis. Resetting the timer breaks the circuit to
the solenoid valve, thus shutting off the nitrogen and venting the
actuator.
When the accumulator pressure drops to a preset value (nominally 20
psi), the pressure switch closes and actuates the RETURN solenoid
valve. The valve opens, admitting nitrogen to the actuator, which
moves the injection valve to the CHARGE position. As the actuator
moves away from the INJECT position, the timer clutch circuit is again
energized restarting the cycle. This also breaks the circuit, which
stops the integrator analysis, and makes another circuit, which starts
the analysis. When the actuator reaches the CHARGE position, a micro-
switch breaks the start analysis circuit because only a momentary con-
tact is required. If contact is maintained, a thermal protective fea-
ture in the printer control circuit is incapacitated. In the CHARGE
position, nitrogen pressure builds up in the accumulator, opening the
pressure switch, closing the solenoid valve, and venting the actuator.
By moving the control switch to MANUAL, power is applied directly to
the INJECT solenoid valve, if the actuator is in the CHARGE position.
The valve then opens and admits nitrogen to the actuator. When the in-
jection valve reaches the INJECT position, the nitrogen charge in the
accumulator pushes the sample in the metering loop into the carrier
gas flow at the entrance to the pyrolysis chamber. In this position,
the circuit to the solenoid valve is broken and a circuit to a timer
and two relays is made. Breaking the solenoid circuit shuts off nitro-
gen and vents the actuator.
The timer indicates time digitally in hundreds of minutes from start of
injection. One relay opens another set of contacts in the circuit that
energized the solenoid valve, but only after the holding contacts have
24
-------
been closed to maintain the timer and relay circuits. The second re-
lay breaks the circuit that stopped the previous integrator analysis
and makes another circuit, which starts the analysis.
Pyrolysis Chamber. The new pyrographic furnace used in the Mark 2
instrument consists of a concentric arrangement of a 0.954 cm (3/8-
inch) chamber tube, a heating element, a refractory, tube, insulation,
and an outer shell. Split end caps support the chamber and provide
easy access for service. The chamber is made of nickel tubing, filled
with 0.159 cm (1/16-inch) diameter by 0.318 cm (1/8-inch) lonp nickel
pellets. The heating element is made of 92.7 cm (36 1/2 inches) of
No. 14 0.165 cm (0.065-inch) Nichrome V wire, supported by the Zircon
refractory tube. Tubular ceramic insulators cover the radial exten-
sions of the heating element. Carborundum Fiber-Frax Type F-13-T
insulation is used both around the refractory tube and in the end
caps.
A chromel-alumel thermocouple is installed in the zone between the
heating element and the chamber. Tests have shown that the depth of
immersion is well within the high-temperature plateau of an axial
traverse of the chamber. A second chromel-alumel thermocouple is in-
stalled radially and bears on the outer surface of the refractory tube.
Temperature Control. The inner thermocouple is connected to both a
digital indicator and a contactless controller. The controller cir-
cuit has a switch to introduce bias with a mercury cell, thus provid-
ing a dual-range capability. Circuit components have been chosen to
provide ranges of 0 to 540 C and 460 to 1000 C. With this increased
sensitivity, the photoelectric controller is able to maintain a fur-
nace temperature within ±2 at 1000 C. (To conserve the mercury cell,
the range switch should always be left in the low-range position when
the furnace is off.) A fail/safe feature shuts off the power if the
thermocouple circuit opens.
Operation of the controller on the same thermocouple as is connected
to the digital indicator does not cause an appreciable error in the
reading. The digital indicator is, of course, the more-accurate in-
strument; therefore, if better agreement of the two readings is re-
quired at the operating point, the zero adjustment of the controller
may be trimmed by the necessary amount. A diode shunt in parallel
with the relay coil has been added to the controller circuit to mini-
mize contact bounce and reduce contact wear.
Other Components. The commercially procured components of this instru-
ment include:
1. Gas chromatograph, dual-flame detector, solid-state electro-
meter Model HP 7612A, Hewlett-Packard Company
2. Digital integrator, Model HP 3371A, Hewlett-Packard Company
25
-------
3. Potentiometer recorder, Hewlett-Packard Company
4. Temperature controller, Model SS-100 001, Simpson Corp.
5. Digital indicator, Model JP, West Corp. of Gulton Indus-
tries, Inc.
6. Sample valve, Model 2013, Carle Instruments, Inc.
Problem Areas
Accurate temperature control of the pyrolysis environment is essential
to the reproducibility of pyrographic data. In the initial design of
the Mark 1 pyrograph, the temperature of the pyrolysis unit was con-
trolled by a variable powerstat and maintained by balancing the heat
input with heat loss. This arrangement was found to be cumbersome
and required constant attention for accurate work. A temperature con-
troller designed by West Corp. was incorporated into the instrument,
and operated satisfactorily for a number of months. Eventually, how-
ever, the relay points became pitted, leading to malfunction of the
unit. The temperature control in the Mark 1 instrument was further
modified by EPA personnel after the instrument was transferred to
Southeast Water Laboratory.
The Simpson Temperature Controller in the Mark 2 instrument proved to
be troublesome and required constant repairs. Finally, it was replaced
by a powerstat, and temperature of the pyrolysis unit was controlled
manually. While this arrangement was found to be functional, automa-
tic control of the temperature of the pyrolytic chamber would be oper-
ationally more satisfactory.
Computer Procedures
Use of the instrument for monitoring complex hydrological systems will
require computerized data handling. As a part of this instrument sys-
tem design, two modes of computer use were evaluated. One mode in-
volved an incorporation of a small computer, such as Model PDP/8 (Digi-
tal Equipment Corporation) into the system on a permanent basis to per-
form the necessary computation functions. The other option involved
subscription to a commercial computer service on a time-sharing basis,
using conventional telephone lines for connection.
Considering the diversity of the potential users for such instrumenta-
tion, the computer time-sharing route was selected, mainly because it
minimizes capital outlay to the user. The prototype instrument, as it
exists today; incorporates the teletype unit and an audio-data set,
and requires only the receiver of a conventional telephone to be con-
nected to any one of a number of computer time-sharing services. In
short, the instrument system can be used, and will perform all the nec-
essary physical operations and the mathematical calculations at any
geographical point that can be reached by'power and telephone lines.
26
-------
The flow diagrams, description, and printouts of computer programs
developed under this program are presented in Appendix A.
Experimental Conditions
Standard conditions for operating the instrument were established.
Pyrolysis Temperature
Methyl isobutyl ketone (MIBK) was used as the initial standard test
solution, and an analysis was performed under the following tempera-
tures of pyrolysis: 550, 575, 600, 625, and 650 C. The reported
temperatures were recorded by a thermocouple located in the proximity
of the pyrolysis tube. The actual temperatures inside the pyrolysis
tube are higher. In this report, however, observed temperatures are
reported, since actual temperatures of pyrolysis can only be estimated.
The CALIFS (i.e., calibrating "fingerprints") for the MIBK aqueous
solution obtained at pyrolysis temperatures of 550, 575, 600, 625,
and 650 C were established; two major peaks are plotted in Fig. 6.
As can be seen, the instrumental response rises almost linearly over
the 550 to 625 C temperature range. Above 625 C, the rate of the in-
tensity rise is reduced, signifying approachment of the plateau for
instrumental response. At higher temperatures, the probability of an
instrument breakdown during periods of unattended operation is in-
creased. To minimize safety hazards, the recorded temperature of
650 C was selected for normal operating conditions.
Gas Chromatographic Column Temperature
The effect of various temperatures imposed on the gas chromatographic
column and the effects of the frequency of sample injections were eval-
uated. These two parameters are interrelated. The frequency of sam-
ple injection determines the amount of moisture retained by a column
at any given time, and this, in turn, affects the retention times of
organic material. The temperature of the column affects both the rates
of elution of organic materials and the quantity of water retained by
a column, which also affects the retention time for analyzed materials.
Effects of temperature were studied first. The separations were car-
ried out at 110, 150, 175, and 200 C. Samples were injected at con-
stant 15-minute intervals. Working at 110 C column temperature, a
number of higher-boiling materials were retained on the column, and
were eluted over a considerable period of time, interfering with the
analysis of subsequently injected samples. A broad ridge of higher-
boiling materials was eluted at 150 and 175 C column temperatures with
some overlap into subsequent injection. At a 200 C column temperature,
the resolution of frontal peaks was poor. Consequently; 175 C was
selected for normal operational conditions.
27
-------
550
575 600
TEMPERATURE. C
Figure 6. Peak Intensity Versus Temperature
28
-------
Sample Injection Frequency
The effect of the frequency of sample injection was studied. The sam-
ples were injected at 15-, 20-, 25-, and 30-minute intervals, while
column temperature was maintained at 175 C. Satisfactory resolution
of frontal peaks and complete elution of higher-boiling fractions
were achieved when samples were injected at 30-minute intervals, and
this frequency of injection was selected for normal operation
conditions.
Standard Instrumental Condition
On the basis of completed experiments, normal operating conditions
for the pyrograph were established as presented below:
Temperatures, C
Pyrolysis 650 recorded
Gas Chromatography Column 175
Detector 200
Gas Flows, cc/min at 40-psig inlet pressure
Helium 25
Hydrogen 40
Oxygen 180
Column
6.1 m (20 feet) long by 0.477 cm (3/16 inch) diameter, packed
with porous glass, 30-50 mesh size
Electrometer
Channel A (single channel)
Range 10
Attenuation 8
Integrator
Noise Suppression 3
Slope Sensitivity Up 0.03 mv/min
Slope Sensitivity Down 0.03 mv/min
Baseline Delay 0.02 minute
Peak Sum 1000 mv/min
Shoulder Control Front Off
Shoulder Control Rear 1000 mv/min
Recorder
0.5 or 1 mv full scale
29
-------
Standard solutions of albumin, glucose, and oleic acid were prepared
and pyrographed. These materials are being considered at this time
as possible model compositions for proteins, carbohydrates, and fats
found in natural waters. CALIFS for all three materials obtained
under normalized conditions of operation are reported in Tables 2
through 4.
Pyrographic Analysis of Known Multicomponent
Synthetic Mixtures
The pyrographic system was evaluated by performing analyses of syn-
thetically prepared aqueous solutions containing known quantities of
albumin, glucose, and oleic acid (as Na-oleate). The instrument was
calibrated by obtaining pyrograms or fingerprints, for separate
aqueous solutions containing 20-ppm albumin, 20-ppm glucose, and
5-ppm oleic acid (Tables 2, 3, and 4).
The comparison of typical patterns are reprinted in Table 5.
TABLE 5. COMPARISON OF PYROGRAMS FOR GLUCOSE,
ALBUMIN, AND OLEIC ACID
Peak
No.
1
2A
2B
3
4
5
6
Average
Retention
Time, minutes
1.79
2.04
2.15
2.79
3.44
5.51
13.32
Peak Intensities as
yV-sec/ppm
Glucose
449
64
165
71
18
9
Albumin
174
263
236
185
17
167
Oleic
Acid
1198
67
149
3154
1699
687
1615
After CALIFS were obtained for each standard material, a number of
mixed solutions of albumin, glucose, and oleic acid were prepared in
various ratios and analyzed pyrographically. The analysis of each
solution was performed three times. The results for six different
solutions are shown in Table 6. Average values and standard devia-
tions are computed from triplicate analyses. A typical BAJA print-
out (i.e., from the computer program which calculates concentrations
of various components in a mixture) is shown in Table 7.
The completed experiment demonstrated that quantitative pyrography is
indeed possible, and that solutions containing a number of various or-
ganic species could be differentiated into component parts by solving
a series of equations in a least-squares mode.
30
-------
TABLE 2. CALIF FOR ALBUMIN, NORMALIZED CONDITIONS
CALIF
10:34 FNB
RUNS
1
2
3
4
5
6
7
B0
PEAK
1
2
3
4
5
6
1 2
180* 215
183 218
182 210
178 215
181 219
182 213
178 214
V. ALBUMIN
AVERAGE
TIME
180.571
214.857
277.286
335
528.286
1292.86
PEAKS
3 4
277 333
281 340
278 338
276 331
275 335
279 337
275 331
20 PPM* 650
STANDARD
DEV.TIME
1 .98806
3.02372
2.21467
3.51188
7.69663
5.61037
567
526 1296 0
536 1293 0
533 1296 0
522 1289 0
523 1288 0
539 1302 0
519 1286 0
PYR0*170 C0L*80 SENS
MAX. MIN. AVERAGE
TIME TIME AREA
183 178 3476.86
219 210 5261.29
281 275 4717.57
340 331 3698.71
539 519 345.857
1302 1286 3345.1
8 9
0 0
0 0
0 0
0 0
0 0
0 0
0 0
,30 MIN.
STANDARD
DEV.AREA
186.1 75
254.61 7
253.038
513.503
190.703
4 277.332
10
0
0
0
0
0
0
0
AREA/
CONCENTRATION
173.843
263.064
235.879
184.936
17.2929
167.257
*Retention times in one-hundredths of a minute
-------
TABLE 3. CALIF FOR GLUCOSE, NORMALIZED CONDITIONS
CALIF
1U49 SYS2 03/24/71 WED
NJ
RUNS
1
2
3
4
20
PEAK
1
2
3
4
5
6
7
8
1 2
173 205
176 200
180 209
1 75 203
PPM GLUC0SE*
AVERAGE
TIME
176
204.25
281 .5
349.5
567.75
1296.5
1066-5
872.75
PEAKS
3 4
267 331
281 349
282 349
296 369
PYR0LISIS
STANDARD
DEV.TIME
2.94392
3.77492
1 1 .8462
15.5242
32.0871
203.818
791 .532
1048.73
567
532 1215 1224
566 1283 1911
563 1582 0
610 1106 1131
650*C0L* 1 75>S-80
MAX. MIN. AVERAGE
TIME TIME AREA
180 173 8979*5
209 200 1286.5
296 267 3303.5
369 331 1418.75
610 532 365
1582 1106 170
1911 1131 110-75
2101 1390 406.5
8 9
1390 0
2101 0
0 0
0 0
STANDARD
DEV.AREA
730.369
495.357
250.948
154.815
1 58.339
36.6151
82-4677
484.086
10
0
0
0
0
AREA/
CONCENTRATION
448.975
64.325
165.1 75
70.9375
18-25
8.5
5.5375
20.325
CP UNITS
-------
TABLE 4. CALIF FOR OLEIC ACID, NORMALIZED CONDITIONS
CALIF
15:36 SYS2 03/26/71 FRI
RUNS
1
2
3
4
5
PEAK
1
2
3
4
5
6
7
8
9
10
1 2
181 205
179 201
181 204
181 203
181 204
5PPM NA 0LEATE
AVERAGE
TIME
180.6
203.4
215.8
279.2
349.4
559
1410.8
1321
999.2
534
PEAKS
3 4
214 279
216 275
216 283
218 280
215 279
*650 PYR0*
STANDARD
DEV.TIME
.894427
1 .51658
1 .48324
2.86356
3.64692
7.681 15
1 7.0059
1221 . 12
1371 .49
1 194.06
567
349 558 1412
344 547 1382
354 568 1426
351 562 1414
349 560 1420
1 75 C0L>80 SENS..30 M
MAX. MIN. AVERAGE
TIME TIME AREA
181 179 5988.6
205 201 333.6
218 214 746.2
283 275 15770
354 344 8499.8
568 547 3435.4
1426 1382 8073.8
2507 1986 806
2632 2364 109.6
2670 2670 52.4
8 9
2507 2632
1986 2364
21 12 0
0 0
0 0
IN.
STANDARD
10
2670
0
0
0
0
AREA/
DEV.AREA C0NCENTRATI0N
148.658
53.3507
159.224
422.788
223.922
147.999
1 592. 1 7
1 127.75
151 .376
117.17
1 197.72
66. 72
149.24
3154
1 699.96
687.08
1 614.76
1 61 .2
21 .92
10.48
CP UNITS:
-------
TABLE 6. ANALYSIS OF MIXED SOLUTIONS
Solution 1
Albumin
Glucose
Oleic Acid
Solution 2
Albumin
Glucose
Oleic Acid
Solution 3
Albumin
Glucose
Oleic Acid
Solution 4
Albumin
Glucose
Oleic Acid
Solution 5
Albumin
Glucose
Oleic Acid
Solution 6
Albumin
Glucose
Oleic Acid
Present,
ppm
20
0
0
10
10
0
20
20
0
20
10
0
4
10
5
10
20
3
Average
Found,
ppm
19.8
-0.1
0
8.0
9.3
0
18.2
19.2
0
19.0
8.7
-0.3
2.9
10.8
5.2
8.3
20.9
3.3
Standard
Deviation,
ppm
0.92
0.52
0.12
0.51
0.20
0.06
2.17
2.13
0.12
0.35
1.20
0.12
0.25
0.56
0
1.33
1.56
0.10
34
-------
TABLE 7. TYPICAL PRINTOUT OF BAJA PROGRAM
cn
hUN*
1
2
3
A V Li -.AGE
ACTUAL
nUN*
1
ii
3
AvEiiAGL
ACTUAL
ALEsDK I N
9.69
7.12
6.04
8-28
10
ALfcUi-:ii\
19.14
18.69
19-52
19-12
20
KG
GLUCOSE
22.47
20.66
19.44
20.93
20
KG
GLUC03;.,
9.3
9.41
9.59
9.43
10
PEu LIlEli
OLE 1C
3.17
3.4
3.3
3.29
?
PH.h LI TEH
OLE 1C
• 92
• 99
.97
.96
1
RESIDUAL
1 17-449
3 '» 3 • 9 7
140.839
HESIDUAL
128.886
149. 194
1 52. 734
%FIT
98
95
9*7
%MT
96
96
96
-------
SECTION V
APPLICATION RESEARCH
The Concept of Multicomponent Pattern
Recognition and Differentiation-
The overall objective of this study was to investigate the applica-
bility of a pyrographic method of analysis to the gross characteri-
zation and simultaneous measurement of organic pollutants present
in water sources, and to evaluate potential uses of this technique
in the field of water pollution control and enforcement of water
quality standards.
Although identification of specific pollutants can be undertaken
with specialized procedures developed for particular situations,
it is not practical in terms of time and manpower to analyze for
the wide variety of organic pollutants that may exist in a water-
way. A gross characterization method is needed in routine moni
toring for control of pollution and maintenance of water quality,
as well as application in specialized situations such as tracing
of particular wastes to their sources and studying the degrada-
tion rates of wastes in a waterway.
Surveillance of pollution, to be an effective tool for maintaining
water quality, must provide answers to the following questions:
What are the volumes of waste effluents discharged into a partic-
ular water system by all the waste discharges? and How much
organic waste does each source contribute to the waterway? tSuch
information should be obtainable by objective means, independent
of information that might be available from records of interested
parties.
To deal with the problem of waste-source identification and quan-
tification of its contribution to overall pollution load of a
waterway, a radically new conceptual approach was proposed. This
concept is based on the theory of multicomponent pattern recogni-
tion and differentiation. It is postulated that the identity and
quantity of matter of discrete composition can be defined by a
pattern formed by all elements of its chemical compositions. The
application of this theory to the pollution surveillance problem
is based on the fact that each industry has a waste with an over-
all chemical composition distinctive to its operation. The
uniqueness of such waste compositions offers a means for their
characterization as separate entities, rather than as a collec-
tion of various chemical species. To take advantage of this con-
sideration, an analytical methodology that can treat and character-
ize the sum of chemicals in an aqueous solution as a single and
unique entity must be used. The method of quantiative pyrographic
analysis has such a capability. Using this technique, it is possible
37
-------
to define a pyrogram as a specific waste standard, and when a number
of such are present in the solution, to determine each, both quali-
tatively and quantitatively.
Water Pollution Surveillance and Monitoring
An extensive field effort was carried out in the southeastern part of
the United States with full cooperation of local industries to eval-
uate practical capabilities of pyrography as a tool for gross charac-
terization and differentiation between various industrial waste efflu-
ents. The waste samples were obtained by EPA personnel in cooperation
with responsible state water quality monitoring agencies. The indus-
tries supplied the samples with the understanding that they will be
used solely for research and will not be considered as a monitoring
of their facilities. The first task of this study was to develop
pyrographic patterns or fingerprints for varied industrial operations
and to establish the degree of differentiation possible between the
same.
The second task was to determine experimentally the feasibility of
qualitative and quantitative differentiation in a solution consisting
of a number of different waste effluents.
The third task was to develop methodology that would permit source
identification and quantitative determination of multiple wastes in
the presence of natural organic water background.
Pyrographic Characterization of Various Industrial Wastes. Waste
effluents for this study were supplied by 20 major industrial opera-
tions in Alabama, Florida, Georgia, Mississippi, and South Carolina,
including: Hercules, Incorporated, Hattiesburg, Mississippi; Air
Products and Chemicals, Incorporated, Escambia Plant, Pensacola,
Florida; Merck and Co., Incorporated, Medicinal Chemical Plant,
Albany, Georgia; Monsanto Company Plant, Pensacola, Florida; and
Delta Airlines, Incorporated, Atlanta, Georgia. The industry sam-
ples included chemical, food, textile, paper and pulp, oil refinery,
brewing, pharmaceutical, and aircraft maintenance.
Samples of industrial waste discharge were collected at the points
of discharge into receiving waters and were preserved by freezing in
dry ice. The collected samples were analyzed pyrographically under
standard conditions.
The calibration constants for 19 industrial waste effluents (one
plant reclaimed its waste, and could not technically be considered
as a waste discharger) were examined with the object of determining
the differences between pyrographic patterns. The data are reported
in Table 8.
38
-------
TABLE 8. PYROGRAPHIC FINGERPRINTS OF WASTE EFFLUENTS
(PEAK MAGNITUDE x 10"3)
Food and Beveraae
Average
Peak Retention
No. Time
1 1.9
2 2.2
3 3.3
4 4.0
5 6.5
6 14.0
Various Industrial
Average
Peak Retention
No. Time
1 1.9
2 2.2
3 3.3
4 4.0
5 6.5
6 14.0
Texti 1 es
Average
Peak Retention
No. Time
1 1.9
2 2.2
3 3.3
4 4.0
5 6.5
6 14.0
Chemical Industrial
Average
Peak Retention
No. Time
1 1.9
2 2.2
3 3.3
4 4.0
5 6.5
6 14.0
Food-A
67.9
2.1
13.3
5.9
0.7
3.8
Paper
and
Pulp
96.6
4.5
32.9
25.0
5.1
29.0
Textile
A
607.0
110.9
576.5
569.2
95.4
156.1
Chemical
A
38.2
4.2
84.3
39.4
10.6
6.5
Food-B
22.5
1.4
31.6
12.0
2.8
8.4
Oil
Food-C
342.1
29.5
147.0
144.6
21.2
61.0
Refining Pharmaceutical
34.6
2.2
46.5
20.9
6.2
16.8
Textile Textile
B C
669.6 2,289.5
83.9 555.2
293'. 4 3,213.5
152.8 2,364.0
279.0 392.8
117.9 1,252.0
Chemical Chemical
B C
85.9 4008.0
7.2 328.0
75.1 330.5
35.4 1444.7
5.8 27.1
38.8 327.0
2320.7
115.7
87.8
19.5
60.5
Textile
D
1,229.0
328.4
2,050.0
1,571.0
210.1
565.4
Chemical
D
49.6
5.0
42.7
26.2
2.2
12.3
Brewing
248.5
39.4
129.9
110.8
15.8
54.4
Aircraft
Maintenance
237.7
17.6
324.6
122.9
7.6
25.4
Tpxt.il e
E
681.5
102.2
363.1
228.2
36.5
171.5
Chemical Chemical
E F
11.7 45.
2.4 2.
8.7 32.
3.6 31.
1.1 1.
3.7 14.
2
3
6
2
6
7
39
-------
To permit qualitative interpretation of data, the intensities of
peaks in the pyrogram of a given waste effluent were reduced to per-
cent representation. Total intensity of a pyrogram was taken as
100 percent, and the magnitude of each peak was expressed as a frac-
tion thereof. Such qualitative display of data for all samples
examined is shown in Table 9.
It was found that the waste effluents from chemical plants produced
significantly different pyrographic patterns. However, a degree of
similarity was noted within the pyrographic patterns for the waste
effluents of textile plants.
In the food and beverage group, brewing and Food "C" plants waste
effluents produced pyrographic patterns that were quite similar.
The waste effluents of the miscellaneous industrial group were dif-
ferent in all cases, both qualitatively and quantitatively.
Generally speaking, there appears to be sufficient differences in
pyrographic patterns between industrial effluents of different types.
It also appears that the majority of industrial effluents within one
category could be differentiated. However, there are also some types
of industries (e.g., textile and those that process natural products)
in which differentiation between patterns of individual plant efflu-
ents within an industrial group will be difficult. Further increase
in complexity of pyrographic patterns would be desirable to increase
the differentiating power of this technique.
Analysis of Industrial Wastes in Mixed Solutions. To maintain water
quality in a stream, the waste discharges from industries and munici-
palities must be controlled. The effective control of waste dis-
charges can be implemented only by the monitoring of pollution source
contributions. A program for such monitoring must recognize two
aspects of water pollution: the volumes of waste effluents dishcharged
from each source, and the waste load contributions of each source.
To obtain needed information for pollution surveillance, computer pro-
gram BAJA was modified. The modified program, which is called BAJA-1,
uses two types of inputs: (1) the calibration constants for each
waste source (CALIF), and (2) the total organic content or mg C/l of
each waste source. When a sample of an aqueous solution containing
a number of different waste compositions is analyzed pyrographically,
two sets of data are computed: (1) the organic waste load contribu-
tion of each waste source, and (2) the percent volumes of waste efflu-
ents present in a sample.
The capability of this program to generate necessary information was
experimentally tested. In the initial test, samples of raw waste
effluents were obtained from poultry, paper, and textile plants. Each
sample was analyzed pyrographically. and calibration constants (CALIF*s)
40
-------
TABLE 9. QUALITATIVE ANALYSIS OF PYROGRAMS FOR
VARIOUS INDUSTRIAL WASTE EFFLUENTS
Chemical A
Chemical B
Chemical C
Chemical D
Chemical E
Chemical F
Pharmaceutical
Textile A
Textile B
Textile C
Textile D
Textile E
Food A
Food B
Food C
Brewing
Paper and Pulp
Oil Refining
Aircraft Maintenance
Total
Counts
(sum of all
peaks)
Percent of Total Counts
for each Peak
1
183067
248285
6.17 x 106
137956
31048
127672
2.60 x 106
1.915 x 106
1.326 x 106
1.007 x 10?
5.953 x 106
1.583 x 106
93650
78759
745430
598688
193080
127250
735796
21
35
65
36
38
35
89
32
48
23
21
43
72
29
46
42
50
27
32
2
3
5
4
8
2
0
6
6
6
6
6
2
2
4
7
2
2
2
46
30
5
31
28
26
4
30
23
32
34
23
14
40
20
22
17
37
44
22
14
23
19
11
24
3
19
11
23
26
14
6
15
19
19
13
16
17
6
2
0
2
3
1
1
5
2
4
4
2
1
4
3
3
3
5
1
4
16
1
9
12
12
2
8
10
12
9
n
4
11
8
9
15
13
3
41
-------
were obtained. The total organic content of each sample was deter-
mined by a gravimetric method. The CALIF's and the total organic
content values were supplied to BAJA-1 program. Then mixtures of the
three wastes were prepared in various concentrations and analyzed
pyrographically. The results were computed by the BAJA-1 program.
A typical printout is shown in Table 10.
As can be seen from this table, one test solution contained 1-percent
waste effluent from a poultry plant, 0.1-percent waste effluent from
a textile plant, and 0.5-percent waste effluent from a paper mill.
The average values (determined and computed by the BAJA-1 program)
which were found for the volumes of each waste present in the solution
were: 1.02 percent for poultry, 0.15 percent for textiles, and 0.45
percent for the paper plant. Each waste effluent, of course, con-
tained different loads of organic matter; Table 10 shows that the
waste load contribution of each waste source is not necessarily pro-
portional to the volumes in the solution.
In this case, the poultry waste contribution is 1.79 mg/liter of
organic waste; the textile, 4.4 mg/liter; and the paper mill,
4.69 mg/liter. As a matter of fact, the textile waste, which con-
stitutes only 6 percent of all waste volumes present in the solution,
comprises nearly half of the organic waste load of the solution. In-
formation of this type, obtained from a single pyrogram from one
downstream sample of river water, could form a logical basis for pol-
lution control enforcement action and could lead to the effective
maintenance of water quality in the waterway.
Additional solutions analyzed included the mixtures of two textile
waste effluents (Textile Processing and Textile Finishing) and one
chemical waste effluent (Hercules, Inc.)- The BAJA-1 program was
supplied with respective CALIFS and TOG numbers, and the results
were-computed. Binary mixtures of Textile Processing and Textile
Finishing also were included in this computation. They were con-
sidered to be ternary mixtures with Hercules, Inc., waste at zero
concentration.
It was important to establish whether the absence of one or more
of the waste effluents from a multiwaste calibrated system will
interfere with computations for materials which were present. Nine
mixtures of the following ternary waste compositions were prepared
in the following ratios:
Textile Processing
Textile Finishing
Hercules, Inc.
Mixtures, volume %
1
1
1
0
2
5
1
0
3
3
3
0
4
1
5
0
5
1
0.5
1
6
1
1
0.5
7
0.5
1
0.5
8
1
0.5
0.2
9
2
0.5
0.5
42
-------
TABLE 10. ANALYSIS OF A THREE-COMPONENT MIXTURE
OF INDUSTRIAL WASTES
BAJA-1
PERCENT OF EACH WASTE
RUN #
1
2
3
4
5
AVERAGE
PRESENT
POULTRY
.77
1.08
1.44
.8
1.03
1.02
1.0%
TEXTILE
.13
.19
.12
.17
.16
0.15
0.1%
PAPER
.51
.43
.47
.44
.41
0.45
0.5%
ORGANIC MATTER FROM EACH SOURCE
(MILLIGRAMS PER LITER)
1
2
3
4
5
AVERAGE
POULTRY
1.34
1.89
2.53
1.4
1.79
1.79
TEXTILE
3.8
5.53
3.29
4.95
4.45
4.4
PAPER
5.27
4.47
4.86
4.55
4.29
4.69
43
-------
The results reported in Table 11 indicate that qualitative and quan-
titative differentiation of the complex industrial wastes in solutions
containing effluents from similar and dissimilar industrial operations
has been achieved. It was further demonstrated that in a multicompo-
nent calibrated system the absence of a given waste does not interfere
with computations for wastes that are present. Specifically, the
presence of the Hercules, Inc., waste CALIF and TOG values in the
BAJA-1 program did not interfere with computations for Textile Pro-
cessing and Textile Finishing concentrations in the solution con-
taining only those wastes.
The completed experiments indicate that the theory of multicomponent
pattern recognition and differentiation are applicable to the char-
acterization and analysis of industrial wastes, and that pyrography
provides the capability for qualitative and quantitative differenti-
ation of mixtures of various industrial wastes.
Further increase in accuracy of analysis is desirable. The increase
in pyrogram complexity will lead to more accurate determinations,
and future research effort should be oriented in that direction.
Identification and Quantitative Determination of Industrial Waste
Effluents in the Natural Waters. Application of the multicomponent
pattern recognition and differentiation theory to the analysis of
industrial wastes in mixed solutions using pyrographic methodology
was demonstrated experimentally. The solutions analyzed were pre-
pared in distilled water. In real-life situations, however, indus-
trial wastes are found in the presence of natural organic background.
The organic background of natural waters, in some cases, can exceed
industrial waste contributions.
To test the performance of the system with such functional con-
straints, a number of industrial waste mixtures were prepared in
water from the Oconee River., collected in the town of Athens,
Georgia, and analyzed. The water analyzed pyrographically for
organic background was found to contain:
Proteins
Carbohydrates
Lipids
Total Natural Organics
ppm
1.22
7.47
0.46
9.15
Approximate
Organic C, mg/1
0.6
3.0
0.4
4.0
Mixtures of two industrial wastes (Hercules, Inc., and Textile Finish-
ing) were prepared in Oconee River water and analyzed pyrographically.
44
-------
TABLE 11. ANALYSIS OF MtXED SOLUTION OF INDUSTRIAL WASTE EFFLUENTS
Present
Found
mg C/l
Present
Found
mg C/l
Present
Found
mg C/l
Present
Found
mg C/l
Present
Found
mg C/'l
Present
Found
mg C/l
Present
Found
mg C/l
Present
Found
mg C/l
Present
Found
mg C/l
Textile
Processing
1.0
0.84
(0.68)
5.0
5.93
(4.8)
1.0
1.52
(1.23)
3.0
3.44
(2.79)
1.0
0.73
(0.5:0
1.0
0.72
(0.58)
0.5
0.46
(0.38)
2.0
2.23
(1.8)
1.0
0.99
(0.8)
Textile
Finishing
1.0
0.9
(2.79)
1.0
1.21
(3.75)
5.0
4.9
(15.29)
3.0
2.69
(8.39)
0.5
0.45
(1.39)
1.0
0.96
(2.99)
1.0
0.87
(2.72)
0.5
0.36
(1.11)
0.5
0.44
(1.36)
Hercules
Inc.
0
-0.02
(0.03)
0
-0.08
(-0.16)
0
-0.06
(-0.13)
0
-0.09
(-0.18)
1.0
1.08
(2.05)
0.5
0.52
(0.98)
0.5
0.49
(0.93)
0.5
0.55
(1.06)
0.2
0.21
(0.4)
NOTE: Numbers referring to present are expressed in percent of specific
waste present in solution. Mg C/l refers to actual organic load
contribution of each waste.
45
-------
The waste concentration (volume percent), waste contribution (as
Organic C), and natural organic background of the water (also as
Organic C) of analyzed solutions are as follows:
Hercules, Inc.
Percent Waste
Organic C, mg/1
Textile Finishing
Percent Waste
Organic C, mg/1
Natural Background
Organic C, mg/1
Solution
!_
1.0
1.9
1.0
3.1
4.0
2
0.5
1.0
1.0
3.1
4.0
3
1.0
1.9
0.5
1.6
4.0
The solutions were analyzed pyrographically. The results were com-
puted using the BAJA-2 program, which is a modification of the prin-
cipal BAJA program and requires the following inputs: (1) CALIF for
natural organic background of water, or CALIF of water sample col-
lected upstream from waste-discharging sources; (2) CALIF's of
industrial wastes; and (3) TOC's of industrial wastes.
The program defines, qualitatively and quantitatively, the character
of natural organics and computes the percent volumes and the organic
waste load contributions of industrial waste effluents.
The results of this experiment are reported in Table 12, and indicate
that the analysis of mixed industrial waste effluents in natural
waters is possible.
Waste Treatment Control and Optimization
Lagooning is a principal means for industrial waste treatment in the
southeast. Simple holding ponds are most commonly used. More advanced
facilities comprise a series of anaerobic and aerobic ponds. Mechan-
ical aeration supplements the oxygen in some cases. During the period
of this program, a broad cross section of industrial waste treatment
operations in five southeastern states were examined. The plants in-
cluded chemical, food, textile, paper and pump, oil refining, brewery,
pharmaceutical, and service industries. Waste water samples were
taken at 24 industrial operations; of those, 13 had waste treatment
facilities and 11 did not. Of the 13 plants with waste treatment
facilities, 11 used the lagooning mode of treatment. One plant re-
claimed its waste and one used the activated-sludge process.
46
-------
TABLE 12. ANALYSIS OF MIXED INDUSTRIAL WASTES
IN NATURAL WATER MATRIX
Natural Background
Proteins
Carbohydrates
Lip ids
1.22 mg/1
7.47 mg/1
.46 mg/1
Volume Percent From
Source J
Hercules, Inc
Textile
Finishing
Source
Hercules, Inc
Textile
Finishing
Run 1 2
. .48 .42
.66 .66
Concentration of
Run 1 2
. .91 .8
2.85 2.67
3
.44
.87
Oraanic
Mg/1
3
.86
2.7
4
.46
.91
Carbon
4.
.87
2.82
Each Source
5 Averaoe
.5 .46
.98 .86
From Each Source
5 Averaoe
.95 .87
3.84 2.67
Actual
0.5
1.0
Natural Background
Proteins
Carbohydrates
Li pi ds
1.22 mg/1
7.47 mg/1
.46 mg/1
Volume Percent From
Source
Hercules, Inc
Texti 1 e
Finishing
Source
Hercules, Inc
Texti 1 e
Run 1 2
. .92 1
.67 .8
Concentration of
Run 1 2
. 1.75 1.9
2.0 2.48
3
.95
.91
Organic
Mg/1
3
1.8
2.82
4
.97
.84
Carbon
4
1.84
2.6
Each Source
5 Average
1 .97
.87 .82
From Each Source
5 Average
1.9 1.84
2.7 2.54
Actual
1.0
1.0
Finishing
47
-------
The principal function of aerobic and anaerobic lagoons is to reduce
the organic waste load in the industrial effluent. Efficiencies of
such treatment operations are judged by the amount of organic matter
removed from the stream. Normally, an indirect measure of organic
content is used, as reflected in the BOD test. Bersem and Ryckman
(Ref. 5) developed a failure scale for lagoons based on BOD and sus-
pended solids in the effluent. The first degree of failure occurs,
according to this scale, when the BOD exceeds 20 mg/1, and suspended
solids exceed 25 mg/1. The second degree of failure occurs at BOD
values above 30 mg/1 and solids values about 35 mg/1. The third-
degree failure occurs when BOD exceeds 40 mg/1 and suspended solids
exceed 45 mg/1.
Investigation of the potential use of the pyrographic method of anal-
ysis for the determination of efficiencies of waste treatment proces-
ses, leading to optimization of these operations, was one of the tasks
of this program. A modern industrial waste treatment facility, typ-
ical for the region and built utilizing standard sanitary engineering
design, was selected for the study. The facility comprised one anaer-
obic and three aerobic ponds. Its process capacity was 3.8 million 1
(1,000,000 gallons) of water per day.
The initial study on efficiencies of anaerobic and aerobic treatment
processes was conducted in June 1969, utilizing the Mark 1 pyro-
graphic analyzer. Using pyrographic data, the BOD equivalents
(Ref. 14) were determined for the raw waste and the effluents from
the anaerobic and three aerobic ponds. The results shown in Table 13
indicate that the principal reduction of BOD values was taking place
in the anaerobic pond. The aerobic ponds, which constitute more than
80 percent of the physical plant of the waste treatment facility,
were highly inefficient as far as reduction of the organic waste load
was concerned. Specifically, the organic load of the incoming waste
was reduced by 1261 mg/1 BOD equivalent in the anaerobic ponds, while
the three subsequent aerobic ponds contributed to an organic load re-
duction of only 223 mg/1 BOD equivalent. Furthermore, the reduction
of organic matter in the second and third aerobic ponds (which con-
stituted more than 60 percent of the physical plant of this waste
facility) amounted to only 44 mg/1 BOD equivalent. From this, it
appeared that the operation of the treatment facility could be sub-
stantially improved. Examination of the literature indicated that
other workers have reported similar results for operation of anaer-
obic and aerobic lagoons. Sollo (Ref. 15) reported on the operation
of anaerobic and aerobic waste treatment lagoons for the meat-packing
industry. His results were similar to those observed in this study.
He reported that waste effluent containing 1060 mg/1 BOD was reduced
by anaerobic treatment to 163 mg/1 BOD or approximately by 85 percent
of the organic load. The aerobic pond further reduced the organic
load to 58 mg/1 BOD. A recent EPA publication (Ref. 5) reports that
these results are typical of the data reported for other meat-packing
48
-------
plants. From this, it is apparent that current design of lagooning
type of waste treatment facilities is far from being optimized, and
a report by Bersen and Ryckman (Ref. 4) concludes that "continued use
of oxidation ponds as presently designed and operated will not enhance,
restore, or maintain the quality of the receiving stream."
TABLE 13. REDUCTION OF ORGANIC LOAD
Poultry Co. Untreated Waste
Anaerobic Pond
First Aerobic Pond
Second Aerobic Pond
Third Aerobic Pond
Waste Effluent
Creek 0.161 km (0.1 mile) Downstream From Discharge
Creek Upstream From Discharge
BOD
Equivalent*
mg/1
1559
298
119
76
85
75
9
7
*Calculated according to a method described in Ref. 14
Recently, Middleton and Bunch (Ref. 3) expressed serious doubt as to
the continued use of oxidation ponds in the future. They believe that
the quality of effluents from oxidation ponds might not be adequate.
Algae seems to be responsible for both high BOD and high suspended
solids in the effluent. To test this suggestion, a differential pro-
cedure for selective analysis for biomass-related organics and waste-
related organic materials in a waste stream was devised utilizing
pyrographic methodology. The conventional techniques for measuring
organic load in water (i.e., BOD, COD, TOC) report nondiscriminating
values, reflecting the total organic load including waste materials
and naturally regenerated organic matter. Consequently, they shed
very little light on the nature and the intensities of heterotrophic
and autotrophic processes which normally take place in conventional
lagoon waste treatment units.
The pyrographic method of analysis, however, has a differentiating
capability, and was used for the analysis of waste treatment streams
in terms of concentration of organic poultry waste and concentration
of biologically produced biomass, reflecting the intensities of auto-
trophic and heterotrophic processes.
The chemical nature of raw waste discharged and the nature of the
treated waste effluent were different both in qualitative and quan-
titative terms, and such differences were reflected in pyrograms of
respective chemical compositions. Because the waste biotransformation
took place in all the units of the treatment facility, it was reason-
able to expect that both starting and product materials were present
in various units of the process, and that their concentration ratios
reflect the efficiency of the process unit.
49
-------
The information recorded on a pyrogram reflects both the qualitative
and quantitative composition of the sample. We had available pyro-
grams for process input matter (raw waste) and process output matter
(treated effluent), the organic composition of which derives princi-
pally from biomass.
The pyrographic data for each unit of the waste treatment plant can
be and was examined mathematically in terms of two such compositions,
Calibration data reflected in the CALIF for raw waste and the CALIF
for treated waste effluent were reduced by percent representations,
and the pyrographic information obtained for each unit of the opera-
tion was subjected to computer analysis BAJA.
The results reported in Table 14 support Middleton and Bunch's
(Ref. 3) proposition. It does, indeed, appear that oxidation ponds
contribute very little to organic load reduction, but rather recycle
organic matter of incoming waste into organic matter of biomass, by
means of photosynthetic reaction.
TABLE 14. DEGRADATION OF ORGANICS IN POULTRY WASTE
Sample Source
Raw waste effluent
Anaerobic pond*
First aerobic pond
Second aerobic pond
Third aerobic pond
Poultry
Waste
1080
3221*
60
65
1
Biomass
75
-859*
336
89
247
Total
1156
396
154
248
*Sample not representative, contains surface foam
Organic Composition of Natural Waters
In natural unpolluted waters, the biomass exists in a near steady-
state condition with water-carried organics. These materials, de-
rived from the decomposition of biomass and excretion from living
organisms, exist in gaseous, liquid, and solid physical forms. It
is estimated that approximately 10 percent of the organics naturally
found in water have been chemically classified. This 10-percent
value probably relates to the principal organic compounds present,
and does not take into account thousands of additional organic com-
pounds present at trace and subtrace levels. In spite of the great
diversity of organic materials present in natural waters, it is
generally agreed that such materials derived from natural processes
belong, to a large extent, to three classes: carbohydrates, nitro-
genous matter, and lipids (Ref. 16). However, the absolute values
for-the concentration of organic materials of each class, as reported
by various authors, vary to a considerable degree.
50
-------
Extensive data on organic composition in inland fresh lakes were col-
lected by Birge and Juday (Ref. 6) almost half a century ago. The
data showed that principal organic compounds of the natural waters
were hydrocarbons, proteins, and lipids. Using this model as an exam-
ple, unpolluted waters of the southeast were analyzed pyrographically.
Included in the study were stream, river; reservoir, and swamp water
sources in the southeast region of the United States. This region is
well-suited for such an investigation because it is not highly indus-
trialized or populated and still contains many unpolluted water sour-
ces. Those waters also represent a variety of hydrochemical types:
1. Coweeta Hydrological Model
The Coweeta Hydrological Station is maintained by the U.S.
Department of Agriculture in the Appalachian Mountains on
the border of North Carolina and Georgia. The water source
has a low mineral and organic content. An extensive hydro-
logical record is available for this watershed.
2. Piedmont Region Streams
Near Athens, Georgia, are a number of swift moving streams,
rivers, and reservoirs. Alcovy River and Clark Hill Reser-
voir were used as water sources in this study.
3. Coast Plains Rivers
The U.S. Department of Agriculture, Hydrological Model,
Tifton, Georgia, was used in this study. The watershed
comprises the Little River and its tributaries. The tribu-
taries flow through cultivated lands, and carry some degree
of agricultural pollution. However, water from the tribu-
taries filters through swamps before it reaches the Little
River.
4. Coastal Swamps
Two principal swamps of the area are the Everglades in
Florida, and the Okefenokee in Georgia. Culturally unpol-
luted waters of swamps contain large quantities of organic
matter.
The hydrochemical characterization of natural waters, in terms of the
three principal classes (carbohydrates, nitrogenous matter, and
lipids), was attempted using pyrographic methodology. The pyrogram
of natural water sample, of course, reflects the chemical nature of
organic compounds present in it at a detectable concentration level.
The problem is to divide the pyrographic information into the three
categories. To do so, calibration constants for these groupings are
needed.
A selection of model materials for calibration was based on the natural
water model. Sufficient quantities of biomass-derived carbohydrates,
proteins, and lipids were obtained from water samples collected in a
pond undergoing algae bloom. The proteinaceous materials present were
51
-------
separated from the water matrix by the ammonium sulfate precipita-
tion method, while the lipids were extracted by methane/chloroform
method. The soluble fraction remaining in the sample was classified
as carbohydrates.
Each of the separated materials was analyzed pyrographically and
CALIF's (i.e., a series of calibration constants) were obtained for
them. The CALIF's for carbohydrates, proteinaceous matter, and lipids
derived from natural water bodies were then compared with CALIF's ob-
tained for a number of compounds of each class. The CALIF's of known
materials, which did approximate the CALIF's of naturally derived
compositions most closely, were used as calibration constants in the
BAJA program. Glucose was selected as a standard for carbohydrates,
albumin as a standard for proteins, and oleic acid as a standard for
fatty materials.
Using derivative calibration constants, pyrographic data from various
unpolluted water samples were examined in terms of carbohydrates, pro-
teinaceous matter, and lipids. The comparison of data obtained by
Birge and Juday for Wisconsin lakes and data obtained pyrographically
for unpolluted waters of the southeast are shown in Table 15.
It appears that, with the exception of low pH, dark-colored waters of
Georgia coastal swamps, the similarity between water types reported by
Birge and Juday (Ref. 6), and water studied under this program does
exist.
52
-------
TABLE 15. ORGANIC COMPOSITION IN NATURAL WATERS
Coweeta IVatershej - IViscor
Coweeta Watershed
Wisconsin Lakes
Birge and Juday, Class 1
Proteins
mg/1
sin La
1.4
1.0
Piedmont Sources - Wisconsin Lak
Alcovy River
Clark Hill Reservoir
Wisconsin Lakes
Birge and Juday , Class 1
1.6
1.0
1.0
percent
kes
38
29
es
31
17
29
Coastal Plains Rivers - Wisconsin Lakes
Little River
Mills Creek
Birge and Juday, Class 3
Moving Waters, Everglades
Everglades, Moving
Streams
Wisconsin Lakes
Birge and Juday, Class 3
Still Ponds, Everglades -
Everglades, ponds
Wisconsin Lakes
Birge and Juday, Class 4
S.I
4.8
3.8
- '.Vise
3.7
3.8
29
29
17.2
Carbohydrates
•TH:/ 1
2.2
2.6
3.4
4.7
2.6
11.7
11.2
18.7
snsin Lakes
15
17.2
Wisconsin Lakes
3.9
4.9
12
14. S
19.7
18.7
27.9
28.6
percent
59
69
65
81
69
67
67
82.2
80
82.2
84
85. 1
Lioids
my/1
0.1
0.1
0.2'
0.1
0.1
0.7
0.6
0.2
1.3
0.2
1.6
0.2
percent
3
3
4
2
3
4
4
0.3
5
0.8
5
0.4
Total
Organic
Content
5.7
3.7
5.2
5.8
3.7
17.5
16.6
1") T
24.7
22.7
33.4
33.6
53
-------
SECTION VI
ACKNOWLEDGEMENTS
We express appreciation to Dr. H. P. Nicholson, Project Officer,
Southeast Environmental Research Laboratory; Athens, Georgia, for
his guidance on this program, and to Mr. W. Taylor, EPA, Athens,
Georgia; and from Rocketdyne, Dr. N. Mann, Mr. L. H. Groeger- and
Dr. B. L. Tuffly for their direct contributions to this program.
55
-------
SECTION VII
REFERENCES
1. McKee, J. E. and H. W. Wolf: Water Quality Criteria, 2nd. Ed.,
Publ. No. 3A, State Water Quality Control Board, Sacramento,
California (1963).
2. Ballinger, D. G.: "Decisions to be Made in the Use of Automatic
Water Quality Monitors," International Symposium or. Identifica-
tion and Measurements of Environmental Pollutants, Campbell Print-
ing, Ottawa, Canada, 158
3. Middleton, F. M. and R. L. Bunch: "Challenge for Waste Water
Lagoons," Proc. 2nd Int. Symp. for Waste Treatment Lagoons,
Kansas City, Mo., 364-366, June 1970.
4. Bersem, G. M. and D. W. Ryckman: "Evaluation of Lagoon Perform-
ance in Light of 1965 Water Quality Act," Proc . 2nd Int . Symp .
for Waste Treatment Lagoons, Kansas City, Mo., 63-80, June 1970.
5. "Waste Treatment Lagoons - State of the Art," Enviornmental Pro-
tection Agency Project #17090 EXH-07/7, Superintendent of Docu-
ments, U. S. Government Printing Office, Washington, D.C. 20402.
6. Birge, E. D. and C. Juday: Ecol. Monographs 4, No. 4, 63-80,
June 1970.
7. Lysyj , I.: Pyrographic Gross Characterization of Water Contam-
inants, Final Report, FWPCA Contract 14-12-424, Report R-7991,
Rocketdyne Division, Rockwell International, Canoga Park,
California.
8. Greenwood, C. T., J. H. Knox, and E. Milne: Chem. Ind., London,
1879-1880 (1961).
9. Winter, L. N. and J. Albro: J. Gas Chromatogr. 2, 1-6 (1964).
10. Reiner, E.: Nature, 206, 1272-1276 (1965).
11. Lysyj, I., K. H. Nelson, and S. R. Webb: Water Res. £, 157
(1970) .
12. Lysyj, I.: "Instrumental Techniques for the Identification of
Pollutants," Organic Matter in Natural Water, University of
Alaska (1970).
57
-------
13. Lysyj, I. and P. R. Newton: "Use of Porous Glass for Gas
Chromatographic Separations," Anal. Chem. 36, 3514 (1964).
14. Lysyj, I., R. H. Nelson, and H. Snell: "Rapid Instrumen-
tal Measurement of Organic Load in Waste Waters," J. Water
Poll. Control Fed., 41, 831 (1968).
15'. Sollo, F. M.: "Pond Treatment of Meat Packing Plant Wastes,"
Proc. 15th Ind. Waste Control, Purdue University, 386-391 (1960)
16. Hood, D. W.: Organic Matter In Natural Waters, University of
Alaska (1970).
58
-------
SECTION VIII
PUBLICATIONS
1. Lysyj, I. and P. R. Newton: Hydrochemical Applications of Pyrog-
raphy, Report R-9086, Rocketdyne Division, Rockwell International,
Canoga Park. California, December 1972.
2. Lysyj, I. and P. R. Newton: "Multicomponent Pattern Recognition
and Differentiation Method: Analysis' for Oil in Natural Waters,"
Anal. Chem. 4£, 2385-2387 (1972).
3. Lysyj, I., P. R. Newton, and W. J. Taylor: "Instrumentation-
Computer System for Analysis of Multicomponent Organic Mixtures,"
Anal. Chem. 43, 1271-1281 (1971).
4. Lysyj, I.: "A Pyrographic Instrument for Analysis of Water
Pollutants," Amer. Lab. 7, 23-25 (1971).
5. Lysyj, I.: "Pyrography - A New Hydrochemical Tool," International
Symposium on Identification and Measurement of Environmental
Pollutants, Campbell Printing, Ottawa, Canada (1971).
59
-------
SECTION IX
APPENDIX
Computer Programs
This appendix discusses computer programs developed under this
program, presents flow diagrams and typical printouts of the
programs, and gives an example of computer and long-hand cal-
culations of program BAJA. All programs are in written BASIC
language.
Computations
A computer program for handling raw data and two principal
programs for data processing and computation were developed.
Program DAY is used for handling and organization of raw data.
Program CALIF is used for computation of calibration constants.
Program BAJA calculates the actual concentrations of individual
components present in the aqueous solution by solving a series
of linear equations in the least-squares mode. Two modifica-
tions of the principal program BAJA were also developed. BAJA-
1 computes, in addition to volume percent of organic composi-
tions present, the contributions of each defined organic com-
position in terms of mg C/liter. The BAJA-2 program, in addi-
tion to all functions of BAJA-1 program, compensates for the
contributions of the natural organic background in the analysed
sample.
DAY Program
The programfDAY handles raw data. In the actual operation of
the instrument, the pyrographic readout in digital form is typed
on a roll of paper and punched on a papertape in ASCII, eight-
level code. Each set of information appears on two lines:
one indicating retention time, the second, peak intensity.
The retention time number is in hundredths of a minute and is
identified by the number 8 appearing on the same line. Num-
ber 6, appearing after the time measurement, indicates termina-
tion of the analysis. Numbers between 0 and 5, appearing af-
ter a number, indicate that it is peak intensity. Those num-
bers are powers of 10 by which the peak intensity is multiplied.
This is shown in Table A-l. After dialing a time-sharing com-
puter center, information from paper tape is transmitted as
audio-frequencies via commercial telephone lines and saved in
program DAY in nonedited form. The data are then edited by
assigning line numbers to each line. The saved, edited file
DAY is then ready to be used for computing results in program
BAJA.
The typical printout of nonedited and edited programs DAY and
associated computer commands are also shown in Table A-l.
61
-------
TABLE A-l. PRINTOUT OF PYROGRAPHIC DATA COMPUTATIONS
NEW:DAY
READY
0171,8
2058,1
0248.8
6791,0
0338,8
1901,1
0515,8
1882,0
0582,8
4089,0
0731,8
1446,0
1500,6
0000,0
KEY
READY
SAV*
READY
EDIT ASSIGN
READY
LIST
DAY
100 0171,8
110 2058,1
120 0248,8
130 6791,0
140 0338,8
150 1901,1
160 0515,8
170 1882,0
180 0582,8
190 4089,0
200 0731,8
210 1446,0
220 1500,6
230 0000,0
READY
14:42 FNE 10/20/70 TUES
tData are inputted from data tape.
Underlined sections are entered from the terminal;
all other material is from computer.
The commands are for system 2, Fulton National
Bank, Atlanta.
62
-------
CALIF Program
The CALIF program reduces the raw data produced by a series of
pyrographic runs on a given standard solution. The printout
lists retention times, average retention time, and standard de-
viation of retention time for each run. It also provides max-
imum and minimum retention times for each peak observed in a
series of calibration runs. The maximum and minimum retention
times form windows, which are used by the computer for the scan-
ning and selecting of peaks in a subsequent analysis of mixed
solutions. The CALIF printout also provides the average area
for the calibration peaks and standard deviation of the same.
Finally, the area per unit concentration in mg/liter is cal-
culated and printed out. This is the proportionality constant
used in subsequent calculation of the mixed solutions. CALIF
also provides noise-filtering capabilities by rejecting peaks
below a certain predetermined intensity.
There are two inputs into the CALIF program. Input DATD is
edited data (in similar fashion to program DAY) received by
the computer from a pyrograph through the telemetering system
and is organized by assigning line numbers. The input DATD
also carries information that is entered in table headings
such as column, carrier gas, and sensitivity setting.
Line 200 in CALIF, the second input, indicates the concentra-
tion of the calibrating solution.
A flow diagram of program CALIF is shown in Fig. A-l. The
figure is divided into sections for ease of explanation of
its various functions:
Section A. The headings are printed and the concentra-
tion is read into the program.
Section B. The data are read into the program from the
data file. It is examined to see if it is
valid.
Section C. The peak retention times are examined.
Section D. The peak times and areas are averaged, and
squared for later standard deviation
calculations.
Section E. Input data are examined to test for the end of
a run or the end of the series of runs.
Section F. The standard deviations are calculated and all
results are printed out.
A listing of CALIF is given in Table A-2.
63
-------
\PRINT /
HEADINGS /
A(Y.2) - 1E6 ''
A(Y,3) - A(Y,3) + U
AtY.d) - A(Y,I|) + U (2
A(Y,5) - A(Y.S) + E
A(Y,6) - A(Y.6) + E|2
_t
\PRINT 7
HEADINGS /
A(Y,3) - A(Y.3)/Z
,1.) - SO.R((A(Y.IO-A(Y.3)2/Z)/(Z-I))
A(Y,5) - A(Y.5)/Z
A(Y,6)-SQR((A(Y.6)-A(Y,5)2/Z)/(Z-0)
A(Y,7)-A(Y.5)/N/Z
PRINT A(Y,3), A(Y,I|),
A(Y,5). A(Y,6).
A(Y.7)
I
Figure A-l. Flow Diagram of CALIF
64
-------
TABLE A-2. COMPUTER PROGRAM CALIF
CALIF
50 DIMK50)
100 REM SET N=T0 C0NCENTRATI0N STATEMENT 200
110 DIM AC50*50>
120 PRINTTAB(20)rfP£AKS"
130 PRINTTA8(7)"r'5TAB( 13>"2"JTAB( I 9 ) "3" J TAB(25 )'M"t TAB(3 1 )"5"J
140 PRINTTA8(37)"6"JTAB(43)"7"JTAB(49)"8"JTAB(55)"9"JTAB(61)"10"
150 PRINT"RUNS"
160 INPUT:DATD:A$
170 F0R I=1TO 50
180 LET A(I,2)=10t6
190 NEXT I
200 LET N=100
210 INPUT:DATD:U*B»C»D
220 IF B=6 THEN 280
230 LET E=C*10tO
240 IF E<100 THEN 210
250 IF 8=8 THEN 380
260 PRINT"HA HA"
270 STOP
280 LET Y=0
290 IF" U<'0 ?!•"•>• 5!0
300 LET Z=Z+1
310 REM --PRINT DATA F0R EACH PEAK
320 PRINTZJTAB(6)I< I >JTAS< 12)1(2) JTAB( 1 8) I C3 ) ITABC24) K4) i TABOO) 1(5) J
330 PRINT TAB(36)I(6);TAB(42)I(7)JTAB(48)I(8)JTAB(54)I(9)
340 F0R X=1T010
350 LET I(X)=0
360 NEXTX
370 60 TO 210
380 LET Y=Y+1
390 IF U>A(Y*1) THEN 420
400 IF U=A(Y*5)+E
490 LET A(Y*6)=A(Y*6)+Et2
65
-------
TABLE A-2. (Concluded)
500 G3T0210
510 PRIMT TAB(3)AS
520 FCR I=1T072
530 PRINT"="J
540 NEXT I
550 PRINT
560 PRINT"PEAK";TABC7VtAVERAGE>tJTABC 1 8 VSTANDARD") TA9< 29 VMAX." JTABC 34 ) ;
570 PRINT"MIN.";TABC39)"AVERAGE";TABC49VSTANDARD";TABC61 VAREA/"
580 PRINTTABC8)"TIME"JTAQC18)"D£V.TIME"JTABC29)"TIME"JTA8C34)"TIME">
590 PRINTTABC40)"AREA"JTAB<49)"DEV.AREAMiTAB(58)MCeiMCENT.RATI0N11
610 F0R Y=1T010
620 IF A(Y»1)=0 THEN670
630 PRINTYITA3C7)A;TAB(39)ACY*5)/Z;TABC50) ;
650 PRINTSQR((ACY*6)-ACY^5)t2/Z)/CZ-l));TAB(61)A(Y*5)/N/Z
660 NEXTY
670 END
66
-------
BAJA Program
BAJA calculates the actual concentrations of individual organic
components present in the aqueous solution. For this, a mathe-
matical model is used that relates all pyrographic peaks to the
concentration of each separate component. It computes estimates
of the concentrations using a least-squares criterion.
There are four inputs into BAJA: (1) File DAY described above
is edited data received by a computer from a pyrograph over the
telemetering system. (2) Input WINDOW gives the dimensions of
the windows as determined by the CALIF for extraction of valid
pyrographic information. Each window consists of two numerals,
representing a lower and higher limit for computation scanning.
The numerals are expressed in hundredths of a minute (Table A-3).
(3) Input DATA, lines 110, 111, and 112, lists the number of the
peaks to be used in computation, the number of components to be
solved, and calibration constants for each peak of each component.
(4) Input PRINT/ lines 130, and 155, provides identification of
the analysis number and the names of the components analyzed as
a table heading.
TABLE A-3. INPUT WINDOW
100 170,190
200 200,220
300 250,290
400 330,360
500 540,560
600 1200,1400
A flow diagram of BAJA is given in Fig. A-2. The diagram is di-
vided into sections for ease of explanation:
Section A. The number of peaks and components present are
read into the programs. The calibration values
are also read in.
Section B. The cross multiplication is performed for equa-
tions listed on page 14.
Section C. The window limits are introduced.
Section D. The experimental data from file DAY are read in.
The data are examined for validity and sorted
to the proper slots.
Section E. The remaining calculations are performed for
equations listed on page 14. The data are fed
into a file.
67
-------
T9 -
V
FART )
T
-,« |
1
HEADING /
N. H /
/
1
• H
H(I.J) /
J
*
- H(|,J)/T9
i
+ H(J,|) * H(K.I)
4
*
2 X N C
-1
1
INDOW: W(l) 7
}
C(l) - 0
J __
R * 1 D
1 « 1
DAY. X /
J\.
$>
T^
JAY: B.C.D /
^^V" >
-^^
[NO
^\.
- 8 ^>
TYES
J * 2
J
JT~7r^ NO
wti-i)^- •
•^^.
W(l) J> 1
"^--^ NO
Tits cd) - c(j) * c
'
T
1 ,
•^
L.
G
L__
* 10 | D
\ PRINT: COEFF: LNH(J) : /
\ E(J K) /
"
LNH(IOOO » J): P(J) /
f \
J 1
M - U-l F
1
RESTORE:
COEFF
1 »- i
\ MAT INPUT: COEFF /
\ A(M,M) /
,
HATB - CON(H,M)
;
HATB - INV(A)
1
1
;
MATZ - CON(H.O)
A
\ HAT INPUT: COEFF /
\ X(H,0) /
i
HATZ - B * X
1
HATY • TRN(Z)
*~
\ PRINT R /
i • i
\ PRINT Y(0,J)/T9 /
\ /
i
V(R,J) - Y(0,J)
^ I
* 1
i 1
• • i
i
PRINT HEADINGS
Z3 •
VU.') |
Z3 - Z3/(R-D/T9
PRINT Z3
Figure A-2. Flow Diagram of BAJA
68
-------
Section F. The DATA from the file are read into a pair of
matrixes that are then solved for the unknown
concentrations.
Section G. The results from each individual run are prin-
ted. The data are saved for later averaging.
Section H. The averages are computed and printed.
A listing of BAJA is given in Table A-4.
69
-------
TABLE A-4. COMPUTER PROGRAM BAJA
BAJA
1 DIM C(10)
90 INPUT:DAY:AS
95 PRINTTAB(5)A$
100 DIM GC15)*S(15)*EC10*10)*VC30*30)*Y<10*10)*W(30)*H<15*15)
110 DATA 6*3*174*263*236*185*17*167*
111 DATA 449*64*165*71*18*9*
112 DATA 1198*216*3154*1699*687*1615
125 PRINT TABOO) "MG PER LITER"
126PRINT
130 PRINT "SAMPLE"JTAB(12)"ALBUMIN"JTAB<26) "GLUC0SE"*
135 PRINT TAB(39)"0LEIC";TAB(52)"£FIT"
160READN*M
170 LET U=M
130REMN=#0FPEAKS*M=00FC0MP0NENTS
190F0RJ=1T0M
200F0RI=1T0N
210READHCJ*!)
220NEXTI
230NEXTJ
240F0RJ=1T0M
250F0RK=1T0M
260F0RI=1T0N
270 LET EC J* K) s=E1E6THEN1050
420 INPUT:DAY:B*C*D
430 IFB=6THEN560
440IFBs8THEN470
450PRINT"INC0RECT PRESENTATI0N 0F DATA"
460ST0P
70
-------
TABLE A-4. (Continued)
470F0RJ=1T0N
480 LET I=J*2
490IFX="5>WCI-1 )THEN510
500G0T0400
510IFX>W( I >TH£N540
520 LET CC J)=C(J)+C*10tD
530G0T0400
540NEXTJ
550G0T0400
560F0RJ=1T0U
570F0RI=1T0N
530LETPC J)=PCJ)+H< J* I )*C( I)
590NEXTI
600NEXTJ
610F0RJ=1T0U
620F0RK=1T0U
630PRINT:C0EFF:LNMCJ> JEC J,K>
640MEXTK
650NEXTJ
660F0RJ=1T0U
670PRINT:C0EFF:LiMM< 1 000 + J) ; PC J >
680MEXTJ
690LETM=U
700REST0RE:C0EFF:
710MATINPUT:C0EFF:A(M*M)
730MAT8=I>MV(A)
740MATY=C0N( 1>M)
750MATZ=C0N(M> 1 )
760MATINPUT:C0EFF:XCM, 1 )
770MATZ=B*X
780MATY=TRNCZ)
790PRINTRJ
800 F0R J=OT0U-1
SIOLETVCR* J)=Y( 1* J+l )
820LETG J
850LETGC J)=YCU J+l )
860 NEXT J
870 F0R J=l T0 M
880 F0R I=1T0 U
890 LET T=T+GCI-1 >*H(I>J)
900 NEXT I
910 LET Z6 = Z6+ABS
930 LET Z5=Z5+-T>t2
940 LET T=0
950 NEXT J
71
-------
TABLE A-4. (Concluded)
960 LET Z5 =
970 LET Z6=INT( 1 00*< 1 -CZ6/Z7 > ))
980PRINTTABC 13*CU+1 »Z6
990 LET Z5=Z6=Z7=0
1000SCRATCH:COEFF:
1010F0RJ=1T0U
1020 LET P(J)=0
1030NEXTJ
1040G0T0350
1050 F0R J=l T0 70
1060PRINT"=";
1070 NEXT J
1080PRINT
1090PRI'NT"AVERAGE§§;
1100F0RI=OT0U-1
1 1 10F0RJ=1T0R
1120 LET Z3=Z3+VC J* I )
1 130NEXTJ
1 140PRINTTABC 12*CI + 1))((INTCC 100*Z3/CR-1 ))•••. 5)))/l 00 1
1150 LET Z3=0
1160NEXTI
1170 END
72
-------
BAJA-1 Program
To deal with problems of characterization and quantitative determina-
tion of waste effluents in natural waters, program BAJA had to be
modified. In addition to information on the volume of waste dischar-
ged, data on the organic load contribution of specific wastes are
needed for effective water pollution surveillance.
Basically, BAJA-1 program determines and computes volume percentages
of individual industrial wastes present in the sample and also pro-
vides information on organic load contributions of each waste. To
calibrate and program BAJA-1 system, two analyses on collected waste
effluents from various industrial discharges are performed: pyro-
graphic patterns of the waste (CALIF) are developed and total or-
ganic content of the sample is determined. In the actual operation
of .the pollution surveillance system, calibration constants of var-
ious industrial wastes are determined from an analysis of samples of
waste effluents collected at discharge points of the respective in-
dustries. The samples are analyzed pyrographically and a set of
calibration patterns are developed for each waste source. In addi-
tion to pyrographic analysis, TOC values are obtained for each sam-
ple. The pyrographic data are used for identification of waste
sources and computation of volume percent of each waste in a solu-
tion. By combining this information with TOC calibration inputs,
contributions of each waste source to a common waterway are computed
in terms of mg C/l.
In addition, this program has a capability for using weighted aver-
ages for the individual peaks. The peaks can be assigned weights
from 1 to 10, with more reliable peaks having a higher weight. The
program requires the following inputs:
1. Number of peaks to be used in computation, number of compo-
sitions to be differentiated, and calibration constants for
each peak. This information is entered in lines 180, 190,
200, and 210.
2. The weights for the individual peaks are entered in line
220.
3. The organic load is calculated on the basis of a calibra-
tion done at the time the standard sample is taken. The
organic load value is entered on line 230 in the units of
mg C/liter.
4. The headings for pollution source are entered on line 270
and 275. All other parameters are the same as in the BAJA
program.
73
-------
The flow diagram of the program is shown in Fig. A-3. To explain var-
ious operations of this program, it is divided into the following
sections:
Section A. The constants for thepeaks of each organic composi-
tion are read into the program.
Section B. The constants are weighted in accordance with
assigned weighing factors.
Section C. The cross multiplication and summing of the constants
are performed for equations listed on page 14.
Section D. The acceptable limits of the peaks are read into the
program.
Section E. The experimental peaks (from DAY input) are examined
for validity. If valid they are sorted into the proper slots.
Section F. This section is a repetition of Section B. This time
the experimentally obtained peaks are weighted instead of the cal-
ibration constants.
Section G. In this section the remaining mathematics from equa-
tions on page 14 are performed to obtain the coefficients for the
right-hand side of the equations.,
Section H. The coefficients are read into the data file for easy
handling and transfer of numbers.
Section I. The coefficients are read back from the data file into
a set of matrixes, which are then solved.
Section J. The solution of the equation is obtained from the
matrix, and the results are printed out.
Section K. The average for a series of analyses is computed.
Section L. The calculations for organic content of the sample
(MgC/1) are performed and printed out.
A listing of the BAJA-1 program is shown in Table A-5.
74
-------
Figure A-3. Flow Diagram of BAJA-1
75
-------
TABLE A-5. COMPUTER PROGRAM BAJA-1
GLFNST 1 5: 1 1 SYS2 12/23/71 THUR
1 BAJA-1 CALCULATES THE PERCENTAGE AND WEIGHT 0F ORGANIC MAT-
2 TER FRO r- A SERIES OF POLLUTION SOURCES. IT DIFFERS FR0M BAJA
3 IN THAT WEIGHTED AVERAGES F0R THE INDIVIDUAL PEAKS CAN BE
A USED. THE PEAKS CAN BE ASSIGNED WEIGHTS FR0M 1T010* WITH
5 THE »"0RE RELIABLE PEAKS HAVING A HIGHER WEIGHT. ONLY INTEGER
6 WEIGHTS ARE ALLOWED.
7 THE WEIGHTS F0R THE INDIVIDUAL PEAKS ARE ENTERED IN LINE 220
8 0R ANY LINE IMMEDIATELY F0LL0WING THE CALIBRATION F0R THE
9 PEAK AREA.
10 THE ORGANIC LOAD IS CALCULATED 0N THE BASIS 0F A CALIBRATION
11 DONE AT THE TIME THE STANDARD IS TAKEN. THE 0RGANIC LOAD
12 IS ENTERED ON LINE 230 IN THE UNI TS OF MG/LITER
13 ALL OTHER PARAMATERS ARE THE SAME AS IN THE BAJA PROGRAM
14 IF THE ORGANIC LOAD IS NOT DESIRED TYPE IN LINE 1 T8= 1
BAJA1
100 REM F0R INSTRUCTIONS LIST GLINST
105 SCRATCHtCOEFF:
110 PRINT" PERCENT 0F EACH WASTE"
120 F0R J=1T053
130 PRINT"-"*
140 NEXTJ
150 PRINT
160 T9=1E4
170REM DATA FILE IS CALLED DAY
180DATA6*3
190DATA669550*83900*293400*152765*27905*117870
200DATA1.229£6*328400*2.05E6*1.571E6*210125*565350
210DATA4.008E6*328800*330500*1.44467E6*27116*327000
220DATA1*!*!* 1*1*1
230DATA81*310*190
240 DIMGC15)*SC15)*DC50*10)*EC10*10)*WC30)*VC10*10)*YC10*10)
250 DIMHC10*60)*C(60)
260 PRINT-RUN*"J
270PRINTTABC12)"BATH"?TABC24)"EXCELSIOR";
275PRINTTA3C36)"HERCULES"
280 READ N9*M9
290 U9=M9
76
-------
TABLE A-5 (Continued)
300REMM=#0FPEAKS*M=#0FC0MP0NENTS
310 F0RJ=tT0 M9
320 F0R I=1T0 N9
330 READ3CJ,I>
340 Q«-HCJ*I>*HCK*I)
SSONEXT!
560 ••• E X T K
570NEXTJ
580 LET J=2*N9
590 F0RI=1T0J
600 INPUT:WINDOW:UCI)
610 NEXT I
620 F0R I=1T0N9
630 P=RCI)=0
640 NEXT I
650REM"R"WILL3EUSEDT0KEEPTRACK0FRUNS
660 R9=R9+1
670 INPUT:DAY:X
680IFX>1E6THEN1230
690 INPUT:DAY:8*C»D
700 IFB=6THEN830
710IFB=8THEN740
720PRINT"INC0RECT PRESENTATION 0F DATA"
730ST0P
740F0R J=IT0N9
750I=J*2
760IFX=>W<1-1)THEN780
770G0T0670
780IFX>W(I)THEN810
790RCJ)=RCJ)+C*10tD
77
-------
TABLE A-5 (Continued)
800G0T0670
810NEXTJ
820G0T0670
830 F0RT1=1T0N9
840T2=L(T1)
850F0RT3=1T0T2
860C*CCI>
940NEXTI
950NEXTJ
960F0RJ=1T0U9
970 F0RKMT0U9
980PRItNT:C0EFF:LNMCJ>;E(J*K)
990NEXTK
1000NEXTJ
1010 F0RJ=1T0U9
1020PRINT:C0EFF:LNMC1000+J)JPCJ>
1030NEXTJ
1040M9=U9-1
1050REST0RE:C0EFF!
1060MATINPUT:C0EFF:A(M9>M9)
1070 MATB*C0N(M9*M9)
1080MATB=INV
1100 MATZ=C0NCM9*0)
1110 MATINPUT:C0EFF:XCM9*0)
1120MATZ=B*X
1 130MATY=TRNCZ)
1MOPRINTR9;
1150 F0RJ=OT0U9-1
1160 yCO>J)=CINTC(10000*CYCO*J)/T9))+.5))/100
1170 PRINTTAB(12*
1310NEXTJ
1320Z3=CINTC ( 1 00*CZ3/C R9- 1 >>>•»•. 5»/.I 00
1330 PRINTTABC12*CI+1)>Z3>
1340Z3=0
1350NEXTI
1360 IF T8=1THEN1650
1370 T8=l
1380 F0R J=1T0U9
1390 F0RK=1T0U9
1400 ECJ*K>=0
1410 NEXTK
1420 P(J)=0
1430 NEXT J
1440REST0RE:DAY:
1450REST0RE:WIND0W:
1460 F0RI=1T0U9
1470 READF(I)
1480 NEXTI
1490F0RJ=1T0U9
1500 F0RI=1T0N9
1510 Q(J>I)=Q
-------
BAJA-2 Program
Program BAJA-2 performs all the functions of program BAJA-1, and has a
capability for compensating for natural organic background of the
sample. It requires the following inputs:
1. Pyrographic data on organic composition of upstream water are
entered on lines 160 and 170. Number of peaks and number of
components to be considered in upstream water are entered on
line 160. Also added on this line are constants for two com-
ponents in upstream water. Line 170 holds the constants for
any remaining components of the upstream water.
2. Number of pollutants in the sample is given on line 180.
3. The pollution sources are identified on lines 190, 200, and
210.
4. The calibration constants for pollution sources are given on
lines 230 through 260.
5. The mathematical weights to be assigned to each peak.are given
on line 270.
6. The calibration data (TOC) for the organic content of each
component are given on line 280.
Normally; the upstream (unpolluted) sample is analyzed pyrographically
and generated data are assigned in DAY-1 file. The contaminated sample
is also analyzed pyrographically and generated data are assigned in
DAY-2 file. The program BAJA-2 subtracts contributions of the DAY-1
file from values recorded in the DAY-2 file and then performs computa-
tions to determine the polluters' contributions.
In cases where the upstream water sample is unpolluted, its organic
composition (in terms of carbohydrates, proteins, and liquids) can be
determined. The flow diagram of BAJA-2 program is shown in Fig. A-4.
To explain various operations of this program, it is divided into a
number of sections:
Section A. The information regarding the upstream water sample is
read into the program.
Section B. Limits for expected peaks are read into the program.
Section C. The data generated by upstream (unpolluted) water
sample are read into the program.
Section D. The peak magnitudes are assigned according to speci-
fied windows.
Section E. Cross multiplication and summation of peaks are per-
formed as specified by the equations on page 14. The results are
read into a data file for later retrieval.
80
-------
J (COWT)
Figure A-4. Flow Diagram of BAJA-2
81
-------
Section F. The data related to upstream water composition are
retrieved and results for natural organic background calculated.
Section G. The calculations for natural organic background are
completed,and averaged.
Section H. The parameters for downstream (polluted) fraction of
the sample are introduced.
Section I. The mathematical weighing of the peaks is accomplished,
The remainder of the program performs functions similar to operations
described in sections C, D, E, F, and G. A notable exception is
point X. At this point, the experimental data of the total sample are
corrected for the contributions of the upstream sample.
A listing of the BAJA-2 program is shown in Table A-6.
82
-------
TABLE A-6. COMPUTER PROGRAM BAJA-2
BAJA2
I THESE ARE THE INSTRUCTIONS FOR FEEDING INFORMATION T0
2 BAJA-2-
3 LINE 160 IN THE PROGRAM ENTERS THE NUMBER OF PEAKS AND THE
4 NUMBER CF COMPONENTS THAT ARE OBTAINED IN THE WATER "BLANK"
5 ALSO ENTERED ON"LINE 160 ARE THE CONSTANTS FOR THE FIRST TWO
6 COMPONENTS IN THE WATER BLANK.
7 LINE 170 HOLDS THE CONSTANTS FOR ANY REMAINING COMPONENTS IN
8 THE BLANK.
9 LINE 180 TELLS THE PROGRAM HOW MANY COMPONENTS THERE ARE IN
10 THE WATER "SAMPLE". THE NUMBER 0F PEAKS WILL BE THE SAME AS
11 WHEN THE "BLANK" WAS RUN.
12 LINES 190-220 CONTAIN THE NAMES 0F THE POLLUTERS IN THE F0RM:
13 190 SS="WATER FOR THE GIN"
14 200 SS(2)="WATER FOR THE WHISKEY"
15 ETC. FOR ALL THE COMPONENTS
16 LINES 230-260 CONTAIN THE CALIF CALIBRATION CONSTANTS FOR THE
17 POLLUTERS GIVEN IN'LINES 190-220.
16 LINE 270 GivES THE MATHEMATICAL WEIGHTS TO BE ASSIGNED T0
19 EACH OF THE PEAKS. THIS ASSINGMENT IS RATHER ARBITRARY AND
20 IS 0N THE BASIS OF THE REPR0DUCIBILITY OF EACH PEAK.
21 LINE 280 CONTAINS THE CALIBRATION DATA FOR THE ORGANIC
22 CARBON FOR EACH OF THE COMPONENTS.
23 THE DATA FILE CALLED DAY1 CONTAINS THE RUNS MADE WITH THE
24 WATER "BLANK".
25 THE DATA FILE CALLED OAY2 HAS THE DATA FOR THE ACTUAL SAMPLE
26 0F WATER. THE DATA FILE CALLED WINDOW TELLS THE COMPUTER WHIVH
27 PEAKS TO ACCEPT.
28 "GOOD LUCK".
100REM===F0RINSTRUCTI0NSLISTGULF99
110PRINTTTY(10>iTTY(10);TTY<10)
120SCRATCH:C0EFF!
1500IMV<10*10>*W<20>*QC10*10)*H<10*50>*FOO>»Y<10*10)*CC60>
160DATA6*3*346*526*470*368*34*334*896*128*330*142*36*24
170DATA2396*432*6308*3340*1374*3228
180DATA3
190SSC1)="HERCULES"
83
-------
TABLE 6. (Continued)
200S$(2)="EXCELSIORtf
210S3<3)=f*BATH"
230DATA4-003E 6* 323800* 330500* 1 . 447E6*271 16*327000
240DATA1 .'>29E6*328400*2.05E6* 1 . 571 E6* 2 1 0 125* 365350
250DATA669 550* 83900* 293400* 152765*27905*117870
270DATA1, 1,1,1*1*1
280DATA190*310
281DATA81
290READM*M
300U=M
320F0RJ=1TCM
330F0RI=1TON
340KEADHCJ* I)
350NEXTI
360NEXTJ
370F0RJ=1TOM
380F0RK=1T0M
390FORI=1T0N
400ECJ*X)=E(J*K)+HCJ*I)*h
500CCI)=0
510NEXTI
520REM"R"WILLBEUSEDT0KEEPTRACK0FRUNS
530R=R+1
540INPUT:DAYl:X
550IFX>1E6THEN1020
560INPUT:DAY1 :3*C*D
570IFB=6THEN700
580IFB=8THEN610
590PRINT"NOGO"
600ST0P
610F0RJ=1T0N
620I=J*2
630IFX=>WCI-1 )THEN650
640G0T0540
650IFX>W< I )THEN680
660C(J)=CCJ>+C*10tD
670G0T0540
680NEXTJ
690G0T0540
84
-------
TABLE A-6 (Continued)
700F0RJ=1T0U
720PC J)=P< J)+H< J*I)*C;ECJ*K>
780MEXTK
790NEXTJ
800F0RJ=1T0U
810PRINT:C0EFFxLNMC 1000+J) JPCJ)
320NEXTJ
830REST0RE:CeEFF:
840M=U-1
350MATINPUT:C0EFF:ACM*M>
860MATB=CGiMCM*M)
870MAT8=INVCA)
880MATY=C0N(0*M)
890MATZsC0NCM*0)
900MATINPUT:C0EFF:X(M*0)
910MATZ=B*X
920MATY=TRN(Z)
930F0RJ=OT0U-1
940V(R> J)=Y<0* J>
950G< J)=YCO* J)
960.MEXTJ
9703CRATCH:C0EFF:
980F0RJ=1T0U
990P( J)=0
1000NEXTJ
1010G0T0480
1020F0RI=OT0U-1
1030F0RJ=1T0R
1040T(I)=T(I )-«-V(J*I>
1050NEXTJ
1060TCI)=(CINTCC100*T(I)/(R-1 ) )+.5)
1070NEXTI
1030PRIMTTABC5)"'NATURAL BACKGR0UND"
1090F0RJ=1T024
1 100PRINT"-"J
1 1 10NEXTJ
1 120PRINT
1 130PRIiMT"PROTEINS"*TCO)
1 140PRINT"CARB0HYDRATES"*T< 1 )
1 150PRINT"LIPIDSft*T
85
-------
TABLE A-6 (Continued)
1160F0RI=1T0N
1170SCI)=SCI)/=ECJ,K)=0
1230NEXTK
1240NEXTJ
1250T9=1E4
1260N9=N
1270KEADM9
1280U9=M9
1290F0RJ=1T0M9
1300F0RI=1T0N9
1310READQCJ,!)
1320Q(J*I)=Q(J*I)/T9
1330NEXTI
1340NEXTJ
13SOF0RJ=1T0N9
1360READLCJ)
1370NEXTJ
1380F0RJ=OT0M9-1
1390READFCJ)
1400NEXTJ
1410F0RT5=1T0U9
1420F0RT1=1T0N9
1430T2=LCT1)
1440F0RT3=1T0T2
1450HCT5*T4+T3)=QCT5>T1)
1460NEXTT3
1470T4=T4+T3
1480NEXTT1
1490T6=T4
1500T4=0
1510NEXTT5
1520F0RJ=1T0U9
1530F0RK=1T0U9
1540F0RI=1T0T6
1550ECJ>K)=ECJ*K)+HCJ*I)*HCK*I)
1560NEXTI
1570NEXTK
1580NEXTJ
1 590F0RI = 1T0.N9
1600PCI)=R(I)=0
1610.MEXTI
1620REM"R"WILLBEUSEDT0KEEPTRACK0FRUNS
1630R9=R9+1
86
-------
TABLE A-6 (Continued)
1640INPUT:DAY2:X
1650IFXME6THEN2180
1660 INPUT:DAY2:B,C>D
1670IFB=6THEN1800
1680IFB=8THEN1710
1690PRINT"Y0U SCREWED UP 0N RUB
1700ST0P
1710F0RJ=1T0N9
1720I=J*2
1730IFX=>W(I-1)THE.M1750
1740G0T01640
1750IFX>W=P*C(I>
1920NEXTI
1930NEXTJ
1940F0RJ=1T0U9
1950F0RK=1T0U9
1960PRINT:C0EFF:LNMCJ>JE(J*K)
1970NEXTK
1980NEXTJ
1990F0RJ=1T0U9
2000PRINT:C0EFF:LNMC1000 + J)5 PCJ)
2010NEXTJ
2020M9=U9-1
2030REST0RE:C0EFF:
2040MATIN?UT:C0EFF:A(M9*M9)
2050MATB=C0.MCM9*M9>
2060MATB=Ii\JV(A)
2070MATY=C0NCO*M9)
2080MATZ=C0NCM9*0)
2090MATINPUTSC0EFFtX(M9>0)
87
-------
TABLE A-6. (Continued)
2100MATZ=B*X
2110MATY=TRN(Z)
2120F0RJ=OT3U9-1
2130YCO* J)=(INTCC10000*CY(0»J)/T9 »«•• 5»/100
2140V(R9*J)=YCO»J>
2150NEXTJ
2160SCRATCH:COEFF:
2170G0T01590
2180F0RI=OTOU9-1
2190F0RJ=1T0R9-1
2200UCn=UCI)+VCJ*I)
2210NEXTJ
2220UCI) = CINTC(100*CLKI>/CR9-l>» + .5>>/100
2230NEXTI
2240PRI.MTTA3C10>"V0LUME PERCENT FR0M EACH SOURCE"
2241PR1NTTABC10)" ----.-- '. ....1...--••
2242PRINT
2243?RINTTAB(10)MRUNS"
2250PRINT"S0URCE"I
2260F0R.J=1T0R9-1
2270PRINTTABC(J-l)*8+ld)Jl
2280NEXTJ
2290PRINTTAB(CR9-1)*8*10)MA\/ERAGE"
2300F0R.J=1T0(CR9-1 )*fi+t7)
2310PRINT"-";
2320NEXTJ
2330PRINT
2340F0RJ=0'»OU9-1
2350PRIMTSS(J+ni
2360F0RI=1T0R9-1
2370PRINTTABCCI-1)*8+10>VCI*J)3
2380NEXTI
2390PRINTTABCCR9-1)*8+10)UCJ)
2400NEXTJ
2410PRINTTTYC10)
2420PRINTTABC10>"CONCENTRATI0N 0F 0RGANIC CARB0N FR0M EACH S0URCE"
2430PRI>JTTA8C20)tIMG/L"
2435PRINTTABC10)" —* k "
2440PRINT
2450PRINTTABC10)"RUMS"
2455PRINT"S0URCE"J
2460F0RJ=1T0R9-1
2470PRINTTAB(CJ-1)*8+10)J«
2480NEXTJ
2490PRI(MTTAB((R9-1>*8*IO)"AVERAG£M
88
-------
TABLE A-6 (Concluded)
2500F0RJT1T0CCR9-1>*8+l7)
•2510PRINT"-"J
2520NEXTJ
2530PRIiMT
2540F0RJ=OT0U9-1
2550PRINTSSCJ+1)J
2560F0RI=1T0H9-1
2570VCI*J)=VCI*J)*FCJ>/100
2530PRINTTABC 8*(I- 1) + 10)CIMT(100* V(l«J) + .5))/100;
2590NEXTI
2600U
-------
Example of BAJA Computations
To illustrate operation of the BAJA program, computer and longhand
calculations were performed on data generated as a result of anal-
ysis of aqueous solution containing 20-ppm albumin, 10-ppm glucose,
and 1-ppm oleic acid. A triplicate analysis was performed, gener-
ating input data (DAY), as shown in Table A-7.
90
-------
TABLE A-7. INPUT DATA (FILE DAY)
90 "MIXTURE 8"
100 0182*8
110 8635*0
120 0214*8
130 5666*0
140 0297*8
150 8742*0
160 0367*8
170 6012*0
180 0596*8
190 0837*0
200 1413*8
210 4968*0
220 2980*6
230 0000*0
330 0178*8
390 8690*0
400 0212*8
410 5578*0
420 0290*8
430 8800*0
440 0359*8
450 6260*0
460 0583*8
470 1100*0
480 1440*8
490 4895*0
500 2983*6
510 0000*0
520 0180*8
530 8930*0
540 0215*8
550 5777,0
560 0290*8
570 8871*0
580 0359*8
590 6189*0
600 0434*8
610 0001*0
620 0578*8
630 1103*0
640 1430*8
650 2637*0
660 1431*8
670 2637*0
680 3000*6
690 0000*0
700 1E10
91
-------
The data, when analyzed by computer program BAJA, yielded the following
results:
TABLE A-8
MIXTURE 8
MG PER LITER
SAMPLE
1
2
3
ALBUMIN
19.15
18.69
19.53
GLUCOSE
9.23
9.34
9.51
0LEIC
.92
.99
.96
%FIT
96
96
96
AVERAGE
19.12
9.36
.96
Data resulting from an analysis of the first sample were also computed
(longhand) by solving the following equations:
n
s a..y.
i=l
n
a. c, + E a..a._ c_ + E a...a._ c
l \i=l ll l2 2 \i-l ll l3/
\ /- n ,\ _ / n
i=i
E a.a.. I 6. + E a._a._ c_ + E a._ c
The analysis of sample yielded six peaks for a mixture containing three
compounds (albumin, glucose, and oleic acid). The following set of
coefficients is taken from the calibration table (page 33):
al
a2
a3
a4
a5
afi
Cl
Albumin
174
263
236
185
17
167
C2
Glucose
449
64
16S
71
18
9
C3
Oleic Acid
1198
216
3154
169$
687
1615
92
-------
The values for "y," as transcribed from the first pyrographic
analysis (lines 110, 130, 150, 170, 190, and 210 from Table A-7)
are :
y1 8635
y2 5666
y3 8742
y4 6012
ys 837
y6 4968
The coefficients for Eq. A-l are calculated
n 2
Z a. = 217,544
n n
Z a., y. = 7,011,865 I a., a.0 = 148,842
. , il 'i ' .,1112 '
=l 1=1
n n
Z a.- y. = 6,168,799 Z a., a., = 1,605,303
l2 x U l3
n n -
Z a y. = 57,953,581 Z a = 238,368
1=1 " x 1=1 ^
n
Z a a = 129,666
J a./ = 17,396,371
93
-------
From this, three linear equations are formed:
7,011,865 = 217,544 C + 148,842 C2 + 1,605,303 C3
6,168,799 = 148,842 C + 238,368 C + 1,219,666 C
J. ^ O
57,953,581 = 1,605,303 GI + 1,219,666 C2 + 17,396,371 C^
Algebraic solution of those linear equations yields:
C1 = 19.2 (Albumin)
C2 = 9.2 (Glucose)
C3 = 0.92 (Oleic Acid)
Which agrees with the results of the computed analysis (Table A-8)
94
*U.S. GOVERNMENT PRINTING OFFICE:1973 514-155/317 1-3
-------
1
Accession Number
w
5
f\ Subject Field & Group
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Rocketdyne, a Division of North American Rockwell Corporal
ion
Canoga Park, California
Title
Pyrographic Gross Characterization of Water Contaminants
10
Authors)
Lysy j ,
Newton.
Ihor
Peter R.
16
21
Project Designation
EPA, Contract No.
14-12-802,
Project No.
16040 EXD
Note
22
Citation
Environmental Protection Agency Report No. EPA-R2-73-227, May 1973
23
Descriptors (Starred First)
*Water Pollution Monitoring,
*Waste Source Identification and Quantification,
Pyrographic Methodology, Analysis of Natural Waters
I Identifiers (Starred First)
*Waste Source Identification and Quantification
27
Abstract A hydrochcmical instrument and methodology were developed for direct analysis of
organic materials in aqueous solutions based on thermal fragmentation followed by gas
chromatographic separation and detection of the resulting derivative compositions. The
applications of the developed technique to water pollution surveillance, optimization of
waste treatment processes, and characterization of natural waters were studied. It was
found that a recorded pattern of pyrolytically produced fragments for a given water sample
reflects the total nature of its organic composition, and can be interpreted and differen-
tiated in a number of ways. Using a pviovi established calibration patters for individual
components to be found in a mixture, the pattern produced by a mixture can be analyzed
mathematically. The system can be calibrated and the data can be interpreted in terms of
pure organic compounds, classes of organic materials, or any other arbitrarily defined
organic mixtures such as those found in industrial waste effluents. Application of this
technique to pollution surveillance is based on the fact that each industry has a waste
whose chemical composition is distinctive to its operation. The uniqueness of such waste
compositions offers a means for their characterization as separate entities, rather than
as a collection of various chemical species. The validity of this postulate was experi-
mentally demonstrated. With this method, both the identity of the source and the quantity
of the waste contributed by each source could be determined.
Abstractor
Ihor Lysyj
Institution
Rocketdyne
WR:102 (REV. JULY 1969)
WRSIC
SEND. WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240
» SPO: 1 9»U-389-930
------- |