EPA
United States Environmental Monitoring and Support EPA-600/4-79-038
Environmental Protection Laboratory June 1979
Agency Cincinnati OH 45268
Research and Development
Functional
Specifications for an
Advanced
Chromatography
Automation System
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/4-79-038
June 1979
FUNCTIONAL SPECIFICATIONS FOR AN
ADVANCED CHROMATOGRAPHY AUTOMATION SYSTEM
by
Jonathan E. Kopke
Southwestern Ohio Regional Computer Center
University of Cincinnati
Cincinnati, Ohio 45220
Contract No. GS-05S-10458
Project Officer
John M. Teuschler
Physical and Chemical Methods Branch
Environmental Monitoring and Support Laboratory
Cincinnati, Ohio 45268
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and Sup-
port Laboratory-Cincinnati, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily re-
flect the views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorsement or
recommendation for use.
n
-------
FOREWORD
Environmental measurements are required to determine the quality of
ambient waters and the character of waste effluents. The Environmental
Monitoring and Support Laboratory - Cincinnati, conducts research to:
+ Develop and evaluate techniques to measure the presence
and concentration of physical, chemical, and radiological
pollutants in water, wastewater, bottom sediments, and
solid waste.
+ Investigate methods for the concentration, recovery, and
identification of viruses, bacteria and other microbiological
organisms in water; and to determine the responses of aquatic
organisms to water quality.
+ Develop and operate an Agency-wide quality assurance program
to assure standardization and quality control of systems for
monitoring water and wastewater.
+ Develop and operate a computerized system for instrument
automation leading to improved data collection, analysis,
and quality control.
This report was developed in the Advanced Instrumentation Section of the
Environmental Monitoring and Support Laboratoary in the interest of
advancing laboratory techniques and quality control through computerization.
Dwight G. Ballinger
Director
Environmental Monitoring and Support
Laboratory - Cincinnati
-------
ABSTRACT
This document contains a project definition, a set of functional
requirements, and a functional design for a system which will link a
commercial chromatography data system to the EPA Laboratory Automation
System.
A Varian 220L Chromatography Data System was selected as the prototype
system to be extended in this project, although these specifications can be
adapted to other commercial systems. The current methods of using the
Varian system are briefly described in this report. The bulk of the report
is a detailed list of the additional functions to be performed by the EPA
Laboratory Automation System. These functions include multi-point
calibration, calculation of concentrations, identification of compounds,
calculation of relative retention times, and calculation of quality control
statistics. A general plan for the proposed system is provided.
This report was submitted in partial fulfillment of Contract No. GS-05S-
10458 by Southwestern Ohio Regional Computer Center. This report covers work
done between November 13, 1978 and January 9, 1979.
iv
-------
CONTENTS
Foreword i i i
Abstract i y
Acknowl edgments vi
1. Introduction 1
2. Project Definition 3
3. Functional Requirements, Overview of the Project
Scientific Principles of Chromatography 4
Instruments to be Automated 4
Description of the Existing System 5
Additional Existing Hardware and
Software to be Utilized 6
Time and Financial Considerations 6
Support of the Completed System 6
Future Considerations 7
4. Functional Requirements, Requirements of the System
Data Transmission and Storage Requirements 8
Description of the Varian Reports 8
Significant Digits 14
Flags for Use in Data Transmission 14
Injection ID Numbers 14
Editing of the Varian Report 15
External Standards Calibration Method 15
Calibration Curve Plotting 16
Concentration Calculations Using
External Calibration 17
Identification of Compounds 17
Relative Retention Times and
Capacity Ratios 19
Processed Data Report 19
Quality Control Statistics
Control (Check) Standards 19
Spiked Samples 20
Duplicate Samples 20
Surrogate Spi kes 20
Dissimilar Analysis Confirmation 20
Replicate Statistics 21
Sample File Control Interaction 21
5. Functional Design
Description of an Automated Run 23
Program Modules 24
File Descriptions 24
6. Signoff Sheet 28
-------
ACKNOWLEDGMENTS
The following people contributed greatly to the compilation of the
functional requirements in this document: Thomas Bellar, Denis Foerst,
William Potter, Leown Moore, and William Budde. Advice on systems develop-
ment methodology was received from John Teuschler and Dennis Ryan. The
assistance of these colleagues is gratefully acknowledged.
vi
-------
SECTION 1
INTRODUCTION
These functional specifications describe a prototype system for linking
a commercially available chromatography data system to the EPA Laboratory
Automation System.
A chromatography data system (such as the Varian 220L, the Perkin-Elmer
Sigma 10, or the Hewlett-Packard 3380 or 3385) is an efficient system for
real time data acquisition from one or more chromatographs, and for identi-
fying compounds and determining their peak areas. However, programming any
of the above commercial systems to perform additional or more sophisticated
data reduction and to communicate with the EPA Sample File Control System
would be a formidable, if not impossible, task.
Therefore, the system described in this document is designed to accept
data from a chromatography data system, perform a variety of processing with
the data, and ultimately report the results to Sample File Control, as shown
in Figure 1. Although these functional specifications are written specifi-
cally for the Varian 220L at EMSL, Cincinnati, they can be applied in
general to other commercial chromatography data systems.
-------
ro
1GAS
CHROMATOGRAPH #1
I GAS
CHROMATOGRAPH #2
GAS
CHROMATOGRAPH #3
GAS
CHROMATOGRAPH #4
GAS
CHROMATOGRAPH ttn
GAS
CHROMATOGRAPH #1
GAS
CHROMATOGRAPH #2
GAS
CHROMATOGRAPH #3
GAS
CHROMATOGRAPH #1
Figure 1. Poter
laboratory autc
system.
\/ A R 1 A M
220L
Chromatography
Data System
SIGMA 10
Data System
H/P 3380
Chromatography
Data System
EMISSION
SPECTROMETER
ATOMIC ABSORPTION
UNIT #1
ti al
ATOMIC ABSORPTION
UNIT #2
ULTRAVIOLET/ VISIBLE
SPECTROMETER
TECHNICON
AUTOANALYZER #1
)mation
TECHNICON
AUTOANALYZER #2
TECHNICON
AUTOANALYZER #3
TOTAL ORGANIC
CARBON ANALYZER *""
Data General
Nova 840
Computer
(Data Acquisition
and Reduction)
Digital Equipment
POP 11/70
Computer
(Sample File Control)
-------
SECTION 2
PROJECT DEFINITION
The goal of this project is to develop a computer automation system
which will accept data from a commercial chromatography data system and do
additional processing of the data. The system will allow for:
+ unattended automatic transfer of data from the commercial system to
the EPA Laboratory Automation System,
+ calibration using regression on multiple internal or external
standards,
+ calculations of concentrations and relative retention times,
+ identification of compounds,
+ timely EPA standard quality control checks,
+ dissimilar analysis confirmation,
+ replicate statistics on samples,
+ detailed printed reports and calibration plots, and
+ transfer of results to the Sample File Control System.
-------
SECTION 3
FUNCTIONAL REQUIREMENTS
OVERVIEW OF THE PROJECT
SCIENTIFIC PRINCIPLES OF CHROMATOGRAPHY
Chromatography is a technique of separating the various components of a
sample so that they can be identified and measured. Basically, a chromato-
graph consists of a tube filled or coated inside with a material (the
stationary phase) through which a steady stream of gas or liquid (the moving
phase) is forced. When a sample is introduced into the tube, the moving
phase forces the sample through the stationary phase and out the end of the
tube. However, the stationary phase impedes the passage of some compounds
more than others, and therefore the different compounds in the sample arrive
at the end of the tube at different times. A detector at the end of the
tube measures the amount of matter leaving the tube, and that variable is
plotted against time on a recording device.
The retention time of a compound is the length of time that the compound
is retained in the stationary phase. A compound may be identified in
chromatography by comparing its retention time to the retention times of
known compounds. Furthermore, the concentration of a compound in a sample
is calculated by comparing the area of its recorded peak to the area of the
peak produced by a known concentration of the same compound.
INSTRUMENTS TO BE AUTOMATED
Tracor 560 Gas Chromatograph, EMSL-CI, Room 588
Infotronics 2400 Gas Chromatograph, EMSL-CI, Room 588
Tracor MT220 Gas Chromatograph, EMSL-CI, Room 586
Tracor MT222 Gas Chromatograph with auto-injector, EMSL-CI, Room 586
Hewlett-Packard 5700 Gas Chromatograph, EMSL-CI, Room 586
Waters Associates 440 Liquid Chromatograph, EMSL-CI, Room 574
Varian Aerograph 1400 Gas Chromatograph, MERL, Room B22
Varian Aerograph 2100 Gas Chromatograph, MERL, Room B22
Two Tracor MT222 Gas Chromatographs, MERL, Room B22
Perkin-Elmer 900 Gas Chromatograph, HERL, Room 670.
All of these instruments are to be automated using the Varian Chroma-
tography Data System 220L in Room 574 for initial processing of data. All
room numbers refer to the Environmental Research Center, Cincinnati.
-------
DESCRIPTION OF THE EXISTING SYSTEM
The chromatographs in EMSL, MERL, and HERL have already been automated
to some extent. Currently, several gas chromatograph-s and a high pressure
liquid chromatograph are interfaced to a Varian computer. To use this auto-
mation system, an analyst must first generate a method in the Varian. This
involves providing the Varian with the retention times and eight-character
identifiers for the compounds which can be expected in the standards and
samples. The method can be retained by the Varian and used repeatedly.
Standards and samples are prepared in three different ways for different
applications. The type of preparation influences the calculations which
must be done to find true concentrations. The three types of preparation
are:
Liquid/Liquid Extraction
The compounds of interest are extracted from the aqueous sample by an
organic solvent. The extract is then concentrated, and a small portion of
it is injected into the chromatograph. (A standard may be extracted from an
aqueous solution, or prepared at a known concentration in the organic sol-
vent.)
Purge and Trap Method
The compounds of interest are extracted from the aqueous sample by an
inert gas. The compounds are trapped in a short column and then backflushed
into the chromatograph. (In this case, a standard is prepared at a known
concentration in an aqueous solution and is treated in exactly the same way
as a sample.)
Direct Aqueous Injection
A small portion of the actual aqueous sample is injected directly into
the chromatograph. (A standard is treated in the same way, although it may
be dissolved in an organic solvent rather than in water.)
After any method of preparation and injection, a sample or standard
travels through the chromatograph tube. The various compounds are separated
and ultimately pass through a detector. The signal from the detector is
transmitted to the Varian and digitized. The Varian determines the reten-
tion time of each peak, the total area of all of the peaks, and the percent
of the total area for each peak. From the retention times provided by the
analyst during method generation, the Varian determines what compounds are
present in the sample or standard, and it prints the eight-character identi-
fier for each compound found. The Varian also provides a variety of other
information including a simple calculation of the concentration of each
compound based on a one-point calibration.
To calculate concentrations more accurately, an analyst must first
develop a calibration curve for each of the compounds covered by the
method. A calibration curve for chromatography relates peak area to the
-------
concentration (or mass) of the compound. To establish a calibration curve,
an analyst normally uses several calibration standards, and may use several
replicates of each of the standards. In a single standard, the concentra-
tions of the various compounds are not necessarily the same. A calibration
curve is most often a straight line; however, other curves are sometimes
employed.
The concentrations of compounds in samples are calculated using the
calibration curves, although other factors must be taken into account such
as the volume of the injection and (in liquid/liquid extraction) the volumes
of the extract and the original aqueous solution.
Calibration standards are also used as control (check) standards. They
are compared to the analyst's previous experience to determine if the
instrument is in control.
There are several other complications unique to chromatography. Some-
times one peak represents the total of several compounds. In such a case,
the analyst must match the eight-character identifier on the Varian report
to the names of several compounds for the final report. Also, in some
cases, one pesticide may be a mixture of several compounds. This implies
that several peaks represent one pesticide. In such a case, the analyst
must total several peaks in the calibration standards, calibrate on that
basis, and then total the corresponding peaks from the samples to calculate
the concentration of one pesticide.
ADDITIONAL EXISTING HARDWARE AND SOFTWARE TO BE UTILIZED
This chromatography project is one branch of a much larger laboratory
automation project. Therefore, some of the basic decisions about system
hardware and software have already been made within the context of the
larger project. Thus, these functional requirements are not strictly hard-
ware and software independent.
This project will use an existing Data General Nova 840 Minicomputer
with the operating system RDOS (Revision 6.2) and the language Extended
BASIC (Revision 4.3). Varian output reports will be transferred to the Nova
for further data processing.
TIME AND FINANCIAL CONSIDERATIONS
The first version of the gas chromatography automation system should be
operational for testing by March 1, 1979. The system described in the func-
tional requirements should be completed and documented by June 30, 1979. No
more than $30,000 should be spent on developing the system.
SUPPORT OF THE COMPLETED SYSTEM
Support of an automated system is divided into software and hardware
support. Since this project will involve two computer systems, the hardware
and software support of each system will be discussed.
-------
The Advanced Instrumentation Section will provide assistance in the
selection and diagnostic analysis of computer terminals used with both
systems; however, when it is cost effective, repairs to the terminals will
be made by the manufacturer's representatives. The purchase order for such
repairs will be completed by and billed to the user's section or branch.
The Varian computer system hardware is supported by the manufacturer
under a maintenance contract which covers all hardware except the terminals.
If additional hardware is necessary to implement this project, the custom
hardware will be documented by the Advanced Instrumentation Section in such
a way as to allow repair on a module-swapping basis, and spare cards will be
provided with any custom hardware installation.
The system software on the Varian has had no maintenance for several
years. The bugs which exist in the system are known and have been avoided
procedurally.
The Nova computer system also has a maintenance contract for all vendor-
supplied hardware. Custom hardware on the Nova system is maintained by
Advanced Instrumentation personnel.
The BASIC software and system software for the Nova computer are sup-
ported by the Data General Corporation. Any custom modifications to the
BASIC software (or assembly language code for a microprocessor) will be
supported by the Advanced Instrumentation Section, as will the applications
software.
The maintenance of the chromatograph instruments is the responsibility
of the users of those instruments.
FUTURE CONSIDERATIONS
The goals of this automation project are limited to those stated in the
functional requirements. However, the system should be designed to allow
for future enhancements to be easily implemented.
A major future consideration is that dedicated computers other than the
Varian will eventually be interfaced with the system. These other computers
include the Perkin-Elmer Sigma 10, the Hewlett-Packard 3380, the Hewlett-
Packard 3385, and the Auto lab chromatography system. The software of this
system will be modularized to facilitate the addition of other dedicated
computers, and the mode of transferring information from the dedicated
computers to the Nova will be generalized.
Another future consideration is that of the internal standards calibra-
tion method. This automation system will be designed to include modules for
the internal standards method. However, the design and implementation of
these modules will be deferred until the rest of the project is operational
and until the internal standards method is more rigorously defined for con-
ventional gas chromatography.
-------
SECTION 4
FUNCTIONAL REQUIREMENTS
REQUIREMENTS OF THE SYSTEM
DATA TRANSMISSION AND STORAGE REQUIREMENTS
Eleven chromatographs will be connected to the Van'an. Servicing all of
these instruments, the Varian will produce at most 115 reports per day
during normal working hours, plus 40 more reports during the night. Each
report will be no more than 100 lines long, and 30 lines will be the average
length.
There are currently about 28 Varian methods in use, although in the
future this number will increase. Each method deals with no more than 60
named compounds, although the average number will be about 15. Presently,
fewer than 1000 different compounds are being identified by chromatograph.y,
although this number will also increase substantially in the future.
Data transmission from the Varian will be accomplished as quickly as
possible, since the Varian will lock a user out of its system until that
user's previous report has been transmitted. This data transmission will
occur without human intervention to accommodate an auto-injector running
during the night.
The Nova will save the 100 most recent Varian reports for each chromato-
graph, and will automatically delete old reports. The Nova will keep a list
of the ID numbers for the existing reports, and this list will be available
for the analyst to examine.
All Varian reports will be saved in one programming area in the Nova,
since they can be distinguished by their ID numbers.
DESCRIPTION OF THE VARIAN REPORT
Normal Report
Figure 2 shows a typical Varian report, as described below.
LINE COLUMN CONTENTS
1 1-2
3-4
5-26
blank
GC
blank
-------
2
3
4
5
6
7
8
27-34 date, left justified
35-47 blank
48-52 time, left justified
1-70 blank
1-2 blank
3-6 INST
7-10 blank
11-12 instrument number, right justified
13-24 blank
25-30 METHOD
31-32 blank
33-36 method
37-39 method
40-47 blank
48-55 operator identification, left justified
number, right justified
number modifier, usually -ES
1-70 blank
1-16 blank
17-46 title, left justified
1-70
1-70
1-2
3-6
7-10
11-14
15-17
18-21
22-28
29-32
33-37
38-41
42-45
46-49
50-56
57-60
61-64
65-70
1-2
3-4
5-28
29-30
31-37
38-41
42-45
46-49
50-56
blank
blank
blank
PEAK
blank
NAME
blank
TYPE
blank
COMP
blank
RETN
blank
CORR
blank
AREA
blank
FACTOR
blank
NO
blank
UG (or
blank
TIME
blank
TIME
blank
other two-character unit abbreviation)
-------
57 %
10 1-2 blank
3-6 peak number, right justified
7-8 blank
9-16 compound identification (or UNEXP)
17-18 blank
19-20 peak type, left justified
21-22 blank
23-33 calculated concentration, right justified
34-35 blank
36-41 retention time, right justified
42-43 blank
44-49 corrected retention time, right justified
50 & (or blank)
51-52 blank
53-62 percent area, right justified
63-64 blank
65-70 response factor, right justified
Lines 11 to 10+N are identical to line 10. N is the number of peaks.
11+N 1-70 blank
12+N 1-22 blank
23-33 total concentration, right justified
13+N 1-70 blank
14+N 1-2 blank
3-12 TOTAL AREA
13-14 blank
15-23 total area, right justified
The Varian is capable of producing a variety of other formats for output
reports, but only the format described above will be used for this project.
Unavoidable Exceptions
Figure 2 shows a typical Varian report, but Figures 3 and 4 show unavoid-
able exceptions to the rules. In Figure 3, line 9 is a message from the
Varian system. Such a message may occur at any point in a report. A
message always begins with the word MESSAGE in columns 1 through 7, and
includes one or more message numbers which may be quite important to an
analyst.
The format of the report in Figure 3 shows a second unavoidable excep-
tion to the rules. When the Varian is unable to apply a method to a sample,
it defaults to this format to report the results. Such a report is not
useful in this project.
A third exception to the rules is the paging which occurs when a Varian
report exceeds 11 inches in length. In Figure 4, after the thirty-ninth
10
-------
GC
INST
PEAK
NO
1
2
3
4
5
6
7
8
9
10
11
12
13
TOTAL
1
NAME
UNEXP
CNCL
UNEXP
UNEXP
11CL2C=C
CCL4
12CLC3+
UNEXP
CL3CH=C
CHBR2CL
UNEXP
2BRCLC3@
CHBR3
AREA
11/1/78
METHOD 100-ES
14:10
AP&T
1 DAY SPIKE #2 CFT
TYPE
P
PB
P
P
P
PB
P
P
P
P
P
P
PB
503990
COMP
UG
0.5065
0.0000
1.8531
2.2701
0.0000
42.9654
8.6511
7.2651
50.6580
61.7029
2.7222
0.0000
2.8358
181.4301
RETN
TIME
764
790
1046
1057
1095
1369
1405
1415
1487
1530
1568
1605
1690
CORR
TIME
42
68
325
336
374&
640&
676&
686
766&
808&
835
872&
958&
AREA
%
1.0049
3.8631
3.6770
4.5044
9.0605
14.0213
17.1650
14.4149
10.7719
11.9360
5.4012
3.8208
0.3587
FACTOR
1000
0
1000
1000
0
6080
1000
1000
9331
10257
1000
0
15685
Figure 2. Normal Varian report
11
-------
GC
INST 2
MESSAGE 00, 23,
PEAK NAME
NO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
40/1000
TYPE
P
P
P
P
P
P
P
P
P
P
P
P
P
P
P
P
P
PB
P
P
P
P
P
11/21/78
METHOD 205- ES
STD #2
RETN CORR
TIME TIME
18
64
93
135
165
211
318
474
542
594
629
675
688
747
807
889
907
941
985
1038
1082
1157
1275
1293
09:17
P&T
AREA
%
16.4304
0.7135
4.3922
0.9373
1.7541
4.2019
3.3525
6.5665
5.2780
5.1760
0.7541
0.2889
0.3839
6.1907
2.0400
5.6351
3.8524
0.7136
2.7393
1.7316
0.1564
7.1984
3.4223
4.8589
Figure 3. Error message and default format.
12
-------
27
28
29
30
31
32
33
34
35
36
37
38
39
PEAK
NO
40
TOTAL
123CLC3
CL4C=C
12CL2C4
UNEXP
CLBZ
UNEXP
UNEXP
UNEXP
UNEXP
UNEXP
UNEXP
UNEXP
UNEXP
NAME
UNEXP
AREA
P
P
P
P
P
P
P
P
P
P
P
P
PB
TYPE
3579640
0.2947
259.4929
92.0854
20.2071
154.2618
12.1952
13.5698
14.2975
10.5835
9.5953
4.4091
5.0859
3.5579
COMP
UG
0.5064
1195.8286
1782
1808
1857
1960
2006
2015
2046
2104
2152
2181
2571
2616
2676
RETN
TIME
2735
1073&
1099&
1145&
1249
1295&
1304
1336
1394
1443
1472
1865
1911
1971
CORR
TIME
2035
4.1181
3.2643
2.3729
5.6450
4.9745
3.4068
3.7908
3.9941
3.9566
2.6805
1.2317
1.4208
0.9939
AREA
%
0.1414
20
22207
10841
1000
8663
1000
1000
1000
1000
1000
1000
1000
1000
FACTOR
1000
Figure 4. Paging feature of the Varian report.
13
-------
peak, the Varian skipped 12 lines, typed a new set of column headings, and
then continued the report. Possibly this paging feature can be disabled
during system generation on the Varian. (See the Varian manual, page 2-5,
i tern 5.)
SIGNIFICANT DIGITS
On the Varian reports, only the three high order digits of each number
are significant.
FLAGS FOR USE IN DATA TRANSMISSION
The following flags will be used to identify the beginning of a Varian
report, the end of the report, and the format of the report.
The first line of a Varian report always begins with the letters GC in
columns 3 and 4.
The last line of a report always begins with the words TOTAL AREA in
columns 3 through 12. If for some unexpected reason, a report does not
contain a line beginning with the words TOTAL AREA, then the next appearance
of the letters GC in columns 3 and 4 would indicate that the end of the
report has been missed, and another report is beginning.
Only one report format will be acceptable in this project. The flag to
be used to identify valid reports is the line of column headings which reads,
PEAK NAME TYPE COMP RETN CORR AREA FACTOR
as described previously. If any report reaches the Nova in a format other
than this, a notation will be made of its arrival, but the contents of the
report will be ignored.
INJECTION ID NUMBERS
Injection ID numbers will be eight-digit numbers of the form IIMMDDSS,
where II = instrument number, MM = month, DD = day, and SS = sequence
number. The Nova will automatically assign an ID number to an injection
when its Varian report reaches the Nova. The instrument number will be
obtained from the Varian report. The month and date will be acquired from
the Nova system clock, except in a case noted below. The sequence number
will be generated by a counter maintained for each instrument by the Nova.
The counter will be automatically reset to 1 when the date of an injection
is different from the date of the previous injection. Thus, for example,
the fifth injection into instrument number seven on January 23 would be
assigned ID number 07012305.
Using this system, ID numbers will be assigned by the Nova without human
intervention and without the analyst having to type them into the Varian.
Nevertheless, the analyst will be able to manually determine the ID number
of each injection immediately when he injects it (or when he loads it into
the sample wheel of the auto-injector).
14
-------
This system requires one further refinement to keep injection ID numbers
organized when a series of injections continues for more than one calender
day (an auto-injector running past midnight, for example). As mentioned
above, the sequence number counter for an instrument will be automatically
reset to 1 when the date changes. However, if an analyst specifies a date
(in the form DATEMMDD) in the "operator identification" field of the Varian
report, the Nova will use that date in the ID numbers it generates, instead
of the date in the system clock. Thus, the sequence number counter will not
be reset until the analyst changes the date specified. Analysts who do not
need this option will simply not begin the "operator identification" field
with the keyword "DATE."
EDITING OF THE VARIAN REPORT
The entire Varian report will be printed on paper and will be transfer-
red to the Nova. However, the report can be pruned considerably before it
is stored in the Nova. The programs in the Nova must store the following
information for each sample:
1. instrument number
2. method number
3. operator identification
4. title
5. total area
The Nova programs must also store the following information for each peak
within a sample:
1. name of compound
2. type of peak
3. retention time
4. percent area
Furthermore, any MESSAGE from the Varian system must be saved in the
Nova and displayed when the report is used so that the analyst does not have
to check the original Varian report for messages.
None of the other information on the Varian report is needed for the
calculations to be done by programs in the Nova.
EXTERNAL STANDARDS CALIBRATION METHOD
The existing calibration procedure in the Varian will not be used in
this project. The Varian will deal with all calibration standards as if
they were ordinary samples, and the Varian reports for calibration standards
15
-------
will be stored in the Nova until the analyst is ready to execute the cali-
bration procedure.
The Nova programs will allow for up to 15 calibration standards,
including replicates. Curve fitting will be accomplished using one of the
following methods, at the option of the analyst: 1) least squares linear
regression using three or more standards, 2) least squares quadratic regres-
sion using four or more standards, or 3) linear interpolation using the
origin and one or more standards. The analyst will be able to recalibrate
at any time by deleting standards and/or introducing new standards.
Two variations of external calibration will be available to allow for
the different methods of sample preparation.
Purge and Trap Method
When an analyst is ready to run the calibration procedure in the Nova
for standards prepared by the Purge and Trap method, it will be necessary to
enter into the Nova from the keyboard the injection ID number for each of
the calibration standards and the concentration of every named compound
within each standard. The analyst will be permitted to enter the concentra-
tion of standards in micrograms per liter. The program in the Nova will
acquire the necessary peak areas from the Varian reports, and will calculate
a set of calibration curve equations relating the peak area for each com-
pound to its concentration. The calibration curve equations will be
developed by linear regression, quadratic regression, linear interpolation,
or linear regression forced through the origin, at the option of the analyst.
A fitting error (percent relative standard error of estimate) will be
calculated so that the analyst can objectively judge the goodness of fit.
Liquid/Liquid Extraction and Direct Aqueous Injection Methods
The calibration procedure for standards prepared by Liquid/Liquid
Extraction or Direct Aqueous Injection differs from the first procedure in
that it is based on the mass of the standards rather than concentration. To
initiate a calibration process, an analyst will enter from the keyboard into
the Nova program the injection ID number of each standard, the injection
volume in microliters for each standard, and the concentration in micrograms
per liter. The program will then multiply the injection volume times the
concentration times 1000 to get the mass in nanograms for each compound.
Next, the program will develop a set of calibration curve equations relating
the peak area for each compound to its mass, and a fitting error will be
calculated as described above.
After either calibration procedure, the set of calibration curves for a
method will be stored in the Nova for repeated use, and the calibration
points will also be stored for plotting purposes.
CALIBRATION CURVE PLOTTING
The analyst will have the option of plotting a calibration curve for
each compound of a method. The plot will contain heading information
16
-------
including the name of the analyst, the name of the compound, the date of
calibration, the type of fit, the fitting error and the regression equation.
The plot itself will be a graph with concentration (or mass) on the hori-
zontal axis and peak area on the vertical axis. Axes will be labeled and
scaled. A table of residuals will also be provided.
All plotting will be done on existing Tektronix 4000 Series Computer
Display Terminals.
CONCENTRATION CALCULATIONS USING EXTERNAL CALIBRATION
Van'an reports for samples will be stored in the Nova until the analyst
is ready to calculate the concentrations for a set of them. Three different
types of concentration calculations will be available for the three methods
of sample preparation.
Purge and Trap Method
When an analyst is ready to calculate concentrations for samples
prepared by the Purge and Trap method, the injection ID numbers will be
entered into the Nova from a keyboard. The program running in the Nova will
calculate the concentrations in micrograms per liter for every named com-
pound within each sample by simply substituting the peak area from the
Varian report into the appropriate calibration equation.
Liquid/Liquid Extraction
Concentration calculations for the Liquid/Liquid Extraction method will
require the analyst to enter from a keyboard not only the ID number for each
injection, but also the volume of the injection in microliters, the volume
of the extract in milliliters, and the volume of the water extracted in
liters. The program running in the Nova will use the peak areas from the
Varian reports and the calibration equations to calculate the mass in nano-
grams of each compound. It will multiply this mass by the volume of the
extract, and then divide by the product of the volume of the injection and
the volume of the water extracted to determine the concentration of each
compound in micrograms per liter.
Direct Aqueous Injection
To calculate the concentrations of samples run by Direct Aqueous
Injection, the analyst will enter into the Nova from a keyboard the ID
number and injection volume in microliters for each injection. Using the
peak area from the Varian report and the set of calibration curves, the
program in the Nova will calculate the mass in nanograms of every compound
in each sample, and will then divide by the injection volume to get the
concentration in nanograms per microliter. Dividing by 1000 will produce
concentration in micrograms per liter.
IDENTIFICATION OF COMPOUNDS
The identification of compounds measured by gas chromatography is cur-
17
-------
rently hampered by the fact that the NAME field on the Varian output report
is only eight characters wide, while the chemical name of a compound may be
30 or more characters long. The following procedure will be used to enable
the Nova program to print the full chemical names on final output reports:
When an analyst generates a method in the Varian, each of the compounds
involved will be identified by a name consisting of eight or fewer alpha-
numeric characters. For example, an analyst may refer to bromodichloro-
methane as CHBRCL2. No conflict will result if two analysts happen to use
the same identifier for different compounds; the only restrictions will be
that 1) within a Varian method, no two compounds will be allowed to have the
same identifier, and 2) within two Varian methods which will be used for
dissimilar analysis confirmation, identifiers will have to be consistent.
The analyst will then create a corresponding method in the Nova. As
part of the Nova method-generation procedure, the analyst will be asked to
type the identifiers used in the Varian, and the matching chemical names and
Chemical Abstracts Services (CAS) Registry numbers. For example, the
analyst would have to tell the Nova program that CHBRCL2 represents bromo-
dichloromethane, and that the CAS Registry number for that compound is
75-27-4. The analyst will not be required to use any particular chemical
name for a compound, but rather will be able to use any synonym he prefers.
For example, dichlorobromomethane could be used instead of bromodichloro-
methane. In the event that one peak in a Varian method represents more than
one compound, the analyst will use one identifier for that group of com-
pounds, and will supply the Nova program with the names and CAS Registry
numbers for all of the individual compounds associated with that identifier.
Each analyst will manually look up the CAS Registry numbers needed in
the Chemical Abstracts Ninth Collective Index (1972-1976). Volumes 76
through 85 of that work are the Formula Index, which seems to be the most
convenient source of CAS numbers currently available. The user's manual for
this system will include a list of EPA priority pollutants and corresponding
CAS numbers.
When an analyst types a CAS Registry number into the Nova program, the
program will recalculate the value of the check digit as described in the
CAS Registry Handbook. This process will be used to check the number for
typing errors, but it will not be used to verify that the analyst has used
the correct CAS number for the compound named.
The final reports from the Nova program will include the full names and
CAS Registry numbers for the compounds measured, but will not include the
identifiers used by the Varian.
Because CAS Registry numbers may often be superfluous, the analyst will
have the option of not entering them into the Nova method at all. However,
CAS numbers may be required for further processing of the data in other EPA
computer systems, specifically the Sample File Control System.
18
-------
RELATIVE RETENTION TIMES AND CAPACITY RATIOS
As part of the method generation procedure on the Nova, the analyst will
elect either to have relative retention times or capacity ratios calculated
for samples. If the analyst chooses relative retention times, it will be
necessary to enter the Varian identifier for the compound to whose retention
time all other compounds are to be compared. Then when the analyst runs the
concentration calculation procedure, the Nova program will automatically
calculate the relative retention time for each compound using the retention
times from the Varian report and the formula:
Rel. Retention Time = Retention Time ,QQ
Retention time of Reference
If during Nova method generation, the analyst chooses to use capacity
ratios instead of relative retention times, it will be necessary to enter
the dead volume time for the method. Then when concentrations are calcu-
lated, the Nova will calculate the capacity ratio for each compound using
the retention times from the Varian report and the formula:
Capacity Ratio = Retention Time ,
Dead Volume Time " '
PROCESSED DATA REPORT
Processing an injection will involve determining the full chemical name
of each compound, calculating the relative retention time or capacity ratio
for each compound, and calculating the concentration of each compound.
After an injection has been processed, a report will be printed. This
report will contain the following heading information: 1) name of the
analyst, 2) the instrument number, 3) the method number, 4) the date of the
report, 5) the title of the injection from the Varian report, if any, and 6)
the injection number and date. The report will also include the following
data for each peak: 1) the full chemical name(s) and CAS Registry number(s)
for the compound(s) associated with the peak, 2) the peak type from the
Varian report, 3) the peak area, 4) the retention time, 5) the relative
retention time,or capacity ratio, and 6) the concentration of the compound
in micrograms per liter. Unexplained peaks will be listed in the report,
but will not be matched to a chemical name nor assigned a concentration.
QUALITY CONTROL STATISTICS
Quality control statistics will be calculated for control (check) stan-
dards, spiked samples, duplicate samples, and surrogate spikes.
Control or Check Standards
The analyst will enter from the keyboard into the Nova program the ID
number of a control (check) standard and the prepared concentration in
micrograms per liter of each compound within the standard. The Nova program
will print a report showing the prepared concentration, measured concentra-
tion, and percent recovery for each compound. Percent recovery for stand-
19
-------
ards will be calculated by the formula:
Percent Recovery = Measured Concentration
Prepared Concentration
Spiked Samples
The analyst will enter from the keyboard into the Nova program the ID
number of the unspiked sample, the ID number of the spiked sample, and the
amount of spike in micrograms per liter for each compound within the sample.
The Nova will print a report showing the measured concentration of the
unspiked sample, the measured concentration of the spiked sample, the amount
of spike, and the percent recovery of the spike for each compound. Percent
recovery for spikes will be calculated by the formula
Percent Recovery = Spiked Cone. - Unspiked Cone. ,m
Spike x IUU>
Duplicate Samples
The analyst will enter from the keyboard into the Nova program the ID
numbers of both members of the duplicate pair. The Nova will print a report
showing both of the measured concentrations and the absolute value of the
difference between them for each compound.
Surrogate Spikes
A surrogate spike is a known amount of a pure compound that is added to
a sample, but which was not previously present in the sample. The percent
recovery of a surrogate spike can be used as an estimate of the recoveries
of the compounds in the sample when actual spikes are not used. The pro-
cedure for processing a surrogate spike will be the same as that for a check
standard, except that only one compound will be involved.
DISSIMILAR ANALYSIS CONFIRMATION
Dissimilar analysis confirmation (also called multi-analysis merging)
involves verifying the presence of a compound in a sample by showing that
the compound has been detected by two or three different methods. This
process must also take into account the fact that a sample may be separated
into as many as three fractions by Florisil Column Adsorption Chromatography
before injection. In such a case, each of the fractions is injected into
the chromatograph separately, but all of the fractions are analyzed by the
same method.
To perform a dissimilar analysis confirmation, the analyst will be
required to enter from the keyboard into the Nova program the ID numbers of
the appropriate injections, and the Nova program will simply print a chart
similar to that in Figure 5. The Nova program will allow for either two or
three methods, and for either no fractioning, two fractions, or three
fractions within each method.
20
-------
The list of compounds will Include every compound appearing in any of
the methods. Into each position of the chart, the Nova program will print
one of the following: 1) the measured concentration, 2) a dash to indicate
that the compound was not detected in the injection, or 3) an asterisk to
indicate that the method does not apply to the compound.
REPLICATE STATISTICS
To calculate replicate statistics on a sample, the analyst will enter
from the keyboard into the Nova program the ID numbers for every replicate
injection of the sample. The Nova will print a report showing the average
(arithmetic mean) concentration of each compound within the sample.
SAMPLE FILE CONTROL INTERACTION
A user of the Gas Chromatography Automation System will be able to get
from the Sample File Control (SFC) system a list of samples which are to be
run and will be able to return to the SFC system the measured concentrations
of compounds in those samples.
The SFC backlog list will be obtained by the user directly from the SFC
POP-11/70 computer in the form of a lineprinter or terminal printout. After
all data processing is complete for the required samples, a module of the
Gas Chromatography system will be used to construct an SFC run results file,
as defined in the document, "EPA Sample File Control Functional Design,"
(specifically, Section I - Foreground BASIC SFC/Nova Design). The method/
parameter code for the header record of the run results file will be entered
manually by the user, who can copy it from the backlog list. The method/
parameter codes for the results records will be the CAS numbers of the com-
pounds measured. There will be one result record for each compound measured.
The user will have to manually match the SFC identification number for each
sample with the corresponding injection ID number. The program will then
collect the necessary data from the processed data files and method file in
the Nova, and fill the run results file.
The run results file will then be transmitted to the SFC system by other
programs which are not part of the Gas Chromatography Automation System.
21
-------
ro
ro
Method One
Method Two
Compound 1
Compound 2
Compound 3
Compound 4
Compound 5
Fraction A Fraction B Fraction C
1.5
* * *
3.2
7.5
0.2
Fraction A Fraction B
1.6
2.7
3.1
7.1
Fraction C
0.2
Compound n
3.1
0.1
Figure 5. Dissimilar analysis confirmation chart.
-------
SECTION 5
FUNCTIONAL DESIGN
DESCRIPTION OF AN AUTOMATED RUN
In operating this system, there will be no need for the analyst to
interact with both the Varian and the Nova at the same time. Rather, it
will be possible for the analyst to work with the Varian alone, and then
sign on to the Nova to do data processing after all of the necessary injec-
tions have been completed.
The Varian will be operated according to its normal procedures with the
only exception being that the Varian calibration procedure will not be used.
As each injection is processed by the Varian, a report will be printed on
the analyst's terminal and also automatically transmitted to the Nova. The
analyst's only responsibility to the Nova at tnjs time will be to note the
ID number of each injection.
When the analyst signs on to the Nova, there will be a choice of 10
options: 1) sign off from the Nova, 2) generate a method in the Nova, 3)
calibrate, 4) plot a calibration curve, 5) process injections, 6) perform
quality control calculations, 7) perform dissimilar analysis confirmation,
8) calculate replicate statistics, 9) list the names of all extant data
files, or 10) interact with Sample File Control. The analyst will be able
to select any one of these options immediately if the prerequisite options
have been executed during a previous session. For example, if a Nova method
has been generated and a calibration curve has been formed, the analyst will
be able to immediately process an injection and evaluate it as a control
(check) standard.
Generating a method in the Nova (option 2) will involve entering heading
information for reports and the Varian identifiers, full chemical names, and
CAS Registry numbers for the compounds covered by the method. Once a method
has been generated in the Nova, it will be available for use or for alter-
ation indefinitely.
Processing injections (option 5) will include calculating concentra-
tions, identifying compounds with their full chemical names and CAS Registry
numbers, calculating relative retention times or capacity ratios, and print-
ing reports.
All of the other options have been described previously. After any
option has been completed, the analyst will again be confronted with the
choice of all ten options.
23
-------
PROGRAM MODULES
There will be 21 programs in this system as shown in Figure 6. Each of
the programs will be independent of the others to the extent that one
program can be altered or replaced without affecting the others. All data
transfer from one module to another will be done via disk files.
FILE DESCRIPTIONS
There will be four types of disk files used by this system.
described as follows:
They are
Arrival File
TYPE:
NUMBER:
CONTENTS:
USES:
EXISTENCE
Data File
TYPE:
NUMBER:
CONTENTS:
Binary sequential
Exactly one
This file will hold the Varian report exactly as it is
transmitted.
This file will hold the Varian report when it is received
from the Varian. A program will then look through this
file for the instrument number and data which are needed
to generate the name of data file. The program will then
compact the information in this file as it copies it into
a data file.
This file will be created when a report arrives from the
Varian, and will exist only until it is overwritten by
the next arrival file.
Binary random access
One for every sample or standard
1. Before processing:
a. Varian method number
b. 30-character title
c. identifier of each compound
d. type of peak for each compound
e. retention time of each compound
f. area of the peak for each compound
g. any error messages from the Varian
2. After processing:
a. all of the above
b. relative retention time of each compound
24
-------
c. concentration of each compound
USES:
EXISTENCE:
This file will store unprocessed data for any sample or
standard. In the case of a calibration standard, this
file will provide the information needed by the cali-
bration routine. In the case of an unknown sample or QC
sample, this file will provide the data needed to
identify the compounds, and to calculate concentrations
and relative retention times for the compounds. After
being processed, this file will contain the information
necessary for doing quality control calculations, dis-
similar analysis confirmation, and replicate statistics,
and for reporting to SFC.
This file will be created from an arrival file. It will
exist in the Nova until 100 more samples have been run on
the same instrument.
ID Number List File
TYPE:
NUMBER:
CONTENTS:
USES:
EXISTENCE:
Nova Method File
TYPE:
NUMBER:
Binary random access
One for each instrument
This file will contain the ID numbers, in reverse chrono-
logical order for the 100 most recent samples run on the
corresponding instrument.
1. When the 101st sample is run on the corresponding
instrument, the Nova will delete the oldest file
recorded in this list.
2. When the Nova assigns an ID number to a sample, it
will check the newest entry in this file to deter-
mine what sequence number should be assigned.
3. If an analyst forgets the ID numbers of his samples
or standards, he will be able to see a copy of this
list.
This file will be automatically created the first time
the Nova encounters a sample from an instrument. The
file will be initialized with 100 dummy entries, and will
then gradually fill with real entries. It will exist
indefinitely.
Binary random access
One for each method in use in the Varian
25
-------
CONTENTS: 1. Analyst's name and other heading information.
2. Calibration points and equations.
3. Table of identifiers, true names, and CAS numbers.
4. Identifier of the compound to whose retention time
all other retention times are relative.
USES: 1. Headings for all reports.
2. Calibration equations are used in calculating
concentrations of unknowns.
3. Calibration points and equations are used in plot-
ting calibration curves.
4. True names of compounds and their CAS numbers are
determined using this file.
5. Relative retention times are calculated from the
retention time of the compound named in this file.
EXISTENCE: This file is created by an analyst during the Nova method
generation procedure. It is altered whenever the analyst
recalibrates, and possibly at other times. This file
exists indefinitely.
26
-------
CREATE DATA
FILE
SAMPLE FILE
CONTROL
LIST EXTANT
DATA FILES
REPLICATE
STATISTICS
SURROGATE
SPIKES
DUPLICATE
SAMPLES
SPIKED
SAMPLES
CONTROL
STANDARDS
MASTER CONTROL
DIRECTAQUEOUS
INJECTION
LIQUID/LIQUID
EXTRACTION
REL. RETENTION
TIMES
IDENTIFY
COMPOUNDS
PURGE AND
TRAP
INTERNAL
STANDARDS
J
CALIBRATION
PLOTTING
L/L EXTRACTION
DIR. AQU. INJ.
PURGE AND
TRAP
GENERATE
METHOD
l/l
D-
10
O
10
O)
i.
ra
.c
O
$-
•r—
O)
tO
(/)
O
T3
O
E
o
(1)
s-
27
-------
SECTION 6
SIGNOFF SHEET
These documents, which describe a proposed automated chromatography data
system but do not constitute an implementation design, are approved by the
undersigned interested parties with the understanding that changes in detail
are likely and that implementation of all features depends upon the avail-
ability of funding and personpower.
Requestors: Thomas Bellar
Denis Foerst
Leown Moore
William Potter
Implementor: Jonathan Kopke
Gas Chromatography
System Project Officer: John Teuschler
Laboratory Automation
Project Officer: William Budde
28
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse bcjo/r i-(imi>lctin)>)
REPORT NO.
EPA-600A-79-038
3. RECIPIENT'S ACCESSIOr+NO.
4. TITLE AND SUBTITLE
Functional Specifications for an Advanced
Chromatography Automation System
5. REPORT DATE
June 1979 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jonathan E. Kopke
8. PERFORMING ORGANIZATION REPORT NO
9. PERF
RFORMING ORGANIZATION NA,ME AND ADDRESS
outhWestern Ohio Regional Computer Center
University of Cincinnati
Cincinnati, Ohio 45220
10. PROGRAM ELEMENT NO.
11. CONTRACT/C4SAWT NO.
GS-05S-10458
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring and Support Lab-Cincinnati
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Extramural
14. SPONSORING AGENCY CODE
EPA/600/06
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document contains a project definition, a set of functional requirements,
and a functional design for a system which will link a commercial Chromatography
data system to the EPA Laboratory Automation System.
A Varian 220L Chromatography Data System was selected as the prototype system
to be extended in this project, although these specifications can be adapted to
other commercial systems. The current methods of using the Varian system are
briefly described in this report. The bulk of the report is a detailed list of
the additional functions to be performed by the EPA Laboratory Automation System.
These functions include multi-point calibration, calculation of concentrations,
identification of compounds, calculation of relative retention times, and calcu-
lation of quality control statistics. A general plan for the proposed system is
provided.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Gas Chromatography
Calibrating
Quality assurance
Computers
09/B
14/B
07/C
8. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport/
Unclassified
21. NO. OF PAGES
35
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
29
ft U.S. GOVERNMENT PRINTING OFFICE: 1V79 — 657-060/5315
------- |