United States
Environmental Protection
Agency
Environmental Monitoring and Support
Laboratory
Cincinnati OH 45268
EPA-600/4-78-064
December 1978
Research and Development
vvEPA
Automation of an
Ultraviolet-Visible
Spectrometer
ft/e
0
600478064
-------
RESEARCH REPORTING SERIES
Research reports ot 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-600/4-78-064
December 1978
AUTOMATION OF AN ULTRAVIOLET-VISIBLE SPECTROMETER
by
Dennis P. Ryan
Southwestern Ohio Regional Computer Center
University of Cincinnati
Cincinnati, Ohio 45220
Contract No. 6S-05S-10458
Project Officer
William L. Budde
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 was reviewed by the Environmental Monitoring and Support
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect 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.
-------
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 wastes.
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 a computerized system for instrument
automation leading to improved data collection, analysis, and
quality control.
This report was developed by the Advanced Instrumentation Section of
the Environmental Monitoring and Support Laboratory in the interest of
distribution of information to aid the advancement of laboratory
techniques, and quality control through computerization.
Dwight G. Ballinger
Director
Environmental Monitoring and Support
Laboratory - Cincinnati
m
-------
ABSTRACT
This report is an overview of the major features of an automated
ultraviolet-visible spectrometer system.
Four functional software modules are described which include the
chlorophyll analysis module, the color analysis module, the multi-option
module, and the quality control module. The hardware interfacing is
described in narrative fashion.
Finally, the general systems design methodology is discussed in
relation to the ultraviolet-visible spectrometer in particular, and the
laboratory automation effort in general.
This report is a result of work done in conjunction with the
laboratory automation project, sponsored by the Environmental Monitoring
and Support Laboratory of the Environmental Protection Agency. This work
was accomplished over the period April, 1977 to June, 1978.
IV
-------
CONTENTS
Page
Foreword i i i
Abstract iv
Acknowledgments vi
I. Introduction 1
II. Chlorophyll Analysis Module 4
III. Color Analysis Module 10
IV. Multi-Option Analysis Module (Environmental Pollutants) 16
V. Quality Control Module 20
VI. Hardware Interfacing and System CALLs 24
VII. Conclusion 26
VIII. References 27
-------
ACKNOWLEDGMENTS
The author and the project officer wish to express their appreciation
to a number of colleagues who provided assistance which was essential to
the successful completion of this project. These colleagues included Dr.
Cornelius Weber, who defined the requirements for the chlorophyll analyses
methodology and provided a chlorophyll calculation and report generation
program, originally written in FORTRAN IV; Larry Lobring, who assisted in
the development of the color analysis module and the overall quality
control requirements; Jack Teuschler, who designed the hardware interface
to the Nova 840 minicomputer and provided assistance with the minicomputer
hardware and software; David Fleck, who also assisted with the minicomputer
hardware and software; Bruce Almich, who gave advice on the software
design; and all of the laboratory personnel of EMSL who provided technical
assistance during the testing phase of the project.
Appreciation is also extended to Louis Taber, Les Rigdor and George
Barton of the Lawrence Livermore Laboratory for their lab automation
software design efforts. Their functional designs served as a framework
for some of the software developed for the UVVIS system.
Special appreciation is due to Dr. George Garland of the Dupont
Company, who supplied several industrial dye samples and test results for a
study using the color analysis module.
-------
SECTION I
INTRODUCTION
The ultraviolet-visible spectrometer is a recent addition to the EPA
laboratory automation system (1). In 1976, a hardware interface was
designed to link a Coleman model 124 dual-beam, scanning ultraviolet-
visible spectrometer to a Data General NOVA 840 minicomputer. This
interface system, which resembled a two-way communication network, was
designed to serve two functions. The first function of the interface was
to control the scanning of the spectrometer. The system was designed so
that signals sent from the computer could start and stop the scanning
mechanism within the spectrometer. The second function of the interface
was to transfer spectrometric signals from the instrument to the computer.
Interface circuitry was designed so that two signals, representing relative
values of wavelength and sample transmittance, could be sent simultaneously
to the computer (See Part VI, Hardware Interfacing and System CALLs, for
more information on this subject).
Once the hardware interface was constructed and installed on a
suitable spectrometer, work began on the development of computer programs
(software) which could make use of the automated spectrometer. The BASIC
programming language was selected as the foundation for this software
system because of its flexibility and versatility in an interactive
laboratory environment (1). The first major software application utilizing
the automated ultraviolet-visible spectrometer was the chlorophyll
computation program. Tests were made in this initial phase to verify the
speed and accuracy of the automated spectrometric analysis of chlorophyll
samples (1). The success of these tests proved the usefulness of the
automated instrument and pointed the way toward expanded software
development utilizing the automated ultraviolet-visible spectrometer.
Starting in April of 1977 a major effort was undertaken to expand and
unify the capabilities of the automated ultraviolet-visible spectrometric
(UVVIS) system. The system was completed in June of 1978 and currently
contains a number of analysis modules as well as a module to aid in
controlling instrument performance (See sections II, III, IV and V of this
report).
The purpose of this report is to describe, in moderate detail, the
software and hardware components of the UVVIS system. However, this report
is intended to do more than just introduce a new laboratory automation
system. It is intended to show the great potential of developing
computerized systems for the automation of even the most common and
relatively inexpensive laboratory instruments.
1
-------
The following overview of the UVVIS system will assist in defining
conventions and processes referred to in the balance of the report. Figure
1 depicts the hardware and software components of the UVVIS system. The
hardware components control scanning and facilitate the transfer of analog
spectrometric signals representing wavelength and transmittance. These
analog signals (voltages) are converted to digital equivalents (counts) in
the analog to digital converter (A/D). These signals are then reduced to
exact wavelength and transmittance values within the central processing
unit (CPU) through the use of system software. It is important to note
that the hardware interface presents only relative signals to the CPU for
processing. It is the responsibility of the user, with the assistance of
system software, to define the real-world equivalents of these signals.
This is done through the process of instrument calibration which must be
performed at the beginning of each computer-assisted analysis. In effect
this process conditions the computer to translate digital signals to their
corresponding equivalents of wavelength and transmittance. Wavelength
signals are always converted to nanometer (nm) equivalents.
Transmittance values are useful in the evaluation of the color
characteristics of solutions in conjunction with the International
Commission on Illumination (CIE) method of color classification. In some
cases, additional reductions may be required to convert transmittance (T)
into absorbance (A) values. This is accomplished through the equation:
A = LOG10(1/T)
Absorbance values are commonly used in the evaluation of concentration
levels of certain environmental pollutants in conjunction with Beer's Law
which states that, at well defined wavelengths, the analyte concentration
is proportional to the measured absorbance of a sample.
The software components of the UVVIS system assist in the automated
acquisition and reduction of spectrometric data for a number of analysis
techniques, and the generation of formatted reports describing the outcome
of each analysis. Control for the various analysis modules resides in an
ultraviolet-visible spectrometric master program called UVVIS. The master
program controls access to a number of program modules which facilitate the
execution of the particular types of spectrometric analysis. Three
analysis modules are currently available within the UVVIS package. The
chlorophyll computation module assists in the determination of the
concentrations of chlorophylls a, b, and c in extracts of algae and other
plants. The color analysis module aids in the evaluation of the color
characteristics and color levels in industrial waste water. The
multi-option module assists in determining concentration levels of a number
of environmental pollutants currently found in wastewater samples.
-------
Chlorophyll
Module
Color
Module
UVVIS
Control
Program
Multi-Option
Module
SOFTWARE
HARDWARE
Quality Control
Module
\
\
\
Future
Modules
CPU
DG NOVA
840
Digital Interface
(Scan Control)
Analog
Interface
A/D
I
WL
Disk I/O
&
Storage
Spectrometer
Data Driver
UV-Visible
Spectrometer
Figure 1. The hardware and software components of the Automated
Ultraviolet-Visible Spectrometric System.
-------
The UVVIS master program contains a set-up/check-out routine which
directs the start-up,procedures for any automated analysis, and which
allows for on-line signal monitoring of the automated instrument.
Finally, the UVVIS package contains a quality control module. This
very important component of the UVVIS system allows for a detailed analysis
of spectrometric precision and accuracy using certified filters available
from the National Bureau of Standards (NBS).
In time, other modules may be designed for inclusion into the UVVIS
system, especially in the realm of ultraviolet analysis. This, however,
does not preclude the fact that the automated ultraviolet-visible
spectrometer is presently a valuable part of the EPA laboratory automation
system.
-------
SECTION II
CHLOROPHYLL ANALYSIS MODULE
The chlorophyll analysis module of the UVVIS software package facili-
tates the determination of the chlorophyll concentrations in environmental
samples utilizing the spectrometric method (2,3).
The chlorophyll is extracted from algae or other plant materials. The
absorbance of the extract is measured between 620 and 760 nm, and the data
are reduced using trichromatic equations. In this way, sample concentra-
tions of chlorphyll a, b, and c are determined quickly and accurately.
Chlorophyll a is considered a useful index of algal standing crop.
Measurable amounts of pheophytin a introduce interference errors in the
evaluation of chlorophyll a concentrations. To correct for pheophytin a
the sample is acidified and the absorbance is remeasured. The new peak
absorbance and the previous peak absorbance are used in the monochromatic
equations to calculate the concentration of pheophytin a and a "corrected"
concentration of chlorophyll a described by CHL a1.
Five trichromatic computational methods have been developed in the
past 25 years to determine concentrations of chlorophyll a, b, and c. Of
these, only the three most recent methods are implemented in the chloro-
phyll analysis module. The newest method is the Jeffrey-Humphrey method
(4), which is still in the development phase. The other two methods
available in the chlorophyll analysis module are the UNESCO Method (2) and
the Strickland-Parson method (5). Of these the most commonly used is the
UNESCO method.
In the trichromatic method, concentrations of chlorophyll a, b, and c
are determined utilizing the absorbance at approximately 663 (peak), 645
and 630 nm, respectively. The precise wavelengths vary with each method.
The absorbance at 750 nm is subtracted from each to correct for sample
turbidity. The adjusted absorbances obtained in these measurements are
used in the chosen trichromatic equations to determine concentrations of
CHL a, b, and c. The measurement of pheophytin a and the subsequent
"correction" of the chlorophyll a concentration (CHL a1) is accomplished by
acidifying the sample and remeasuring the absorbance at 665 nm (peak).
Once again the 750 nm measurement is used to correct for the turbidity of
the acidified sample and the adjusted absorbance is used in the monochro-
matic equation to determine the concentration of pheophytin a and
"corrected" chlorophyll a.
-------
Table 1 shows the precise wavelengths at which absorbances are
required for each method. In Table 1, the B indicates measurements taken
before sample acidification and the A indicates measurements taken after
sample acidification.
TABLE 1. WAVELENGTHS (NM) USED IN THE CHLOROPHYLL ANALYSIS COMPUTATIONAL
METHODS
Method CHL a CHL b CHL c PHEO a
(peak)
UNESCO 663B 645B 630B 665A
Strickland-
Parson 665B 645B 630B 665A
Jeffrey-
Humphrey 664B 647B 630B 665A
In the program, absorbances are acquired at the appropriate
wavelengths and are immediately adjusted for turbidity. The values are
then substituted into one of the following sets of equations:
UNESCO Method
CHL a = F*(n.64*A663B-2.16*A645B+0.10*A630B)
CHL b = F*(-3.g4*A663B+20.g7*A645B-3.66*A63oB)i
CHL c = F*(-5.53*A«63B-14.81*A645B+54.22*A530B)
PHEO a = F*(26.73*(1.7*A6§5A-AcMB))
CHL a' = F*(26.73*(A663B-A665AJJ
Strickland-Parsons Method
CHL a = F*(11.64*A665B-1.31*A645B-0.14*A63oB)
CHL b = F*(-4.34*A665B+20.7*A645B-4.42*A
CHL c = F*(-4.64*A§65B-16.30*A645B+55.00*A63oB)
PHEO a = F*(26.73*[1.7*A665A-A*65B))
CHL a1 = F*(26.73*A665B-A665A)}
Jeffrey-Humphrey Method
CHL a = F*(11.85*A664B-1.54*A647B-0.08*A630B)
CHL b = F*(-5.43*A664B+21.03*A647B-2.66*A630B)
CHL c = F*(-1.67*Afi64B-7.60*A647B+24.52*A630B)
PHEO a = F*(26.73*(1.7*A664A-AW4B))
CHL a' = F*(26.73*(A664B-A664A))
-------
F is an analysis constant defined in the following way:
F = Extract Volume * Dilution Factor
Cell path Length (cm) * Sample Volume * 1000
Figure 2 shows the spectra from an unacidified and acidified standard
chlorophyll sample. Note that each method required a "peak" reading at a
different wavelength (Table 1). This dichotomy will be examined in more
detail in the following description of the automated chlorophyll computa-
tion module.
Data acquisition, data reduction and report generation are the three
major segments of the chlorophyll analysis module (Figure 3). Data
acquisition is accomplished through on-line measurements of chlorophyll
samples using the automated spectrometer. Precision may be established by
allowing the user to accumulate data on up to five replicates of the
sample. Line printer plots may be generated for each sample scan showing
absorbance vs. wavelength from 620 to 760 nm.
Once the absorbance data has been acquired, the computation segment
reduces the data to chlorophyll concentrations using one of the computa-
tional methods discussed above. Certain ratios and indexes are also
calculated in this segment.
After the data is reduced, replicate statistics are generated. These
statistics include the mean, standard deviation and relative standard
deviation for each set of replicate parameters. The reduced data and the
replicate statistics are then output to the lineprinter in the form of a
final report. Once an automated run has been completed and the output
report has been generated, the user may elect to delete one "bad" replicate
and regenerate updated replicate statistics in another report.
Two additional features are available to the user who is making auto-
mated sample scans. The first of these is a simplified NBS filter test to
determine the relative accuracy of the instrument, and the second feature
is an optional peak search routine for determining the wavelength location
of the CHL a peak absorbance.
The relative accuracy of the automated spectrometer may be checked
through a specialized NBS filter test option. This test simply compares
the absorbance values obtained from a scan of an NBS linearity filter
(SRM930b, Filters 1-282,2-282 and 3-282) with those obtained from another
"more accurate" instrument such as the Beckman ACTA-V or the Gary 14
spectrometer. The wavelengths sampled in this test correspond to those
used in the particular computational method (663, 645, 630 and 750 nm), and
as yet absorbances at these wavelengths have not been certified by the
National Bureau of Standards for these filters. (See Section IV, Quality
Control Module for more information on the evaluation of spectrometric
precision and accuracy using NBS filters.)
It was previously noted that the three computation methods described
above differ with respect to the exact location of the sample peak.
-------
.81
00
u -c
z
eo
O .4^
CO
.2^
• Before Acidification
Acidified
-A A
620 640
660 680 700 720
WAVELENGTH (NM)
740 760
Figure 2. Visible spectra from unacidified and acidified standard chlorophyll sample.
-------
Calculate
Replicate
Statistics
Instrument
Baseline
Calibration
1
Enter Data
from Keyboard
i
r
Calculate
Concentrations
and Ratios
Option
i
Scan Samples
and Plot
Figure 3. Flow chart of the chlorophyll analysis module.
-------
Normally the errors introduced by this deviation are minor compared to
other method errors; however, a peak search routine has been installed in
the automated segment of the module in order to systematically verify this
hypothesis. A peak search is made on every automated sample scan and the
user is supplied with the "true" sample peak as seen by the automated
spectrometer. At the user's discretion, this peak absorbance can be used
in place of the peak defined by the particular computation method. For
example, if a scan is made on a chlorophyll sample and the peak occurs at
662 nm, the user may elect to retain the absorbance at 662 nm in place of
the method-defined peak found at 663, 664 or 665 nm. Presently this
feature is considered a research tool and caution should be exercised in
order to avoid inconsistencies inherent in this type of option.
The discussion above has dealt only with measurements obtained from an
automated spectrometer; however, it is likely that chlorophyll samples will
sometimes be measured using an unautomated instrument, and strip chart
recordings will be generated from these scans. The chlorophyll analysis
module is able to accept data from these charts through the manual mode.
The analyst must type in the absorbances recorded at the appropriate
wavelengths, and when data entry is complete, the computer will calculate
and display the results in the same manner as described above for the
automated run.
The chlorophyll analysis module is a fast, flexible, and accurate tool
for the evaluation of chlorophyll concentrations in environmental samples.
As computational methods are perfected and algae families are characterized
in terms of chlorophyll a, b, and c, the chlorophyll analysis module will
assist in quantifying ocean food supplies and monitoring the ecosystems
which facilitate the growth of these algae.
10
-------
SECTION III
COLOR ANALYSIS MODULE
The color analysis module of the UVVIS software package incorporates
newly developed techniques for the evaluation of color levels in environ-
mental samples. For this reason a more descriptive discussion will be
presented on the color analysis module.
The American Public Health Association (APHA) color method has, in the
past, been adequate for the evaluation of yellow color levels in river and
stream waters. The method required the preparation of a number of platimum
cobalt standard solutions with known APHA colorimetric units. An unknown
sample was then visually located between two of these standards. In this
way an approximate APHA number could be assigned to the unknown. This
method was not only inexact but it was also unusable on industrial wastes
which had colors significantly different from the yellow of the platinum
cobalt standards. The American Dye Manufacturers Institute (ADMI), through
its Ecology Analytical subcommittee, developed an alternative method for
measuring color levels in water (6). Their goal was to devise a method
that would meet four criteria:
1. The method should be applicable to any color (hue).
2. The method should be sensitive to small color differences.
3. The method results should be related to APHA colormetric units
(platimum cobalt standards).
4. The method should require relatively inexpensive instrumentation.
The method devised by the subcommittee will be referred to as the ADMI
method for the determination of color in water. The following is a brief
summary of the evolution of this method.
The committee first assumed that the color of water has a negative
impact on the environment only in an esthetic sense. For this reason, it
was decided to determine color in water as a strictly visual perception.The
International Commission on Illumination (CIE) system of specifying color
based on tristimulus values X, Y, and Z has had wide acceptance in relating
physical measurements to the stimulus perceived by the normal observer, and
can be used to specify the color of environmental samples in terms of
dominant wave length (hue), purity, and luminosity (7).
11
-------
Tristimulus values are hue dependent, however, and cannot be used
directly to determine the uniform color difference between the color of a
solution and that of a colorless solution. By transforming the tristimulus
values to Munsell coordinates (8), a uniform color difference equation can
be used to produce a single number which represents the vector length
between the color of the sample and colorless water.
One of the simpler equations used to evaluate this color difference is
that known as the Adams-Nickerson equation (9) which combines the chromatic
value transformation of the CIE chromaticity diagram proposed by Adams with
the lightness or luminosity modification proposed by Nickerson. This
single number color difference is known as Delta-E or the change in sensi-
tivity from the colorless state.
All non-phosphorescent solutions regardless of hue can be labeled with
this Delta-E (or DE) number. The DE number represents the uniform color
difference from colorless. By measuring platinum-cobalt standard solutions
and calculating DE by the Adams-Nickerson formula, a calibration curve
relating DE to standard ADMI values can be used to obtain accurate ADMI
values to describe color level in water of any hue. A plot of measured
values of DE as a function of ADMI values for six APHA standards is shown
in Figure 4. Any colored solution whose DE number has been calculated can
be related to the standards curve, and an ADMI number can be assigned to
the solution. Table 2 shows the method for preparation of the standards,
the definitions of ADMI values, and the measured DE values used in Figure 4.
The ADMI method for the determination of color in water described
above can be divided into five parts.
1. Measurement of the visible spectrum of the sample on suitable
instrument.
2. Calculation of CIE tristimulus values X, Y, and Z.
3. Conversion of tristimulus X, Y, and Z to Munsell coordinates Vx,
Vy, and Vz.
4. Calculation of Adams-Nickerson color difference (DE).
5. Conversion of DE to ADMI values using standards curve.
By Federal law section 304g, PL 92-500, the EPA was required to
develop a procedure for the measurement of color in water. The ADMI method
described above was adopted as a means to measure color levels in waste-
water (10). However, the steps involved in the ADMI method have tradi-
tionally involved complicated and lengthy measurement and computation. To
solve this problem the color analysis module was developed for the UVVIS
system. The module was designed to make repetative measurements, hand
calculations, and table referencing unnecessary.
12
-------
RECORDS SHELF LIST
FROM:
DATE:.
4cy 7?
Name
Building/Room Number
Telephone Extension
Bureau or Office
Division
CONTAINER
NO.
DESCRIPTION & DATE
(Please list each folder)
ESTIMATED
DISPOSAL
DATE/NARA
SCHEDULE
NUMBER
?u/»
-------
(NID)
570 --
/fa
o
_.,
-------
TABLE 2. ADMI VS. DE FOR PLATINUM COBALT STANDARD SOLUTIONS
Milliliters of Color in Measured Standard
Standard Solution ADMI Units DE Values2 Deviation
Diluted to 100 ml
with distilled water1
5.0 25 0.090 0.003
10.0 50 0.180 0.004
20.0 100 0.342 0.006
30.0 150 0.512 0.004
40.0 200 0.666 0.006
50.0 250 0.810 0.008
1 Preparation of Standard Solution:
Dissolve 1.246g of potassium chloroplatinate, K2PtCl4 (equivalent
to 500 mg metalic platinum) and l.OOg crystallized cobaltous chloride,
COC12 6^0 (equivalent to about 250 mg metalic Cobalt) in
distilled water with 100 ml concentrated hydrochloric acid and dilute
to 1000 ml with distilled water. This stock standard has a color of
500 ADMI units.
2 Represents ten non-consecutive replicates of each standard.
13
-------
6)
in-
cu
©
s>-
cu
©
in-
PLATINUM-COBALT STANDARDS CURVE
in
I 1 1 1 1 1 1 1 1
.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
D E
Figure 4. A plot of ADMI values for six platinum-cobalt standard solutions as a function
of measured DE values.
-------
The five steps of the ADMI method are stated in terms of this auto-
mated procedure as follows:
1. A dual-beam scanning spectrometer was selected and interfaced to
a Data General NOVA 840 mini-computer. The BASIC computer
language was chosen to allow for maximum computer-operator
interaction.
2. The module was designed to make a spectrometric scan of a sample
between 400 and 700 nanometers, acquire and store transmittance
measurements through the range, and reduce these values to the
tristimulus values X, Y, and Z (7).
3. A simple iterative algorithm was used to convert the tristimulus
values X, Y, and Z to Munsell values Vx, Vy and Vz (8).
4. The DE value could then be calculated from the following
Adams-Nickerson equation:
DE ={(0.23*AVy)2+ C A( Vx-Vy)]2 + C0.4* A(Vy-Vz) ]2} "2
Where Vx, Vy, and Vz are the Munsell value equivalents of
tristimulus values X, Y, and Z respectively, and where Delta-V
is the difference between the Munsell value of a sample (Vxs,
Vys, and Vzs) and the Munsell value of a colorless solution
(Vxc, Vyc and Vzc)(9).
5. The ADMI value of the sample could then be determined from a
calibration equation of ADMI as a function of DE for the
standard solutions. In the computer program this was
accomplished by fitting the standard points to a second degree
equation using the least squares regression technique. In this
way the relative ADMI value of any sample could then be obtained
as a function of DE (Figure 4).
Because of the excellent reproducibility obtained in repetative
measurements of the platinum cobalt standards, it was decided that the
entire set of standards need not be run each time samples are analyzed.
Instead the user is required to run at least one standard, commonly
referred to as a check standard, to verify the validity of the calibration
curve. If the outcome of this measurement is within statistical bounds,
the user may begin measuring color levels in environmental samples.
However, if the check standard values are repeatedly unacceptable, the user
may update the standard equation.
Figure 5 shows an overall flow chart for the color analysis module.
The Color Analysis Module serves as a versatile research tool for the
evaluation of the color characteristics of virtually any uniform-nonopaque
substance or solution, while still fulfilling the legal requirements for
determining the ADMI color levels in environmental samples.
15
-------
Instrument
Baseline
Calibration
Scan Standards
and Update
Statistics
Update
Standards
7
Check
Standards
7
Scan Sample
Scan Standard
and Display
Color Data
Another
Sample
Evaluate Quality
Control
Scan Final
Check Standard
Evaluate QC
Figure 5. Flowchart of the Color Analysis Module.
16
-------
SECTION IV
MULTI-OPTION ANALYSIS OF ENVIRONMENTAL POLLUTANTS
As was stated in the introduction of this report, there are many
compounds, ions, and elements that exhibit spectrometric properties in
compliance with Beer's law. In these cases, the concentration of an
analyte is proportional to the absorbance of the sample measured at a
specific wavelength. Table 3 shows the list of analytes, methods, units of
measurement and wavelengths for several methods currently approved for EPA
monitoring programs (10). The module may also be used to evaluate newly
proposed methods for determining pollutant concentration levels in environ-
mental samples.
The multi-option analysis module of the UVVIS system was developed as
a general purpose approach to this type of spectrometric measurement. In
all instances where this methodology is used, it is necessary to measure
the absorbances (at the specified wavelength) of standards with known
concentrations before an analysis of environmental samples with unknown
concentrations can be made. This process entails the development of a
standards calibration equation relating the absorbance of the standard to
its concentration. This is done by calculating coefficients for the first
and second order equations using the least squares regression technique.
Once an accurate calibration curve has been developed, unknown sample
absorbances are measured, and these absorbances are substituted into the
standard calibration equation in order to determine the analyte concen-
tration in the unknown sample.
The multi-option module normally facilitates data acquisition through
an automated mode. Absorbance data may be input manually through keyboard
entry in cases where standards and samples were measured using a non-
automated instrument. This facility speeds up data reduction, and produces
a formal report describing the outcome of the analysis.
In the on-line or automated mode, data is collected directly from the
spectrometer. The user is responsible for processing and labeling
standards, unknowns, and quality control samples including spiked samples,
check standards, and duplicate unknowns. The user inputs various run
parameters, sets the appropriate wavelength, and then loads standards with
known concentrations. The computer records each standard absorbance.
Standard calibration curves are created using the least squares regression
technique. The program displays tables showing the results of this
calibration. The multi-option module allows for a great deal of
17
-------
flexibility in manipulating standards. After standard values have been
input in the initialization segment of the run, standards can be loaded and
readings can be taken in any order. Divergent points can be eliminated and
calibration equations can be updated. Internally generated error commands,
warnings, and suggestions aid the operator in developing a standards equa-
tion in compliance with analysis specifications. With this kind of flexi-
bility, the effective concentration range for a given test can be deter-
mined easily, and the least squares fit parameters can be calculated to
optimize measurement accuracy.
Samples of unknown concentration are then loaded into the spectro-
meter, and the absorbances measured. By utilizing the calibration
equation, unknown concentrations are determined. Quality control samples
can be measured during the samples run to document the accuracy and
precision of the unknown measurement. As many as ten replicate measure-
ments can be made on any standard or sample to evaluate precision and
assure measurement reliability.
Interim reports are displayed during the course of a run, and consist
of: equations obtained by least squares fit of standards data; standard
calibration tables; concentrations of samples; concentrations and
statistical data for spiked samples, check standards, and unknown dupli-
cates; and other miscellaneous messages.
Hard copy reports may be obtained during the course of a run or at the
end of a run on the line printer.
The multi-option analysis module performs each of the functions
described above through the execution of user-supplied instructions. The
module is therefore termed a command structured system. Since flexibility
is the key to a command structured system, the internal processes do not
exhibit a uniform flow. Because of this, Table 4 is presented to describe
the numerous commands available in the multi-option module.
The multi-option analysis module is a very flexible and highly
adaptable element of the UVVIS package and as such will serve the future as
well as the present EPA analytical monitoring requirements.
18
-------
TABLE 3. MULTI-OPTION ANALYSIS TECHNIQUES CURRENTLY AVAILABLE USING THE
UVVIS SYSTEM (10)
Analyte Method
Kjeldahl
Nitrogen Nesslerization
Ammonia Nesslerization
Nitrate Brucine
Nitrate-
Nitrite Cadmium Reduction
Phosphorus Total Phosphorus
Phosphorus Ortho-Phosphorus
Phosphorus Hydrolyzable Phosphorus
Arsenic Silver Diethyldithiocarbamate
Chlorine DPD Colorimetric
Cyanide Pyridine-Pyrazolone
Cyanide Pyridine-Barbituric Acid
Sulfide Methylene Blue Photometric
Fluride SPADNS
Silica Molybdosilicate
Phenol 4AAP Direct Photometric
Phenol 4AAP CHCL3 Extraction
COD Colorimetric
Units
Mi ligrams/liter
Miligram/liter
Mi ligrams/liter
Miligrams/liter
Miligrams/liter
Miligrams/liter
Miligrams/liter
Mi crograms/liter
Miligrams/liter
Miligrams/liter
Miligrams/liter
Miligrams/liter
Miligrams/liter
Miligrams/liter
Micrograms/liter
Mi crograms/liter
Milligrams/liter
Wavelength
425
425
410
540
650
650
650
535
515
620
578
625
570
410
510
460
600
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
19
-------
TABLE 4. SAMPLE TYPES AND COMMAND SUMMARY FOR THE MULTI-OPTION
ANALYSIS MODULE
Sample
Type ($$) Sample Description
S Standard in Calibration Curve
U Unknown Sample
RB Reagent Blank (Quality Control)
CS Check Standard (Quality Control)
DU Unknown Duplicate (Quality Control)
SP Spiked Sample (Quality Control)
Command Description
$$x/y Measure $$ number x for the yth time.
C Develop calibration equations for current standards.
B Measure instrument blank.
$$x/A Show concentration of $$ number x using a 1s*
degree equation.
$$x/B Show concentration of $$ number x using a 2nd
degree equation.
$$x/I Show concentration of $$ number x by interpolating
between two bracketting standards.
$$x/yD Delete replicate y of $$ number x.
$$x/E Erase all replicates of $$ number x.
$$x/L Look at current replicate data on $$ number x.
R Report on completed run results.
20
-------
SECTION V
QUALITY CONTROL MODULE
Clearly from the previous sections, the automated ultraviolet-visible
spectrometer can be used in a variety of ways. These include broad
spectrum measurements as used in the chlorophyll application, static
transmittance/absorbance measurements as used in the multi-option module,
and tristimulus integrator measurements for evaluating the color
characteristics of samples as used in the color analysis module.
In order to verify the precision and accuracy of the automated
instrument, a quality control module has been developed for the UVVIS
software package. The module allows for two types of filter tests using
National Bureau of Standards (NBS) certified filters (11). These tests
facilitate the evaluation of the automated spectrometer as a color
tristimulus integrator system or as a transmittance/absorbance measuring
device. Each type of test is made by scanning the appropriate NBS glass
filter, acquiring data on the scan, reducing the data, and comparing the
results with certified NBS standard values. An additional feature allows
for the comparison of the test results with historical test results
obtained on the same instrument.
The filter test to evaluate the performance of the automated
spectrometer as a color tristimulus integrator system uses NBS filters
2101, 2102, 2103, 2104 and 2105 (12). Each filter is a 2-inch square of
transparent colored glass. A chart of tristimulus values for CIE source C,
representing average day light, is furnished with each set of glasses, and
these values are certified by the National Bureau of Standards. Each
filter exhibits unique color characteristics as shown in Figure 7. Table 5
shows the NBS certified tristimulus values for each filter, along with the
mean and standard deviation acquired with eight non-consecutive replicates
of each filter made on the prototype UVVIS system.
After measurements are made on a particular tristimulus filter from
400 to 700 nm, the computer automatically calculates the tristimulus and
trichromatic values using the thirty ordinate CIE spectrometric method
(7). No internal corrections are made for back-reflectance, slit-width or
inertia! errors, but the accuracy and precision are very good, especially
for filter 2105 which exhibits color characteristics likely to be found in
environmental samples.
The filter test to evaluate the performance of the automated
spectrometer as a transmittance/absorbance measuring device uses NBS
linearity filters 1-282, 2-282 and 3-282 (SRM 930b). Each of these neutral
21
-------
glass filters is mounted in a standard 1 centimeter holder and exhibits
transmittance of approximately 10, 20, and 30 percent over the scanning
range. Each filter is individually calibrated and certified by the
National Bureau of Standards for absorbance and transmittance values at
wavelengths of 440, 465, 564.1, 590 and 635 nanometers. Table 6 shows the
NBS certified transmittance values for each filter (in %T) along with the
mean and standard deviaton acquired with eight nonconsecutive replicates of
each filter made on the prototype UVVIS system.
Note that the standard deviation remains relatively constant as the
transmittance (%T) increases. This implies that for environmental samples,
which normally exhibit much higher transmittance values, the precision will
improve proportionally. Another point worth noting is that the test on
these filters is dynamic since the filter is scanned over the wavelength
region. The accuracy improves under static conditions where the wavelength
is set before a measurement is taken.
A typical quality control run using the NBS filters can be broken down
into three parts: spectrometric baseline calibration, filter scans
including calculations and interim reports on each filter, and the summary
output report including precision and accuracy evaluations on all filter
scans. Test filters are run in any order, and a single filter can be run
up to 10 times if desired. Test results on each filter will be displayed
after each filter has been scanned, and, at the operator's discretion, a
more detailed report can be generated on the lineprinter, including a
spectral plot of the filter. Once the desired filters have been run, a
summary report is output to the lineprinter detailing the accuracy of the
tests relative to NBS standards and the precision of the test relative to
historical statistics maintained in the system.
In general, the quality control module of the UVVIS software package
can be used to rigorously test the precision and accuracy of the automated
spectrometer under extreme conditions. This points the way to the use of
the system as a tristimulus integrator and transmittance/absorbance
measuring device for environmental samples which exhibit more normal color
characteristics.
22
-------
NBS 21«1 137
NBS aiZa 137
r
4S«. saa. ssa. sae.
WIWELENSTH
ssa. 7ae.
-i r
sue. sse. 6««.
WflVELEN«TH
NBS 31«3 137
sa«. ss«.
WflVELENSTH
ss«.
8-
NBS aiOM 137
sae. ss«. eat.
UflVELENSTH
ssa. 7aa.
§31
5
NBS 210S 137
soe.
.
UflVELENSTH
ss«.
Figure 7. Color characteristics for five National Bureau of Standards
Tristimulus Filters.
23
-------
TABLE 5 - CERTIFIED AND MEASURED TRISTIMULUS VALUES FOR FIVE NATIONAL
BUREAU OF STANDARDS FILTERS
Filter X Y Z
2101
NBS value 45.0 25.3 0.0
Measured mean 46.83 26.73 0.20
SD 0.57 0.48 0.13
2102
NBS value 51.5 48.9 5.6
Measured mean 52.16 49.50 5.98
SD 0.21 0.27 0.14
2103
NBS value 3.6 11.3 2.6
Measured mean 3.38 10.90 2.84
SD 0.17 0.19 0.13
2104
NBS value 17.2 9.1 84.3
Measured mean 17.08 9.05 83.34
SD 0.18 0.21 0.34
2105
NBS value 51.8 56.1 75.4
Measured mean 51.75 56.00 75.43
SD 0.05 0.11 0.28
24
-------
TABLE 6. NBS CERTIFIED TRANSMITTANCE VALUES FOR THREE LINEARITY FILTERS
Filter %T at Wavelength (NM)
440.0 465.0 546.1 590.0 635.0
1-282
NBS Value 9.15 10.86 9.78 8.68 9.70
Measured Mean 9.26 10.85 9.89 8.81 9.82
SD 0.15 0.14 0.18 0.17 0.16
2-282
NBS Value 18.90 21.26 19.78 18.22 19.67
Measured Mean 18.83 21.01 19.73 18.24 19.69
SD 0.13 0.12 0.16 0.17 0.16
3-282
NBS Value 29.16 32.27 30.73 27.71 27.83
Measured Mean 29.33 32.25 30.93 27.97 28.07
SD 0.12 0.11 0.16 0.17 0.17
25
-------
SECTION VI
HARDWARE INTERFACING AND SYSTEM CALLS IN THE UVVIS SYSTEM
The major hardware components of the automated ultrviolet-visible
spectrometric system include the following items: the Data General Nova 840
mini-computer, fixed and moving head disks, medium or high speed line-
printer, hard copy or CRT-display computer terminal, digital and analog
interfaces, spectrometer remote interface box, and the scanning ultraviolet-
visible spectrometer.(Figure 1) The computer terminal serves as the
user-computer interface which allows the analyst to access the UVVIS
software package for spectrometric analysis. After the analyst chooses a
particular option, the CPU loads the appropriate module from disk to core
memory. The chosen BASIC program then questions and directs the operator
through well defined steps in order to complete the analysis. In cases when
data are acquired from the on-line spectrometer, a BASIC CALL is used to
access assembler language routines which control the start/stop scanning
mechanism, and which acquire wavelength and transmittance signals from the
spectrometer. Scanning is controlled through the digital interface, and
data acquisition is controlled in the analog interface. Once the analog
signals representing wavelength and transmittance are acquired by the analog
interface, they are converted to relative digital counts in the analog to
digital converter. These digital signals are then manipulated by system
software to give values of wavelength and transmittance for use in software
data reduction routines. As data is reduced, it is stored in disk data
files and is retrieved as output for lineprinter reports during the course
of the analysis.
Two BASIC data acquisition CALLs are used in the UVVIS system (13).
CALL 6 is used in static measurement applications such as the multi-option
module. Static measurements do not require a spectrometric scan. Instead,
the operator sets the wavelength on the spectrometer and the computer
repeatedly measures the transmittance signal of the sample at predefined
time intervals (2/15 sec). These signals are then averaged and presented as
a single signal to the system software. CALL 6 parameters are shown as
follows:
CALL 6, C, N, A
where C is the single A/D channel number for the spectrometric trans-
mittance signal, N is the number of points to be averaged and A is the
returned average of the converted digital transmittance signals.
CALL 7 is used in dynamic measurement applications such as the
chlorophyll computation module and the color analysis module. In the
26
-------
dyanmic mode, the operator sets the wavelength at which the scan is to
begin and the computer starts the scan, collects the wavelength and trans-
mittance signals at specified intervals, converts the signals to digital
values and stores them in a sequential software array. CALL 7 parameters
are shown as follows:
CALL 7, C, Dl, D2, N, V(l)
where C is the composite A/D channel number for scan control, wavelength,
and transmittance channels of the spectrometer, Dl is the delay time (in
l/60th second intervals) between when the CALL is made and when scanning
and data acquisition begin (this parameter may be 0), D2 is the delay time
(in l/60th second intervals) between data point pairs, N is the total
number of points to be taken, and V(l) is the first element of the array
used to store the converted digital signals.
CALL 6 is a rudimentary single channel CALL and, as such, requires no
further explanation. CALL 7, on the other hand, involves a much more
complicated interplay of hardware components which require a more detailed
description. To do this, the hardware system used to acquire wavelength
and transmittance signals must be examined in more detail. Then we will
show how these signals are passed to the computer, converted to digital
values, and stored in the software array.
Remember that CALL 7 is used in the scan mode. This implies that at
certain discrete intervals, a pair of signals representing transmittance
and wavelength is acquired at the spectrometer by the computer and trans-
lated to digital equivalents in the A/D converter. The wavelength signal
is supplied by a ten turn (.005%) potentiometer which is permanently
attached to the shaft of the scanning motor. Compensation for potentio-
meter non-linearity is made through a piecewise linear transformation
within the systems software. Transmittance signals are taken directly from
the spectrometer output electronics. The transmittance voltage may range
from about 1 volt for opaque solids to about 9 volts for air. Interim
voltage levels are nearly proportional to the transmittance of substances
within this range. Certain instrument idiosyncrasies which interfere with
transmittance signal linearity have been compensated for in the system
software. The wavelength and transmittance signals are continuously
available as long as the instrument is turned on.
When CALL 7 is invoked and scanning begins, the computer begins to
acquire signal pairs - one voltage level representing wavelength and one
voltage level representing transmittance at that wavelength. The transfer
of these voltages to the A/D converter occurs simultaneously. They are
converted to 14 bit digital equivalents and stored in the software array.
Points continue to be taken in discrete intervals until all of the signal
pairs required by the CALL (N/2) have been acquired. System software then
utilizes internally stored reference values to reduce the digital array
values to the actual wavelength and transmittance values. These wave-
length/transmittance pairs represent the spectral response of the sample
throughout the scanning range.
27
-------
SECTION VII
CONCLUSION
This report was intended to succinctly describe the UWIS system
without being overly simplistic. No progress will ever be made in lab
automation if system developers sacrifice user awareness for the sake of
expediency. On the other hand an attempt has been made to be descriptive
without burdening the reader with minute technical details. Let there be
no doubt that any automated system abounds with technical complexities at
all levels. With the right balance of order and detail, a foundation can
be built through which the user can understand even the most sundry
elements of a system. New patterns of communication arising from a
systematic approach to lab automation will benefit both the user community
and the systems developer. On the one hand, the designers will be pressed
by the user to explain rather than complicate. Conversely, the users will
be asked to organize and solidify their thinking rather than to postulate
nebulous needs.
The UWIS system has been shown to be a potentially powerful tool in
the lab automation arsenal, yet its primary value may reside in its form
rather than in its content. Because of advances in lab methodology and
because of an expansion in lab instrument complexity, it is necessary to
build automated instrument systems with qualities of multi-functionality.
The UWIS system is a simple example of this approach. The main
elements of the system are viewed as modules, and these modules are broken
down into easily understandable components. In the UWIS system the
software modules are also reduced to program segments which perform
individual tasks. More complicated systems require more levels of task
differentiation. It is the duty of systems designers and the user
community to work together to structure understandable segments within each
system level. Certainly the current gap which exists between the user
community and the systems designers will not close overnight, but a common
adherence to a general system design methodology may do much to reduce
misunderstanding and encourage productive communication.
28
-------
SECTION VIII
REFERENCES
1. W.L. Budde, B.P. Almich, and J.M. Teuschler, The Status of the EPA
Laboratory Automation Project, EPA report 600/4-77-025, April 1977.
2. C.I. Weber (Editor), Biological Field and Laboratory Methods for
Measuring the Quality of Surface Waters and Effluents, EPA report
670/4-73-001, July 1973.
3. C.I. Weber, et. al., "The Status of Methods for the Analysis of
Chlorophyll in Periphyton and Plankton", U. S. Environmental
Protection Agency, Environmental Monitoring and Support Laboratory,
Cincinnati, Ohio (in preparation).
4. S. W. Jeffrey and G. F. Humphrey, Biochem. Physiol. Plazen (BPP), 167,
pp. 191-194 (1975).
5. T.R. Parsons and J.D.H. Strickland, J. Mar. Res. 21, pp. 155-163
(1963).
6. W. Allen, et. al., Proceedings of the 28th Industrial Waste
Conference, Purdue University, 28, pp. 661-675 (1973).
7. "Standard Methods for the Examination of Water and Waste Water", 14th
Edition, APHA, Washington, D.C. (1976).
8. W.C. Rheinbolt, J. P. Menard, J. Opt. Soc. Amer. 50 80Z (1960).
9. K. McLaren, J. Soc. Dyers Colour. 86, 354 (1970).
10. Methods for Chemical Analysis of Water and Wastes, EPA Report
625/6-74-003, 1974.
11. Nat. Bur. Stand. (U.S.), Spec. Publ. 260-1975-76, catalog, p. 55 (June
1975).
12. H. J. Keegan, J. C. Schleter, D. P. Judd, J. Res. Nat. Bur. Stand.,
A., 66A 203 (1962).
13. L. Taber, et. al., Real Time Code for Laboratory Automation, Lawrence
Livermore Laboratories, UCRL-52, 392 (1978).
29
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/4-78-064
2.
3. RECIPIENT'S ACCESSIOWNO.
4. TITLE AND SUBTITLE
Automation of an Ultraviolet-Visible Spectrometer
5 REPORT DATE
December 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Dennis P. Ryan
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southwestern Ohio Regional Computer Center
University of Cincinnati
Cincinnati. Ohio 45220
10. PROGRAM ELEMENT NO.
1BD621
11. CONTRACT/GRANT NO.
GS-055-10458
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring and Support Lab. - Cinn, OH
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 report is an overview of the functional description and major features of
an automated ultraviolet-visible spectrometer system intended for environmental
measurements application. As such, it defines functional specifications and require-
ments which are divided into the chlorophyll, color, multi-option, and quality contro
modules. The general system design methodology is discussed with regard to the EPA
laboratory automation project. The interfacing hardware requirements are included in
general terms only.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Computer programming
Environmental tests
Spectrophotometers
Algorithms
Ultraviolet Spectrophotometers
09/B
1A/B
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
36
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
30
-------
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Environmental Research Information Center
Cincinnati, Ohio 45268
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE, S3OO
AN EQUAL OPPORTUNITY EMPLOYER
POSTAGE AND FEES PAID
U S ENVIRONMENTAL PROTECTION AGENCY
EPA-335
Special Fourth-C lass R ate
Book
If your address is incorrect, please change on the above label
tear off; and return to the above address.
If you do not desire to continue receiving these technical
reports, CHECK HERE I~I; tear off label, and return it to the
above address.
EPA-600/4-78-064
------- |