&EPA
United States
Environmental Protection
Agency
Environmental Monitoring and
Support Laboratory
Cincinnati OH 45268
EPA 600/4-80-040
Auyust 1980
Research and Development
Development of
Oil-in-Water Monitor
Phase I
-------
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-600/4-80-040
August 1980
DEVELOPMENT OF OIL-IN-WATER MONITOR
PHASE II
by
H.S. Silvus, Jr.
P.M. Newman
J.H. Frazar
Southwest Research Institute
San Antonio, Texas 78284
Grant Number R805817-01
Project Officer
Fred K. Kawahara
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 46268
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and Support Laboratory, U.S. Environ-
mental 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:
o Develop and evaluate techniques to measure the presence and
concentration of physical, chemical, and radiological pollutants in
water, wastewater, bottom sediments, and solid waste.
o 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.
o Develop and operate an Agency-wide quality assurance program to
assure standardization and quality control of systems for
monitoring water and wastewater.
o Test and investigate automated monitoring devices which will be
both anticipatory as well as responsive to the needs of the Agency
whose mandate is to restore, enhance, and protect the quality of
the environment.
This report describes the preliminary investigation of an on-line
continuous flow oil-in-water monitor developed by the Southwest Research
Institute.
Dwight G. Ballinger
Director
Environmental Monitoring and Support
Laboratory - Cincinnati
HI
-------
ABSTRACT
A novel approach to quantitatively monitoring suspended hydrocarbons in water was conceived at
Southwest Research institute in 1975, and feasibility was subsequently demonstrated under sponsorship of
the U. S. Environmental Protection Agency through Grant R804368-01. This new oil-in-water monitor tech-
nique brings together for the first time two previously unrelated technologies: U) reversed-phase liquid
chromatography and (2) fiber optics. A special organophilic optical fiber, created by a chemical treatment
process routinely used in reversed-phase liquid chromatography, collects and concentrates suspended hydro-
carbon materials on its surface. Collected material alters optical transmission through the fiber in such a way
that the logarithm of the time rate of change of the logarithm of optical transmission through fiber is linearly
proportional to the logarithm of contaminant concentration, provided that concentration exceeds a certain
detection threshold.
Various methods of chemically treating an optical fiber and various treatment reagents were evaluated.
Numerous treated optical fibers were tested individually, each with a variety of separate aromatic hydrocar-
bon contaminants, in a newly designed capillary-tube sensor cell. Additionally, a laboratory demonstration
instrument employing the organophilic optical fiber hydrocarbon-in-water monitor technique was fabricated
and tested. For such aromatic hydrocarbons as n-hexylbenzene, cyclohexylbenzene, heptadecylbenzene, 3,
3'-dimethylbiphenyl, and 1-phenylnaphthalene and for crude oil, detection thresholds of less than 3 mg/L
were observed, and system response was linear over a contaminant concentration range of greater than 2.5
decades (i.e., greater than 300:1). Data indicate that the high end of the range can be extended if desired.
The reported work was sponsored by the Environmental Monitoring and Support Laboratory, Office of
Research and Development, U. S. Environmental Protection Agency, Cincinnati, Ohio 45268, under Grant
R805817-01. Portions of the reported work were conducted in the facilities of the U.S. Army Fuels and Lubri-
cants Research Laboratory which is located at Southwest Research Institute and is operated under contract
for the U. S. Army Mobility Equipment Research and Development Command, Ft. Belvoir, Virginia.
Iv
-------
CONTENTS
Page
Foreword in
Abstract iv
Figures vi
Acknowledgements vii
I. INTRODUCTION 1
II. OBJECTIVES 3
III. SUMMARY 4
IV. CONCLUSIONS 5
V. RECOMMENDATIONS 6
VI. DISCUSSION 7
Review of Concept 7
Process Characterization 10
Improved Apparatus 10
Procedures and Coating Experiments 10
Fiber Evaluation 13
Test Apparatus 13
Test Suspension Preparation and Materials 15
Fiber Materials 18
Test Procedure 18
Test Results 18
Comparison of Performance of Two Sensor Cell Designs 19
Evaluation of Fiber Treatment Processes 19
Effects of Various Treatment Reagents 19
Sensitivity to Various Contaminants 21
Prototype Instrument 24
Sensor Cell 24
Optics 24
Signal Processing 25
Fluid Flow 25
Package 28
Operation 28
Test Data 28
Organic Polymer Fibers 28
Fluoride Glass Investigation 28
Evaluation of Light Sources 31
References 33
APPENDIX A. Discussion of Sensitivity 34
APPENDIX B. Description of Crude Oil Used in Tests 36
-------
FIGURES
umber Page
1 Typical Construction of Optical Fiber 8
2 Cross-Section of Optical Fiber Showing Acceptance Angle and Typical Ray Path 9
3 Possible Structure of Octadecylsilyl Group Bonded To Glass Surface 11
4 Close-Up View of Capillary-Tube Sensor Cell and Fluid Connections 14
5 Schematic Diagram of Test Apparatus Including Coiled Capillary-Tube Sensor Cell .. 16
6 Photograph of Test Apparatus Employing Capillary-Tube Sensor Cell 17
7 Comparison of Responses of Sensor Cells of Two Different Designs to Tetralin 20
8 Response of Coiled Capillary-Tube Sensor Cell to Contaminants of Varying Solubility
in Water 22
9 Response of Coiled Capillary-Tube Sensor Cell to Crude Oil in Water 23
10 Block Diagram of Signal Processing System 26
11 Flow Diagram of Prototype Instrument 27
12 Hydrocarbon-in-Water Monitor, MK. I 29
13 Response of Prototype Hydrocarbon-in-Water Monitor Instrument to
n-Hexylbenzene 30
VI
-------
ACKNOWLEDGEMENTS
The authors of this report wish to thank the Project Officer, Dr. Fred K. Kawahara of the Environmental
Monitoring and Support Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio, for his en-
couragement, guidance, and valuable suggestions.
Portions of the reported work were conducted in the facilities of the U.S. Army Fuels and Lubricants
Research Laboratory located at Southwest Research Institute. This laboratory is operated under contract for
the U. S. Army Mobility Equipment Research and Development Command, Ft. Belvoir, Virginia.
VII
-------
SECTION I
INTRODUCTION
Waste water from such sources as refineries, shale-oil recovery plants, coal-conversion operations,
chemical plants, ships, offshore drilling platforms, petroleum-handling facilities, and other industries may
contain harmful quantities of hydrocarbon materials in suspension. Through programs coordinated and ad-
ministered at the Federal level, a significant effort is underway to eliminate or substantially reduce the quan-
tity of hydrocarbon material discharged in waste water. Treatment processes which remove suspended hy-
drocarbons from waste water have been placed in service by many organizations cooperating in water
pollution control. While installation and operation of such treatment processes is a major step toward elimina-
tion of hydrocarbon pollution, to achieve maximum effectiveness it is necessary to monitor the effluent of a
waste-water treatment process to insure that the system is functioning correctly. A simple, inexpensive, au-
tomated, on-line instrument is needed for this important monitoring function.
A new approach to quantitatively monitoring suspended hydrocarbons in water was conceived at
Southwest Research Institute in 1975 (Ref.1). Feasibility demonstration and early development of this new
concept was accomplished under sponsorship of the U. S. Environmental Protection Agency through Grant
R804368-01 (Refs. 2 and 3). Briefly, this new oil-in-water monitor system employs a sensor cell through which
a continuously flowing sample of the stream to be monitored is diverted. The sensor cell is similar to a re-
versed-phase liquid chromatographic column; however, in place of conventional column packing, the sensor
cell contains a continuous unclad optical fiber, the surfaces of which have been made organophilic by chemi-
cal treatment. Suspended oil in the fluid flowing through the sensor cell is adsorbed on the optical fiber and
(2) oil adsorbed on the fiber surfaces produces a change in the through-transmission attenuation factor of the
optical fiber. Rate of decrease in optical fiber transmission is related to concentration of suspended hydrocar-
bons contaminating the stream being monitored.
The ultimate goal of developing the organophilic optical fiber oil-in-water monitor technique is to pro-
duce a reliable, simple to operate, easy to maintain and relatively inexpensive instrument for analysis of sus-
pended hydrocarbons in water. Such an instrument could be employed in plants, on offshore platforms,
aboard ships or in other facilities which generate hydrocarbon polluted waste water to monitor the effective-
ness of treatment systems or to determine the quantity of pollutant being discharged. In the future, availabil-
ity of an economical instrument for monitoring suspended hydrocarbons in water should result in increased
awareness of the quantity of hydrocarbon material being discharged into streams or other bodies of water.
This awareness should, in turn, increase industrial concern with water treatment and pollution prevention.
Thus, the ultimate impact and benefit of developing an analytical instrument employing the organophilic opti-
cal fiber technique should be a decrease in hydrocarbon pollution in areas where this problem is already se-
vere and prevention of such pollution in new areas.
This report covers (1) development of the organophilic optical fiber oil-in-water monitor concept, (2) im-
provement of the chemical treatment process and investigation of various treatment reagents, (3) extensive
test of variously treated organophilic optical fibers with a wide range of hydrocarbon contaminants, (4) sub-
stantial improvement in sensor cell performance, and (5) fabrication of an engineering model instrument em-
ploying the organophilic optical fiber hydrocarbon detection principle. Substantial increase in sensitivity (by a
-------
factor of 260) above that obtained in Phase I (Ref. 2) was realized during this program. Additionally, a detec-
tion threshold of less than 3 mg/L of crude oil (and some pure hydrocarbon compounds) in water was ob-
served. Further details of the research and development work performed in accomplishing these results are
presented in the remainder of this report.
-------
SECTION II
OBJECTIVES
The objectives of the reported research work were (1) to further investigate experimentally the organo-
philic optical fiber concept as applied to measurement of suspended hydrocarbon material in water, and (2) to
fabricate an engineering model instrument suitable for demonstration and test purposes having sensitivity at
least as good as that achieved during early development in Phase I (i.e., 50 to 100 mg/L detection threshold
fortetralin).
-------
SECTION III
SUMMARY
1. A new sensor cell employing an organophilic optical fiber contained in a coiled stainless steel capillary
tube was designed, fabricated, and tested; sensitivity of this new sensor cell to tetralin exceeded that of
the U-tube sensor cell used in Phase I by a factor of 260, and a detection threshold of less than 3 mg/L
of crude oil in water was observed.
2. Modifications were made in the test apparatus to improve sensor cell performance and reduce the time
required to conduct tests of various combinations of fiber material, treatment reagent and process, and
hydrocarbon contaminant.
3. Improved chemical treatment apparatus was developed which facilitated faster processing with smaller
quantities of treatment reagents and less potentially damaging handling of the optical fiber.
4. The chemical treatment process was improved by reducing the number of steps required and by con-
verting to a flowing-reagent system.
5. Performance of the organophilic optical fiber sensor cell was characterized by conducting a large num-
ber of tests with a wide range of aromatic hydrocarbon compounds, crude oil and diesel fuel.
6. Contacts were made with manufacturers of fluoride glass to determine whether that material was po-
tentially useful in the organophilic optical fiber oit-in-water monitor system.
7. A laboratory demonstration hydrocarbon-in-water monitor instrument was fabricated.
-------
SECTION IV
CONCLUSIONS
1. Use of a coiled capillary-tube sensor cell to (1) increase active (i.e., curved) length of fiber, (2) eliminate
optical cross-coupling and (3) improve contact between the fiber and test fluid increased system
sensitivity by a factor of 260 for tetralin, but did not significantly change the detection threshold for this
contaminant.
2. Response of the system to some contaminants, such as crude oil and n-hexylbenzene, was linear over a
2.5-decade range of concentration with a detection threshold of less than 3 mg/L.
3. A linear functional relationship between (1) the logarithm of the time rate of change of the logarithm of
optical transmission through the organophilic optical fiber and (2) the logarithm of contaminant concen-
tration was observed.
4. The organophilic optical fiber sensor cell is not uniformly sensitive to all hydrocarbon contaminants
within its detection range, and no correlation was found between sensitivity and (1) refractive index of
the contaminant, (2) ratio of refractive indexes of contaminant and fiber material, or (3) difference be-
tween refractive indexes of contaminant and fiber material.
5. Use of a fused-silica fiber facilitates detection of the widest range of hydrocarbon compounds; most
aromatics can be detected by a fused-silica fiber, but all normal paraffins are excluded from detection.
6. Threshold of detection for a particular contaminant appears to be slightly greater than the solubility of
that material in water.
7. An organophilic optical fiber can be cleaned and regenerated repeatedly without degradation.
8. Fluoride glasses are not presently practical in the organophilic optical fiber oil-in-water monitor because
of solubility and chemical treatment problems.
9. Organic polymer fibers of types which have suitable surface properties for oil-in-water monitor applica-
tions are not available with sufficient optical clarity for practical use.
10. A low-power laser is the best illumination source for the organophilic optical fiber sensor cell because of
the high efficiency with which light can be coupled into the fiber.
-------
SECTION V
RECOMMENDATIONS
1. Optical fibers, treated with many different coating reagents used both singly and in combination, should
be tested with a wide range of contaminants to establish definitive performance data which can provide
a firm basis for selecting the optimum coating reagent for a particular application.
2. Use of non-visible radiation in the organophilic optical fiber should be investigated so that the infrared
absorption, ultraviolet absorption, and ultraviolet-stimulated fluorescence effects characteristic of aro-
matic hydrocarbons may be evaluated with the objective of increasing sensitivity and specificity of the
system.
3. The possibility of applying special coatings to optical fibers to facilitate adsorption of dissolved (in addi-
tion to suspended) contaminants from water should be investigated; sophisticated coating reagents
which have surfactant, coagulant, or anticoagulant properties may be required.
4. The effects of increasing optical fiber diameter should be explored since such an increase would make
coupling of light into the fiber easier and, perhaps, facilitate practical use of a non-laser illumination
source.
-------
SECTION VI
DISCUSSION
REVIEW OF CONCEPT
The organophilic optical fiber hydrocarbon-in-water monitor concept has been described fully in previ-
ous reports (Refs. 2 and 3). However, a brief review is included in this report for completeness.
Optical fibers are usually of coaxial construction as illustrated in Figure 1. The central part of the fiber,
referred to as the "core", is surrounded by a tight-fitting sheath, referred to as the "cladding", which has
lower index of refraction than the core. Sometimes a protective jacket of plastic or other material is applied
over the cladding to increase mechanical strength of the fiber and to make the fiber easier to handle. A light
ray entering the core of the fiber within a cone defined by the acceptance angle (illustrated in Figure 2) strikes
the boundary between the core and the cladding at an angle greater than the critical angle so that total inter-
nal reflection occurs. Total internal reflection is highly efficient, so such a ray is transmitted through the core
with low attenuation by successive reflections at the core-cladding interface. The acceptance angle is a func-
tion of the indexes of refraction of (1) the core material, (2) the cladding material, and (3) the medium in
which the fiber is immersed (often air).
In the organophilic optical fiber oil-in-water monitor, an unclad optical fiber is used, that is, a fiber which
has only a core and no cladding or protective jacket. The optical transmission properties of an unclad fiber are
highly sensitive to the medium in which the fiber is immersed since that medium effectively functions as the
fiber cladding. If the medium has lower index of refraction than the core material, then light entering the fiber
within the acceptance angle cone will be retained within the fiber by total internal reflection and will be trans-
mitted with relatively low attenuation to the opposite end of the fiber. If, on the other hand, the surrounding
medium has greater index of refraction than the core material, then total internal reflection does not occur,
and light entering one end of the fiber is rapidly dissipated into the medium. To provide for detection of hy-
drocarbon materials suspended in water, an unclad optical fiber is utilized which has greater index of refrac-
tion than water but lower index of refraction than the hydrocarbon materials to be detected. When such a
fiber is in contact with water, it is effectively clad and, therefore, transmits light from one end to the other
with low attenuation. However, when higher refractive index hydrocarbons deposit on the fiber surface, total
internal reflection is locally degraded, and some of the light propagating in the fiber is lost through the fiber
walls. As the quantity of hydrocarbon material accumulated on the fiber surface increases, efficiency of total
internal reflection is further decreased, and more light escapes into the medium, thereby reducing intensity of
the light arriving at the output end of the fiber. The degree to which total internal reflection is destroyed is
related to the quantity of contaminant (i.e., thickness and extent of adsorbed layer) deposited on the fiber
surface. It was experimentally determined during Phase I (Ref. 2) that the optical fiber must be curved for
significant light loss to occur.
To make this phenomenon useful in an analytical instrument, it is merely necessary to provide a con-
stant-intensity light source at the input end of the fiber and a photosensor at the output end to convert light
emitted from the fiber into a proportionate electrical signal. A flowing sample of the stream to be monitored is
then passed over the optical iiber. If this stream contains only water with no hydrocarbon contamination.
-------
OPTIONAL
PROTECTIVE
JACKET
oo
CLADDING
-CORE
Figure 1. Typical Construction of Optical Rber
-------
CLADDING
ACCEPTANCE
ANGLE
GREATER
CRITICAL ANGLE
CO
CORE
INCIDENT RAY
Figure 2. CroM-Section of Optical Fiber Showing Acceptance Angle and Typical Ray Path
-------
then optical transmission through the fiber is unaffected. On the other hand, if there is hydrocarbon contami-
nation present in the water, optical transmission through the fiber will decrease at a rate which is related to
contaminant concentration.
To maximize the previously described effect, it is necessary to concentrate the available hydrocarbon
contaminant on the optical fiber surface. This function is provided by making the optical fiber surfaces orga-
nophilic by use of a chemical treatment process similar to that employed in preparation of column-packing
materials for reversed-phase liquid chromatography. Such treatment chemically bonds organic functional
groups to the glass or fused-silica fiber material as illustrated conceptually in Figure 3.
PROCESS CHARACTERIZATION
Improved Apparatus
Chemical treatment of optical fibers was carried out by three methods designed to convert systemati-
cally from the method of Phase I to a method compatible with the capillary-tube sensor cell test apparatus
used in this program (i.e., Phase II).
The first method employed a process developed during Phase I (Ref. 2) in which batch treatment was
conducted in a tubular, jacketed, glass reaction vessel. This apparatus was used for the first two fiber sets
treated during Phase II to allow comparison between Phase I and Phase II systems.
The second method was compatible with the capillary-tube sensor cell configuration and facilitated
coating the optical fiber in a capillary column. The capillary column containing the coated fiber was coiled and
installed into the test apparatus after chemical treatment with minimal handling. Treatment apparatus in-
cluded a stainless steel reservoir equipped with a dump valve and a flow-regulating needle valve. Stainless
steel tubing conducted the various solvents and treatment reagents to a long, straight capillary tube con-
taining the optical fiber to be processed. Two fiber sets were coated with this apparatus to develop the proper
reagent/solvent solutions for compatibility with stainless steel.
The third method employed the same capillary column apparatus described above; however, the un-
coated fiber was installed in a straight segment of capillary tubing which was coiled to fit the test apparatus
before the fiber was coated to further minimize handling of the treated fiber. The treatment apparatus was
also modified to coat three fibers simultaneously by installing a four-way cross in the flow system.
Procedures and Coating Experiments
By coating the fiber in the same capillary column used in the oil-in-water monitor device, possibility of
damage to the optical fiber is minimized. Another benefit of treating fibers in a capillary tube is that reduced
quantities of reagents are required. Further, more complete coating of the optical fiber occurs because
(1) fresh reagent is constantly flowing through the system and (2) contact between the treatment reagent
and the optical fiber is increased as a result of reduced reaction vessel volume.
The first set of fibers treated in Phase II comprised several 0.8-m-long crown-glass optical fibers which
were coated by the glass-reaction-vessel method employed in Phase I of the program using a 100-mL-capac-
ity glass chromatography column into which was placed a glass rod with fittings to hold the optical fibers.
The various solutions were added to the column, allowed to stand for a predetermined time period and then
drained.
After installation of the fibers, the column and fibers were cleaned by soaking with solvents and were
dried overnight with heated helium. The next day, the column was allowed to cool, and the coating solution
was added. After a 2-hr reaction time, the solution was drained. The coating solution was followed by several
solvent sets to wash away excess coating reagent and reaction by-products. The fibers were then dried over-
10
-------
Figure 3. Possible Structure of Octadecylsilyl Group Bonded To Glass Surface
-------
night by heated helium to remove the last traces of solvent. The coating reagent used for this fiber set was a
v/v mixture of 80% octadecyltrichlorosilane and 20% octadecyltriethoxysilane. Fibers treated by this method
were tested in the redesigned sensor-cell apparatus described later. By starting with a fiber simitar to the ones
used in Phase I of the program, performance of the new detection unit could be compared with results
achieved in Phase I.
The second fiber set was coated in the glass chromatography column utilizing the solution sets planned
for use in the metal capillary-tube system. The coating reagent used was octadecyltrichlorosilane. This exper-
iment was run to determine if the change in solvent systems would make a difference in fiber performance
and to set the baseline for further experimentation with octadecyltrichlorosilane as a coating reagent.
The third fiber set was coated with octadecyltrichlorosilane in the straight metal capillary column appa-
ratus. Operationally, the optical fiber was installed in a suitable length of capillary tubing, and this tubing was
coupled into the treatment apparatus. The reservoir was filled with a cleaning solution which was then forced
through the capillary tube until the supply was exhausted. Rinsing solutions were passed through the system
in a similar manner. The interior of the treatment apparatus was then dried overnight by passage of heated
helium through the tubing. The next day, the silane treatment reagent and subsequent rinsing solutions were
forced through the apparatus. The interior of the treatment apparatus and fiber were dried overnight by pas-
sage of heated helium. To assist in overcoming heat loss through the metallic walls of the apparatus, the res-
ervoir was wrapped with heating tape.
The fourth fiber set also was coated with octadecyltrichlorosilane in the straight metal capillary column
apparatus. In this treatment, however, the hydrogen chloride by-product was scavenged by tri-n-octyl amine
in the coating solution. The standard scavenger, pyridine, was unsuitable because the pyridine-hydrogen
chloride reaction product was insoluble in toluene and would have clogged the capillary. Several amines were
screened, and tri-n-octyl amine was found to be very satisfactory. This amine does not have an active hydro-
gen, so it does not react with the organosilane compound, and it also has sufficiently long alkyl groups that
the tri-n-octyl amine-hydrogen chloride salt Is soluble in toluene despite its polarity. The organosilane coating
solution contained enough amine to neutralize the hydrogen chloride formed.
Eight sets of optical fibers were chemically treated by various reagent solutions in the coiled capillary
system. The first fiber set treated in the coiled capillary column yielded three fibers which had low response in
the test apparatus believed to be caused by insufficient coating reagent. For the second set, concentration
and amount of coating reagent were increased so that each fused-silica fiber in the apparatus received the
same amount of coating reagent as one fiber had in the earlier straight-capillary column method. Coating rea-
gent flow rate was also decreased to allow Increased reaction time.
Additional fused-silica fibers were coated with various reagent/solution systems to test the effects of
different chemical coatings and methods of application. The various coating solution systems are listed
below. One fiber was coated with each.
1. Octadecyltriethoxysilane (coating reagent)
Toluene (solvent)
2. Octadecyltriethoxysilane (75% v/v, coating reagent)
Octadecyltrichlorosilane (25% v/v, coating reagent)
Tri-n-octyl amine (hydrogen chloride scavenger)
Toluene (solvent)
3. Octadecyltriethoxysilane (95% v/v, coating reagent)
Octadecyltrichlorosilane (5% v/v, coating reagent acting mainly as catalyst)
Tri-n-octyl amine (hydrogen chloride scavenger)
Toluene (solvent)
4. Octadecyltriethoxysilane (95% v/v, coating reagent)
Trimethylchlorosilane (5% v/v, coating reagent acting mainly as catalyst)
Tri-n-octyl amine (hydrogen chloride scavenger)
12
-------
Toluene (solvent)
5. Octadecyltriethoxysilane (coating reagent)
Isopropanol (solvent)
Water (catalyst)
The octadecylsilyl group was applied to most of the fibers treated during the program. This alkyl group
can be applied with ethoxy or chloride as the leaving group. The chloro and ethoxy reagents were used both
singly and in mixtures. Trimethylchlorosilane (TMCS) has been used successfully as a silylation catalyst in
analytical chemistry. In this program, octadecyltrichlorosilane was used instead of TMCS as a catalyst to aid
in fiber coating (Fibers C and 0, Table I). Also, isopropanol was tried as a solvent for Octadecyltriethoxysilane
because this material is compatible with the alcohol where the chlorosilane is not. Additionally, water was
tried as a catalyst for Octadecyltriethoxysilane in alcohol.
TABLE I. IDENTIFICATION OF TEST FIBERS
Fiber Coating Reagent Composition
A Octadecyltrichlorosilane
B Octadecyltriethoxysilane
C Octadecyltriethoxysilane 75%
Octadecyltrichlorosilane 25%
D Octadecyltriethoxysilane 95%
Octadecyltrichlorosilane 5%
E Octadecyltriethoxysilane
isopropanol and water as solvent)
F Diphenyldichlorosilane
G n-Decyltrichlorosilane
FIBER EVALUATION
Test Apparatus
It was experimentally determined (Ref. 2) that an organophilic optical fiber must be curved to be highly
active in attenuating transmitted light as hydrocarbon material collected on its surface. To maximize the ef-
fect of this property, attempts were made in the previous program (Ref. 2) to fabricate a multi-turn optical
fiber coil so that a much greater curved length could be obtained. Tests with this device indicated increased
sensitivity, but the degree of increase was not as great as expected. Optical cross-coupling between turns of
the coil and inefficient contact between the flowing fluid and the fiber surface were postulated as possible
explanations for these observed results. In designing the sensor cell to be used in the test apparatus for this
program, particular consideration was given to (1) increasing the active (i.e., curved) length of optical fiber,
(2) eliminating optical cross-coupling between turns of the fiber coil, and (3) achieving substantial increase in
contact between the flowing fluid and the fiber surface.
The resulting experimental sensor cell, illustrated in Figure 4, comprised a coil of 1.59-mm (0.0625-in.)
O.D. by 0.58-mm (0.022-in.) I.D. stainless steel capillary tubing containing a 0.13-mm (O.OOB-in.)-diameter
optical fiber in its interior. Approximately 5% of the cross-sectional area of the capillary tube lumen was occu-
pied by the optical fiber. The fluid sample to be tested was flowed through the capillary tubing so that it came
in close contact with the fiber surface. By coiling the fiber-containing capillary tube, any desired active length
could be achieved. Additionally, with the fiber contained inside an opaque tube, cross-coupling between
turns of the coil was entirely eliminated.
13
-------
Figure 4. Close-Up View of Capillary-Tube Sensor Cell and Fluid Connections
-------
To facilitate quantitative testing, the coiled capillary tube sensor cell containing the optical fiber was
installed in the test apparatus shown schematically in Figure 5 and illustrated in Figure 6. The capillary tube
was connected to the fluid handling apparatus by means of compression-type tube fittings, and the ends of
the fiber were brought out through seals. The beam from a low-power helium-neon laser was reflected by a
90° prism through a microscope objective and brought to focus on the input end of the organophilic optical
fiber under test. Optical energy emitted from the output end of the fiber was directed on the sensitive area of
a silicon PIN (positive-intrinsic-negative) photodiode. A power supply was provided to bias the photodiode,
and a commercial electrometer was used to measure diode photocurrent. Electrometer output was recorded
on a commercial strip-chart recorder having logarithmic amplitude response.
A reservoir was provided to contain the test fluid. Compressed air, obtained from a 700-kPa (100-psi)
supply through a filter-separator and pressure regulator, was admitted to the sealed reservoir above the liquid
surface. Various valves provided for (1) pressurizing and venting the reservoir, (2) draining and flushing liquid
from the reservoir, and (3) controlling fluid flow through the capillary-tube sensor cell. A graduated cylinder
was used to collect system effluent for the purpose of determining average flow rate. Tubing, fittings, valves,
etc. in the fluid-flow path were made of 300-series stainless steel.
Test Suspension Preparation and Materials
Test suspensions used to evaluate chemically treated fibers in the test apparatus were prepared using
an ultrasonic disperser. A 1000-mg sample of the hydrocarbon contaminant to be used was added to 1L of
deionized water in a glass vessel. Cavitation produced by a Sonicator horn broke the contaminant into very
fine particles which remained suspended for several hours. Dilutions were made from the prepared stock to
obtain a series of suspensions in half-decade steps of concentration (i.e., 1000, 300, 100, etc. mg/L). Be-
cause contaminant concentrations greater than 1000 mg/L were beyond the range of primary interest, con-
centrations greater than this value were not prepared or used. The lowest concentration was either 10 mg/L
or 1 mg/L depending upon the particular contaminant.
A list of the contaminants used during this program is presented in Table II.
TABLE II. CONTAMINANTS EXAMINED BY TEST APPARATUS
Chemical Used as Contaminant Refractive Index
Heptadecylbenzene 1 -4798
Dodecylbenzene 1.4820
n-Hexylbenzene 1.4900
tert-Butylbenzene 1.4927
p-Xylene 1 -4958
Ethylbenzene 1 -4959
m-Xylene 1 -4972
o-Xylene 1 -5055
Chlorobenzene 1.5241
2,6-Dimethylstyrene 1.5315
Cyclohexylbenzene 1.5329
1,2,3,4-Tetrahydronaphthalene (Tetralin) 1.5414
Bromobenzene 1.5597
3,3'-Dimethylbiphenyl 1.5946
Phenanthrene 1.5973
1-Methylnaphthalene (alpha isomer) 1.6170
1-Phenylnaphthalene (alpha isomer) 1.6646
Diesel Fuel —
Crude Oil —
15
-------
FILTER, REGULATOR
AND GAUGE
o>
3-WAY
AIR-CONTROL
VALVE
PRISM/I
NEUTRAL-DENSITY
FILTER (IF DESIRED)
RESERVOIR
MICROSCOPE
OBJECTIVE '
f
SUPPLY
CUT-OFF
VALVE
DRAIN
VALVE
LOW-POWER
LASER
POWER
SUPPLY
0
PHOTODIODE
SAMPLE
CUT-OFF
VALVE
COMPRESSED
AIR
SUPPLY
COILED CAPILLARY
TUBE CONTAINING
OPTICAL FIBER
ELECTROMETER
STRIP-CHART
RECORDER
frLOW-REGULATION
VALVE
GRADUATED
CYLINDER
Figure 5. Schematic Diagram of Test Apparatus Including Coiled Capillary-Tube Sensor Cell
-------
Figure 6. Photograph of Test Apparatus Employing Capillary-Tube Sensor Cell
-------
Fiber Materials
During the time that the chemical treatment apparatus and process were being perfected, 0.13-mm
(0.005-in.(-diameter crown-glass optical fibers were used because a large quantity of this material was left
from Phase I. However, because the index of refraction of crown glass (1.5224) is relatively high, only a lim-
ited range of contaminants could be tested.
After preliminary experiments with the chemical treatment apparatus and procedure were completed,
all subsequent experiments were conducted using 0.13-mm (0.005-in.)-diameter fused-silica optical fibers
which had a refractive index of 1.4585. With this reduced value of refractive index, it was possible to use a
wide range of hydrocarbon contaminants as described in the previous subsection.
Test Procedure
A chemically treated optical fiber contained in a coiled stainless steel capillary tube (i.e., the sensor cell)
was installed in the test apparatus previously illustrated in Figures 4 and 6. Typically, the stainless steel cap-
illary tube was approximately 0.8 m (31 in.) long and was formed into a 3.5-turn coil of 70-mm (2.8-in.) inside
diameter. The ends of the optical fiber projecting from the ends of the capillary tube were threaded straight
through the tee fittings of the test apparatus and were passed through silicone rubber compression seals
which retained the test fluid in the sensor cell.
After sensor cell installation, the apparatus was cleaned with successive washes of approximately
150 ml each of acetone and methanol. Cleaning solvents and test suspensions were forced through the sen-
sor cell by 300-kPa (44-psi) compressed air introduced into the reservoir. Following the two solvents, 300 ml
of deionized water was forced through the sensor cell to insure that all solvent was flushed from the system;
flow was terminated before ajl of the water had been used so that the flow path remained filled with deionized
water.
Test suspensions of a particular hydrocarbon contaminant in water were prepared in various concentra-
tions as previously described. These test suspensions were introduced sequentially into the test apparatus
reservoir beginning with the highest concentration. Each test suspension was forced through the sensor cell
at a rate of approximately 0.5 mL/s until the fiber saturated or all of the test suspension was used. While the
test suspension was flowing through the sensor cell, current through the photodiode was recorded on a strip-
chart recorder. After use of each test suspension, the three-step clean-up procedure previously described
was repeated. In most cases, this clean-up procedure restored optical transmission of the fiber to the value
observed prior to introduction of any hydrocarbon contaminants.
Transmission loss occurring over a specified time interval was measured on the strip-chart record. This
value of transmission loss in decibels (dB) was divided by the time interval in seconds (s) to obtain the slope
of the line on the strip-chart record in units of decibels per second IdB/s). Slopes obtained with this proce-
dure for a particular hydrocarbon contaminant were plotted on logarithmic graph paper as a function of con-
taminant concentration. Typical curves are presented later.
TEST RESULTS
Tests conducted during this program were intended to cover four main points: (1) comparison of the
new capillary-tube sensor cell with the glass U-tube configuration used in Phase I (Ref. 2), (2) determination
of the effects of various chemical treatment processes on performance of an organophilic optical fiber,
(3) evaluation of various treatment reagents, and (4) measurement of sensor cell response to a variety of hy-
drocarbon contaminants at various concentrations. As previously noted, raw data from a series of tests in-
volving a particular fiber and a single contaminant at various concentrations were translated into a graph re-
lating the time rate of decrease of optical transmission through the fiber to contaminant concentration. This
18
-------
graph, when plotted on logarithmic coordinates, typically was linear over a wide range of contaminant con-
centration. Examples of such curves are presented in later paragraphs to illustrate specific points.
Because the mathematical function which plots as a straight line on logarithmic coordinates is nonlinear
if plotted in linear Cartesian coordinates, sensitivity (usually defined as the derivative of the function relating
output to input) is not a constant and, hence, varies as a function of the independent variable (in the case of
the organophilic optical fiber oil-in-water monitor system, the independent variable is contaminant concentra-
tion). Thus, for comparison purposes, sensitivity must be measured at a specific value of the independent
variable. Throughout this report, sensitivity figures are specified for a contaminant concentration of 1000
mg/L. Details of the method used in calculating sensitivity figures given in this report are presented in
Appendix A.
Comparison of Performance of Two Sensor Cell Designs
In designing a new sensor cell for this program, three significant changes were made with respect to the
sensor cell previously employed in Phase I (Ref. 2) of the program: (1) contact between the test fluid and the
organophilic optical fiber was greatly enhanced, (2) active {i.e., curved) length of the optical fiber was in-
creased, and (3) optical cross-coupling between turns of the coiled optical fiber was eliminated. Effects of
these factors were not evaluated individually, but their combined influence on sensor cell sensitivity was de-
termined. The curves in Figure 7 show data acquired with the glass U-tube sensor cell employed during Phase
I (Ref. 2) and the new coiled capillary-tube sensor cell used in this program. In both cases, the contaminant
was T,2,3,4-tetrahydronaphthalene (tetralin). It is apparent that the response of the new sensor cell is sub-
stantially greater at any specified value of contaminant concentration and that the slope of the response
curve for the new sensor cell is greater than that of the old sensor cell. Using the definition of sensitivity given
in Appendix A, sensitivity of the new sensor cell at a contaminant concentration of 1000 mg/L is approxi-
mately 260 times the sensitivity of the U-tube cell at the same concentration. This increase in sensitivity
should substantially reduce the effects of system noise.
Evaluation of Fiber Treatment Processes
As described in an earlier section, a transition from the chemical treatment apparatus and process em-
ployed in Phase I (Ref. 2) to the improved process and apparatus used during Phase II was achieved by a
systematic progression of changes. The objectives of transition were to (1) simplify the process, (2) reduce
the quantities of reagents consumed, (3) minimize handling of the fiber, and (4) reduce the time required to
chemically treat a fiber. Optical fibers produced by each step in the transition progression were evaluated
using the previously described test apparatus and procedure. The various chemical treatment processes and
apparatus variations produced optical fibers which had equivalent sensitivities, indicating that the four objec-
tives had been achieved.
Effects of Various Treatment Reagents
Several coating reagents, used singly and in combination (discussed in a previous section), were tried.
Analysis of test data indicates that the method of applying the coating reagent and the reagent itself affect
the capability of the optical fiber to adsorb hydrocarbons.
Fibers A, B, C, D, and E (refer to Table I for identification) were prepared to facilitate comparison of
octadecyltrichlorosilane and octadecyltriethoxysilane coating reagents. The hydrocarbon chain left on the
glass fiber is the same with these reagents, but the leaving group is different. With n-hexylbenzene as the test
contaminant, better response was obtained with Fibers B, C, and E which were treated with octadecyltrie-
thoxysilane; however, response of Fiber D, which was also treated with a triethoxy-based silylating reagent,
was about the same as for Fiber A which was treated with a trichloro-based silylating reagent. With 1-methyl-
naphthalene as the test contaminant, Fibers B and C exhibited steeper response curve slopes than did Fiber
A; however, the lower limit of detection of 1-methylnaphthalene was approximately 30 mg/L for fibers
19
-------
10
M
CO
tn
o
V)
i
-------
treated with the trichloro-based reagent and between 30 mg/L and 100 mg/L for fibers treated with the trie-
thoxy-based reagents. Sensitivity differences could not be correlated with the reagent leaving group. Further,
sensitivities to various hydrocarbon contaminants were not the same for fibers treated with reagents con-
taining different leaving groups.
Experiments were also performed with additional treatment reagents which left other hydrocarbon
chain groups on the fiber. A diphenylsilane reagent was used on Fiber F, and an n-decylsilane reagent was
used on Fiber G (see Table I for fiber identification). Test data indicate a potentially greater response curve
slope for fibers treated with the diphenylsilane reagent. The lower limit of detection is approximately the same
for fibers treated with diphenylsilane and octadecylsilane reagents. Fibers treated with the n-decylsilane rea-
gent were not as sensitive as fibers treated with octadecyl or diphenylsilane reagents when n-hexylbenzene
was the testing contaminant, but were more sensitive than fibers treated with the octadecylsilane reagent
when the contaminant was crude oil.
For the present, the best coating reagent for use in the hydrocarbon-in-water monitor instrument ap-
pears to be octadecyltrichlorosilane. This reagent produces the best overall response to a wide variety of hy-
drocarbon contaminants. Because different contaminants do not produce uniform responses in fibers with
various coatings, there does not appear to be an optimum coating reagent, among those tested, for all con-
taminants of interest.
Sensitivity to Various Contaminants
Numerous pure hydrocarbon compounds, previously listed in Table II, were employed as contaminants
during evaluation and test of the new coiled capillary-tube sensor cell. The curves of Figure 8 illustrate sensor
cell response to three such contaminants: (1) n-hexylbenzene, (2) tetralin (1,2,3,4-tetrahydronaphthalene),
and (3) p-xylene. The curve for tetralin in Figure 8 is typical of the responses observed for many different con-
taminants; the curve is linear (on logarithmic coordinates) above a particular value of contaminant concentra-
tion (100 mg/L for tetralin) and falls off sharply below that value. The value of contaminant concentration at
which system response (i.e., rate of transmission loss ) is 0.001 dB/s is defined for purposes of this report as
the "detection threshold". From the curves in Figure 8 it is evident that the detection threshold for n-hexyl-
benzene is about 2 mg/L; for tetralin, approximately 30 mg/L; and for p-xylene, slightly greater than 400
mg/L. Because the greatest contaminant concentration used in this program was 1000 mg/L, it is not known
how far the linear range extends beyond this value. Data acquired during Phase I (Ref. 2) indicate that the
response curve for tetralin is linear to a contaminant concentration of at least 7000 mg/ L.
Although most tests of the organophilic optical fiber hydrocarbon-in-water monitor system were con-
ducted using pure hydrocarbon compounds as previously noted, a few tests were conducted using diesel fuel
and crude oil as the contaminant. The system was capable of detecting diesel fuel at a concentration of
17 mg/L. However, as illustrated in Figure 9, the system was highly sensitive to crude oil and had a detection
threshold of less than 3 mg/L. Composition of the crude oil used in this experiment is described in
Appendix B.
Sensitivity (defined in Appendix A) and detection threshold (i.e., contaminant concentration for which
rate of transmission loss is 0.001 dB/s) varied from contaminant to contaminant. From the very limited data
available, it appeared that detection threshold for a particular contaminant was related to solubility of that
compound in water. A reasonable search of the literature produced very little specific solubility information
because most hydrocarbon compounds which are less than 1% (10,000 mg/L) soluble in water are reported
as "insoluble". Specific values of solubility in water at 25°C were found for only five of the hydrocarbon com-
pounds used as contaminants during this program:
1-methylnaphthalene 30mg/L
m-xylene 170mg/L
o-xylene 200 mg/L
21
-------
n-HEXYLBENZENE
0.001
10 100
CONTAMINANT CONCENTRATION (mg/L)
1000
Figure 8. Response of Coiled Capillary-Tube Sensor Cell to Contaminants of Varying Solubility in Water
22
-------
M
CD
1/5
tf)
O
o
(f)
-------
p-xylene 200 mg/L
ethylbenzene 210mg/L
For each of these compounds, the detection threshold (as defined above) was slightly greater than the solu-
bility figure as illustrated by the curve for p-xylene in Figure 8. Although still speculative, it may be possible to
utilize this effect to actually measure solubility of various hydrocarbon compounds in water in the low part-
per-million range.
Sensitivity (as defined in Appendix A) of the coiled capillary-tube sensor cell containing an organophilic
optical fiber treated with octadecyltrichlorosilane was not the same for all contaminants tested. This is illus-
trated in Table III for contaminants which had detection thresholds of 100 mg/L or less. Sensitivity ranged
from 0.15 x 1Q-3 for 2,6-dimethylstyrene and 1-phenylnaphthalene to 2.2 x 1Q-3 for 1-methylnaphthalene. Be-
cause sensitivity is not the same for all contaminants, it is necessary to calibrate the instrument in which the
sensor cell is used for the particular contaminant or combination of contaminants expected.
TABLE III. SENSITIVITIES AND DETECTION THRESHOLDS OBSERVED FOR
VARIOUS AROMATIC HYDROCARBON CONTAMINANTS
Approximate
Detection Threshold
Contaminant Sensitivity* (mg/L)
n-Hexylbenzene
Cyclohexylbenzene
Heptadecylbenzene
3,3'-Dimethylbiphenyl
1 -Phenylnaphthalene
1,2,3,4-Tetrahydronaphthalene(Tetralin)
1 -Methylnaphthalene
2,6-Dimethylstyrene
tert-Butylbenzene
1.97x10-3
0.95 x 1C-3
0.34x10-3
0.18x10-3
0.15x10-3
1.58x10-3
2.12x10-3
0.15x10-3
0.17x10-3
2
<10
<10
<10
<10
30
40
40
40
'Calculated using the formula derived in Appendix A.
PROTOTYPE INSTRUMENT
Sensor Ceil
The sensor cell used in the previously described test apparatus was modified to provide a different
mounting configuration and was incorporated in the prototype instrument. This cell employed a 0.13-mm
(0.005-in.(-diameter fused-silica optical fiber approximately 1 m (39 in.) long. The fiber was treated with octa-
decyltrichlorosilane while installed in a coiled stainless steel capillary tube in accordance with the previously
described process. Further details of the sensor cell are presented in a previous section.
Optics
Illumination for the optical fiber in the sensor cell is provided by a 2-mW helium-neon laser operating at
the 632.8-nm (visible red) wavelength. The collimated beam from the laser is focused to a small spot by a
10X, 0.33-numerical aperture (N.A.) microscope objective. The objective lens is mounted in a 3-axis posi-
tioner to facilitate focusing and centering the small spot of light on the input end of the optical fiber.
24
-------
Light emitted from the output end of the optical fiber falls on the sensitive area of a PIN (positive-intrin-
sic-negative) silicon photodiode. Ends of the optical fiber were cleaved using a sharp diamond scribe to initi-
ate the break; this process produces smooth, flat surfaces at the ends of the fiber.
Signal Processing
It was determined experimentally that the logarithm of contaminant concentration is a linear function of
the logarithm of the time rate of change of the logarithm of the light intensity emitted at the output end of the
optical fiber. Accordingly, the signal processing system shown in block diagram form in Figure 10 was
devised.
Photodiode output current, which is linearly proportional to intensity of light emitted from the output
end of the optical fiber, passes into a logarithmic amplifier which delivers an output voltage proportional to
the logarithm of the input current. The logarithmic signal is scaled and biased in a linear amplifier. Output
voltage from this amplifier is impressed upon an analog differentiator which delivers an output voltage pro-
portional to the time rate of change of its input voltage. Time constant (i.e., gain factor) of the differentiator
is set by means of a range switch which facilitates selection of one of four full-scale sensitivity ranges. Differ-
entiator output drives an analog meter mounted on the front panel of the prototype instrument. The meter is
calibrated in units of decibels per second (dB/s) over the range from 0 to 1.0 dB/s, and the range switch pro-
vides multipliers of 0.01, 0.1, 1, and 10. System noise establishes a lower limit of 0.001 dB/s on the operating
range of the system. In operation, two known concentrations of the contaminant to be analyzed are passed
through the prototype instrument and scale readings are noted. Using a procedure described in the instruc-
tion manual for the prototype instrument, other concentrations of the same contaminant can be derived from
subsequent meter readings.
Fluid Flow
A flow diagram of the prototype instrument is illustrated in Figure 11. The central component of the
fluid handling system is the Reservoir which is used for temporary storage of the sample and regeneration
solvents. Materials may be introduced into the reservoir through (Da dip tube which fills the reservoir
smoothly from the bottom or (2) a spray nozzle which directs the incoming material against the walls of the
reservoir to improve cleaning action. Entry path into the reservoir is determined by the Reservoir Entry Selec-
tor, and the material entering the reservoir (i.e., Sample, Solvent No. 1, Solvent No. 2, or Deionized Water) is
determined by the Source Selector. Containers are provided for each of the four materials which can be ad-
mitted to the reservoir.
A compressed-air-operated Aspirator creates a partial vacuum in the reservoir, thus providing motive
force for transferring fluids from their respective containers to the reservoir. The Air Control provides for di-
recting compressed air into the aspirator during filling of the reservoir and into the reservoir for the purpose of
forcing fluids through the Sensor Cell. An Air Supply Inlet fitting and an Air Supply Cutoff are provided for
convenience in handling the compressed air supply, and gauges are provided for monitoring Air Supply Pres-
sure and Reservoir Pressure.
Fluid flow out of the reservoir can be cutoff, directed into the sensor cell, or directed into the Water-
Based Waste Receptacle by the Flow Control. Direction of flow through the sensor cell is determined by the
Flow Direction Control which incorporates two mechanically ganged valves. During analysis and flushing the
sensor cell with deionized water, flow through the sensor cell is in the forward direction, and waste fluid flows
into the water-based waste receptacle. During regeneration, when either Solvent No. 1 or Solvent No. 2 is
passing through the system, flow in the sensor cell is in the reverse direction, and waste fluids are delivered to
the Hydrocarbon-Based Waste Receptacle. A Filter in the forward-flow inlet side of the sensor cell retains
paniculate matter which could clog the cell. During regeneration this filter is back-flushed along with the sen-
sor cell.
25
-------
PHOTODIODE
LOGARITHMIC
AMPLIFIER
RANGE
SWITCH
METER
DIFFERENTIATOR
LINEAR
AMPLIFIER
SET-UP
CONTROL
Figure 10. Block Diagram of Signal Processing System
-------
AIR SUPPLY
PRESSURE
GAUGE
RESERVOIR
PRESSURE
GAUGE
RESERVOIR ENTRY
SELECTOR
AIR
SUPPLY
INLET
COILED
SENSOR
CELL
FLOW DIRECTION CONTROL
(MECHANICALLY GANGED)
VENT •*
WATER-BASED
WASTE
RECEPTACLE
HYDROCARBON-BASED
WASTE
RECEPTACLE
Figure 11. Flow Diagram of Prototype Instrument
-------
Package
The prototype instrument is packaged in a metal cabinet which is 64 cm wide, 38 cm high, and 81 cm
deep (25 x 15 x 32 in., respectively) as illustrated in Figure 12. Valves for controlling the flow of samples and
cleaning solvents, electrical controls, and the output meter are mounted on the front panel of the cabinet.
Containers for the cleaning solvents rest on a shelf attached to the rear panel of the cabinet. Services required
for operation of the system include compressed air and electrical power, connectors for which are provided
on the rear panel of the instrument. The sensor cell, laser, optical components and electrical components are
housed within the cabinet.
Because the beam of light emitted by the laser is completely enclosed within the cabinet, the prototype
instrument qualifies as a Class I laser system (Ref. 4) provided that the cabinet is not opened. A Class I laser
system may contain a laser of a higher classification, but is considered exempt from operational controls pro-
vided that the laser is completely enclosed and laser light is not accessible to the skin or eyes of a human
during normal equipment operation.
Operation
Procedures for operating the prototype instrument, designated Hydrocarbon-in-Water Monitor, Mk. I,
are given in a separate instruction manual provided with the instrument. Briefly, the signal processing circuit
is placed in the set-up mode, and preliminary electronic adjustment is made. Using the sequence of valve ma-
nipulations specified in the instruction manual, the reservoir is filled with the sample material, and the sample
is caused to flow through the sensor cell. The signal processing circuit is placed in the run mode, and a meter
reading is obtained. A calibration procedure, specified in the instruction manual, facilitates conversion of the
meter reading to contaminant concentration.
After a stable meter reading has been obtained, the remaining sample is dumped to a waste receptacle,
and regeneration is performed by passing two solvent systems sequentially through the sensor cell and,
finally, flushing with deionized water.
Test Data
Tests of the prototype instrument were performed using n-hexylbenzene to which the experimental test
apparatus responded well. A performance curve showing contaminant concentration as a function of meter
reading in dB/s is presented in Figure 13. Threshold of detection for this material is between 3 and 10 mg/L
which is not quite as good as obtained with the laboratory apparatus described earlier. The increased detec-
tion threshold was caused by system noise generated by (1) intensity fluctuations in the laser output and (2)
variations in optical fiber transmission resulting from irregularities in flow through the sensor cell; such noise,
even though of relatively low magnitude, is greatly accentuated by the differentiator in the signal processing
circuit and makes accurate reading of the meter very difficult. Even so, a detection threshold of less than 10
mg/L is quite good for a first-generation portable instrument employing a new technology.
Organic Polymer Fibers
Several types of organic polymer fibers, including conventional polyethylene, linear (high-density) poly-
ethylene, polypropylene and nylon, were obtained and tested to determine their potential as optical transmis-
sion devices and as oleophilic surfaces. None of these fibers exhibited adequate optical properties for use in
an oil-in-water monitor device. Fibers of the same generic material from different manufacturers had similar
properties. At present, an organic polymer fiber is not a satisfactory alternate to fused silica.
Fluoride Glass Investigation
For a hydrocarbon contaminant to be detectable by the organophilic optical fiber sensor, its index of
refraction must be greater than the refractive index of the fiber material. Fused silica, with a refractive index
28
-------
Figure 12. Hydrocarbon-in-Water Monitor, MK. I
-------
1000
O 100
<
oc
UJ
o
o
o
10
o
o
I
0.001
0.01
O.I
METER READING (dB/s)
10
Figure 13. Response of Prototype Hydrocarbon-in-Water Monitor Instrument to n-Hexytbenzene
-------
of 1.4585, has the lowest value of any commonly available material from which optical fibers can be made.
This relatively high refractive index excludes from detection all paraffinic hydrocarbons, but most aromatic
hydrocarbons are detectable by a fused-silica fiber.
Patents awarded in 1949 and 1950 (Refs. 5-8) describe a class of optical materials, referred to as "fluo-
ride glasses", which were composed of fused fluorides of beryllium, aluminum, barium, magnesium and
other metals in varying ratios. Indexes of refraction of fluoride glasses range from 1.3718 to 1.5140. If a fluo-
ride glass having refractive index of 1.3718 could be satisfactorily formed into a durable optical fiber and orga-
nophilically treated, paraffinic hydrocarbons down to n-hexane could be detected.
Because of the potentially useful characteristics of fluoride glasses, attempts were made to obtain addi-
tional information. Contact was established with Eastman Kodak Company to which the referenced patents
were assigned. The original inventor is no longer with the company, and because the patents have expired,
Eastman Kodak has no interest in pursuing development of fluoride glass. Contact was also established with
Corning Glass Works which has a current Department of Energy contract to develop and manufacture fluo-
ride and fluorophosphate glasses for laser-fusion experiments at Lawrence Livermore Laboratories. This con-
tract had been in existence for approximately 24 months, during which time Corning has established a plant
and has begun experimentation with these materials. However, work is still in the research phase, and pro-
duction has not been initiated; in fact, it will be at least one year before Corning is interested in other applica-
tions for fluoride glass.
Contact was also established with Galileo Electro-Optics Corp. which has personnel with limited experi-
ence in fluoride and fluorophosphate glasses. This company has a model shop for producing small quantities
of special optical fibers, and interest was expressed in manufacturing sample fluoride and fluorophosphate
glass optical fibers.
The principal constituent of fluoride glass is beryllium fluoride which is highly soluble in water. When
this material is fused with fluorides of other metals in appropriate ratios, solubility of the resulting glass in bulk
form is low. However, in the form of a small-diameter optical fiber immersed in flowing water, it is probable
that fluoride glass would dissolve rapidly. In general, the fluorophosphate glasses, which have refractive in-
dexes near 1.43, are significantly less soluble than fluoride glasses, and the higher-index fluoride glasses are
less soluble than the low-index types.
The presently used organophilic treatment reagents react with oxygen atoms at the surface of the glass
and probably are not suitable for treating a fluoride glass which has no oxygen-containing compounds in its
composition. Thus, a completely new treatment system, which could accommodate fluoride glass chemistry,
is necessary. From the proceeding, use of fluoride glass in an organophilic optical fiber oil-in-water monitor
system is judged to be impractical at this time.
Evaluation of Light Sources
In the test apparatus used during Phases I and II of this program, a low-power helium-neon laser
emitting red light at a wavelength of 632.8 nm was employed. A laser source was selected because it simpli-
fied coupling light into the input end of the optical fiber. However, there is no theoretical restriction on the
type of light source. Accordingly, during this program, comparison was made of the relative intensity of light
emitted from the output end of an optical fiber when the input end was illuminated by four different light
sources. The fiber used in these tests was similar to that used in the sensor cell, that is, fiber diameter was
0.13 mm (0.005 in.), fiber length was 1.0 m (39.4 in.), and fiber material was fused silica. The sources evalu-
ated included (1) a Hewlett Packard Type 5082-4658 red-light-emitting diode, (2) a Monsanto Type ME 7124
infrared-emitting diode, (3) a General Electric Type 1631X high-intensity incandescent lamp, and (4) a Spectra
Physics Model 120 helium-neon laser. Important characteristics of each source are presented in Table IV.
Each source was tested in a number of different ways including (1) the input end of the fiber placed in
contact with the exterior surface of the light source in the region of most intense emission, (2) the input end
31
-------
TABLE IV. LIGHT SOURCE DATA
Manufacturer
Type Number
Output Power
Luminous Intensity
Power Input
Power Efficiency
Luminous Efficiency
Half-Power Beam Angle
Unit
—
—
mW
Im
W
%
Im/W
O
Light-
Emitting
Diode
Hewlett Packard
5082-4658
0.31
—
0.066
0.47
—
35
Infrared-
Emitting
Diode
Monsanto
ME 7124
2.2
—
0.19
1.2
—
12
Incandescent
Lamp
General Electric
1631X
—
290
18
—
16
180
Laser
Spectra
120
5.0
—
50
0.01
—
0.057
Physics
of the fiber at the focal point of a 30-mm focal-length, f/1.6 plano-convex lens focused to produce minimum
spot size, (3) the input end of the fiber at the focal point of a 12,-mm focal-length f/0.67 aspheric condensing
lens with focus adjusted for minimum spot size, and (4) the input end of the fiber at the focal point of a lens
system comprising both of the previously described lenses with the aspheric lens nearest the source. It was
found that light intensity emitted from the output end of the fiber was greatest by a factor of 100 when the
laser source was used with any of the lenses.
Because light emission from a laser is highly coherent, it appears to originate from an almost dimension-
less point source. As a result, energy from the laser can be focused to a very small point approximating the
diffraction limit. All of the other sources were relatively extended and, therefore, energy from these sources
could not be focused to a very small spot. Since the input aperture of the optical fiber was only 0.13 mm
(0.005 in.) in diameter, capability of focusing a high percentage of the energy available from the source into a
spot sufficiently smalt to enter the optical fiber aperture is extremely important. Even though some of the
non-laser sources tested emitted sufficient energy to be useful, under the best conditions this energy could
be focused to a spot no smaller than 1 mm (0.04 in.) in diameter; as a result, less than 2% of the available
energy actually passed through the entrance aperture of the optical fiber so that the source was very ineffi-
ciently utilized. Accordingly, an inexpensive low-power helium-neon laser was selected as the light source to
be included in the prototype instrument.
32
-------
REFERENCES
1. H.S. Silvus, Jr. and F. M. Newman, "Development of Oil-in-Water Monitor," Proposal No. 15-1487,
Southwest Research Institute, San Antonio, Texas, 13 October 1975.
2. H. S. Silvus, Jr. , F. M. Newman and G. E. Fodor, "Development of Oil-in-Water Monitor," Interim
Technical Report, Grant R804368-01, Environmental Monitoring and Support Laboratory, U. S. Envi-
ronmental Protection Agency, Cincinnati, Ohio, 11 November 1977.
3. H. S. Silvus, Jr., F. M. Newman, G. E. Fodor and F. K. Kawahara, "Development of a Novel Hydrocar-
bon-in-Water Monitor," presented at the Symposium on Energy and Resource Development of Conti-
nental Margins, University of Southern California, Los Angeles, California, 8-10 March 1978.
4. American National Standard for Safe Use of Lasers, American National Standards Institute, New York,
NY, Standard ANSI Z136.1-1976.
5. K.-H. Sun and T. E. Callear, "Method of Making Fluoride Glasses," U. S. Patent 2,466,506, 5 April
1949, Assigned to Eastman Kodak Company, Rochester, New York.
6. K.-H. Sun, "Method of Making Fluoride Glass," U. S. Patent 2,466,507, 5 April 1949, Assigned to
Eastman Kodak Company, Rochester, New York.
7. K.-H. Sun, "Fluoride Glass," U. S. Patent 2,466,509, 5 April 1949, Assigned to Eastman Kodak Com-
pany, Rochester, New York.
8. K.-H. Sun and M. L. Huggins, " Fluoride Glass," U. S. Patent 2,511,244, 13 June 1950, Assigned to
Eastman Kodak Company, Rochester, New York.
33
-------
APPENDIX A
DISCUSSION OF SENSITIVITY
When the response function of a system plots as a straight line on linear Cartesian coordinates, sensitiv-
ity of that system is defined as the slope of the straight line or, in other words, the derivative of the dependent
variable (output) with respect to the independent variable (input). Since such a linear function has constant
slope, sensitivity is a well defined, constant number. However, in the organophilic optical fiber oil-in-water
monitor system discussed in this report, output (i.e., time rate of optical transmission change) as a function
of input (i.e., contaminant concentration) plots as a straight line on logarithmic coordinates. Such a function,
when plotted on linear Cartesian coordinates, is nonlinear and, therefore, does not have constant slope; thus,
sensitivity is not constant and must be specified at a particular value of input. In this report, sensitivity figures
are specified for a contaminant concentration of 1000 mg/L. The remainder of this Appendix derives the for-
mula used for calculating sensitivity when the system response function plots as a straight line on logarithmic
(i.e., log-log) coordinates.
A function of the form
y = abxm (1)
plots as a straight line on logarithmic coordinates. By taking the logarithm (to the base a) of both sides of
Equation (1), this functional form can also be expressed as
logaV = m loga x + b (2)
This is the equation of a straight line. The slope m can be determined from the coordinates (x1( yj) and (X2,
y2> of two points which lie on the straight line, and the intercept b can be determined from the coordinates
(x,y) of any point on the line according to the following equations:
b =
(4)
Note that the slope m is independent of the logarithmic base a; thus, m can be determined using either natu-
ral (base e) or common (base 10) logarithms.
Sensitivity is usually defined as the slope or derivative of the system response function. As will be
shown in subsequent steps of this derivation, this concept of sensitivity can be applied to a function of the
form indicated by Equation (1) or Equation (2) provided that sensitivity is calculated for a specified value of
the input variable. Therefore, define sensitivity S as the derivative y' of Equation (1).
34
-------
S = y' = mabxm-1 (5)
The slope m is a constant which can be calculated from Equation (3) above and thus requires no further dis-
cussion. The second term can be evaluated [using Equation (4)] as
ab = a exp (loga yx~m) = yx-m (6)
Substituting Equation (6) in Equation (5) yields
, my ,7%
S = m (yx-m) xm ~ ' =— "'
This result indicates that a value of sensitivity S can be calculated for any point (x,y) on the system response
curve. To achieve comparable sensitivity figures, the selected point should always have the same value of the
independent variable x (in this case, contaminant concentration).
EXAMPLE: Calculate sensitivity of sensor cell for tetralin using data given in the curve of Figure 8 in
the body of the report.
SOLUTION: Select two points at extremes of the linear range of the curve:
Xl = 1000mg/L
y, = 1.32dB/s
x2=100mg/L
x2 = 0.084 dB/s
Compute slope from Equation (3)
logliN '09
ly7/
\
/
..., 084 / log(15.7)
m= ;—r= 7 v = = 1.196
x2\ MOOO \ log (10)
Then compute sensitivity from Equation (7) using the values of xj and yj for x and y.
my 1.196x1.32
s= x - 1000 - '-53*10-3
35
-------
APPENDIX B
DESCRIPTION OF CRUDE OIL USED IN TESTS
The crude oil used in evaluating the organophilic optical fiber sensor cell described in this report was
taken from Sample AL6846-C. This sample is a black, opaque, viscous crude oil from Pearsall, Texas. The
sample was separated into saturates, aromatics, and polars using a gravimetric liquid chromatographic proce-
dure for oils. Olefins could not be obtained by this method. Saturates comprised 42% of the sample by
weight, aromatics 25%, polars 18%, and 15% were lost. Other tests on the crude oil sample indicated a sulfur
content of 1.7% by weight and a water content, determined by the Karl Fischer method, of 0.11 % by weight.
Specific gravity of the sample was 26.0° API at 15.6°C(60°F), and gross heat of combustion was 43.4 MJ/kg.
Ash content was 0.02% by weight. Pour point was determined to be -10°C(14°F), and viscosity at 24°C
(75° F) was 97 mm2/s (97 centistokes).
36
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/4-80-040
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Development of Oil-in-Water Monitor
Phase II
5. REPORT DATE
AUGUST 1980 ISSUING DATE.
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
H. S. Silvus, Jr.
F. M. Newman, 0. H. Frazar
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
Southwest Research Institute
San Antonio, Texas 78264
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
Grant No. R805817-01
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Research 5/78-5/79
14. SPONSORING AGENCY CODE
EPA/600/06
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A novel approach to quantitatively monitoring suspended hydrocarbons in water was conceived at
Southwest Research Institute in 1975, and feasibility was subsequently demonstrated under sponsorship of
the U. S. Environmental Protection Agency through Grant R804368-01. This new oil-in-water monitor tech-
nique brings together for the first time two previously unrelated technologies: (1) reversed-phase liquid
chromatography and (2) fiber optics. A special organophilic optical fiber, created by a chemical treatment
process routinely used in reversed-phase liquid chromatography, collects and concentrates suspended hydro-
carbon materials on its surface. Collected material alters optical transmission through the fiber in such a way
that the logarithm of the time rate of change of the logarithm of optical transmission through fiber is linearly
proportional to the logarithm of contaminant concentration, provided that concentratior\exceeds a certain
detection threshold.
Various methods of chemically treating an optical fiber and various treatment reagents were evaluated.
Numerous treated optical fibers were tested individually, each with a variety of separate aromatic hydrocar-
bon contaminants, in a newly designed capillary-tube sensor cell. Additionally, a laboratory demonstration
instrument employing the organophilic optical fiber hydrocarbon-in-water monitor technique was fabricated
and tested. For such aromatic hydrocarbons as n-hexylbenzene, cyclohexylbenzene, heptadecylbenzene, 3,
3'-dimethylbiphenyl, and 1-phenylnaphthalene and for crude oil, detection thresholds of less than 3 mg/L
were observed, and system response was linear over a contaminant concentration range of greater than 2.5
decades (i.e.. greater than 300:1). Data indicate that the high end of the range can be extended if desired.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Pollution, Crude Oil, Petroleum,
Wastes, Shale oil, coal liquid wastes,
Oil Pollution, Monitoring
Monitor, Liquid
Chromatography, Fiber
Optics, Hydrocarbon
Chromatography
14C
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
45
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
37
•fr U.S. GOVERNMENT PRINTING OFFICE: 1980--657-165/0089
------- |