vvEPA
United States
Environmental Protection
Agency
Municipal Environmental Research"
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-206 Oct. 1981
Project Summary
Development of a
Multidetector Petroleum
Oil-ln-Water Monitor
Robert W. Melvold
This research report describes an
effort to develop a prototype petroleum
oil-in-water monitoring system that
will continuously measure oil (whether
free, suspended, dissolved, or emul-
sified) in water carrying a variety of
potential interfering substances. An
extensive desk-top survey of com-
mercially available oil monitors was
carried out. Three devices were
selected for thorough laboratory
evaluation: (1) Sigrist Fluorescence
Monitor Model FU, * (2) CE Invalco
D.O.W. Monitor, and (3) Horiba
OCMA-25 Oil Content Monitor.
The laboratory evaluation of the
three oil detectors forms the major
emphasis of this report. The work is
described in detail, including the
construction of a flow loop and its
operation, the installation of the
selected detectors in the breadboard
system, the development of a data
retrieval system, the calibration
methodology employed, the initial
checkout of the detectors, and the
extensive multidetector evaluation of
each detector's performance for
quantifying petroleum oil in the
presence of impurities and other
interferents. Included is a compre-
hensive discussion of results and
conclusions derived from the data-
reduction phase of the work. The
interferent effect and the sensitivity,
linearity and accuracy, repeatability,
'Mention of trade names or commercial products
does not constitute endorsement or recommenda-
tion for use.
response speed, and reliability of each
detector are described in detail.
Recommendations fora preliminary
design for the prototype system and a
suggested program plan to develop it
are also given.
This Project Summary was devel-
oped by EPA's Municipal Environ-
mental Research Laboratory, Cincin-
nati, OH, to announce key findings of
the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
No satisfactory commercially available
instrument or concept presently exists
for continuously measuring all types of
oil in various physical states (free,
suspended, dissolved, or emulsified) in
water containing various types and
concentrations of matrix materials. Oil
monitors currently available or being
developed rely on single detection
concepts. Each concept has inherent
advantages and disadvantages that may
significantly influence the accuracy of
results. Factors such as oil type, water
characteristics, particulates, and non-
petroleum contaminants all influence
the accuracy and may result in errors of
a factor of 2 or more.
The complex chemical composition of
petroleum-derived oils and the similar-
ities of oil components to nonpetroleum
water contaminants dictate the use of a
multidetector monitoring system for the
improved measurement of petroleum oil
-------
in water containing nonpetroleum
contaminants. The concept is that a
combination of different detectors
would be integrated into a viable system
that, once calibrated for a particular oil,
responds to the various compounds it
detects but acknowledges the ratio that
is characteristic of the petroleum oil
being monitored. Accurate and specific
measurements can therefore be per-
formed despite the presence of a limited
number of interfering nonpetroleum
contaminants by subtracting the con-
tribution of the interfering materials
from the total response.
The multidetector system will have
the potential for monitoring the level of
petroleum oil present in discharge
bilgewater from marine cargo vessels
and in discharge effluents from offshore
oil platforms, oil tankers, and ballast
treatment plants. Applications may also
include detection of crude oil leaks/
seepages into the ocean from drilling
operations around and related to marine
oil drilling rigs.
Oil Monitor Survey
To determine the feasibility of devel-
oping a multidetector monitoring system
for the continuous quantification of
petroleum oil in water, Rockwell EMSC
conducted a survey of commercially
available oil monitors. This study
involved a desk-top analysis of currently
available oil monitors and an evaluation
of their detection concepts. This survey
resulted in the recommendation of
several highest-rated, off-the-shelf oil
monitors for inclusion in a detection
system. (The oil monitor survey is
included in the appendix to the report.)
More than 40 manufacturers were
contacted by telephone to identify oil-m-
water monitoring devices that might
qualify for the analytical evaluation.
From the product literature and telecons,
devices produced by 13 firms appeared
worthy of evaluation. Technical informa-
tion deemed pertinent to the evaluation
of the 13 devices was obtained and
tabulated with respect to 25 technical
specifications and qualifications. Rating
factors were assigned to each particular
parameter, and then each device was
rated according to criteria established to
reflect differences in the various
devices.
In addition, a user's evaluation of the
top-rated devices in each of three
different design categories (based on
principle of operation) was conducted to
provide additional data for the study.
Asa result of the oil monitor survey of
manufacturers and users, three com-
mercially available oil-in-water monitors
were recommended for inclusion in the
multidetector petroleum oil-in-water
monitoring system to be evaluated
under the experimental study. These
monitors were the Sigrist Fluorescence
Monitor Model FLJ (UV fluorescence),
the CE Invalco D.O.W. Monitor (ultra-
violet absorbance), and the Horiba
OCMA-32A Oil Content Monitor (ex-
traction with infrared absorbance).
These specific devices were selected on
the basis of their reported ability to
quantify petroleum oil in water in the
presence of various interferents. An
important factor in the selection of
these three detectors was that each
device should utilize a different principle
of operation, thus potentially providing a
better-defined signature of the known
petroleum oil than a group of detectors
operating on the same principle.
Experimental System
A flow loop was fabricated for the
purpose of evaluating the assembled
detectors in a breadboard, preprototype
system configuration. The flow loop was
designed to distribute simultaneously a
suitable oil-in-water sample to each of
the various detectors. The preprototype
or breadboard system is thus capable of
introducing oil (with or without various
other contaminants) into water in such
a manner that reproducible and repre-
sentative oil-in-water dispersions can
be provided to the selected monitoring
devices for evaluation. The selected
instrumentation was assembled at
Rockwell EMSC according to the con-
ceptual flow loop design depicted in
Figure 1.
Two polyethylene-lined, 0.21 -m3 (55-
gal) drums serve as the water reservoir,
providing water to a Price HP175
centrifugal pump. The pump circulates
the water through a subloop, wherein
the water flows through an eductor (Din
Figure 1), aspirating oil and, in some
cases, an interferent from individual
stainless steel containers. By suitable
manipulation of the metering valve
associated with each rotameter (E in
Figure 1), the amount of fluid desired is
A —HZ0 reservoir t
B —Centrifugal pump "
Cx —Pressure gauge
D —Eductor
E —Rotameter
F —Oil reservoir
G —Interferent reservoir
— Temperature probe
—Sigrist Fluorescence Meter Model FLJ
—CE Invalco D. O. W. Monitor
—Horiba Model OCMA-25 Oil Content
Monitor
H
OCx
—tXHX T
Drain
Sampling
Port
Drain
Figure 1. Flow loop (conceptual design) used in evaluation testing.
-------
regulated to produce specific concen-
trations of the fluid in water. While a
portion of the water stream is diverted to
the three detectors (shown in Figure 2
along with associated readout gear), the
remainder is recycled through the oil
eductor subloop, where it is joined by
fresh quantities of water and oil. At the
high linear flowrates produced in the
flow loop, a steady-state concentration
level is rapidly reached. By appropriate
adjustment of the detector metering
valves and the drain shutoff valve (see
Figure 1), the manufacturer-specific
sample flowrate to each detector and
the desired total system flowrate can be
achieved. This adjustment is accom-
plished by monitoring the detector
pressure gauges (Ci, Cj, and CK) and the
system pressure gauge (Cs) for specific
pressures that wilt produce the desired
flowrates, as determined from previous
flow calibration data. A sampling port is
included so that samples can be drawn
for analysis to corroborate oil concen-
trations in water as indicated by
preselected rotameter settings. Also, a
Digitec digital thermometer (Model
2770A) was installed in the flow system
for monitoring the temperature of the
fluid medium during the course of the
laboratory experiments.
The data retrieval system consisted of
a Digistrip 15-channel digital recorder
along with a Linear Instruments Corp.
(LIC) two-pen recorder. All monitoring
readouts were recorded on the Digistrip
and, additionally, the response outputs
from both the Invalco and the Sigrist
devices were interfaced with the LIC
dual-channel strip-chart recorder Model
485.
Multidetector System
The three detectors evaluated during
the experimental study are described in
the following paragraphs.
The Sigrist Fluorescence Meter is
based on the principle that the aromatic
compounds present in nearly all petrol-
eum oils characteristically emit radiation
of a higher wavelength than the
incident radiation when excited by
ultraviolet radiation. Through the use of
a suitable optical system and detector,
this emitted radiation can be measured
and correlated to the concentration of
petroleum oil present in a sample.
A mercury light source emits a light
beam through a calibrated fluorescence
comparison standard and a flow cell.
The measuring beam produced in the
flow cell by fluorescent material strikes,
Figure 2. Selected detectors and associated recorders (on right) used in labora-
tory evaluation.
(together with the light originating from
the fluorescence comparison standard),
the oscillating mirror, which alternately
emits (about 600 times per second) the
measuring beam and the comparison
beam onto a photocell. The latter
therefore receives alternately at identi-
cal frequency a measuring beam and a
comparison beam of different light
intensity. The photocell transforms the
varying intensities of the two light
beams into a current, which after
subsequent amplification drives a
synchronous servomotor. This motor
uses a mechanical shutter to regulate
the intensity of the comparison beam to
such an extent that both light beams
directed onto the photocell have an
equal intensity. The shutter is coupled
mechanically to a measuring drum. The
indicated value obtained thus directly
depends on the position of the shutter,
which will change only if fluorescence
fluctuations influence the intensity of
the measuring beam.
The Invalco Dispersed Oil in Water
(D.O.W.) Monitor is based on the
principle that clean, clear water will
transmit ultraviolet light with very low
absorption, whereas most petroleum
oils and their derivatives will partially or
completely absorb the ultraviolet light
because of the aromatic compounds
present in petroleum oils. Variations in
absorption of ultraviolet light by the
carrier stream provides a measurement
of aromatic content to which the oil
content may be related.
Flowing through the sample cell, the
sample is continuously monitored by
the filtered ultraviolet and visible light.
The visible light maintains a background
reference for the detector circuit. When
any contaminant enters the chamber,
the ultraviolet light is absorbed to a
corresponding degree and the measur-
ing cell (a photodiode tube) senses the
reduction in the radiant light and
produces a corresponding current
output. This current is then converted to
a voltage that is amplified by the cell
amplifier. At the same time, the ref-
erence photodiode tube senses the
filtered visible light and produces a
current output that is converted to a
voltage and amplified by the reference
amplifier. The logarithms of these two
voltages are taken and then subtracted
to yield a ratio output signal that should
be linearly proportional to oil concen-
tration in the sample stream.
The Horiba Model OCMA-32A orig-
inally recommended was replaced by its
successor, the Horiba Model OCMA-25.
This detector employs solvent extrac-
tion/nondispersive infrared absorption.
Basically, the oil in the sample water is
extracted in a solvent, the absorption of
-------
which is measured at the wavelength
region of 3.4 to 3.5 micrometers by a
nondispersive infrared analyzer; the
reading is converted into an equivalent
value of oil. The extraction solvent is
identified by the manufacturer as
fluorochlorocarbon solvent S-316.*
Infrared radiation from matched light
sources is converted into an intermittent
light beam by a rotating chopper and
enters a reference cell and a sample
cell, passing alternately through each of
the cells to a detector cell. If any of the
infrared radiation entering the sample
cell is absorbed by oil in the solvent, it
creates a difference in the amount of
light alternately reaching the detector
cell. Because the detector is filled with a
gas that absorbs radiation in the
infrared region required for detection of
oil content, a thin membrane provided in
the detector cell flexes with the resulting
change in pressure and an electrical
signal corresponding to the difference
in the amount of light absorbed is
developed. This electrical signal is thus
proportional to the oil content of the
sample stream.
Calibrations
Before flow testing of the detectors
* Trademark registered by
Honba, Ltd., Kyoto, Japan.
could begin, certain calibration data
were required. Data on the centrifugal
pump flowrates with pressure and data
on rotameter flowrates with various
fluids were obtained by measurements
conducted in the laboratory. Centrifugal
pump flowrates were measured at
several flow loop pressures, and a flow
curve was developed from the data.
Thereafter, readings of the main flow
pressure provided the necessary water
flowrates for calculating oil or inter-
ferent concentrations.
The primary method for determining
the concentration levels of the various
fluids used either as known oils or
interferents involved the use of labor-
atory-calibrated rotameters in the flow
loop. Except for a few cases in which
only a crude indication of flowrate was
required (such as in some of the
detector checkout tests), rotameter
tube-float combinations selected for a
specific fluid were calibrated with that
fluid under controlled conditions at
room temperature. Verification was
obtained from subsequent analytical
determinations using gas chromatog-
raphy. The analytical instrumentation
involved a Perkin Elmer Model 3920
(with an OV17 column and an FID
detector) and a Hewlett-Packard
HP3380A integrator.
Multidetector Evaluation
A comprehensive evaluation of the
selected oil meters was critically
conducted under laboratory conditions
to identify their ability to detect accu-
rately and specifically and to quantify
petroleum oil in fresh, brackish, or sea
water over the temperature and con-
centration ranges listed by the manu-
facturer. The selected devices were
tested with water contaminated with
petroleum oil and possible interfering
materials such as organics, vegetable
oils, dissolved solids, and detergents to
demonstrate that it was feasible for the
integrated package to quantify petroleum
oil in water despite the presence of such
interfering substances. The experi-
mental effort also provided information
on the sensitivity, linearity and accuracy,
repeatability, response speed, and
reliability of the selected oil monitors.
Useful data from all three devices
were obtained with DF-2 fuel blend (a
commercial grade produced by Chevron)
at a number of concentration levels and
in the presence of benzene and iso-
octane. Tabulation of the data from a
single run conducted with stagnant tap
water at room temperature followed by
fresh tap water is presented in Table 1.
All important flow system parameters
appear in the table. The table also lists (
Table 1 Detector Responses to DF-2 - Run A *
Indicated Response
Rotameter Concentration
Corrected Response
Time
8:43:00
8:47:45
8:51:00
8:54:30
8:59:30
9:03:30
9:06:15
9:06:30
9:07:15
9:12:15
9:15:45
9:19:30
9:27:00
9:37:30
9:47:15
9:56:45
10:06:45
10:21:45
10:28:00
10:35:30
Temperature
°C
20.7
20.5
20.6
20.6
20.6
20.6
20.6
Start tap
20.6
18.3
Invalco
%FS
8.5
11.7
15.5
19.3
24.7
67.1
86.1
Sigrist
%FS
31.5
35.3
39.0
42.0
48.0
48.0
48.4
Horiba
ppm^
0.0
1.3
3.36
4.86
7.44
8.04
7.78
DF-2
ppm
4.2
9.8
16.5
23.5
23.5
23.5
Benzene
ppm
—
—
—
—
54.9
115.0
Isooctane
ppm
—
—
—
—
—
—
Total
ppm
4.2
9.8
16.5
23.5
78.4
138.5
Invalco
ppm
0.0
4.8
10.5
16.2
24.3
87.9
116.4
Sigrist
ppm
0.0
0.19
0.37
0.52
0.82
0.82
0.84
Horiba
ppmv
0.0
5.2
13.4
19.4
29.8
32.2
31.1
water replenishment
88.0
25.5
48.0
44.5
8.00
7.32
23.5
23.5
117.2
—
—
—
140.7
23.5
119.3
24.0
0.82
0.79
32.0
29.3
Horiba manual reset
17.6
17.5
17.5
17.5
17.6
17.6
17.6
17.6
17.6
35.0
40.0
79.0
94.0
36.5
41.0
22.5
27.5
9.5
41.0
38.4
380
38.5
38.5
37.5
41.5
46.5
26.0
12.08
16.06
16.92
17.26
13.68
16.14
7.40
9.50
0.0
23.5
23.5
23.5
23.5
23.5
23.5
23.5
31.6
—
—
—
54.9
113.5
—
—
—
—
—
19.5
33.6
34.5
33.3
26.6
33.3
—
—
—
43.0
57.1
113.0
170.4
50.1
56.8
23.5
31.6
—
38.3
45.8
104.3
126.8
40.5
47.3
19.5
27.0
0.0
0.75
0.62
0.60
0.62
0.62
0.57
0.77
1.02
0.0
48.3
64.2
67.7
69.0
54.7
64.6
29.6
38.0
0.0
* Run A was conducted with STA GNANT (24 HRS) TAP WA TER A T ROOM TEMPERA JURE, followed by FRESH TAP WA TER A T ROOM
TEMPERATURE. Flow system parameters were: Sigrist ref. cell. FLJ-5; rotameters, DF-2, R-2-15AA-sapphire, benzene, R-2-15AAA-SS.
isooctane, R-2-15-AAA-glass; flow loop pressure, 3.24 x 10s Pa (32.5 psig); inlet pressure, Invalco. 1.5 x 10s Pa (7.0 psig),
Sigrist. 1.5 x 10s Pa (7.0 psig). Horiba. 1.1 x 10* Pa (1.5 psig).
it No acid injection.
-------
the time, temperature, indicated re-
sponse, calculated rotameter concen-
tration, and corrected response for each
individual flow test (i.e., each different
set of rotameter conditions). Data tables
such as Table 1 were developed for each
of seven runs that were conducted to
investigate the effects of temperature,
dissolved solids (synthetic sea water),
vegetable oil, and detergent. A plot of
the data showing the effect of water
type and temperature on the Sigrist
response to DF-2 is depicted in Figure 3.
Similar data plots were developed for
the other two detectors.
Useful data from all three devices
were also obtained with Gulf crude oil
(from Shell platform SM130B in the
Morgan City, Louisiana, grid) at a
number of concentration levels and in
the presence of benzene and isooctane.
A total of six runs investigating the
effects of temperature, dissolved solids
(synthetic sea water) and vegetable oil
were conducted. These data are pre-
sented in tables and data plots showing
the effects of water type and temperature
and of time on the responses that the
three detectors had to Gulf crude oil.
Conclusions
1. The multidetector concept (that is,
the combination of several dif-
ferent detectors integrated into a
system that, once calibrated for a
specific oil, quantifies that oil) has
been demonstrated as a viable
technique for the measurement of
petroleum oil in water containing
nonpetroleum contaminants.
2. A preprototype multidetector sys-
tem has been assembled and
tested with petroleum oil and has
resulted in definite improvement
over any single commercially
available detector; but it will
require additional research to
become truly effective.
3. Oil monitors currently available or
being developed rely on single
detection concepts, each of which
has inherent advantages and
disadvantages that may signif-
icantly influence the accuracy of
the oil concentration as determined
by the individual detector. Factors
such as oil type, water character-
istics, and nonpetroleum contam-
inants all influence the accuracy.
4. On an individual basis, the three
selected oil-in-water detectors fail
to measure up to their reputed
ability to quantify petroleum oil in
2.00
7.75
7.50
1
I '-25
D
a 7.00
0.75
0.50
0.25
0
(D
(2)
(3)
(4)
(5)
(6)
Legend
+ Stagnant tap water at R. T. (Run A)
Q Fresh tap water at 30°C (Run B)
H Fresh tap water at 10°C (Run C)
CD Synthetic sea water at 30°C (Run D)
Q Synthetic sea water at 15°C (Run E)
Fresh tap water at R. T. (Runs A, C.D, and E)
(31
10 75 20 25 30
Rotameter Concentration, ppm
35
40
Figure 3. Effect of water type and temperature on Sigrist response to DF-2.
water. Although the devices detect
diesel fuel and crude oil, each fails
to discriminate (to some extent)
when interferents are present.
a. With respect to overall capabili-
ties, the Sigrist Fluorescence
Monitor Model FLJ appears to
be more capable of quantifying
petroleum oil in water with
interferents present than the
other two detectors tested. But
its capabilities are still too
limited to do an effective job by
itself.
b. The CE Invalco D.O.W. Monitor
does a good job of quantifying
petroleum oil-in-water without
interferents present. In fact, the
Invalco responds to most of the
chemicals tested on a one-for-
one basis, unlike the other two
detectors tested. Unfortunately,
its inability to discriminate suf-
ficiently is a major drawback to
a multidetector petroleum oil-
in-water monitoring system.
c. The Horiba OCMA-25 Oil Con-
tent Monitor does a good job of
quantifying petroleum oil in
water with aromatic hydrocar-
bons present. On the average,
the Horiba responds on a
somewhat less than one-for-
one basis to crude oil and
aliphatic hydrocarbons, and on
a somewhat greater than one-
for-one basis to diesel fuel.
Recommendations
We recommend that a subsequent
effort be undertaken to develop a
prototype multidetector petroleum oil-
in-water monitoring system that will
include the Sigrist Fluorescence Monitor
Model FU, the Horiba OCMA-25 Oil
Content Monitor, and a replacement for
the Invalco D.O.W. Monitor. The de-
velopment effort should include the
following items:
a. A limited review of the commer-
cially available detectors (based on
the original oil monitor survey)
should be performed to identify a
suitable substitute for the Invalco
D.O.W. Monitor, which is marginal
for inclusion because of its general
-------
inability to discriminate between a
known oil and various interferents.
If this search does identify a
suitable substitute, the identified
detector should be evaluated
experimentally before inclusion in
the multidetector system. If the
search is unsuccessful and a
determination is made to use the
Invalco, then the source of the
zero-shift problem must be found
and eradicated.
b. The Sigrist response suppression
by aliphatic hydrocarbons should
be investigated further so that a
satisfactory resolution may be
found.
c. The Horiba OCMA-25 should be an
original 0- to 100-ppm range unit
or a factory-converted 0- to 100-
ppm range unit.
d. A system flow loop should be
constructed according to a pro-
posed design that allows for
calibration of the prototype multi-
detector system as well as repre-
sentative sampling of the fluid
stream for the three detectors.
e. A data retrieval system consisting
of a minicomputer and associated
software should be integrated into
the prototype multidetector system
that will reduce the outputs from
the three detectors to a concentra-
tion value for the petroleum oil in
water. A minicomputer considered
acceptable for the assignment is
the Radio Shack TRS-80 Model III
or its equivalent. The software
language considered suitable for
this application is BASIC.
We further recommend that a followup
program be conducted to evaluate the
prototype multidetector system first in
the laboratory and then in full-scale
field trials.
The full report was submitted in
partial fulfillment of Contract No. 68-
03-2648 by Rockwell International
under the sponsorship of the U.S.
Environmental Protection Agency.
Robert W. Me/void is with Rockwell International, Newbury Park. CA 91320.
Uwe Frank is the EPA Project Officer (see below).
The complete report, entitled "Development of a Multidetector Petroleum Oil-ln-
Water Monitor," (Order No. PB 82-105 206; Cost: $9.50, subject to change)
will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, v'A 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Oil and Hazardous Materials Spills Branch
Municipal Environmental Research Laboratory—Cincinnati
U.S. Environmental Protection Agency
Edison, NJ 08837
U. S. GOVERNMENT PRINTING Off ICE: 1981/559-092/3330
-------
Postage and
United States Center for Environmental Research Fees Paid
Environmental Protection Information Environmental
Agency Cincinnati OH 45268 Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
PS OOOU329
RETURN POSTAGE GUARANTEED U 5 "
------- |