United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30605
EPA-600/4-80-015
February 1980
Research and Development
oEPA
New Liquid
Chromatographic
Detection System for
Environmental
Pollutants
-------
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-600/4-80-015
February 1980
NEW LIQUID CHROMATOGRAPHIC DETECTION SYSTEM
FOR ENVIRONMENTAL POLLUTANTS
by
L. A. Carrelra and L. B. Rogers
University of Georgia
Athens, Georgia 30602
Grant No. R804155-03-0
Project Officer
Leo Azarraga
Analytical Chemistry Branch
Environmental Research Laboratory
Athens, Georgia 30605
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30605
-------
DISCLAIMER
This report has been reviewed by the Environmental Research Laboratory,
Athens, Georgia, 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.
ii
-------
FOREWORD
Nearly every phase of environmental protection depends on a capability
to identify and measure specific pollutants in the environment. As part of this
Laboratory's research on the occurrence, movement, transformation, impact and
\
control of environmental contaminants, the Analytical Chemistry Branch develops
and assesses new techniques for identifying and measuring chemical constituents of
water and soil.
Early in the development of gas chromatography-Fourier transform infrared
spectroscopy methods at this Laboratory, the need for an on-line spectroscopic
technique to measure vibration spectra of liquid chroma tog raphic eluates was
recognized. One of the responses to this need was the funding of a project to
develop the coherent anti-Stokes resonance Raman spectroscopy liquid chromatog-
raphy system. This report describes the development of such a system, which should
prove useful in the analysis of relatively nonvolatile organic compounds.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
-------
ABSTRACT
Resonance enhanced coherent anti-Stokes Raman spectrometry (CARS) has
been demonstrated as a specific identification system for liquid chromatography.
To achieve this work in the areas of CARS, liquid chromatographic preconcentra-
tion and separation and computer control of the liquid chromatograph/ultraviolet-
visible/CARS experiments were undertaken.
Under the optimum resonance condition for CARS, spectra have been
obtained on five microliter volumes with detection limits of one micromolar or
less under favorable conditions. Excellent fluorescence rejection and ultraviolet
CARS have also been shown.
A number of different species have been surveyed by liquid chromatography
with the selectivity of different stationary phases and special solvent effects exam-
ined. Observations useful in relating experimental variables in the preconcentra-
tion step to the recovery of a given species have been made.
A system consisting of a liquid chromatograph, a scanning ultraviolet-
visible spectrometer and the CARS experiment have been automated using a DEC
PDF 11/34 minicomputer. Column selection, flow rate, and solvent composition
were monitored and controlled. The ultraviolet-visible spectrometer was used at
a fixed wavelength until a species was detected. The flow was stopped, the absorp-
tion spectrum was scanned, and flow was resumed. From the absorption spectrum the
optimum laser frequencies for resonance CARS were set along with the proper crossing
angle, detector stage angle, and monochromator wavelength for the Raman signal.
The Raman spectra was then scanned under control of the computer with on-line
signal averaging and data smoothing.
This report was submitted in fulfillment of Grant No. R804155-03-0 by the
University of Georgia under the sponsorship of the U.S. Environmental Protection
Agency. This report covers the period January 1, 1976 to December 31, 1978, and
work was completed as of December 31, 1978.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables xi
Executive Summary xii
Acknowledgments xvi
1. Introduction 1
2. Conclusions 3
3. Recommendations 5
4. Liquid Chromatography 6
General Apparatus 6
Chemicals and Packing Materials 7
Separation of Polynuclear Aromatic Hydrocarbons (PAH) and
Aroclor 1242 on a Qg Column 7
Retention Behavior of Dinitrobenzene (DNB) and Dihydroxyben-
zene (DHB) Isomers on Phenyl-Bonded Phase Columns 9
Comparison of Two Recovery Methods on Bonded Phase Packings.. 16
Effect of Change in Solvent on Chromatographic Peak Shapes .... 18
Comparison of the Recovery on Two Different Bonded Phases 19
5. Coherent Anti-Stokes Raman Spectroscopy 21
Introduction to CARS 21
Initial Setup and Solvent Studies 23
Sample Cell Development 23
Signal Enhancement in the Rigorous Resonance Region 24
Photomultiplier Detection 33
Raman Spectra of Fluorescent Compounds 34
UV Experiments 35
Lineshapes in CARS Spectra 36
Preresonance Enhancement of Fluorescent Compounds 37
Resonance Enhancement of Fluorescent Compounds Near Their
Absorption Maximum 38
6. Computer Interfaces 43
References 51
Collaborators and Publications 53
-------
FIGURES
Number Page
1 Chromatogram of nine polynuclear aromatic hydrocarbons
using Lichrosorb RP 18 packing and isocratic elution. Column:
30 cm x 0.5 cm i.d., Lichrosorb RP18, 10|a; solvent compo-
sition: 80% methanol/20% water; flow rate: 1 ml/min; temp:
ambient; detector: UV (254), 0.1 absorbance, full scale ... 8
2 Chromatogram of 42% chlorinated biphenyls (Aroclor 1242)
using Lichrosorb RP 18 packing and gradient elution. Column:
30 cm x 0.5 cm i.d., Lichrosorb RP 18, lOu. ; mobile phase,
linear gradient: 30% methanol/70% water to 80% methanol/
20% water; gradient time: 60 min; temp: ambient; flow rate:
1.5 ml/min; detector: UV (254), 0.1 absorbance, full scale ... 9
3 Separation of dinitrobenzene isomers on the phenyl-bonded
phase packing at different solvent composition. Column:
60 cm x 0.5 cm i.d., Phenyl Bondpak/Porasil B, 37~75u.;
temp: ambient; detector: UV (254), 0.1 absorbance, full
scale 11
(A) Solvent composition: 20% methanol/80% water for
the first 120 minutes and 45% methanol/55% water
after 120 minutes; flow rate: 4.5 ml/min
(B) Solvent composition: 35% methanol/65% water;
flow rate: 3 ml/min
(C) Solvent composition: 60% methanol/40% water;
flow rate: 2 ml/min
4 Separation of dinitrobenzene isomers on the octadecyl-
bonded phase packing at different solvent composition.
Column: 30 cm x 0.5 cm i.d., Lichrosorb RP 18, 37~75u.;
temp: ambient; detector: UV (254), 0.1 absorbance,
full scale 12
(A) Solvent composition: 20% methanol/80% water;
flow rate: 3 ml/min
(B) Solvent composition: 35% methanol/65% water;
flow rate: 2 ml/min
VI
-------
Number Page
(C) Solvent composition: 60% methanol/40% water
flow rate: 1 ml/min
5 Comparison of the separations of dinitrobenzene isomers on
two different bonded phase packings under the same condition
except the particle size. Column: 30 cm x 0.5 cm 5 ,d.,
temp: ambient: flow rate; 2 ml/min; solvent composition:
45% methanol/55% water> detector: UV (254), 0.1 absorbance,
full scale 13
(A) The octadecyl-bonded phase (Lichrosorb RP 18),
particle size: 10 ^
(B) The phenyl-bonded phase (Phenyl Bondpak/PorasiI B),
particle size; 37~75 ^
6 Separation of dihydroxybenzene isomers on the phenyl-
bonded phase packing at different solvent composition.
Column: 60 cm x 0.5 cm i.d., Phenyl Bondpak/Porasi I
B, 37~75 ,j> temp: ambient-, detector; UV (254), 0.1
absorbance, full scale 14
(A) 30% methanol/70% water; flow rate; 2 ml/min
(B) 15% methanol/85% water; flow rate; 2 ml/min
(C) 5% methano 1/95% water; flow rate: 2.5 ml/min
(D) 0% methanol/100% water; flow rate: 2.5 ml/min
7 Separation of dihydroxybenzene isomers on the octadecyl-
bonded phase packing. Column: 30 cm x 0.5 cm i.d.,
Lichrosorb RP 18, 10^; temp: ambient; detector: UV (254),
0.1 absorbance, full scale; flow rate: 2 ml/min; solvent
composition: 100% water 15
8 A portion of the preresonance coherent anti-Stokes Raman
spectrum of trans-beta-carotene. u^ is 575 nm. The concen-
tration is 10-3 M 25
9 Observed (...) and calculated (solid line) CARS excitation
profile for the j/j vibrational mode of beta-carotene 28
10 Observed (...) and calculated (solid line) CARS excitation
profile for the v2 vibrational mode of beta-carotene 29
11 Calculated spontaneous Raman excitation profile for the ^
vibration of beta-carotene using the same parameters as in
Figure 3 30
VII
-------
Number Page
12 Calculated spontaneous Raman excitation profile
for the ^2 vibration of beta-carotene using the
same parameters as in Figure 4 31
13 Plot of the log of the power in ul versus the log
of the power in 03 for trans-beta-carotene (•)
and benzene (O) 32
14 Preresonance CARS spectrum of beta-carboline 34
15 Preresonance CARS spectrum of beta-naphtha I 35
16 UV-CARS spectrum of the CH stretch region
in methanol 36
17 Automated UV-CARS spectrum of toluene 37
18 CARS spectrum of perylene in benzene taken
at a pump wavelength of 450 nm. The quantum
efficiency of fluorescence for this compound
is .94 39
19 CARS spectrum of rhodamine 6G in ethanol
taken at a pump wavelength of 535 nm.
The quantum efficiency of fluorescence for
this compound in ethanol is .95 39
20 CARS spectrum of rhodamine B in ethanol
taken at a pump wavelength of 555 nm.
The quantum efficiency of fluorescence for
this compound in ethanol is .40 40
21 CARS spectrum of sodium fluorescein in
water at an alkaline pH taken at a pump
wavelength of 510 nm. The quantum
efficiency of fluorescence for this compound
is .90 40
22 CARS spectrum of acridine orange
in water at a pump wavelength of
510 nm. The quantum efficiency
of fluorescence for this compound
is above .50 41
VIII
-------
Number Page
23 CARS spectrum of coumarin 495 in
ethanol at a pump wavelength of
425 nm. The quantum efficiency
of fluorescence for this compound
is above .70 41
24 Excitation profiles for the ring breathing
mode of tetracene as a function of the
pump wavelength and displacement
parameter. The solid lines represent
the theoretical profile at various
displacement parameters. The circles
represent the experimentally obtained
values 42
25 Experimental CARS setup used in the
studies . 44
26 Schematic of the peripheral bus
(peribus) used to extend the number
of possible I/O ports. The symbols
are; OB - output bit, DA - data
available, DT - data transmit, IB -
input bit and REQ - request. The
pull up resistors are 1 k. The
resistor-capacitor circuit employed
a 51 k resistor and a 560 pF capacitor 45
27 Simplified schematic of the basic
operation of the scan unit inter-
face. Both the input and output
ports of the DR1 Ik were required
for communication with the computer 46
28 Simplified schematic of the basic
operation of the stepping motor
interface. The DR11C I/O port-
was used for controlling this inter-
face 47
29 More detailed schematic of the
stepping motor interface under
local and computer control 48
IX
-------
Number Page
30 Schematic of the actual circuit
used for the stepping motor inter-
face. The symbols are; OW0 -
output word bit 0, SA - computer/
local switch, SB - direction switch,
SC - scan switch, SD - single step
switch, OSC -oscillator, MSI -
limit micro-switch for the CW direction
and RUN turns on an external
oscillator 49
-------
TABLES
Number Page
1 Elution orders of dinitrobenzene isomers on the octadecyl
and the phenyl bonded phase packings using different
solvent compositions 10
2 Results of recovery studies on the CIg column using two
different methods 17
3 Results of recovery studies of some compounds by injection
method using two different packings 20
XI
-------
EXECUTIVE SUMMARY
The goal of this research was to explore the use of resonance enhanced
coherent anti-Stokes Raman spectrometry (CARS) as the basis for a specific
identification system for liquid chromatography. The overall system, which also
includes a liquid chromatograph and a UV-visible spectrometer, not only provides
sensitive detection but also multiple means of identification through retention
times (volumes), UV-visible absorption spectra, and vibrational Raman
spectra.
Work was concentrated in these areas: CARS spectroscopy, liquid
chromatographic preconcentration and separation, and computer control of the
LC/UV-visible/CARS experiments. The overall goal has been accomplished
successfully, and we have demonstrated using a model chemical sample that the
system is capable of: (a) separating a complex mixture using liquid chromatography,
(b) detecting individual components at a fixed, preselected wavelength, (c) obtain-
ing the entire absorption spectrum near the point of maximum absorption using a
stopped-flow mode, (d) analyzing that spectrum to determine the optimum condition
for exciting the Raman spectrum, (e) starting the flow to pass the solution from the
absorption cell to the Raman emission cell, and (f) stopping the flow once again to
obtain the anti-Stokes Raman spectrum under optimum resonance conditions. Steps
(b) to (f) can be completed in approximately 10 minutes per species. By starting the
flow once again, successive species e luting from the liquid chromatograph can be
characterized.
Now that the feasibility of the system has been demonstrated, one can easily
predict how to modify the existing system so as to reduce the time required to
characterize the single species in less than two minutes. In addition,
XII
-------
one can predict that the versatility of the system could be expanded to include
other types of measurements.
The bases of those predictions can best be understood by following the
stepwise developments that occurred during the past three years. In the CARS
area, it was first necessary to purchase the powerful lasers and assemble the
spectrometer. Using only manual adjustments, an approach for attaining
the optimum resonance condition was developed following which it was possible
to show that spectra could be obtained on volumes of less than five microliters
while at the same time having a detection limit under favorable circumstances
of less than one micromolar. Furthermore, low limits were obtained using
highly fluorescent compounds, e.g., rhodamine 6G, thereby demonstrating
good rejection of unwanted fluorescence. In addition, it was found that
laser power had to be decreased when working with very dilute solutions
so as to avoid saturation of the excited state. Finally, work was extended
into the ultraviolet range.
In chromatography a number of different species were surveyed but
most of the effort was devoted to comparing selectivities of different packing
materials and exploring special solvent effects. For example, if the sample
is dissolved in one solvent and eluted from the chromatographic column using
a second, a single substance was sometimes found to produce two peaks instead
of one. Finally, observations were made relating experimental variables in
the preconcentration step to the recovery of a given species.
The final step involved overall automation of a system that consisted
of a liquid chromatograph, a scanning ultraviolet-visible spectrophotometer,
and the Raman spectrometer. In order to speed up the process of interfacing
as well as to increase its flexibility, a modular approach was taken.
xiii
-------
Furthermore, the designs were such that control and data acquisition could
be easily effected using the real-time Fortran provided by the Digital Equip-
ment Corporation for its PDF 11/34 minicomputer.
With respect to the liquid chromatograph, the system permitted selection
between two columns so that, if desired, a sample could be preconcentrated
in one column and then moved to a second column for analysis. In addition,
operating parameters such as flow rate of the eluent, solvent composition,
and temperature could be monitored and controlled.
A UV-visible spectrophotometer was set at a pre-selected wavelength
in order to be used as a detector. Shortly after the maximum absorption for
a species had been detected, the flow was stopped and the spectrophotometer
was scanned in order to obtain a spectrum. After the flow had been resumed,
the absorption continued to be measured so that the peak could be integrated
for quantitative purposes.
Meanwhile, from the ultraviolet-visible spectrum, the optimum wave-
length was calculated for exciting the Raman emission. After the optimum
pumping frequencies had been set, mirror positions were adjusted so as to
obtain the proper crossing angles and the emission was directed at a spectrometer
that served to reject stray light, including that from fluorescence. The latter
was done by adjusting the angle of the detector relative to the sample cell.
It is important to note that the computer adjusted the instrument before each
measurement.
The software was developed independently for each major function
for controlling each of the instruments, for detecting and integrating chro-
matographic peaks, and for calculating results such as those for the optimum
xnv
-------
wavelength for irradiation. In addition, there was on-line signal-to-noise
enhancement using signal averaging and statistical treatment of data to reject
out-I/ing results. Finally, there was digital smoothing and plotting of data.
xv
-------
ACKNOWLEDGMENTS
The helpful discussions with Dr. Leo Azarraga, Environmental Protection
Agency, Office of Research and Development, Environmental Research Laboratory,
Athens, Georgia, are gratefully acknowledged.
XVI
-------
SECTION 1
INTRODUCTION
There is a rapidly growing need for detection systems for liquid chromatography
(LC) that provide a capability comparable to that obtained from the combination of a
gaschromatograph with a mass spectrometer. Although attempts have been made to
adapt mass spectrometry to liquid chromatography, that marriage is an unrealistic one
in the sense that liquid chromatography can best be used (in preference to gas chroma-
tography) for compounds that are unstable to heat or are non-volatile. While those
compounds can sometimes be handled by adding another step, such as derivatization
or pyrolysis, extra work is required, and the capability for on-line detection and
quantitation is usually lost. Furthermore, the question of how quantitatively or repro-
ducibly the derivatization or pyrolysis can be done has often been a limiting factor.
Hence, a direct analysis is highly desirable.
A major goal of our work has been to make a detailed evaluation of coherent
anti-Stokes resonance Raman spectroscopy using relatively non-volatile organic
compounds, mostly aromatics, that are typical of those encountered in different phases
of coal utilization and in the pesticide-insecticide area.
Since the coherent anti-Stokes Raman signal can be measured with little
interference from fluorescence because: (a) the CARS signal occurs at a higher energy
than the excitation (whereas fluorescence occurs at a lower energy) and (b) the fluoresc-
ence is emitted in all directions so the fraction in the small solid angle occupied by the
spatially coherent CARS signal will be very small. We have been able to record CARS
spectra of molecu les whose fluorescence woul d prohibit their measurement us ing the spontan-
eous Raman effect.
-------
As will be shown later, use of resonance conditions along with ordinary
CARS enhancement has pushed the detection limits for many compounds below
the micromolar level. In order to achieve this level of detection, much
has been done to optimize cell design and to determine solvent limitations
on the detection of very dilute solutes.
Since most molecules of interest will have absorptions in the UV region
of the spectrum, it was advantageous to carry out resonance enhancement
studies in this range. By using a frequency doubling technique and photo-
multiplier detection, we have been able to record strong CARS signals in the
CH stretch region of the polar methanol molecule.
-------
SECTION 2
CONCLUSIONS
Our results indicate that coherent anti-Stokes Raman spectroscopy
(CARS) may be used as a specific detector for liquid chromatography (LC).
The two major problems involved in applying CARS as a specific LC detector
have been overcome. The small internal volume (1 to 10 microliters) require-
ment has been solved by using capillary cells. Based on the shape (cylindrical)
and inside diameter (1.2 mm), we have calculated and experimentally verified
that flow rates as high as two meters per second can be used without turbulent
flow disturbing the beams. This is over an order of magnitude faster than a fast
LC flow rate.
The second problem of being able to detect species at concentrations of
10~ to 10 M has also been solved. The previous drawback to using CARS for
low concentrations had been the fact that the solute signal was obscured by the
strong non-resonant background signal of the solvent and allowed lower limits of
_2
detection of only 5x10 M. The solution was to increase the solute signal over
the solvent signal by pumping the system with a frequency resonant absorption fre-
quency of the solute. This adjustment increased the relative power of the solute
signal by several orders of magnitude and has allowed solute detection at concen-
trations of 5 to 7 M. This has allowed empirical establishment of the lower molar
limit of detection by CARS to be (10 • E) M where E is the decadic molar absorp-
tion coefficient of the solute.
In order to apply the resonance enhancement technique to a wide range
of compounds, it is necessary to be able to pump in the ultraviolet region of
the spectrum. We have shown that by using frequency
-------
doubling techniques we can obtain UV beams of sufficient intensity to produce
strong CARS signals. This means that the CARS technique will be applicable
for a great number of compounds of interest.
In the liquid chromatographic studies, a phenol-derivatized silica packing
was found to be more selective than a C,g-derivatized packing for highly polar
aromatics, so it should provide better chromatographic separations of polar pollu-
tants in water. In addition, an injection method for testing recovery, in which the
enriched sample is dissolved in one solvent and then eluted with a second suitable
solvent has proven to be an easy, fast approach for testing the relative recoveries
of water pollutants. In a related study, double peaks for a single component were
found under some conditions when two solvents were employed. Finally, prelimi-
nary schemes have been developed for the liquid chromatographic separation of
polynuclear aromatics and pal/chlorinated biphenols. The separations of components
were usually large enough to allow individual species to be investigated by CARS.
The design of the computer interface has been based on a modular approach.
This design allowed the development of each controlled experimental function as
a unit. Each unit can be addressed individually using the software. The overall
result is that, by using this apparatus one can now obtain a much better (i.e.,
higher S/N) CARS spectrum in six minutes than one could obtain in two hours by
manual operation. It is now possible to characterize isolated solute peaks (both
UV-visible absorption and Raman emission) in approximately 10 minutes. The only
additional time is that required for the chromatographic fractionation itself.
-------
SECTIONS
RECOMMENDATIONS
Now that the basic approach has been demonstrated to be feasible, one
can predict how parts of the development process could be speeded up. First,
both UV-visible absorption and Raman emission could be measured on a time-
scale smaller by a factor of 10 by using vidicon (multi-cathode) detectors. Those
detectors permit simultaneous measurements to be made over a relatively wide
range of wavelengths. Second, the system could also be made much more com-
pact by using newer, smaller lasers having equivalent power. Hence, a more
manageable, table-top unit could be constructed by a user who wishes to dupli-
cate the present capability.
-------
SECTION 4
LIQUID CHROMATOGRAPHY
Our main efforts in the liquid chromatographic study can be described in
three different areas. First, quantitative separations of water pollutants are the
1-3
chief goal. In the past, different absorbents such as XAD resins and charcoal
have been Osed for preconcentration. Recently, an on-column method based on
4
a C)8 column has been developed for both concentration and separation. The
second area involved in our studies was to characterize the phenyl-bonded phase
which has not yet been used by others. We have studied dinitrobenzene and dihy-
droxylbenzene isomers.
Besides separation, the efficiency of the column to collect these trace
organics from water is very important for the present project. Therefore, the third
area covered in our studies was the recovery of the sample from a given column.
The recovery efficiencies of the C,g and the phenyl-bonded phase packings were
compared for some aromatic compounds. In this study, both the pumping method
and the Injection method were used.
GENERAL APPARATUS
The liquid chromatograph used for the present work is a Micromeritics Model
7000 B. This instrument is equipped with a 6000 psi solvent delivery system with
constant flow and different gradient capabilities. It is also equipped with a UV
detector operating at 254 nm. Later, a GCA McPherson, Model EU-700, UV-
visible spectrophotometer was substituted for the original UV detector. Sample in-
jections were accomplished by means of the Universal Sample Injector (Micromeritics)
with 100 microliters as the maximum capacity of the sample loop.
-------
CHEMICALS AND PACKING MATERIAL
All the chemicals used were reagent grade. Water used was deionized
followed by distillation. In all cases, the mobile phase used was water,
methanol, or their mixture. These solvents were degassed in an ultrasonic
vibrator before use. Columns of C1&/porasil B and phenyl/porasil B (Waters
Associates, 37-75 |j.m) were dry packed. Columns of Lichrosorb RP 18 (E.
Merck Company, 10 |j.m) were packed by using the high pressure slurry method.
The sample solutions, unless otherwise indicated, were prepared by dissolving
each compound in methanol.
SEPARATION OF POLYNUCLEAR AROMATIC HYDROCARBONS (PAH) AND
AROCLOR 1242 ON A C18 COLUMN
The main purpose of this portion of the study was to show the capability
of new chromatographic packing for the analysis of some pollutants in water.
Polynuclear aromatic hydrocarbons (PAH) and polychlorinated biphenyls were
chosen for our study.
The experimental conditions in our work were very similar to what had
been done in the literature5^7. The reverse phase mode of separation based
on the C18 column (Lichrosorb RP 18) and water/methanol as the mobile phase
were used. The column dimension was 30 cm x 0.5 cm. A UV detector was
used and the attenuation was adjusted to 0.1 absorbance unit. The temperature
was ambient and the chart speed of the recorder was 12 cm/hr. The flow
rate in this case was adjusted to 1 ml/min and the solvent composition was
made of 80% methanol and 20% water. From the above experimental condition,
the separation of nine PAH compounds was achieved as shown in Figure 1.
These PAH compounds were identified from their retention times.
-------
Figure 2 is the chromatogram of Aroclor 1242 (PCB mixture) separated
by the same system using gradient elution from 30% MeOH/r^O to 80%
MeOH/H2O in 60 minutes. Here the flow rate was adjusted to 1 .5 ml/min.
Though each peak in this chromatogram was not identified, this separation
indicates that the Cjg column is a very efficient tool for the analysis of PCB
compounds.
The CARS technique should be able to provide structural information
for the above compounds. From the above separations, it is evident that
1 • Benzene
2. Naphthalene
3. Biphenyl
4. Phenonthrene
5. Anthracene
o. 2-PhenylnapMhalene
7. l.l'-Binaphthyl
8. 2,2'-Binophrtiyl
9. Pyrene
20
40
Time (min)
60
80
1. Chromatogram of nine polynuclear aromatic hydrocarbons
using Lichrosorb RP 18 packing and isocratic elution.
Column: 30 cm x 0.5 cm i.d., Lichrosorb RP 18, 10(j;
solvent composition: 80 <#, methanol/20% water* flow rate:
1 ml/min* temp: ambient* detector: UV (254), 0.1 absorbance,
full scale .
8
-------
the coupling of LC and CARS should be a powerful method for the monitoring
of these pollutants in water. In addition, since the reverse phase mode of separation
is used here, it may be possible to concentrate these pollutants and resolve them
in the same column without tedious sample preparation.
RETENTION BEHAVIOR OF PI NITROBENZENE (DNB) AND DIHYDROXY-
BENZENE (DHB) ISOMERS ON PHENYL BONDED PHASE COLUMNS
The phenyl-bonded phase is often overlooked because a much better
separation of PAH can be done on the Cj8 column. However, it has been
mentioned that the phenyl-bonded phase has a strong interaction with the
aromatic compounds.8 In addition, this packing material seemed to retain
r
40
60
T~
80
Time(min)
~T
100
120
2. Chromatogram of 42% chlorinated biphenyls (Aroclor 1242)
using Lichrosorb RP 18 packing and gradient elution. Column:
30 cm x 0.5 cm i.d., Lichrosorb RP 18, lOu •, mobile phase,
linear gradient: 30% methanol/70% water to 80% methanol/
20% water; gradient time: 60 min; temp: ambient; flow rate:
1.5 ml/min; detector: UV (254), 0.1 absorbance, full scale.
-------
some polar compounds longer than the Cj8 column.8 Thus, It is of interest
to determine the retention behavior on the phenyl-bond phase with some
model compounds and compare it with the Cjg column . Isomers of dinitro-
benzene (DNB) and dihydroxybenzene (DHB) were chosen for this study because
they are aromatic and have different polarities. Solution of these isomers
were prepared in methanol just before use. The experimental setup for this
study was basically the same as that described above. Both the phenyl-bonded
phase and the C18 phase were used for the separation of DNB isomers or DHB
isomers. Four different eluent compositions, 20%, 35%, 45%, and 60%
MeOH/h^O, were used in comparing the two bonded phases.
Results of the study of the DNB isomers are shown in Figures 3 through
5 and in Table 1 . Figures 3 and 4 are the chromatograms of two columns
for the separation of DNB isomers using different solvent compositions. It
is interesting to note that the elution order of DNB isomers on the Cj8 column
varied from one solvent composition to another while it was constant on the
phenyl-bonded phase. In addition, the resolution of these DNB isomers on
Table 1. Elution orders of dinftrobenzene isomers on the octadecyl and the phenyl
bonded phase packings using different solvent compositions.
The octadecyl bonded
phase packing,
Lichrosorb RP 18
The Phenyl bonded
phase packing,
Bondpak Phenyl/Porasil B
Solvent
Composition
% Methanol/% Water
20
35
60
20
35
60
Flow
Rate
(MI/Min)
3
2
1
4.5
3
2
Elution
Order
p
-------
the pheny I-bonded phase increased when the percentage of methanol decreased.
This, however, was not generally true if the C18 column was used. Figure 5
is another comparison of these two bonded phases based on the same conditions
except for the particle size. It is apparent that the phenyl-bonded phase
was superior to the Q8 column for the separation of DNB isomers even when
the particle size was much larger.
(A)
i 1—i—i
10 20
Mil
IB]
—1—
10
20
Mil
10
Mil
Separation of dinitrobenzene isomers on the phenyl-bonded
phase packing at different solvent composition. Column:
60 cm x 0.5 cm i.d., Phenyl Bondpak/Porasil B, 37~75u.j
temp: ambient; detector: UV (254), 0.1 absorbance, full
scale .
(A) Solvent composition: 20% methanol/80% water for
the first 120 minutes and 45% methanol/55% water
after 120 minutesj flow rate: 4.5 ml/min
(B) Solvent composition: 35% methanol/65% waterj
flow rate: 3 ml/min
(C) Solvent composition: 60% methanol/40% waterj
flow rate: 2 ml/min
11
-------
The results of the separation of DHB isomers on these two bonded phases
can be seen in Figures 6 and 7. It is interesting to note that the elution order
(P temp: ambient^ detector; UV (254), 0.1 absorbance,
full scale .
(A) Solvent composition: 20% methanol/80% water>
flow rate: 3 ml/min
(B) Solvent composition: 35% methanol/65% waterj
flow rate: 2 ml/min
(C) Solvent composition: 60% methanol/40% water
flow rate; 1 ml/min
12
-------
a stronger interaction between the solute and the stationary phase. Further
evidence can be seen from the longer retention times of these isomers compared
with those from the C18 column under the same condition. (See Figure 5.) It
is also expected that the phenyl-bonded phase packing of a smaller particle
size (5 ~10 ^m) will give a much better separation and much higher capacity.
(A)
O + p
(B)
10
Time(min)
I
20
T~
10
—I—
20
30
Tlme(min)
Comparison of the separations of dinitrobenzene isomers on
two different bonded phase packings under the same condition
except the particle size. Column: 30 cm x 0.5 cm i .d.,
temp: ambient: flow rate; 2 ml/minj solvent composition:
45% methano 1/55% waterj detector; UV (254), 0.1 absorbance,
full scale .
(A) The octadecy I-bonded phase (Lichrosorb RP 18),
particle size: 10 ^
(B) The phenyl-bonded phase (Phenyl Bondpak/Porasil B),
particle size; 37~75jj.
13
-------
The separation of DHB isomers can be done by both C18 and the phenyl-bonded
phase using the same composition of mobile phase. It is interesting that these
highly polar isomers could be well separated in a reverse phase system.
The separation mechanism of these isomers on the phenyl-bonded phase
can be understood from the consideration of dipole-dipole interaction. This
A
B
I m t
10 20 30
TIMC(ni»
Separation of dihydroxybenzene isomers on the phenyl-
bonded phase packing at different solvent composition.
Column: 60 cm x 0.5 cm i.d., Phenyl Bondpak/Porasil
B, 37 ~ 75^ temp; ambient* detector; UV (254), 0.1
absorbance, full scale .
(A) 30% methanol/70% water* flow rate; 2 ml/min
(B) 15% methanol/85% water> flow rate: 2 ml/min
(C) 5% methanol/95% water) flow rate; 2.5 ml/min
(D) 0% methanol/100% water; flow rate; 2.5 ml/min
14
-------
can be proved from the elution order of DNB isomers or DHB isomers. O-DNB,
with the strongest dipole moment among three isomers, has the strongest inter-
action with the dipole developed on the surface of the phenyl-bonded phase.
Thus, it was the last species eluted out of the column. M-DNB,which has
the second strongest dipole moment,came out second and P-DNB, which has
smallest dipole momenfrcame out first. The same principle is also valid for
the separation of DHB isomers. This interaction, however, was not a dominant
force on the Cj8 column because of the low polarity of the C18 and the residual
hydroxyls on the silica. Therefore, a relatively weak dipole-dipole inter-
action will occur in the Cj8 column, and the change of solvent composition
may overcome this force to change the elution order. If this is true, the
phenyl-bonded phase should have a better separation for the DHB isomers
than the C18 column. Besides, the phenyl-bonded phase may have better
selectivity for those polar pollutants such as carbamates and nitrosoamines.
Most of this work has been published.9
TIME (min>
Separation of dihydroxybenzene isomers on the octadecyl-
bonded phase packing. Column: 30 cm x 0.5 cm i.d.,
Lichrosorb RP 18, 10^ temp: ambient; detector: UV (254),
0.1 absorbance, full scale; flow rate: 2 ml/min; solvent
composition: 100% water .
15
-------
COMPARISON OF TWO RECOVERY METHODS ON BONDED PHASE PACKINGS
As mentioned above, both the pumping method and the injection method
were used for recovery studies. In the pumping method, a very dilute solution
of the solute was passed through the column following which the retained
solute was stripped by using a good solvent such as methanol. This is very
similar to the procedure that would be-followed in the analysis of a real sample.
The injection method, however, is a different approach. In this method, a
small volume of a concentrated sample was injected into the column followed
by a large volume of wash-water prior to stripping with methanol. The second
method has the potential advantage of minimizing adsorption losses onto walls
of vessels and tubing before the sample enters the column.
The apparatus used for the recovery study is essentially the same as
that described above. The column used for this study was the bonded C18
packing (20 cm x 0.5 cm). Sevin and naphthalene were the model compounds
for these two different methods. The concentration of the sample prepared
in methanol was 10~3to 10~4 M. In the pumping method, the solution was
made by adding a small amount of sample (about 50 jj) to 25 ml of water,
then the experiment was carried out according to the following procedure.
(a) Wash the column and the intensifiers with 50 ml MeOH.
(b) Wash the column and the intensifiers with 50 ml h^O.
(c) Pump the sample solution (25 ml) directly through the column (C18)
at a flow rate of 1 ml/min.
(d) Pump 10 ml of clean water through the whole system to clean the sample
residue in the intensifiers.
(e) Elute the sample, retained in the column, by using pure MeOH.
16
-------
The result of this method for Sevln and naphthalene can be seen in
Table 2. It is obvious that the measurements obtained from this experiment
were not very reproducible and thus no useful data were obtained. The
reason for these poor recoveries is adsorption losses as described above.
This adsorption problem must definitely be studied further in order to know
the exact recovery for a given sample.
In the injection method, the experimental procedure is listed as the
following.
(a) Wash the column and the intensifiers with 50 ml MeOH.
(b) Wash the column and the intensifiers with 50 ml H2O.
(c) Pump water continuously across the column as the mobile phase. Flow
rate is adjusted to 1 ml/min.
(d) Inject the sample (~50 |J) into the column.
Table 2. Results of recovery studies on the C.. Column using two different methods*
Compound
Studied
Naphthalene
•
Sevin
Pumping
Trials
1
2
3
4
1
2
Method
Recovery
42%
37%
10%
57%
20%
80%
Trials
1
2
3
4
1
2
Recovery
80%
87%
84%
87%
100%
100%
"Recovery was measured from the ratio of peak heights
17
-------
(e) Wash the column with h^O for 25 minutes or more.
(f) Change the mobile phase to pure MeOH and elute the sample from
the column.
The result based on this method is shown in Table 2. It is obvious that
these recoveries were very high and reproducible, especially if we compare
them with those previously obtained. However, it is important to note that
the samples must be soluble in the total amount of washwater. Otherwise,
the data are meaningless. It is also obvious from this study that the adsorption
losses due to the vessels or tubings are a crucial factor for compound
recover/.
Another factor that affects the recovery is the retention volume of
the species. If, for example, a species can be eluted from a column using
200 ml of water, such a column cannot be used to preconcentrate that species
from 200 ml or more of sample. Although that observation may be obvious
when stated in those terms, it appears that low recoveries reported in the
literature for some species may have resulted from passing too large a volume
of sample through the column used for preconcentration. Phenol, for example,
falls into that category.
EFFECT OF A CHANGE IN SOLVENT ON CHROMATOGRAPHIC PEAK
SHAPES
Using solutions containing single dihydroxybenzene isomers as well
as a solution containing all three, produced an unusual case of peak-
splitting. In spite of the fact that water and methanol are completely miscible,
each species, when dissolved in alcohol, gave two peaks (or a shoulder plus
a peak) when injected into a column and eluted with water. Similarly, an
18
-------
aqueous sample Injected into a methanolic eluent gave the same type of be-
havior. However, a sample dissolved in water injected into an
aqueous eluent or a sample in alcohol injected into an eluent of alcohol
gave only a single, sharp peak for each species.
The above observations are important for two reasons. First, care must
be exercised in matching solvents if retention times (volumes) of peaks are
to be used for identification. Second, column efficiencies can be adversely
affected, leading to true overlaps of two different species as a result of peaks
for individual species being wider than normal. These results have been pub-
lished. 1°
COMPARISON OF THE RECOVERY ON TWO DIFFERENT BONDED PHASES
Since two bonded phases were used for the present project, it would
be interesting to know which of these two had the better recovery efficiencies
for the aromatic compounds tested. At the present time, this study is still
not complete^ therefore, only preliminary results are available.
The injection method was basically used for both the C18 column (20 cm
x 0.5 cm) and the phenyl-bonded phase column (30 cm x 0.5 cm). Sevin,
matacil, furdan, and naphthalene were tested for the C18 column. For the
phenyl-bonded phase, 2-naphthol was the only compound used. 2-Naphthol
was the only compound commonly used^ however, the recoveries of this com-
pound were about the same in these two different systems. It is interesting
to note that 2-naphthol and the carbonates (e.g., Sevin, furdan, and matacil)
were strongly retained in the C18 column even though these compounds are
polar. Thus, the C18 column may be widely used for different types of com-
pounds. (See Table 3.)
19
-------
Table 3. Results of recovery studies of some compounds by injection method using
two different packings*
Compound Recovery from the Recovery from the
octadecyl-bonded phenyl-bonded
phase packing phase packing
Sevin 90%
Matacil 83%
Furdan 93%
Naphthalene 85%
2-Naphthol 86%
* Recovery was measured from ratio of peak heights
From the above studies, we know the phenyl-bonded phase has a stronger
Interaction than C|g with some polar compounds (e.g., DNB or DHB). There-
fore, the phenyl-bonded phase may be a better column for the concentration
and separation of water pollutants like carbamates, herbicides, and nitrosamines.
Further study in this area will extend to those nonpolar compounds such as
PAH.
20
-------
SECTION 5
COHERENT ANTI-STOKES RAMAN SPECTROSCOPY
INTRODUCTION TO CARS
There is a rapidly growing need for specific identification systems for liquid
chromatography (LC) that can provide a capability comparable to that obtained from
the combination of a gas chromatograph with a mass spectrometer. Although attempts
have been made to adopt mass spectrometry to liquid chromatography, that marriage is
an unrealistic one in the sense that LC can best be used for compounds that are unstable
to heat or are non-volatile. In addition to being a specific detector, this technique
should be reasonably sensitive and useful with a wide variety of solvents including water.
One technique that would satisfy these requirements is Raman spectroscopy. The
major problem with Raman spectroscopy is its inherent inefficiency (1 photon scattered to
8 .
10 incident). Although resonance Raman spectroscopy allows many orders of magnitude
enhancement, this enhancement is often negated by strong fluorescence. Another Raman
technique that shows promise is coherent anti-Stokes Raman spectroscopy.
A major goal of our work has been to make a detailed evaluation of coherent anti-
Stokes resonance Raman spectroscopy using relatively non-volatile organic compounds,
mostly aromatic, that are typical of those encountered in different phases of coal utili-
zation and in the pesticide-insecticide area.
Coherent anti-Stokes Raman spectroscopy (CARS) is a nonlinear optical technique
that has been applied to obtain Raman spectra with very high efficiencies9"22 CARS
involves the use of two tunable dye lasers set at frequencies u1 and u., respectively.
When these two light beams cross in the sample at the phase matching angle 0, coherent
anti-Stokes emission at u. = 2O, - u_ is generated through the third order nonlinear polar-
ization. This laser-1 ike beam, u,, appears on the opposite side of the pump beam, u]F
21
-------
from the Stokes beam uy at the angle 0. The intensity of the signal is greatly enhanced
where the frequency interval, o, - a>2 = A, is equal to a Raman active molecular vibrational
frequency. Raman spectra are normally obtained by fixing the frequency of u. and
varying the frequency of u. and, therefore, changing A .
The major drawback to the CARS technique has been the intensity of the back-
ground emission resulting from the non-resonant third order susceptibility of the solvent.
This usually allows detection limih for solutes of only about 0.05 M. Recently, Chabay,
Klauminzer and Hudson23 have shown that when the u, beam approached an allowed
electric dipole transition, the signal generated was greatly enhanced whereas the back-
ground emission remained virtually unchanged. This preresonance enhancement allowed
the authors to record a very low noise high intensity spectrum of diphenyloclatetraene at
—3
concentrations of 1 .4 x 10 M. By going closer to resonance, this lower limit can be
reduced to the micromolar level.
Because CARS starts and ends in the ground state/both energy and momentum must
be conserved. For the energy
u1 - u2 + u1 - u3 = 0
where u>1 is the pump frequency, u. the stimulating frequency and u the signal frequency.
In this scheme o. = 2o, - u_. Since the u, signal increases drastically when y. - u = u ,
o I L o I i p
the output peaks will occur at u, = u, + u or anti-Stokes shifted. For the momentum
j i r
h . 2w 2irv
p
-------
This means that to obtain a signal the beams u, and u. will have to be crossed at an
angle 0. Also, the signal beam, u3, will be removed from these beams. Therefore,
the CARS technique offers not only higher efficiencies but also essentially complete
rejection of fluorescence since the signal beam u, is both spatially and frequency
removed from the fluorescence signal. That is, the CARS signal at u occurs at
•j
higher frequency than either of the exciting frequencies. Hence, very efficient
9
(up to 1 in 10 ) discrimination of the natural fluorescence of the molecule is possible
because fluorescence involves the emission of energy at lower frequencies than the
excitation process. Also, fluorescence occurs over 4rr steradians whereas the CARS
signal is emitted in a laser-like beam. The fluorescence can be reduced greatly by
observing the signal through a pin hole well away from the sample. In our particular
setup, the beam passes through an iris 30 cm away from the sample. This corresponds
to an f of~100.
INITIAL SET UP AND SOLVENT STUDIES
In evaluating the applicability of CARS as a specific detector for liquid
chromatography, we have had to modify the system to fit our needs. Since each sol-
vent will have a different index of refraction and dispersion, the crossing angle needed
to be changed drastically from solvent to solvent. To avoid constant repositioning of
the collection optics, a rotation stage was constructed. This stage holds all of the
collection optics and detectors and pivots about the crossing point of the u^, u2, and
u. beams. When the crossing angle is changed, the stage is merely rotated to the new
a
angle and all of the optics are in alignment.
Since the dispersion is a function of frequency, the crossing angle and, therefore,
the collection stage will need to be varied during the scan. In order to tabulate both
crossing angles, and their frequency dependence, the original u2 steering mirror was
replaced with a vernier drive mirror mount. This mirror allows one to determine a
scanning algorithm for computer control of the crossing angle.
SAMPLE CELL DEVELOPMENT
In order to be compatible with high resolution liquid chromatography, the sample
cells had to be of small volume (1 to lOu I). Our preliminary results showed that this prob-
23
-------
lem could be solved by using standard melting point capillaries. This proved more con-
venient and considerably less expensive than the standard cuvettes used in past CARS
experiments. The capillary was placed in the plane of u, and u. and perpendicular to
their bisector. The capillary was then raised or lowered until the u, beam passed
undeviated. This placement was critical since the rounded walls greatly displaced
the beams unless they travelled along the diameter. This displacement actually made
the alignment easier. Once the u, beam passed undeviated, the steering mirror for
the u, beam was adjusted to make it pass undeviated. This visual alignment is much
more sensitive than the knife edge test for vertical placement. Once aligned verti-
cally,the steering mirror for the u. beam was adjusted horizontally to obtain the CARS
signal. The loss in signal compared to a cuvette is only on the order of 10%.
In order to test the feasibility of observing CARS signals in a flowing system,
we have simulated the flow conditions experienced in a liquid chromatograph by using
an infusion pump. Even at maximum flow rates, no additional noise was observed.
Based on the cylindrical shape and inside diameter (1.2 mm) we have calculated and
experimentally verified that flow rates as high as two meters per second can be used
without turbulent flow disturbing the beams.
SIGNAL ENHANCEMENT IN THE RIGOROUS RESONANCE REGION
Studies of the CARS intensity as a function of excitation wavelength had not
been previously made in the rigorous resonance region. We have attempted to com-
pare the experimental resonance profiles with those calculated using the Albrecht
theory2*, 25 modified for CARS by Hudson et al.2* /3-Carotene was chosen as a
model system because of the extensive work carried out by Miyazawa and coworkers27
on the excitation profiles for the spontaneous Raman effect. The parameters derived
by these workers were directly transferable for our experiment.
A portion of the CARS spectrum of 0-carotene in benzene solution is shown in
Figyre 8. The pump frequency used, 575 nm, is in the preresonance region. The spectrum
observed compares favorably in both frequency and relative intensity with that reported
by Rimai et al.28 The weak peaks at 1193 and 1215 cm" reported in this reference are
evident in Figure 8.
24
-------
_
Since Miyazawa and coworkers27 found that v,, at 1527 cm , the CC double
bond stretching mode, and v2/ at 1158 cm , the CC single bond stretching mode,
accounted quite well for the electronic absorption progression in the region between
400-500 nm, we chose to study the coherent anti-Stokes Raman excitation profiles for
these two lines.
1158
1527
630
j /
620
615
X(NM)
8 A portion of the preresonance coherent anti-Stokes Raman
spectrum of trans-beta-carotene. wj is 575 nm. The concen-
tration is ID'3 M .
25
-------
The intensity of a CARS signal is proportional to the square of the third
order susceptibility. This susceptibility is the sum of two terms,
x(3) - x(3> + x<3^ 0)
whereX c "s tne electronic susceptibility andX D !s tne Raman susceptibility.
(3) E
X p is a slowly varying function of the frequency difference u] - u2 = A and the
sum is dominated by the Raman susceptibility when Ais equal to an allowed Raman
18
frequency ( A = u ) of the molecule. As shown by Hudson et al . when A = u>
(3) r r
Xn 's given by
2 f
where N is the number density and f the Raman linewidth. The treatment of the
spontaneous resonance Raman data for (3-carotene by Miyazawa et al. followed the
development of Tang and Albrecht. The expression for the scattering tensor, a, is
given by
(0(
< g I R I bXblh la XaU I g>
(4)
< I R lg>
E c^ £_!_
"bv + "2 -I" ' rb
26
-------
X< i IvXv I Q Ij >/ (u, - u )
q bo ao
w v^ r
"bv
B 3_ > "X
bv + "2 + [ rb
< i I Q I v >< v I i >/u, - u
q bo ao
where the terms are identified as follows.
g = ground electronic state
i, j = vibrational states in the ground electronic states
a, b = excited electronic states
v = vibrational state in the excited electronic state
p, or = x, y or z
R = a th component of the electric dipole moment operator
h = vibronic coupling operator 8 H/8 Q with H = electronic Hamiltonian
q q
Q = q'th normal coordinate of the ground electronic state
q
f, = halfwidth of the g -> b electronic transition
Miyazawa et a I., determined that the contribution of the B term was negligible for the
symmetric modes, ^ ar|d v-p in the rigorous resonance region. They calculated the sepa-
rations between the equilibrium positions of the ground and electronic excited states for
the vj and v2 modes by treating these as adjustable parameters to fit the experimental ex-
citation profiles. These values then reproduced the observed low temperature absorption
spectrum when a Lorentzian halfwidth of 250 cm was used for the electronic absorption
I ines.
27
-------
In order to calculate the theoretical CARS excitation profiles, we have used the
separation parameters of Miyazawa et al. and have increased the electronic halfwidth
_1 21
F, , to 500 cm since our measurements were made at room temperature . This halfwidth
reproduced the electronic absorption spectrum at room temperature. The separation
o
parameters used were 1.1 for v ,/ and 1.4 for V2 corresponding to an elongation of 0.02A
o
for the C = C bonds and contraction of -0.031 A for the C-C bonds. The frequencies in
the ground and excited states were assumed to be equal and the summation over v was
II
taken to v = 8. Using this procedure, we have calculated relative values of \\& I anc'
these are the solid curves shown in Figures 9 and 10. For comparison, we have calculated
the excitation profiles for the spontaneous Raman effect for the same experimental condi-
tions. These are shown in Figure 11 and 12. By comparing Figures 9 with 11 and 10 with
12 it is evident that there is a significant difference between the CARS and spontaneous
Raman excitation profiles.
o
o
o
CO CD
z>
>-..
CD
CO
"400.00 420-00 440-00 460.00 480.00 500.00 520-00 540-00
LflMBDfl INM)
Observed (...) and calculated (solid line) CARS excitation
profile for the ^ vibrational mode of beta-carotene .
28
-------
Several of the CARS experimental points were obtainable with different sets of dyes.
In particular, the point at 570 nm could be done using coumarin dyes 485 and 495 or
coumarin dyes 500 and 485. When first attempting these experiments, we noted serious
discrepancies between the ratio of the intensity of v2 TO the benzene 992 cm line
depending on which set of dyes were used. The major difference in the two experiments
was the power in u]. Since the CARS signal of both the solute and solvent lines should
vary with the square of the power in u. and linearly with the power in u_, this ratio should
have remained constant. A plot of log Pu, vs log Pu. should be linear with slope 2.
Figure 13 indicates that the solvent signal (992 cm benzene line) does have slope 2 over
a wide power range in u?. At very low powers, the solute signal (v2 1158 cm ) did
follow the expected slope 2 curve. However, it deviated markedly from this line at
higher powers. In fact, at very high powers, the signal was lost in the background emis-
sion. This explains the anomalous results obtained at 510 nm with the two sets of dyes.
"400.00
420.00
440.00
460-00 480.00
LflMBDfl (NM)
500.00
520.00 540-00
10
Observed (...) and calculated (solid line) CARS excitation
profile for the v2 vibrationdl mode of beta-carotene .
29
-------
The solution to this problem was to lower the laser power in u., 1) by using the nitrogen
laser at lower power, 2) by attenuating u by the use of neutral density filters. This
resulted in peak powers in u. of ~1 kW.
It is possible to gain some insight into the cause of this phenomenon by considering
a simple calculation. At a power of ~100 kW in u. with a coherence length of ~1 mm,
15
there will be ~10 photons/ pulse (pulse length ~6 to 8 nsec) passing through the focal
volume. In this volume at 10 M, there are ~10 solute molecules. Since the decadic
molar extinction coefficient of p-carotene is ~ 10 liter mole cm , it is quite possible
that the system saturates, and increasing the intensity of Uj has no effect on the solute
CARS signal but only increases the background. Assuming u. is intense enough to saturate
and steady state populations of the vibrational levels of the ground electronic state re-
sult, then the intensity of (•>„ is independent of u.. It may be that the power of u>, for
<5 I 0
(3-carotene is asymptotically approaching a limit (Figure 13).
400.00
11
420.00 440.00 460.00 480.00
LflMBDfl (MM)
500.00 520.00
540.00
Calculated spontaneous Raman excitation profile for the v\
vibration of beta-carotene using the same parameters as in
Figure 3 .
30
-------
Once we discovered the necessity of lowering the laser power, it was possible to
determine when we were operating in the linear region.
p-carotene lines to the benzene 992 cm line and continuing to lower the power in
u1 until no change in the ratio resulted. We had already switched to photomultiplier de-
tection for experiments at low power in the UV region using frequency doubling. Due
to the low power required in the resonance experiments reported here, ~1 kW peak power,
photomultiplier detection was also essential. Once at low power, the determination of
the excitation profiles proceeded rapidly.
The ratios of the intensity of (3-carotene lines to the benzene 992 cm line were
corrected for the power differences and absorption differences at the various wavelengths.
This corrected ratio for v. i s 1, 2 is given by
R. =
(6)
"400.00 420.00 440.00 460.00 480.00 500.00
520.00 540.00
12 Calculated spontaneous Raman excitation profile
for the V2 vibration of beta-carotene using the
same parameters as in Figure 4 .
31
-------
where
P (v.) is the observed power at frequency u for the R - carotene
(A) I
P ($) is the observed power at frequency u for benzene
U)
A (v.) is the absorbance at u for R - carotene
A ((j>) is the absorbance at u for benzene.
Since P°, = P
UK o
10 u2 Eq. (6) reduces to
R. =
i
P ,v.
U2 1
where P - is the power measured in u. after passing through the sample. This also proved
a3
o
O
13
LOG R
UJ,
Plot of the log of the power in U] versus the log
of the power in 03 for trans-beta-carotene (•)
and benzene (O) .
32
-------
to be a very convenient method since it was possible to determine the ratio P 2(<£)/P 2(v,)
simply by measuring the relative powers at the three frequencies without the necessity of
moving the sample. It was, however, absolutely necessary to adjust the crossing angle
to maximize the signal at each of the three frequencies (i.e. v ,, v , and v ) in order to
' *
obtain reliable ratios. The experimental points for v, and v2 are indicated in Figures 9
and 10. The agreement between the experimental points and the curves generated using
the known parameters determined from the spontaneous resonance Raman effect is quite
reasonable.
These experiments show that the CARS technique is applicable to the determination
of excitation profiles. Since the method is also fluorescence rejecting, it may be applied
to molecules whose fluorescence prohibits the observation of spontaneous resonance Raman
profiles. We have shown that resonance CARS experiments may be performed with con-
considerably lower power laser systems than the system used here. lf,however, one wishes
to extend these experiments to wavelengths shorter than 360 nm, in which inherently in-
efficient frequency doubling must be used, then high power lasers are necessary. This
extension is desirable because of the much wider range of molecular systems accessible to
resonance experiments.
With the program developed and the UV-visible absorption spectrum, it will
be possible to predict the pump frequencies which will give optimum signal generation.
For our model system this allowed us to detect concentrations of 5 to 7 x 10 M. Since
this CARS setup can be used with cells of 1 u.1 volume, we can easily detect 1-2 nano-
grams of sample.
PHOTOMULTIPLIER DETECTION
For reasons mentioned above, it was necessary to use low laser power in order
to obtain meaningful results. This was accomplished by the use of neutral density
filters in the to, beam and/or by lowering the power of the nitrogen pump laser. The
resultant decrease in signal strength was more than compensated by changing from pin
diode to photomultiplier detection. So as to be compatible with the wide wavelength
range available using nitrogen laser pumped dye lasers (260 to 740nm) we have used an RCA
33
-------
C31034 gallium arsenide photomultiplier tube having ~10yo quantum efficiency through-
out this range. Since target size of this tube is fairly large (4 x 10mm),the image-object
movement was negligible.
RAMAN SPECTRA OF FLUORESCENT COMPOUNDS
The real advantage of resonance enhanced coherent anti-Stokes Raman spectroscopy
over spontaneous resonance Raman spectroscopy is its ability to reject fluorescence. We
have observed resonance enhancement in several fluorescent compounds whose spectra
(even in preresonance) are partially or totally obscured by background fluorescence.
Figurel4shows a pre-resonance enhanced spectrum of /3-carboline. This compound shows
annoying fluorescence even when irradiated in the yellow and red portions of the spec-
trum. Another spectrum which is particularly difficult to obtain using the spontaneous
1/3
WAVENUM8ER (CM'1)
1500
1000
425
415
14
420
X(NM)
Preresonance CARS spectrum of beta-carboline.
34
-------
Raman effect is /3-naphthol. The spectrum shown in Figure 15 was recorded using a
pump frequency of 365 nm—-a region too highly fluorescing to obtain a spontaneous
Raman spectrum.
UV EXPERIMENT
In order to obtain maximum detection sensitivity for most aromatic compounds,it
will be necessary to pump their IT - ir* transitions in the UV. To work in this region
(below 360 nm) it is necessary to frequency double the visible output of the dye lasers.
Since this is a relatively inefficient process (H0% conversion), we expect to lose
3 .
about 10 in signal strength. This loss is more than compensated by switching to photo-
multiplier (PM) detection. PM detection is ~10 more times sensitive than conventional
pin diode detection. By switching to PM detection and using fused silica optics, we
have been able to record the first CARS spectra recorded in the UV region. The lines
WAVENUMBER (CM'1)
1500
1000
1
385
380
375
15
X (NM)
Preresonance CARS spectrum of beta-naphthol.
35
-------
in the CH stretch region of CH.OH are shown in Figure 16- A complete automated
O
CARS scan in the UV region is shown in Figure 17.
LINESHAPES IN CARS SPECTRA
Under certain conditions, when the laser pump frequency approaches an
electronic transition frequency, negative peaks have been observed in coherent anti-
Stokes Raman spectra. Two conflicting explanations of this phenomenon have been
offered. We have noted that these two explanations lead to a different dependence of
the intensity of the negative peaks on the power in the pump beam. We have experi-
mentally determined this power dependence in the case of a 5 x 10 M solution of
2837
B
2942
289.9 290.1
i i
X(NM)
16 UV-CARS spectrum of the CH stretch region
in methanol .
36
-------
N, N-diethyl-p-nitrosoaniline in benzene and found that the intensity of the negative
peaks varies with the square of the power in the pump beam. This result is consistent
with the explanation that the negative peaks arise from a cross term between the back-
ground susceptibility and the imaginary part of the Raman susceptibility and not from the
Inverse Raman Effect.
PRERESONANCE ENHANCEMENT OF FLUORESCENT COMPOUNDS
In order for CARS to be a totally versatile technique, the CARS spectra of
fluorescent compounds must be able to be resonance enhanced in the presence of strong
background fluorescence. In order to show that CARS is a versatile technique, we have
undertaken the investigation of the preresonance excitation profiles of the fluorescent
compounds /3-naphthol and the /3-naphtholate ion. The fluorescence rejection of the CARS
V00_
5 320.
o:
cr
o:
- 2V0J
CD
cr
cr
160.
CO
2
LU
80.
0.
CflRS - TOLUENE W1=295.0 NM
700
8130 900 1080 H00 1200
FREQUENCY (CM-1)
17 Automated UV-CARS spectrum of toluene
37
-------
technique allowed Raman spectra to be obtained in a region where fluorescence would
obscure the spontaneous Raman signal. A method of taking the ratio of the Raman reso-
nant CARS signal to the background signal allowed for a more convenient and reliable
method of generating CARS excitation profiles. The observed excitation profiles are in
excellent agreement with those predicted using a single line approximation.
RESONANCE ENHANCEMENT OF FLUORESCENT COMPOUNDS NEAR THEIR
ABSORPTION MAXIMUM
In these studies, we have investigated the resonance enhancement of highly
fluorescent compounds near their absorption maxima. An empirical rule for deter-
mining the minimum detectable concentration has been found to be given by
C . = (10 • E) where E is the decadic molar absorption coefficient. For most
min .5 .6
compounds studied thus far, C . falls between 10 M and 10 M, well within
r mm '
the range necessary for CARS to be useful for providing specific structural information
on liquid chromatographic eluents.
The CARS spectra of several of the compounds are shown in Figures 18 to 23.
These spectra have been recorded using concentrations ranging from 1x10 M to
7x 10~5M.
We have conducted an in-depth excitation profile study for the pol/aromatic
hydrocarbon tetracene (naphthacene). The experimental results agree ex-
tremely well with those predicted using CARS resonance theory . The comparison of
the experimental results and theory is given in Figure 24.
38
-------
A,
U)r
485
480
475
470
1500
1000
-i
18 CARS spectrum of perylene In benzene taken
at a pump wavelength of 450 nm. The quantum
efficiency of fluorescence for this compound
Is .94
X,= 535 NM
Ai
580
570
560
550
nm
em-1
1500
1000
19
CARS spectrum of rhodamine 6G in ethanol
taken at a pump wavelength of 535 nm.
The quantum efficiency of fluorescence for
this compound in ethanol is .95
39
-------
X,=555 NM
Xi
1500
1000
20
CARS spectrum of rhodamine B In ethanol
taken at a pump wavelength of 555 nm.
The quantum efficiency of fluorescence for
this compound in ethanol is .40*
Xi=510 NM
Ai
Ur
560
21
-^A
w~.
550
540
530
1500
1000
nm
em-'
CARS spectrum of sodium fluorescein in
water at an alkaline pH taken at a pump
wavejength of 510 nm. The quantum
efficiency of fluorescence for this compound
is .90.
40
-------
X,= 510 NM
560
UJi
550
540
22
150O
1000
CARS spectrum of acridine orange
in water at a pump wavelength of
510 nm. The quantum efficiency
of fluorescence for this compound
is above .50.
530
OM
CM'1
X, = 425 NM
UJ<
(60
i
1500
450
1000
440
nm
cm-1
23 CARS spectrum of coumarin 495 in
ethanol at a pump wavelength of
425 nm. The quantum efficiency
of fluorescence for this compound
is above .70.
41
-------
1.0 J
I
J I
I / /
470
490 500
510
24
Excitation profiles for the ring breathing
mode of tetracene as a function of the
pump wavelength and displacement
parameter. The solid lines represent
the theoretical profile at various
displacement parameters. The circles
represent the experimentally obtained
values.
42
-------
SECTION 6
COMPUTER INTERFACE
The CARS experiment is a multi-parameter process that requires precise control of
mirror and micrometer movements. To obtain large frequency scans in short periods of
time, it is necessary to control the various parameters by a computer. The minicomputer
chosen for this purpose was a PDF 11/34 with such options as dual floppy disk, two gen-
eral purpose I/O ports (DR11 K and DR11C), analog to digital converter (ADI IK) with
multiplexer (AMI IK), and digital to analog converter (AA11K). Included with this
package was Digital's RT-11 realtime FORTRAN software package for real time control.
Figure 25 shows the CARS setup employed for these experiments. It consists of a
nitrogen laser split into two beams by a 2:1 beam splitter. These beams are then used to
pump twoMolectron DL200 dye lasers that create the pump ( Uj) and probe (u2) beam for
the CARS experiment. Scanning of the dye lasers is achieved by a sine-bar-driven grating.
The u2 dye laser came equipped with a stepping" motor drive for continuous scanning. The
controller for this dye laser stepping motor is Molectron's DL-040A scan control unit. The
two beams are crossed and focused in the sample by lens 2. A Raman signal results when
the difference in frequency of the two lasers differ by an allowed Raman transition. The
resulting CARS signal (a>3) is generated at an angle G with respect to the Ujbeam.
In order to detect this signal the detector and monochromator network had to be
able to follow the 0 angle as the scan was taken. This was accomplished by mounting
them on a movable stage pivoted about the crossing point of the pump and probe beams.
This stage allows the angle of collection to be varied synchronously with the crossing angle.
The signal beam, u,, passes through an iris 30cm from the sample. The beam is then focused
O
onto the entrance slits of the monochromator network consisting of two Jobin-Yvon model
HR10 units mounted front to back. The signals from the photomultiplier and the reference
detector (Molectron LP-141) are fed to a Molectron LP-20 boxcar integrator-ratiometer and
sent to a analog to digital converter where it is stored and processed by a PDP-11/34
minicomputer.
43
-------
Computer control of the CARS experiment required the incorporation of a
minimum of five stepping motors. Three were required for variation of the cross
angle. They controlled mirror translation, vertical and horizontal rotation. The
horizontal and vertical rotation stepping motors were 400 steps/revolution while all
other motors employed were 200 step/revolution. The collection stage position as well
as the monochromator scanning also had to be controlled by stepping motors.
Three types of interfaces had to be built in order to make the CARS unit com-
patible with the minicomputer. They were: A) busing system for the I/O ports
(peribus), B) DL-040A Molectron scan unit interface, and C) stepping motor interfaces.
The design was from the modular concept developed at Purdue?? . The peribus
(peripheral bus) was constructed to increase the number of possible I/O ports. This is
achieved by busing all of the bits from the DRII port to a network of plug-in ports.
This allows more devices to have access to the same I/O interface. In order for the
PHOTOMULTIPUER
25 Experimental CARS setup used in the studies.
44
-------
computer to determine which device was being controlled by the I/O interface, each
device or operation had to have its own address. If one uses 6-bits of the output word
(16-bits) for addressing the various instrumental functions, there are 64 possible addresses
and thus operations which can be controlled by a single I/O interface. Figure26 shows
the schematic for the peribus system. It consists of open collector drivers which boost
•Cfe
From
CPU
SimpM
8-bil
I/O
Into
CPU
Tlmino. at—-
\ CQ^I| F=
From
CPU
Computer
BufUr
Modul*
Into
CPU
26 Schematic of the peripheral bus.
45
-------
and distribute the signal from the DR11 I/O port to the various interfaces. It also con-
tains a monostable multivibrator for lengthening various pulses (such as the data-available
pulse) put out by the computer. The data available pulse is lengthened to 4M sec by the
peribus and used in the address decode. By doing this, the information of the output word
is insured of being valid.
The scan unit has three types of operations that had to be interfaced to the
computer. They included data input (latching), data output and pulse generation. A
simplified scheme of how these operations are handled is shown in Figure 27. The opera-
tions that required latching were those in which a level (high or low) had to be held
for long periods of time. Data output was accomplished by two input open collector
output wo rd
address
decode .
scan unt t
f uncti on
scan unit
V
address
3-
input WQ Fd
scan unit
read ou t
27
Simplified schematic of the basic
operation of the scan unit inter-
face. Both the input and output
ports of the DR1 Ik were required
for communication with the computer
46
-------
ncmd gates in which one input was the address decode and the other the actual data
being sent to the computer. Those from the scan unit were BCD outputs of the wave-
length of the dye laser.
The stepping motor interface is also controlled by the DR11 I/O port. This
interface consist of two operations: control and status monitoring. The general scheme
for this interface is shown in Figure 28- The information that is required to make a
particular stepping motor run is the address of that motor, the number of steps to be made
and the direction in which to step. Thus a single output word can run a stepping motor
leaving the computer free for other operations. The run and limit status of a motor are
monitored by the status card and sent to'the computer through the input word.
input word
28
Simplified schematic of the basic
operation of the stepping motor
interface. The DRUG I/O port
was used for controlling this inter-
face.
47
-------
The stepper control card can operate in one of two possible modes, local and
computer. Part A of Figure 29 depicts a simplified scheme of what is involved in the
local mode. The operations that are controlled are run (free run), direction (clock-
wise or counter clockwise),and step (single step). The run switch turns on an oscillator
that puts out a pulse train to the motor either in the clockwise or counter clockwise
direction. The step switch puts a single pulse to the motor. If a limit has been reached
by the stepping motor, a micro switch for that direction is shorted to ground blocking the
pulse train at the three-input nand gate. The computer mode of operation is depicted in
Part B of Figure 29and operates in the following manner. An output word containing the
address, direction,and number of steps is put on the output bus. The address decodes
allowing a binary up-down counter to be loaded. The counter is used only in the down
count mode for this interface. Once the counter has been loaded, an external oscillator
is switched on and a pulse train sent to the count down of the counter. The binary counter
counts the number loaded down with its output being first coupled with the direction and
A. LOCAL CONTROL
direct! an
B. COMPUTER CONTROL
step
address
decode
binary
counter
•CE
"n7s~> Cms )
29 More detailed schematic of the
stepping motor interface under
local and computer control.
48
-------
then sent to the motor. The maximum number of steps that can be loaded at one time is
200 (one revolution). The actual circuit that was used is shown in Figure 30. The run
signal of each motor is monitored by the motor status card. When the correct address is
given, the information is sent to the input port and subsequently to the computer. The
micro switches are not individually monitored but monitored collectively. This is accom-
plished by an eight-input nand gate. If a limit is reached by a motor, program control
can stop all operations and ring a bell, warning the user that a limit has been reached.
> |LC£-
V_li.Y!Ll>'
fKW-T
t~OW4> -
nrws>
nrw?>
(Owio> •
COwH)-
C£ST2>-
COW13)- •
(_OWl4>
CSTZ) — I
JJ
CHS—
C'SCO
C 5C2J
d=
—
— u
C SPlj
C5I£>
'
— |
74151
['-
4
r~
• f
1
H
I
-I
1
74193
' « 0 0
1
7475
U * * j ° *
i
30
f
U-
^hlfi
1
74 OO
hP
'? j
r-
I
j_
t
7
H
1
Schematic
r
74193
TV
Irf
« • *
t . — i
4279
"
"f
1
J
0 » i
74157
» ^ *
l 1
740O
_T
» • * * •
1
1
n
i
»— 1
[f
irhnu'
7410
^f
I
of the actual circuit
i- ^
— v
"A~w_
— rcvo
-(RUN;
used for the stepping motor inter-
face. The symbols are: OW0-
output word bit 0, SA - computer/
local switch, SB - direction switch,
SC - scan switch, SD - single step
switch, OSC - oscillator, MSI -
limit micro-switch for the CW direction
and RUN turns on an external
oscillator.
49
-------
Using this hardware, complete automation of the CARS instrument has been achieved.
With the liquid chromatograph, U-V visible spectrometer and CARS instrument under
computer control, software was developed to affect the separation, detection and identification
of compounds. The first demonstration of the automated LC-CARS instrument was
conducted for several chemists from the Environmental Research Laboratory, Athens,
Georgia. Using a mixture of two model compounds: (3-carotene and N-N diethyl
paranitrosoaniline. The entire process was conducted under computer control:
chromatographic fractionation, UV-visible detection and spectral scan, and CARS
spectrum.
50
-------
REFERENCES
1. G. A. Junk, J. J. Richard, M. D. Grieser, D. Witiak, J. L. Witiak,
M. D. Arguello, R. Vick, H. J. Svec, J. S. Fritz, and G. V. Colder,
J. Chromatog., 99 (1974)745-762.
2. J. J. Lichtenberg, FCB Newsletter, Aug. 20 (1970).
3. B. Ahling and S. Jensen, Anal. Chem., 42 (1970) 1483.
4. W. E. May, S. N. Chesler, S. P. Cram, B. H. Gump, H. S. Hertz,
D. P. Enagonio, and S. M. Dyszel, J. Chromatogr. Sci., 13 (1975) 535.
5. J. J. Kirkland, Anal. Chem., 43 (1971) 37A.
6. Application Highlights No. 24-PCB's, Waters Associates, Framingham, Mass.,
1972.
7. J. A. Schmit, R. A. Henry, R. C. Williams, and J. F. Dreckman, J. Chromatogr.
Sci., 9,645 (1971).
8. R. E. Majors, American Lab., October (1975).
9. J. A. Armstrong, N. Bloembergen, J. Queuing, and P. S. Pershan, Phys.
Rev., 127, 1918 (1962).
10. P. D. Maker and R. W. Terhune, Phys. Rev., A137, 801 (1965).
11. J. J. Wynne, Phys. Rev. Lett., 29, 650 (1972).
12. E. Yablonovitch, C. Flytzanis, and N. Bloembergen, Phys. Rev.
Lett., 29, 285 (1972).
13. M. D. Levenson, C. Flytzanis, and N. Bloembergen, Phys. Rev., B6, 3962
(1972).
14. P. R. Regnier and J. P. E. Taran, Appl. Phys. Lett., 23, 240 (1973).
15. M. D. Levenson, IEEE J. Quantum Electron., QE-K), 110 (1974).
16. M. D. Levenson and N. Bloembergen, J. Chem. Phys., 60, 1323 (1974).
17. R. F. Begley, A. B. Harvey, and R. L. Byer, Appl. Phys. Lett., 25, 387
(1974).
18. R. F. Begley, A. B. Harvey, R. L. Byer, and B. S. Hudson, J. Chem. Phys.,
61,2466 (1974).
51
-------
19. R. F. Begley, A. B. Harvey, R. L. Byer, and B. S. Hudson, Am. Lab., 6(11),
11 (1974).
20. I. Itzkan and D. A. Leonard, Appl. Phys. Lett., 26, 106 (1975).
21. F. Mova, S. A. J. Druel, and J. P. E. Taran, Opt. Commun., K3, 169 (1975).
22. J. J. Barrett and R. F. Begley, Appl. Phys. Lett., 27, 129 (1975).
23. I. Chabay, G. K. Klauminzer, B. S. Hudson, Appl. Phys. Lett.,
28, 27 (1976).
24. A. C. Albrecht, J. Chem. Phys., 34, 1476 (1961).
25. J. Tang and A. C. Albrecht, in "Raman Spectroscopy," Vol. 2 (H. A.
Szymanski, Ed.), Plenum Press, New York, 1970, and references cited
therein.
26. B. Hudson, W. Heatherington, 111, S. Kramer, I. Chabay, and G. K.
Klauminzer, Proc. Natl. Acad. Sci. (in press).
27. F. Inagaki, M. Tasumi, and T. Miyazawa, J. Mol. Spectrosc., 50,
286(1974). ~
28. L. Rimqri, R. G. Kilponen, and D. Gill, J. Amer. Chem. Soc.,
92, 3824 (1970).
29. J. E. Davis and E. D. Schmidlin, Chem. Inst., 4, 169 (1973).
52
-------
COLLABORATORS
FACULTY
T. B. Mailoy (Mississippi State University)
J. D. Stuart (University of Connecticut)
R. von Wandruszka (University of Georgia)
POSTDOCTORAL RESEARCH ASSOCIATES
R. M. Irwin
G. W. Martin
P. K. Tseng
GRADUATE STUDENTS
R. R. Antcliff
L. P. Goss
PUBLICATIONS
1. "Excitation Profiles for the Coherent anti-Stokes Raman Spectrum of Beta Carotene,"
L. A. Carreira, T. C. Maguire and T. B. Mailoy, Jr.,J. Chem. Phys.,66,2621 (1977).
2. "Micresampling and the Use of a Flow Cell for Coherent Anti-Stokes Raman
Spectroscopy (CARS)," L. B. Rogers, J. D. Stuart, L. P. Goss, T. B. Mailoy, Jr.,
and L. A. Carreira, Anal. Chem., 49, 959 (1977).
3. "On the Possibility of Coherent Anti-Stokes Raman Spectroscopy in the UV,"
L. A. Carreira, L. P. Goss, and T. B. Mailoy, Jr., J. Chem. Phys., 66,2762 (1977).
4. "Experimental Evidence on the Source of Negative Peaks in Coherent Anti-Stokes
Raman Spectra," L. A. Carreira, L. P. Goss, and T. B. Mailoy, Jr., J. Chem.
Phys., 66,4360 (1977).
5. "Preresonance Enhancement of the Coherent Anti-Stokes Raman Spectra of
Fluorescent Compounds," L. A. Carreira, L. P. Goss, andT. B. Mailoy, Jr.,
J. Chem. Phys., 68, 280 (1978).
6. "Raman Spectrum and Torsional Potential Function for Vinyl cyclopropane,"
L. A. Carreira, T. G. Towns,andT. B. Mailoy, Jr., J. Amer. Chem. Soc., 100,
385 (1978).
7. "Comparison of the Coherent Anti-Stokes and Coherent Stokes Raman Lineshapes of
the v. Line of Beta-Carotene Near a One Photon Resonance," L. A. Carreira,
L. P. Goss,andT. B. Mailoy, Jr., J. Chem. Phys., 69,855 (1978).
8. "Coherent Anti-Stokes Resonance Raman Spectra of Highly Fluorescent Compounds,"
L. A. Carreira, Advances in Laser Chemistry, A. H. Zewafl,ed., Springer Series
in Chemical Physics, Springer, Berlin, Heidelberg, New York, 1978.
53
-------
9. "Conformational Barriers and Interconversion Pathways in Some Small
Ring Compounds," T. B. Malloy, Jr., L. E. Bauman, and L. A. Carreira,
Topics in Stereochemistry, N. L. Allingerand Ernest L. Eliel, editors,
1978.
10. "Low-Frequency Vibrations in Small Ring Compounds," L. A. Carreira,
R. C. Lord, and T. B. Malloy, Jr., Springer-Verlag Publishing Co.,
Heidelberg, 'Germany, 1978.
11. "Resonance Enhanced CARS," L. A. Carreira, Proceedings of the Sixth
International Raman Conference, 1978.
12. "Effect of a Change in Solvent on Chromatographic Peak Shapes,"
Paul K. Tseng and L. B. Rogers, J. Chromatogr. Sci., ]6, 436 (1978).
13. "Retention Behavior of Dinitrobenzene Isomers and Dihydroxybenzene
Isomers on Octadecyl and Phenyl Bonded Phase Packings," Paul K.
Tseng, J. Chromatogr. Sci., ^6, 438 (1978).
14. "Coherent Anti-Stokes Resonance Raman Excitation Profiles of Tetracene,"
L. A. Carreira, T. C. Maguire, and T. B. Malloy, Jr., J. Chem. Phys.,
00, 0000 (1979).
54
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/4-80-015
2.
3. RECIPIENT'S ACCESSION-NO.
. TITLE AND SUBTITLE
New Liquid Chromatographic Detection System for
Environmental Pollutants
5. REPORT DATE
February 1980 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
L.A. Carreira and L.B. Rogers
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemistry
University of Georgia
Athens, Georgia 30602
10. PROGRAM ELEMENT NO.
A37B1D
11. CONTRACT/GRANT NO.
R804155-03-0
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Research Laboratory—Athens GA
Office of Research and Development
U.S. Environmental Protection Agency
Athens, Georgia 30605
13. TYPE OF REPORT AND PERIOD COVERED
Final. 1/76-12/78
14. SPONSORING AGENCY CODE
EPA/600/01
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Resonance enhanced coherent anti-Stokes Raman spectrometry (CARS) has been demon'
strated as a specific identification system for liquid chromatog raphy. To achieve this
liquid chromatographic preconcentration and separation and computer control of the
liquid chromatograph/ultraviolet-visible/CARS were undertaken. Under the optimum reso-
nance condition, spectra were obtained on 5 microliter volumes with detection limits o
1 micromolar or less under favorable conditions. Excellent fluorescence rejection and
ultraviolet CARS were also shown. A number of different species were surveyed by 1 iqu
chromatography with the selectivity of different stationary phases and special solvent
effects examined. Observations useful in relating experimental variables in the pre-
concentration step to the recovery of a given species were made. A system consisting
of a liquid chromatograph, a scanning ultraviolet-visible spectrometer and the CARS
was automated using a minicomputer. Column selection, flow rate, and solvent composi-
tion were monitored and controlled. The ultraviolet-visible spectrometer was used at
a fixed wavelength until a species was detected. The flow was stopped, the absorption
spectrum was scanned, and flow was resumed. The optimum laser frequencies for resonan
CARS were set from the absorption spectrum. The Raman spectra were then scanned under
control of the computer with on-line signal averaging and data smoothing.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Chemical analysis
Chromatography
Spectrometers
07B
68D
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
71
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
55
U.S. GOVERNMENT PRINTING OFFICE: 1980-657-146/5591
------- |