Document text
APPROVED FOR PUBLIC RELEASE. DISTRIBUTION UNLIMITED.
if'
jsl
THE DEVELOPMENT OF A
TEST SYSTEM FOR THE EVALUATION OF
REVERSE OSMOSIS WATER PURIFICATION MEMBRANES
Final Report Submitted in June 1984
Capt. Stephen J. Walker, Jr.
Robert E. Martin
Vincent P. Olivieri
lfM
^ .f
Supported By
FORT BELVOIR RESEARCH & DEVELOPMENT COMMAND
Contract No. DAAK70-82-K 7
The Johns Hopkins University
School of Hygiene & Public Health
Division of Environmental Health Engineering
Baltimore, Maryland 21205
i
SIX >?C i 1503 1
DT1C
jELECTE]
k JUN 2 6 1985
85 06 10 11
APPROVED FOR PUBLIC RELEASE. DISTRIBUTION UNLIMITED.
THE DEVELOPMENT OF A
TEST SYSTEM FOR THE EVALUATION OF
REVERSE OSMOSIS WATER PURIFICATION MEMBRANES
Final Report Submitted in June 1984
by
Capt. Stephen J. Walker, Jr.
Robert E. Martin
Vincent P. Olivieri
Supported By
FORT BELVOIR RESEARCH & DEVELOPMENT COMMAND
Contract No. DAAK70-82-K
The Johns Hopkins University
School of Hygiene & Public Health
Division of Environmental Health Engineering
Baltimore, Maryland 21205
TABLE OF CONTENTS
Page
Introduction . 1
Literature Review . . 3
Methods . 15
Results . 26
Discussion . 67
Conclusions . 82
Recommendations . 83
Literature Cited . 84
<
Bibliography . 86
Appendix 1: Simulant Data Sheets CROC
Appendix 2: Detailed Chemical Properties of Simulants
Accession For
NTIS GRA&I
DTIC TAB
#
Unannounced
Justification —
□
By -
Distribution/
Availability Codes
{Avail and/or
Dist
M
Special
»>h> fcV V.-
FIGURES
Number Page
1 Schematic of three models of membrane transport . 4
2 Schematic of construction of spiral wound membrane module . . 8
3 Solute concentration profile in a spiral wound membrane
module . ...... . . 9
4 Schematic of extraction procedure . .... 17
5 Schematic of bench scale test cell . 23
6 Schematic of the reverse osmosis test stand . 24
7 Stability of pH over time for water with 35,000 mg/1 NaCl . . 27
8 Stability of pH over time for water with 5,000 mg/1 NaCl. . . 28
9 Stability of pH over time for tap water . 29
10 Variation in offset for the nitrogen-phosphorous detector . . 31
11 Effect of intentional change in offset on the response of the
nitrogen-phosphorous detector . . . 32
12 Selected calibration curve for expected concentrations of
DIMP in acetone . . 40
13 Calibration curve for low concentrations of DIMP in acetone . 42
14 Example calibration curves of DIMP in acetone prepared on
different days . 43
15 Stability of DIMP in acetone at 25°C stored in the light. . 46
16 Stability of DIMP in acetone at 25°C stored in the dark . . 47
17 Stability of DIMP in acetone at 4°C stored in the dark . . 48
18 logarithm of the X DIMP remaining in dechlorinated tap
water - JHU . 58
19 Logarithm of the X DIMP remaining in brackish (5000 mg/1
NaCl) water - JHU . 59
FIGURES (cont'd)
Number Page
20 Logarithm of the Z DIMP remaining in dechlorinated tap water
- CRDC . - 64
21 Logarithm of the Z DIMP remaining in brackish (5000 mg/1 NaCl)
water - CRDC . . 65
22 Percent DIMP removal by reverse osmosis on expanded scale
versus time for all runs .... .
66
TABLES
Number Page
1 20 percent range of selected agent properties . 12
2 Candidate nerve agent simulants . .......... 14
3 Response of the n-p detector with different collectors for a
range of DIMP concentrations . 34
4 Response to 0.1 mg/1 DIMP in acetone from day to day . 35
5 Response of the n-p detector for replicate injections
of DIMP in acetone . 37
6 Response of the n-p detector for six replicate extractions
of a product sample and a feed sample . . 39
7 Recovery of DIMP from aqueous solutions . . 44
8 Stability of DIMP in dechlorinated tap water . 50
9 Stability of DIMP in dechlorinated tap water with 5000 mg/1
NaCl . 51
10 Rejection of DIMP by bench scale test cells . 52
11 Rejection of DIMP in dechlorinated tap water by reverse
osmosis in the Johns Hopkins University test stand (trial
1 ) membrane . 54
12 Rejection of DIMP in dechlorinated tap water by reverse
osmosis in the Johns Hopkins University test stand (trial
2) membrane . 55
13 Rejection of DIMP in brackish water (5000 mg/1) by reverse
osmosis in the Johns Hopkins University test stand
(trial 1) . 56
14 Rejection of DIMP in brackish water (5000 og/1) by reverse
osmosis in the Johns Hopkins University test stand
(trial 2) . 57
15 Rejection of DIMP in dechlorinated tap water by reverse
osmosis in the CRDC test stand (trial 1) . 60
TABLES (cont'd)
Number Page
16 Rejection of DIMP in dechlorinated tap water by reverse
osmosis in the CRDC test stand (trial 2) . 61
17 Rejection of DIMP in brackish water (5000 mg/NaCl) by reverse
osmosis in the CRDC test stand . 62
18 Selected properties for candidate nerve agent simulants ... 67
INTRODUCTION
As a resulc of ten years of extensive research and development, _£he
US Army has developed a trailer-mounted reverse osmosis water purification
unit (ROPU) which effectively treats brackish water, sea water, and
chemically contaminated freshwater. The treatment processes consist of
high rate filtration followed by the reverse osmosis system. Under the
present concept there will be two units: one will produce 600 gallons per
hour and the other 3,000 gallons per hour of potable water.
\
The smaller unit was designed to operate for 20 hours a day at a
production rate equivalent to 600 gallons per hour on freshwater and 40Q
#
gallons per hour on sea water. A climatic requirement to operate at
temperatures ranging between 1.6°C and 40°C (35°F and 105°F) at
relative humidities as high as 90 percent was established to provide a
world wide operational capability. The water quality standards which the
product water had to meet^wert- established by the US Army Surgeon General.
. . . — " ”
"‘The use of simulants for chemical warfare agents and the search for
better simulants have greatly escalated because of the ban on open-air
testing of agentsT' No compound can exactly match all the properties of an
agent and yet be non-toxic because of the interrelationship between
toxicity and chemical structure. The structural features that determine
toxicity may also uniquely determine the chemical and physical properties
of the compound. The properties to be matched in any application are those
that determine the specific parameter under investigation (dissemination,
decontamination, detection, removal from water, etc.).
The specific mechanisms of the rejection of chemical compounds by
reverse osmosis have not been firmly established. Selection of a simulant
for this process must therefore be based on empirical investigations. The
removal rates of chemical compounds is partly a function of the
configuration and composition of the specific membrane employed. To
evaluate all membranes for all possible chemical contaminants of water
wo*»ld be an excessively expensive and laborious task. A test system of
indicator compounds would prove cost-effective as a preliminary evaluation
before extensive testing is undertaken. A lower initial testing cost
should also expand the competition of suppliers, possibly producing an
overall reduction in unit cost and a more effective membrane.
-2-
LITERATURE REVIEW
REVERSE OSMOSIS
Reverse osmosis (RO) is a membrane process in which the input water is
pressurized to a value above the osmotic pressure. Pure water passes
through the membrane leaving most of the soluble salts behind. At the same
time, a large part of particulate matter, including microorganisms and
suspended colloids, is removed.
The mechanisms by which RO membranes allow the transport of certain
solutes are still a matter of conjecture. Various theories have been
i
proposed, but as yet, no one theory has enjoyed universal acceptance.
Three of the more common approaches to membrane transport are shown in
Figure 1 (Blais 1977) representing the solution-diffusion (Panel A), the
sieve transport (Panel B) and the preferential sorption (Panel C) models.
The solution diffusion model (Lonsdale et al., 1965) envisions a
transport corridor in the interstices between the molecules composing the
"thin film" rejection area of the membrane. A size estimate of these
0
species places it in the range of 6-20 A for membranes with high salt
rejections, which is of the same order of magnitude as intermolecular
distances in swollen polymers (Blais 1977). In this mode, actual passage of
solute and solvent first requires a dissolving into the membrane followed
by diffusion through the rejection layer. A modified form of this theory
includes imperfections in the casting process to allow for some pore
transport (Pusch 1977; Sherwood etal . , 1967).
3-
Passage o£ water through the membrane would be governed by diffusive
transport according to the following equation, which relates to permeate
quantity:
F - K^Pa - Po) U>
where :
F * Product (permeate) water flux in gal/(6q ft of membrane area)
( day )
Kj * Constant in gal/(sq ft) (day) (psi)
Pa ■ Applied pressure in psi
Po * Osmotic pressure in psi
«
According to equation (l), no product water is produced when the applied
pressure is less than the osmotic pressure. However, above the osmotic
pressure, the more the pressure, the more product water. Seawater, for
example, has an osmotic pressure of approximately 350 psi and would require
pressure greater than 350 psi to yield permeate.
Permeate quality would be governed by equation (2):
S - K2 (Cr - Cp) (2)
where :
S “ Salt flux in grams/sq ft of membrane area/day
K2 » Constant in gal/(sq ft) (day)
Cr ■ Concentration of salt in raw water in grams/gal
Cp ■ Concentration of salt in product (permeate) water in grams/gal
The constants Kj and K2 depend on a variety of factors including
temperature, viscosity, electrical resistance, diffusion and partition
coefficients as jell as membrane potential. As yet, the specific
interrelationships and mechanisms are not clearly understood. (Spiegler and
Lavial, 1980; Lindsten, 1972)
The sieve transport model (Banks and Sharpies, 1964) detailed a
non-interactive method cf membrane rejection based on steric exclusion. In
the sieve approach, it is the membrane matrix with its associated pore
structure that governs the rejection of solute and solvent. As shown in
Figure 1 (Panel B), the distribution of pore sizes allows for varying
rejection between solute and solvent. Compounds such as phenol, however,
which would be rejected at higher rates than sodium and chloride ions based
on their size, tend to penetrate membranes at much higher rates. This type
of information supports the premise that transport processes also depend on
membrane chemistry and its interaction with the solute. The steric
parameters of a molecule are important for larger species, for whatever the
model, there must be a physical space for movement.
The preferential sorption model (Sourirajan, 1970) evaluates three
molecular parameters as determining factors of solute rejection by RO
membranes: molecular size and the molecule's polar and nonpolar
characteristics. The steric factor determines passage based on the bulk of
a molecule in relation to the size of the transport corridor. The polar
and nonpolar parameters attempt to quantify the chemical interaction
between the solute and the membrane. With membranes thought to have
specific polar and nonpolar regions (Chian et al., 1975) these two
parameters determine both aqueous type reactions as well as hydrophobic
interactions. This model theorizes a sorption at the membrane-solution
interface that would show marked differences in a concentration profile
across the membrane solution junction.
There are obviously common elements among these theories. For certain
membrane configurations all may apply but one may be most appropriate.
The diffus; -i coefficient used in the Lonsdale model is determined by the
interaction between solute and membrane, interaction that the Scurirajan
model attempts to quantify with certain solute characteristics. . The models
account for unexplained rejections by including factors for membrane
imperfections.
The configuration of the membrane will have a direct bearing on the
relationship between the rate of rejection and the amount of potable water-
produced. The U.S. Army has adopted membranes in the spiral wound
configuration as shown in Figure 2.
The pressurized water passing along the length of a RO element is
continuously "dewatered." Therefore, the feed becomes more concentrated
and the quality of the product continually deteriorates through the system
as more salt migrates through the membrane and less water passes through to
dilute it. At the end of the system, the concentrated feed is discharged
as the waste stream. A graphic representation of this cross flow process
is shown in Figure 3. Alleviation of the concentration problem is
achievable by operation at a low "water recovery," i.e., maintaining a high
feed rate so that the product output is a small fraction of the feed.
However, when a highly concentrated waste stream is desired, such as when
processing wastewater, low "water recovery” is undesirable. Also, low
"water recovery" results in a comparatively high energy requirement. A
drop in flux as a function of time is a commonly encountered occurrence.
It is believed that this phenomenon is a direct result of increased flow
resistance due to any or all of the following reasons: (a) compaction of
the porous membrane substructure; (b) release of tiny pinpoints of rtir or
dissolved gas on and in the membrane; (c) electrical charge buildup due to
streaming potential; (d) deposition of ra / water turbidity (including
microorganisms, clay, organic turbidity, suspended iron and manganese, and
colloidal particles); (e) deposition of scale due to the precipitation of
sparingly soluble dissolved salts; (f) growth of biological films; and (g)
accumulation of ions adjacent to the membrane surface, which is responsible
for "concentration polarization." (Nusbaum 1981, Lindsten, 1972)
SIMULANTS
The chemical agent data center at the Edgewood area of Aberdeen
Proving Ground was used to perform a search of the literature based on
chemical properties that would be related to membrane rejection (Coon £t
al . 1982). The chemical and physical properties used and their units
were:
1. Molecular weight
2. Vapor pressure, mm Hg at 25°C
3. Molecular diffusion coefficient, cm^/sec at 25°C
4. Solubility in water, gm/1 at 25°C
3 1/2
5. Hildebrand solubility parameter, (cal/cm )
-10-
The additional considerations and general guidelines outlined in the
contract proposal listed below were also considered during the selection of
simulants.
1. Stability vn aqueous solutions
2. Similar molecular structure
3. Simple analytical methods
4. Reasonable detectable limits
5. Non-toxic characteristics
6. Past use as indicator compounds
The criteria used for the selection of these compounds were that
values of the specific chemical properties fall within a plus or minus 10
percent range around the value for a volatile nerve agent (GB) and a
non-volatile nerve agent (VX). The range wes expanded to plus or minus 20
percent when limited output was generated for various combinations of these
properties. The 20 percent bracket is shown in Table 1 for both GB and VX.
The data base checked the Technical Library at the Chemical Research and
Development Command (CRDC) and various Department of Defense literature
surveys to identify the approximately 800 references contained in its
files. Typical of the material contained in this data base are two
Department of Defense Publications. Arthur D. Little, Inc. (1982) reported
chemical properties, toxicological data, and analytical procedures for
various simulants. Bagley et al. (1977) at Dugway Proving Ground,
reviewed simulants and compared chemical properties of simulants to agents.
-11-
TABLE 1. A 20 PERCENT RANGE OF SELECTED PROPERTIES
OF A VOLATILE NERVE AGENT (GB) AND A NON-VOLATILE NERVE
AGENT (VX)
Property
AGENT
Molecular Weight
GB
112.0 - 168.1
VX
213.9 - 320.9
Vapor Pressure
mm Hg at 25°C
2.32 - 3.48
0.00050 - 0.00074
Molecular Diffusion Coefficient
cm /sec at 25 °C
0.049 - 0.073
0.028 - 0.041
Solubility in Water
gm/1 at 25°C
Miscible
0.05
Hildebrand^Sj^bility Parameter
7.24 - 10.85
6.40 - 9.60
The candidate simulants developed from the literature are listed in
Table 2. While a literature search was a useful tool, a careful evaluation
of the results must be performed. Highlighting the requirement of this
follow up was that one of the compounds selected by the search, diethyl
pthalate was insoluble in water even though solubility was one of the
parameters to be matched on. Diisopropyl methyl phosphonate (JIMP) was
chosen for initial study.
-13-
TABLE 2. CANDIDATE NERVE AGENT SIMULANTS
1. Bis (2-ethyl hexyl) nhosphonate
2. Diethyl glycol dimethyl ether
3. Diethyl phosphonate
4. Diethyl phthalate
5. Diethyl sebacate
6. Diethyl sulfite
7. Diisopropyl methyl phosphonate (DIMP)
8. Dimethoxy methyl phosphonate (DMMP)
9. Ethyl dimethyl phosphite
METHODS
PREPARATION AND ANALYTICAL METHODS
Sample Preparation, Handling and Storage
Glassware Preparation —
All glassware was washed with detergent, rinsed with distilled water
and maintained at 400°C for oue hour to remove organics.
Preparation of Simulant Standards —
DIMP standards in water and acetone were prepared from 1,000 mg/1
stock solutions. An aliquot of 0.200 ml DIMP was added to 200 ml of
t
solvent in a* volumetric flask. DIMP in water standards of 10.0, 1.00 and.
0.100 mg/1 were made up in 2,000 ml volumetric flasks. DIMP in acetone
standards of 20.0, 10.0, 5.00, 1.00 and 0.100 mg/1 were made up in 100 ml
volumetric flasks. A 0.050 mg/1 standard and a 0.025 mg/1 standard were
made by diluting the 0.100 mg/1 standard.
Aqueous Samples—
Aqueous samples were collected and stored in 150 ml screw cap bottles.
Caps were lined with aluminum foil which had been heated at 400°C for at
least one hour to remove organics. Samples were stored at 4°C.
DIMP in Acetone Samples —
DIMP in acetone samples included standard solutions and extractions of
aqueous samples. Standard solutions were stored in 10 ml serum bottles
with teflon faced septa. Sample extractions resulted in a 2 ml volume of
DIMP in acetone. These were transferred to 1.8 ml screw cap vials with
open top caps and teflon faced septa. All septa were scrubbed with acetone
before placement on vials.
•-* \." •/ Ca
r '
M
«
r"
4
4
Sample Extraction Procedure
A schematic illustrating the extraction procedure appears in Figure 4.
Aqueous DIMP samples were poured through silica gel columns (Baker 10 3PE
disposable reversed-phase extraction columns, octylsilane bonded silica
gel) under low vacuum to collect and concentrate the DIMP on the sorbent
bed. The retained DIMP was then eluted with acetone into a volumetric
flask. A detailed description is given below. Multiple extractions were
conducted simultaneously on a vacuum manifold.
Sample Extraction Steps
1. Column Preparation
i
1.1 Place column on manifold (one column per extraction).
1.2 Fill column with HPLC grade methanol.
1.3 Turn on vacuum and draw methanol through column.
1.4 Turn off vacuum immediately to avoid drying the column.
2. Extraction
2.1 Using a volumetric pipet, apply desired volume of sample to the
column and draw through with vacuum.
3. Elution
3.1 Remove column from manifold and place on a volumetric flask of the
appropriate size.
3.2 Using a volumetric pipet, add chromatography grade acetone to the
column and force it through with compressed air.
3.3 Remove the column from the flask and adjust the volume to the mark
with acetone.
PREPARATION _ _ EXTRACTION _ ■ ELUTION
Figure 4. Schematic of extraction procedure with Baker 10 SPE disposable
reversed ~ phase extraction columns. Sorbent bed: octylsilane
bonded silica gel.
4. Storage
4.1 Transfer extracted sample to an appropriately sized vial and seal
with septum cap.
Gas Chromatographic Analysis
Equipment —
Gas chromatographic analysis was performed with a Hewlett-Packard
5830A gas chromatograph equipped with a carbowax column and a
nitrogen-phosphorous detector (N-P detector). The specific chromatographic
conditions are listed below.
Chromatographic Conditions
1. Detector: Hewlett-Packard nitrogen-phosphorous flame ionization
detector (HP 18847A/8A) with long wide bore jet.
2. Column: 10Z Carbowax 20M on 80/100 chromosorb W-HP. Type: glass.
Length: 2 meters. Outside diameter: 1/4 inch. Inside diameter: 2 mm.
3. Carrier gas: Helium, 99.995Z minimum purity with inline molecular
sieve drying.
4. Support gases: Hydrogen, 99.995Z minimum purity and 'dry' quality air,
both with inline molecular sieve traps.
5. Injection port: On column injection. Septum: Thermogreen LB-1
(Supelco 2-0659).
Operating conditions —
All results were obtained under the following operating conditions:
Operating Conditions
1. Gas flows, measured with soap bubble flow meter: helium 30 ml/min,
hydrogen 3 ml/min, air 60 ml/min;
-18-
2. Oven temperature: 165°C, isothermal;
3. Injection port temperature: 220°C;
4. Detector temperature: 300°C.
5. Offset: set at approximately 100 mm at the start of each series of
analyses.
Injections were performed manually by the solvent flush technique using
acetone as the solvent. Injection volumes were approximately 2
microliters. In order to minimize the effects of detector sensitivity
variations, samples were grouped according to approximate DIMP
concentration and the groups were analyzed in order of increasing
concentration. A DIMP in acetone standard of approximately the same
concentration was injected with each set of samples in a fixed sequence.
The sequence was repeated for three to five replicate injections. Any
variations in sensitivity thus did not exert an inordinate influence on any
one sample.
For each injection, the peak area and injection volume were recorded.
The response was then calculated as peak area per microliter injected and
averaged for replicate injections. Calculation of the corresponding DIMP
concentration was based on a least squares line of best fit for calibration
data. The calibration curve was adjusted for day to day variations in
sensitivity on the basis of responses to standards analyzed at the same
time as the samples. Reported concentrations were adjusted for the density
and purity of DIMP.
EXPERIMENTAL PROTOCOL
Stability of pH
The stability of the pH of tap water from the Edgewood area of Aberdeen
Proving Ground, Building 1956, was evaluated over a thirty hour period
under various conditions of pH and salt concentration. A five gallon
sample of filtered tap water was collected in a carboy which had been
washed and rinsed three times with triple distilled water. The sample was
dechlorinated by aeration for 48 hours followed by the addition of sodium
thiosulfate sufficient to remove the remaining residual. The absence of
chlorine residual was confirmed by regular determinations with
«
N,N-diethyl-p-phenylene-diamine (DPD) according to Standard Methods for
the Examination of Water and Wastewater (1981).
Aliquots of the dechlorinated tap water and salt solutions containing .
35,000 and 5,000 mg/1 NaCl were dispensed in brown glass bottles. For each
trial, the pH was adjusted to 5, 7 or 9 with 0.1 N solutions of sulfuric
acid or sodium hydroxide. The bottles were stored at room temperature.
The pH waa determined electrometrically with a Beckman Zeromatic Model
11 pH meter according to Standard Methods for the Examination of Water
and Wastewater (1980). Samples were agitated with a magnetic stirrer
during the measurement.
Evaluation of Extraction Procedure
Efficiency —
The efficiency of the extraction procedure was evaluated for 10.0, 1.00
and 0.100 mg/1 DIMP ic dechlorinated tap water and brackish water (5,000
mg/1 NaCl in dechlorinated tap water) solutions at pH values of 5, 7, and
9. These solutions were prepared as described above. Dechlorination of the
tap water and pH adjustment were described in the previous section. The
efficiency was determined by comparing the response (area per microliter)
for an extracted aqueous sample with that for a DIMP in acetone standard of
the same nominal concentration.
Variability —
The variability of the extraction procedure was assessed by performing
six replicate extractions each of a feed water sample and a product water
sample from the Johns Hopkins University reverse osmosis test stand. Five
replicate injections of each extraction were made.
«
Stability of ‘DIMP in Acetone
To determine the stability of DIMP in acetone, solutions containing
10.0, 1.00 and 0.100 mg/1 DIMP were prepared as described above. Aliquots
of each were stored under three different conditions: 1) at 25°C, with
normal diurnal variations in light; 2) at 25°C, in the dark; and 3) at
4°C, in the dark. On each day that the samples were analyzed, the gas
chromatograph was calibrated with fresh standards. Analyses were performed
on days 0, 1, 6 and 20.
Stability of DIMP in Aqueous Solution
The stability of DIMP in aqueous solution was determined for 10.0, 1.00
and 0.100 mg/1 DIMP in dechlorinated tap water and brackish water at pH
values of 5, 7 and 9. Extractions were performed on days 0 and 14. Having
established the stability of DIMP in acetone, the day 0 extractions were
stored and analyzed at the same time as the day 14 extractions. The
percent change in response between the two* provided an indication of the
stability.
Rejection of DIMP by Reverse Osmosis
Bench Scale Test' Cell —
The RO test cells were assembled with 47 mm diameter pieces of UOP
TFC-801 membrane material. The flow system for these smaller units was a
once through system. The feed solutions were prepared, as needed,
in 100 liter containers with tap water filtered through a 10 micron cotton
filter for removal of rust and scale. The tap water was dechlorinated with
the addition of 15 mg/1 of sodium thiosulfate. Dechlorination was
confirmed by the determination of chlorine concentration by the DPD
technique as described above. Required levels of NaCl were added to the
100 liter batch and mixed for one half hour by a chemical mixer. A
schematic of the bench scale test cell is shown in Figure 5.
The test cell system incorporated high pressure pumps capable of
operating the system at pressures up to 800 psi. Complete mixing was
obtained by rotating magnetic stirrers at 400 rpm. On exiting the cells,
the waste stream was returned to atmospheric pressure through a pressure
relief valve.
Four Inch Module Test Stand —
The four inch module reverse osmosis test stand was supplied by the
Ft. Belvoir Research and Development Command. The evaluation of rejection
was performed using a recirculation mode of operation shown in Figure 6.
The reservoir was a 500 gallon water storage tank. The energy dissipation
on return to atmosphere caused an increase of temperature of the feed water
ever the course of a run, necessitating the recording and inclusion of
temperature as an additional variable. The tank was cleaned when received
-22-
from Fort Belvoir Research and Development Center and after each sample
run. The water used for each run was tap water filtered through a 10
micron cotton filter wound cartridge (Filterite # C10A10A). The tap water
was dechlorinated and the residual was measured as described in the
previous oection. Upon addition of the solute and required level of sodium
chloride to the feed water, two Lightning heavy-duty stirrers operating at
1750 rpm were engaged. The mixing action coupled with the flow from the
high pressure discharge line provided the blending of the return of product
and waste streams. The return product line was fixed eighteen inches above
the surface of the tank to avoid solute contamination of the product
4
sample. At the start of a test run, the feed pump operated for
approximately 1 minute before the high pressure pump was engaged. The
system temperatures could not be controlled but were measured to ensure
that the membrane was not exposed to temperatures greater than 34°C.
Three 4 inch UOP spiral wound elements were provided, one with previous use
and two new.
The module with previous operational use was employed for tests to
bring the system on line and for personnel training. Of the two remaining
modules, one was flawed in some manner as shown by very high product water
flow rate (in excess of three gallons per minute). The only remaining
module proved satisfactory and w?* used for all testing.
RESULTS
STABILITY OF pH
Three samples of dechlorinated tap water containing 35,000 mg/1 NaCl,
5,000 mg/1 NaCl and no salt were tested for pH stability over time at pH
values of 5, 7 and 9. Plots of pH versus time are shown in Figures 7, 8,
and 9. The samples were kept at room temperrture and sample temperature
was recorded when pH determinations were made. The temperatures ranged
from 23 to 25°C. Five pH measurements were made over 24 hours. Although
small fluctuations occurred, no trends were observed over this period.
ANALYSIS OF DIMP
Characteristics of the Nitrogen-Phosphorous (N-P) Detector
Tne level of sensitivity of the N-P detector can be varied by applying
different voltages to the ceramic bead in the collector. The offset can be
measured from a trace on the chart and provides a measure of this
sensitivity. It is desirable to maintain a constant offset over the course
of an analysis in order to avoid distortion of the results due to
variations in sensitivity. The manufacturer's literature on this detector
indicated that it should be stable over an eight hour period, but that
variations from day to day and from one collector to another can be
expected. No quantitative description of the relationship between offset
variations and sensitivity were provided by the manufacturer however.
Prior to analysis of samples in this study, an attempt was made to
characterize the variation in offset with time and the effect of changing
offset on the detector response. In the course of the study, variations in
-26-
detector response independent of offset were observed and considerable
differences in behavior between two collectors were noted.
Offset and Detector Response —
To observe the variation in offset over time, the offset was monitored
continuously for a twelve hour period without injecting any samples.
Smooth peaks and valleys were seen in the resulting trace. Figure 10 shows
the offset at high and low points over the period starting with the first
measured value. The maximum value was 90.5 mm at 0.42 hr and the minimum
was 66.5 mm at 11.53 hr, a difference of 27%. Overall, the plot showed a
downward trend with irregular fluctuations. ‘
In order to assess the significance of offset variations with respect
to the response of the detector, three replicate injections of 1.00 mg/1
DIMP in acetone were made at offset values ranging from 74.0 mm to 123.0
mm. Afuer each injection, the offset was higher than it had been before
the injection. Time was allowed for the offset to restabilize at the
original setting before the next injection. At offsets over 100 mm, this
was not always possible to accomplish in a reasonable period. A plot of
the detector response (area per microliter) versus offset is shown in
Figure 11 with the least squares line of best fit. The response varied
from about 21,000 to 25,000, that is, by 193! for a 66% increase in offset
over the range tested.
Variation of Response Independent of Offset —
The variation in the response is the result of numerous factors other
than the offset. These include irregularities in gas flows, uncertainty of
of injection volume, fluctuations in temperature at the injection port, in
-30-
/■*» O <-s
BWH
a g
O B
•rl •• M
UBS
O « A
« u u
•n B O
CUB
■H Tt
a
OJ Q
rl rl O
O. Sx Q
8 m
3 SO
03 • f-t
03
O O g
C oj |
w iH 3
bOr-l
4> O
4S°
44
U2 I W
u oj
UN U
» 9
K 4J
H «
O B u
4J 0)
UNO.
2 .8
« ftj H
T3 2
I
03 a •
9 H
O JO Tt
M U CO
0)0 0
sc js ® a
3 D. O *rl
2 0) a B
3- O O ^
A U H
ui 0--C 0
1 i ° _ ‘
m C O
H- 4) O sO
60 O
o ■— « »
M ^ C
*j O 0)
i-t 00 60
a _ o
a) o -
5 a r
UN B
O -rj
^ 5-5
03 O
SM M CO
o <3 *
e m i
O «rt
*"4 fH
a a
o JS
-r4 »•
•u fl C
<8 1 *4
•HS0
U H>.
BOH
>OB
Oh.li.IAUh- . XX
injection port 220°C, detector 300°C
2B000-
<SIU< 0. UJ E IHUIO JHhIUI
0 vm
vi O
u
o
4J
u
ai
jj
95
•o
i
«
Vi
95
Q.
a) « e
U 0- 41
a s h
0) i
■o 3
a
3
O
Vi
O
X
a.
m
o
x
a 9)
i x
C v>
V
60 CO
O i-l
Vi
vi oi
•H c
C H
0)
X -O
vi *H
r-l
ivi o
o CD
g) 91
se
o
D.
01 •
aj 9J
M C
CO
91
u S
d
3 O
cr
05 ^
m
VI 05
05
«
<u
WAV
Vi iH
O <8
co -
o c
B -H
o e
Vi
X I-l
95 U S
U O O
O vO
95 -w •
5- • U
a © c o
00 05 O
a boo
95 B O CO
a o u
-H 0 U
-I J5 >v O
O X VI
•O CM O
95 C 95
X X iH V)
C0 d e 95
d ? -a
•O OH
X B «
95 5-1 a
X
CO CO O
H
O • o
« CM
M E CM
•
O 3
3
r-4 «H iJ
3
r-; Vl
95 O
•
•• X Cl
O
1 o a
1
OHO
-32-
the oven and at the detector, condition of the column and age of the
collector. In addition, variations due to the extraction procedure can be
expected.
Two collectors were used in the course of the study; the first was
expended after three months of use. The observations reported above in
relation to the offset were obtained with the first collector. The same
type of variation may be expected with any collector. No effort was made
to characterize the second so extensively, but important differences were
seen. The first collector gave higher responses than the second. Typical
responses to DIMP in acetone are shown in Table 3. The second appeared to'
be more stable since there was considerably less change in offset following
injections and a more rapid recovery to pre-injection levels.
Table 4 shows the initial offset values and the response (area per
microliter) for the analysis of 0.100 mg/1 DIMP in acetone performed on
different days. Although the offset values were all approximately the
same, the average responses on different days showed considerable
variation. The minimum and maximum average responses were l,4i0 and 1,690
respectively, corresponding to offsets of 101.5 and 100.0 mm. The overall
mean response was 1,550 with standard deviation of 111.
The variation between injections on a single day {or a series of DIMP
concentrations was observed by performing ten to twenty replicate
injections for each concentration. Table 5 shows the response (area per
microliter) for DIMP concentrations from 0.025 to 20.0 mg/1. The
coefficient of variation was highest for the U.025 mg/1 concentration at
about 7%. For concentrations of 0.050, 0.100 and 20.0 mg/1, the
•-isii
mi
.* .* j
!v.-v>
• • * * «
. % * « * «
;w>v
•/" • / s'" . v/_#. T.“/:
. • .** *• .
/• • . .
*« T VtVl.
’• ;• .*• .*• % ,
TABLE 3. RESPONSE OF THE N-P DETECTOR WITH DIFFERENT COLLECTORS
FOR A RANGE OF DIMP CONCENTRATIONS. COLUMN: 102 CARBOWAX
20 M ON 80/100 CHROMOSORB V-HP, 2 M x 2 MM ID GLASS.
FLOW RATES: 30 ML /MIN HELIUM, 3 ML/MIN HYDROGEN, 60 ML/MIN
AIR. TEMPERATURES: COLUMN 165°C (ISOTHERMAL), INJECTION
PORT 220°C, DETECTOR 300°C
Response (area/microliter)
DIMP Concentration (mg/1) 0.100 1.00 10.0
1
Collector
2
2,200
1,500
23,000
16,000
220,000
163,000
TABLE 4. RESPONSE TO 0.100 MG/L D1MP IN ACETONE FROM DAY TO
DAY WITH INITIAL OFFSET APPROXIMATELY 100 MM. COLUMN: 10Z
CARBOWAX 20 M ON 80/100 CHROMOSORB W-HP, 2 M x 2 MM ID GLASS. FLOW
RATES: 30 ML/MIN HELIUM, 3 ML/MIN HYDROGEN, 60 ML/MIN AIR.
TEMPERATURES: COLUMN 165°C (ISOTHERMAL), INJECTION PORT 220°C,
DETECTOR 300°C
Date
Initial Offset
(cm)
Response
(area/microliter)
Average Response
(area/microliter)
2/2/84
101.5
1,465
1,412
1,345
1,407
2/3/84 '
8:50 to 9:15 am
99.5
1,545
1,434
1,449
1,476
11:10 to 12:15
1,418
1,426
1,443
1,391
1,420
2/9/84
102.0
1,634
1,505
1,584
1,538
• 1,540
1,560
2/23/84
109.0
1,666
1,701
1,672
1,591
1,607
1,647
2/25/84
100.0
1,713
1,649
1,672
1,715
1,687
2/29/84
103.0
1,686
1,725
1,564
1,618
TABLE 4. continued
Date
Initial Offset
(nn)
Response
(area/microliter)
Average Response
( area /microliter)
3/1/84
101.5
1,634
1,734
1,667
1,549
1,696
1,669
1,658
3/8/84
102.0
1,407
1,416
1,430
1,483
1,524
1,481
1,495
1,377
1,388
1,407
1,441
Mean
Standard Deviation
1,550
111
TABLE 5. RESPONSE OF THE N-P DETECTOR FOR REPLICATE INJECTIONS
OF DIMP IN ACETONE. COLUMN: 102 CARBOWAX 20 M ON
80/100 CHROMOSORB W-HP, 2 M x 2 MM ID GLASS. FLOW RATES: 30
ML/MXN HELIUM, 3 ML /MIN HYDROGEN, 60 ML /MIN AIR. TEMPERATURES:
COLUMN 165°C (ISOTHERMAL), INJECTION PORT 220°C, DETECTOR
300 C
Concentration
Injection
0.025
0.050
0.100
20.0
1
255
638
1473
338842
2
284
657
1397
312105
3
246
633
1419
319282
4
246
641
1407
335282
5
231
689
1416
321179
6
246
624
1430
339800
7
263
624
1483
322513
8
239
638
1524
321000
9
287
640
1481
338316
10
260
637
1495
329100
11
ND
ND
1377
347282
12
ND
ND
1388
341474
13
ND
ND
1407
327700
14
ND
ND
ND
330051
15
ND
ND
ND
338256
16
ND
ND
ND
340718
17
ND
ND
ND
345744
18
ND
ND
ND
327333
19
ND
ND
ND
333053
20
ND
ND
ND
319590
Mean
255
642
1438
331481
Standard Deviation
18.3
18.9
47.0
9826
Coefficient of Variation (2)
7.18
2.94
3.27
2.96
coefficients of variation were substantially the same at about 32. The
mean response for 0.025 mg/1 DIMP was 255. Greater variation is to be
expected at such low levels.
The assay of DIMP in water by gas chromatography requires the
extraction of DIMP from the water by adsorption on treated silica gel and
elution with acetone. To assess the variability associated with th*
extraction procedure, six extractions each of a single product water and a
single feed water sample were made. Five replicate injections of each
extraction were performed. The data appear in Table 6. Mean values for
the response (area per microliter) obtained with the product water
extractions ranged from 323 to 419. Standard deviations ranged from 14 to
91. The mean of all responses was 376 and the overall standard deviation
was 55. The coefficient of variation was 14.62. For the feed water
extractions, mean responses ranged from 292,000 to 306,000 with standard
deviations from 5,100 to 9,210. The mean of all responses was 297,000, and
the overall standard deviation was 7,810. The coefficent of variation was
2.62.
Calibration of Response to DIMP in Acetone Standards
Before the reverse osmosis testing began, it was anticipated that
concentrations of DIMP in the extractions would range from 0.10 to 20 mg/1.
Figure 12 presents a calibration curve over this range with collector #2
installed. The response over, the range of concentration tested was linear.
The least squares line of best fit is given by the equation:
Response = 18,900 x Concentration + 1,080
The correlation coefficient, r, was 0.997. It was subsequently found that
TABLE 6. RESPONSE OF THE N-P DETECTOR FOR SIX REPLICATE EXTRACTIONS OF A
PRODUCT SAMPLE AND A FEED SAMPLE. COLUMN: 102 CARBOWAX 20 M ON
80/100 CHROMOSORB W-HP, 2 M x 2 MM ID GLASS. FLOW RATES: 30 ML/MIN
HELIUM, 3 ML/MIN HYDROGEN, 60 ML/MIN AIR. TEMPERATURES: COLUMN 165°C
(ISOTHERMAL), INJECTION PORT 220 C, DETECTOR 300°C
Extraction Number
Product Water
1
2
3
4
5
6
444
398
405
348
423
332
476
460
377
356
335
318
554
389
367
365
324
295
373
344
378
352
408
301
0
296
409
403
384
351
352
Mean
419
400
386
361
368
323
Standard
91
41
17
14
44
27
Deviation
Overall Mean 376
Overall Standard 55
Deviation
Coefficient of 14.6
Variation (2)
Feed Water
306263
300895
302368
305526
299895
294579
297179
286500
285421
305892
298900
294684
296769
304051
279282
306923
293282
307316
295026
301026
294632
299210
284950
292211
292308
300821
296526
313250
290667
285684
Mean
297504
298658
291645
305760
291558
294894
Standard
Deviation
5256
6931
9210
5096
7687
7849
Overall Mean 297000
Overall Standard 7810
Deviation
Coefficient of 2.6
Variation (2)
400000
"0 *
O £
» JS 3
Cm x
X O X
X -a 0)
x x c
•o 0) o o
T-t X -* C T- (
X H X X
o E o
io •• *•*. y
•
s
H
-r->
y
ON
E
£
C
X
ON
3
•H
H
O'
H
o
•
O
CO
o
U
/•N
o
•
f— <
cw
Cl
1
s*
CO
c
•
CD
B
8)
o
u
44
a)
C
44
«r«
44
CD
a
‘ u
44
CO
X
o
c
w
44
o
eo
o
44
O
**
•H
«
)
CO
|
0
a
o
«H
N
•H
X X
44
P4
&«
44
44
a
2
y
o
<
0
M
o
•
m
a
y
y
a
vO
c
CD
*-4
m
C
CO
O
O
H
iH
c
00
g
(0
44
a
3
c
a x
« X
o
rl
X
M
o
*H
y
y
m
44
>4
•o
8
-J
Q
u
c
M
\
U
o
3
01
o
4J
y
O
CM
CD
z
c
X
a
F-- 5
r — —
: i-
L <
P- O (t
r — v—
t Z
f Ili
- 01 o
<U fl (0 X M
y a
C X X a X
o x y a
y t4 > Ol M
? 0)
~a y « o.
y « u cu s
x x c x y
y x n i h
y vm 5
a. y
x x o x •
tl » C h h
y y o x
v> x •o « a
o -HO
44 44 44 B C
o e o x
If)
y
o
U 60
w
>
y
y
X -40
|4
c
y x o
3
X
M
6 o
m
y
X
«/•>
o n
o>
o o
e
a
X SO >4
o
y
y o
X
u
4=
O - X
X
«
44
oo e y
CO
3
y y
n
u
0* 44
c to x
X
Q
c
o o y
X
V
K T3
01
X
44
•to
£ -a
a
00
y
O >% -
y a «4 <s x u
vo. o
-o rH y X C O
y u y x <n
x v 3 S cm
y x » o *'«.
y X V X r4 x
X C W B w
y n 4-i y o
C/J X r4 O t'f a
zhuico jhi-uc
v' O
Che product vaCer extracts gave results below this range. Figure 13 gives
a calibration curve for the range of 0.025 to 0.100 mg/1. The response was
again linear; the equation of this line was:
Response ■ 15,000 x Concentration - 143
with r ■ 0.9980. No attempt was made to force the curve through the origin
or extrapolate below the data, since it seemed likely that some threshold
amount of DIMP would be required to produce a response. Calculations of
concentrations on the basis of response utilized these equations.
Figure 14 shows four calibretion curves generated by a single set of
DIMP in acetone standards on four different days with collector #1
installed. As discussed above, variation was to be expected from day to
day. In order to relate the response of an unknown sample to a
concentration it was therefore necessary to incorporate calibration samples
in each analysis. Such variation may have accounted for the difference in
slope between the low range and high range calibration curves.
Evaluation of Extraction Procedure
The recovery efficiency of the extraction procedure was evaluated by
comparing responses to DIMP in acetone standards and to extracted DIMP in
water standards. The aqueous systems included brackish (5000 mg/1 NaCl)
and dechlorinated tap water at pH values of 5, 7 ana 9. Responses and
percent recoveries are given in Table 7. Overall, recoveries ranged from
86.0 to 100%. It appeared that recoveries were higher for 1.00 mg/1 than
for 10.0 mg/1 and that for the 10.0 mg/1 standards recovery increased with
increasing pH. The results must be treated with caution, however, since in
the course of analyzing the 10.0 mg/1 samples, a significant drop in the
-41-
1900
<KW< (LUIS XhUIO JHhUIT
24000
2 mo ID glass. Flow rates: 30 ml/min helium, 3 ml/min hydrogen, 6
t: column 165°C (isothermal), injection 'port 220°C, detector 300°C.
TABLE 7. RECOVERY OF DIMP FROM AQUEOUS SOLUTIONS WITH AND WITHOUT
5000 MG/L NaCl at pH 5, 7, AND 9. COLUMN: 102 CARBOWAX 20 M ON
80/100 CHROMOSORB W-HP, 2 M x 2 MM ID GLASS. FLOW RATES: 30 ML/MIN
HELIUM, 3 ML/MIN HYDROGEN, 60 ML/MIN AIR. TEMPERATURES: COLU^ 165°C
(ISOTHERMAL), INJECTION PORT 220 C, DETECTOR 300°C
Response
area per microliter (percent recovery)
Sample
1.00 mg/1
10.0 mg/1
DIMP in Acetone Standard 21,400 218,000
DIMP in Dechlorinated Tap Water
pH 5
pH 7
pH 9
21,400 (99.9)
20,900 (97.4)
21,000 (98.0)
187,000 (86.0) •
194,000 (89.2)
211,000 (96.9)
DIMP in Dechlorinated Tap Water with 5000 mg/1 NaCl
pH 5
pH 7
pH 9
21,000 (98.2)
21,600 (100)
20,200 (94.1)
193,000 (38.8)
200,000 (92.0)
209,000 (96.0)
-44-
/
-L
V v\
offset occurred. It was readjusted, but whether the sensitivity returned
to the same level as when the DXMP in acetone standards were analyzed was
uncertain. There was also the possibility that the extraction columns were
overloaded for the 10.0 mg/1 aqueous standards since relatively large
volumes of 15 and 25 ml were passed through. In extracting feed water
samples during the reverse osmosis tests, only 2 ml of sample were
extracted.
Variation between extractions was discussed above. As expected, the
relative variation was lower for the feed water extractions which yielded
responses nearly 1,000 times greater than those of the product water
r. •.
extractions.
STABILITY OF DIMP
In Acetone
The stability of DIMP in acetone was evaluated under three different
storage conditions over a period of twenty days. The conditions were 1)
room temperature (23-25°C), with normal diurnal variations in light, 2)
room temperature with no light and 3) 4°C with no light. For no light
storage, the vials were wrapped in foil and kept in the dark except during
analysis. Figures 15 through 17 present the results for the different
storage conditions for the three concentrations of DIMP tested: 0.100,
1.00 and 10.0 mg/1. No degradation in the response with time was observed.
One of the more useful implications of this result was that once extracted,
samples could be stored before analysis. Also, the same standards could be
kept and reused over long periods, obviating the burden of preparing fresh
standards for each analysis.
*, -. % •. *. *, «. V
• .• ,» v* „■
** . •*-. **. ***» ■* >
• . ■ .'s','.*,*.*. •
■» ” » “ • • » * .
0000*2
<u
<kuj< ctu-^aozuj ou. ow^-wukoc:
. 1
* C# o
s • o.
o a
M jo a e
V H rl O
M () H t(
BOBU
J3 O
^ M a)
• Ofin
° ^ S -S
-o ^
C •• r-j *
C8 C p ^
/-v.3 o«
E
w O 0)
•• X
H ra u
*>%. • a) o
M JJ AJ ffl
fi J2 « -H
m h ^
O *H
. H > O
o o
C I-H m
» ri h vO
M -H CO O
a u _
i-H a
a n a
o tv,
2e^
o o 3 ^
u u <o
CD .O
CJ T3 M C
« CD O -H •
O to a O
COO ~-~.o_
t! ftEHO
* o a o
Cm 4) l-c CO
2 -E O
H 08 O vD I-*
P CD O
» O » *J
\w o a «
o ti m ii v
COM
>, 0.0 O «
U E CO D T)
•rl 3 tl
HOC >> «
■cH O J3 O
m3 X O
« o s c o
D Cfl O *H N
COUN 0N
r. V*. r. jr.» « .
* r»>t v"jr» *"* ^
v.v.y.v.'». r.v. .*.
In Aqueous Solution
The stability of DIMP in aqueous solution was tested for brackish
water and dechlorinated tap water at pH values of 5, 7 and 9 over a 14 day
period. The tap water results appear in Table 8. The brackish water
results are presented in Table 9. Overall, average responses (area per
microliter) decreased from 1.5X to 19% between day 0 and day 14 for the
0.100 and 1.00 mg/1 DIMP concentrations. No consistent trends related to
pH, salt or concentration were seen. A significant portion of the
variation in these results was probably due to sources discussed above.
The responses for day 14 extractions of the 10.0 mg/1 DIMP samples were
0
higher than the day 0 values by 9X to 34. 4% for all but one sample which
showed a decrease of 1.62%. The same precautions in interpretation of the
results for the 10.0 mg/1 solutions discussed in relation to extraction
efficiency apply here as well. During the reverse osmosis trials, samples
were generally extracted within 24 hours and all were extracted within 48
hours.
REJECTION OF DIMP BY REVERSE OSMOSIS
Johns Hopkins University Trials
Preliminary trials of DIMP rejection by reverse osmosis were conducted
with the bench scale test cell. The results for both brackish (2,404 mg/1
NaCl) water and salt (33,540 mg/1 NaCl) water trials are presented in Table
10.' The salt water rejection was 75% with a corresponding 88.7% DIMP
rejection in the salt water test. For the brackish water test, the salt
rejectiou was 81.0% with a DIMP rejection of 98%.
TABLE 8. STABILITY OF DIMP IN DECHLORINATED TAP WATER AT 35°C
AT pH 5.0, 7.0, AND 9.0 OVER A 14 DAY PERIOD. COLUMN: 102 CARBOWAX
20 M ON 80/100 CHROMOSORB W-HP, 2 M x 2 MM ID GLASS. FLOW RATES: 30
ML /MIN HELIUM, 3 ML/MIN HYDROGEN, 60 ML/KIN AIR. TEMPERATURES: COLUMN
165°C (ISOTHERMAL), INJECTION PORT 220°C, DETECTOR 300°C
pH
Nominal DIMP
Concentration
(mg/1)
A.erage
(area per
Day 0
Response
microliter)
Day 14
Z Change
5
0.100
2820
2630
-6.74
1.00
20400
19000
-6.91
10.00
324000
385000
+18.6
7
0.100
2750
2260
«
-17.9
1.00
20900
16900
-19.3
10.0
308706
349673
+13.3
9
0.100
2580
2140
-16.9
1.00
20500
19300
-5.77
10.00
323000
352000
+8.96
-50-
TABLE 9. STABILITY OF DIMP IN DECHLORINATED TAP WATER WITH 5000 mg/1 NaCl
AT 35°C AT pH 5.0, 7.0, AND 9.0 OVER A 14 DAY PERIOD COLUMN: 102
CARBOWAX 20 M ON 80/100 CHROMOSORB W-«P, 2 M x 2 MM ID GLASS. FLOW RATES:
30 ML/MIN HELIUM, 3 ML/MIN HYDROGEN, oO ML /MIN AIR. TEMPERATURES: COLUMN
165°C (ISOTHERMAL), INJECTION PORT 220 C, DETECTOR 300°C
pH
Nominal DIMP
Concentration
(mg/1)
Average
(area per
Day 0
Response
microliter)
Day 14
2 Change
5
0.100
3100
2670
-13.9
1.00
19700
19000
-3.48
10.0
307000
302000
-1.62
7
0.100
2910
2470
-15.2
1.00
19700
19400
-1.53
10.0
301000
350000
+16.5
9
0.100
2970
2710
-8.98
1.00
19800
19200
-3.31
10.0
325000
437000
+34.4
-51-
TABLE 10. REJECTION OF DIMP AND SALT IN BRACKISH AND SALT WATER BY BENCH
SCALE REVERSE OSMOSIS TEST CELLS
Sale Water Brackish Water
Temp
pH
Flow
Pressure
24°C + 1°C
7.5 + 0.2
SO ml7min .
800 PSI + 50 PSI
24°C + 1°C
7.4 + 0.2
85 ml7min.
200 PSI + 20 PSI
Feed
NaCl mg/1 33,540 mg/1 2404 mg/1
DIMP mg/1 19.5 mg/1 25.0 mg/1
Product
NaCl mg/1 8,430 mg/1 445 mg/1
DIMP mg/1 2.2 mg/1 2.5 mg/1
X Reduction
NaCl
DIMP
75.0
88.7
81.0
98.0
The rejection of D1MF by reverse osmosis was evaluated for DIMP in
dechlorinated tap water at pH 7 and in brackish water (5000 mg/1 NaCl) at
pH 7. Two trials were performed for each. The tap water results are
presented in Tables 11 and 12. Product water concentrations ranged from
0.041 to 0.091 tag/1 for feed water concentrations ranging from 14.6 to 17.8
mg/1. The percent removal varied from 99. 52 to 99. 72. Results for the
brackish water are in Tables 13 and 14. For the two trials, product water
concentrations ranged from 0.023 to 0.063 mg/1 for feed concentrations from
16.2 to 18.7 mg/1. Percent removals ranged from 99. 62 to 99.82, with most
of the samples resulting in the latter value.
Figure 18 shows a plot of the logarithm of the percent DIMP remaining
against cumulative flow through the membrane for the tap water trials.
There is little difference between the two trials with the rejection
remaining essentially constant over the course of a run. Figure 19 is the
same plot for the brackish water trialr. The same pattern was observed.
CRDC Trials
The trials performed at CRDC comprised two with DIMP in dechlorinated
tap water and one with brackish water. The tap water results appear in
-Tables 15 and 16. Product water concentrations ranged from 0.080 to 0.223
mg/1 with feed water concentrations from 22.0 to 22.9 mg/1. Rejection
ranged from 99.02 to 99.62. Results for the brackish water tria*. are in
.Table 17. The product water concentration ranged from 0.057 to 0.143 mg/1
with the feed water concentration from 21.3 to 22.5 mg/1. The percent
-53-
removal was from 99.42 to 99.72
TABLE 11. REJECTION OF DIMP IN DECHLORINATED TAP WATER BY REVERSE
OSMOSIS IN THE JOHNS HOPKINS UNIVERSITY TEST STAND (TRIAL 1) MEMBRANE
Time
Temperature
pH
Pressure
(psi)
DIMP (mg/1)
Feed Product
Z Removed
0.25
23
7.1
580
15.6
0.0756
99.5
1.25
24
7.1
550
15.8
0.0625
99.6
2.25
25
7.1
510
15.6
0.0622
99.6
3:25
26
7.1
500
15.3
0.0558
99.6 •
10:25
33
7.1
500
14.6
0.0413
99.7
TABLE 12. REJECTION OF DIMP IN DECHLORINATED TAP WATER BY REVERSE
OSMOSIS IN THE JOHNS HOPKINS UNIVERSITY TEST STAND (TRIAL 2) MEMBRANE
Time
Temperature
<°C)
PH
Pressure
(psi)
DIMP (mg/1 )
Feed Product
0.25
20
7.2
600
17.0
0.0912
1.92
22
7.2
550
17.0
0.0488
2.42
23
7.2
550
17.6
0.0493
2.92
* 24
7.2
525
16.9
0.0472
3.42
25
7.2
500
17.4
0.0432
3.92
25.5
7.2
500
17.8
0.0456
TABLE 13. REJECTION OF DIMP IN BRACKISH WATER (5000 mg/1) BY REVERSE
OSMOSIS IN THE JOHNS HOPKINS UNIVERSITY TEST STAND (TRIAL 1)
DIMP (mg/1)
Temperature Pressure Z Salt
Time ( C) pH (psi) Feed Product Z Removed Removed
0.25 21 7.2 680 17.0 0.0538 99.7 99.0
1.75 24 7.2 650 17.1 0.0353 99.8 99.3
2.75 25 7.2 600 16.2 0.0376 99.8 99.4
3.75 26 7.2 575 18.7 0.0337 99.8 99.3 *
4.75 27 7.2 575 16.5 0.0330 99.8 99.2
5.75 28 7.2 550 18.7 0.031C 99.0 99.3
TABLE 14. REJECTION OF DIMP IN BRACKISH WATER (5000 mg/1) BY REVERSE
OSMOSIS IN THE JOHNS HOPKINS UNIVERSITY TEST STAND (TRIAL 2)
DIMP (mg/1)
Temperature Pressure Z Salt
Time
(°C)
pH
(psi)
Feed
Product
Z Removed
Removed
0.25
18
7.3
750
18.4
0.0634
99.6 ’
98.7
1.25
20
7.3
700
18.0
0.0437
99.7
99.1
2.25
21
7.3
720
18.2
0.0353
99.8
99.2
2.75
22
7.3
650
17.8
0.0383
99.8
99.3 •
3.25
23
7.3
650
17.9
0.0327
99.8
99.2
3.75
24
7.3
625
18.3
0.0266
99.8
99.4
4.25
25
7.3
625
17.9
0.0318
99.8
99.3
4.75
25.5
7.3
600
18.3
0.0327
99.8
99.5
-57-
Figure- 18. Logarithm of the Z DIMP remaining in dechlorinated tap water after treatment by reverse osmosis
in the Johns Hopkins University test stand. Average feed concentration 15.4 to 17.3 mg/1
DIMP. Trial 1 (*) and trial 2 (0).
reverse osmosis in the Johns Hopkins University test stand. Average feed concentration 17.3
to 17.4 mg/1 DIM?. Trial 1 (^) and trial 2 (*) . Average salt rejection 99.2%.
TABLE 15. REJECTION OF DIMP IN DECHLORINATED TAP WATER BY REVERSE
OSMOSIS IN THE CRDC TEST STAND (TRIAL 1)
Time
DIMP
Feed
Concentration,
Product
(mg/1)
Waste
X Removed
1030
22.6
0.101
24.3
99.6
1200
22.9
0.0952
25.0
99.6
1330
22.0
0.103
24.7
99.5
1500
22.6
0.0849
24.2
99.6
-60-
I*;.’/.
!•
«
m
i*. .
-Si
• •*. •*. *.• *,* *.•
* . * * * •« 1 M 41 - ,*
/ V* ■
.‘■s':
- .** .*«V- /•
* V *■ H ’ > - > ' . ' . * . • . . • I
• *,* » \ ■ m* «■ ■ * . *
»- ^ m mj. +JL »Jf *.1
TABLE 16. REJECTION OF DIMP IN DECHLORINATED TAP WATER BY REVERSE
OSMOSIS IN THE CROC TEST STAND (TRIAL 2)
DIMP
Concentration ,
(mg/1)
Time
Feed
Product
Waste
Z Removed
1015
22.8
0.223
25.3
99.0
1215
22.5
0.0928
24.8
99.6
1415
22.5
0.0796
26.3
99.6
1615
22.5
0.0819
25.5
99.6
Figure 20 shows a plot of the logarithm of the percent DIMP remaining
against time for the tap water trials. Figure 21 is the same plot for the
brackish water trial. The pattern is quite similar to that observed in the
JHU tests with little change in the percent remaining over the course of a
run.
It should be noted that in both the JHU and the CRDC trials the
removal for the first sample was lower than for subsequent samples. This
is illustrated in Figure 22 which shows the percent removal of DIMP on an
expanded scale against time from the first sample. One CRDC run was an
exception. The time between start-up of the reverse osmosis unit and the ‘
first sampling in the CRDC trials was not known, however. It appears from
these results that a warm-up period should be provided in order to obtain
optimum removal of the chemical.
* TIME. HOURS
Logarithm of the % DIMP remaining in brackish (5000 mg/1 NaCI) water after treatment by
reverse osmosis in the CROC test stand. Average feed concentration 22.0 mg/1.
©
Figure 22. Percent DIMP removal by reverse osmosis on expanded scale versus time for all runs. Time
corresponds to time of first sample. The Johns Hopkins University tap water trial 1 (*)
trial 2 (^); CRDC tap water trial 1 (x>), trial 2 ( ); Johns Hopkins University brackish
(5000 mg/1 NaCl) water trial 1 (+) , trial 2 (O); CRDC brackish (5000 mg/1 NaCl) water
DISCUSSION
SIMULANT CRITERIA
The choice of candidate simulants was based on the desirable
characteristics of an ideal simulant and a comparison of chemical
properties that might influence the rejection of compounds by membranes.
The primary purpose of the simulant was to develop a system that could be
used to provide a preliminary evaluation of the ability of membranes to
reject important toxic organics. Such a system could also be used to
develop operational and evaluation procedures under field conditions.
Consequently, the toxicity, stability, analytical methodology and membrane
rejection were prime criteria. Unfortunately, information for these
parameters was limited and a more pragmatic approach was employed to make
the initial selection of the compounds to be tested. After the compounds
were chosen for chemical, physical and operational characteristics, the
short list was subjected to further review. The initial list of simulants
developed in the literature search and additional selected criteria are
shown in Table 18. The list was quickly reduced based on molecular weight,
presence of phosphorous, and solubility in water. Bis (2-ethyl hexyl)
phosphonate was rejected because the molecular weight was near the
exclusion limit suggested for RO membranes (Reid, 1966). Diethyl phthalate
was insoluble in water. Most agents of concern contain the phosphorous.
The presence of phosphorous would be a useful characteristic for the
simulant and provide a common characteristic that could be used to develop
analytical methods. Diethyl glycol dimethyl ether, dietyl phthalate,
-67-
TABLE xd. SELECTED PROPERTIES FOR CANDIDATE NERVE AGENT SIMULANTS
Molecular
Water
Weight
Phosphorous
Solubility
1.
Bis (2-ethyl hexyl) phospnonate
306.4
+
NL
2.
Diethyl glycol dimethyl ether
134.2
-
S
3.
Diethyl phosphonate (DEMP)
138.2
+
S
4.
Diethyl phuhalate
222.2
-
-
5.
Diethyl sebacate
255.4
-
NL
6.
Diethyl sulfite
138.2
-
S
7.
Diisopropyl methyl phosphonate (DIMP)
181.1
+
S
8.
Dimethoxy methyl phosphonate (DMMP)
129.1
+
S
9.
Ethyl dimethyl phosphite
138.1
+
S
NL * Not listed.
-68-
diethyl sebacate and diethyl sulfite did not contain phosphorous and were
not considered further. Of the remaining compounds DIMP, dimethoxy methyl
phosphonate (DMMP) and diethyl phosphonate (DEMP) were commercially
available. A detailed description of the chemical properties, chemical
reactivity, and toxicity for DIMP and DMMP compiled by the U.S. Army CRBC
Environmental Technical Division on Current Chemical Agent Simulants can be
found in APPENDIX 2.
DIMP was chosen for the initial evaluation because of the ancillary
information available and previous experience with this compound in related
situations. While DIMP did not match within 20Z of the selected chemical ‘
and physical properties, the other attributes were considered important for
initial trials. Except for a minor inhalation hazard, little evidence of
acute or chronic toxicity has been reported and at present a threshold
limit value has not been established. DIMP is a hydrolysis product of the
agent GB and contains phosphorous. Outterson and Prociv (1980) reported
the applicability of DIMP as a simulant. The compound was stable in water
and believed to be non toxic. The U.S. Army at Rocky Mountain Arsenal
found DIMP to be a useful indicator of groundwater contamination and
exhaustion of activated carbon used to treat wastes before recharging to
the groundwater (Civil Engineering, 1981). As a result of the use of DIMP
for these investigations, analytical methods have been investigated for the
determination of DIMP in watej: (Fasime, 1982; Broaders, 1982).
DEVELOPMENT OF ANALYTICAL METHODS
A major task of the present project was to develop a method of
analysis for the assay of the simulants. Conventional colorimetric methods.
-69-
while easy and simple to perform, may suffer from interferences in actual
systems, may not be flexible enough to evaluate different simulants and may
not possess the necessary sensitivity to be of value in detecting the low
concentrations of materials expected in the reverse osmosis product water.
The methods employed for the assay of the simulants were selected by the
criteria listed below:
1. SENSITIVITY
2. FLEXIBILITY
3. SPECIFICITY
4. ADAPTABILITY
5. SIMPLICITY
6. COST
Considerable effort was expended to develop a basic analytical approach
that would provide a workable, flexible, adaptable technique that could be
applied to a variety of simulants. DIMP was used as a prototype simulant
to develop the analytical methodology.
The method for the analysis of DIMP involved the quantitative
adsorption of the DIMP on octasilized silica gel cartridges, the
quantitative elution with acetone, and subsequent analysis by gas liquid
chromatography .
The adsorption on the silica gel was performed with commercially
available cartridges and served to remove the compound. The material could
then be eluted, seperated by gas liquid chromatography and analyzed by a
specific detector sensitive for the elements nitrogen, phosphorous and
sulfur. The adsorption step also provides a practical technique to
concentrate the simulant to increase the sensitivity of the assay. In this
study, small volumes of water were passed through the cartridge since the
increased sensitivity was not needed. For other simulants, the sensitivity
could be increased by several orders of magnitude by simply increasing the
volume of sample applied to the cartridge. The volume of sample applied
would not be unlimited. It would be a function of the adsorbant employed
and the compound to be adsorbed. A large number of adsorbents are
available and can be chosen to increase or decrease the specificity and
selectivity of the determination.
The elution step can be carried out with a variety of solvents. The ‘
choice of solvent will be influenced by the efficiency of elution of the
simulant from the adsorptive agent, the column and detector used on the gas
chromatograph and the stability of the simulant in the solvent. The
solvent employed can also add specificity and selectivity to the
determination.
Gas chromatography is becoming the standard method for the analysis of
volatile organics in water and a considerable body of literature is
available on specific procedures, methods and techniques. It separates and
provides quantitative information on complex mixtures that would be found
in any trials conducted under field conditions. The column may be selected
to provide the capability of a wide variety of separations. The detector
may be chosen to yield the necessary selectivity and sensitivity.
Gas Chromatographic Analysis of DIMP
A considerable amount of time was spent establishing satisfactory
operation of the gas chromatograph (GC) and the nitrogen-phosphorous
-71-
detector. As discussed in the Results section, the offset provides a
measure or the detector sensitivity. Wide variations in the offset were a
major point of concern. Variations may be symptomatic of several problems
including detector performance, column performance and more banal
difficulties such as column and septum leakage. A discussion concerning
the nitrogen-phosphorous detector and various columns based on experience
and technical information from Hewlett-Packard (HP) follows.
Nitrogen-Phosphorous Detector
The nitrogen-phosphorous detector is a thermionic emission detector,
also known as an alkali flame ionization detector. It employs a rubidium
silicate bead positioned above the flame tip by the collector assembV.
The bead is heated by an adjustable current and the sensitivity of the
detector is a function of the bead temperature. When the bead has been
consumed or damaged, the collector must be replaced.
The beads are highly hygroscopic. New collectors must therefore be
conditioned according to the instructions provided by HP. This procedure
drives out moisture slowly and avoids cracking or chipping of the bead.
Likewise, if the GC has been left on standby, the bead may have had an
opportunity to taka up moisture. When starting up, the voltage should be
increased slowly until the desired offset is reached. None of the HP
literature was explicit on this point. It was possible that some
collectors suffered damage frpm rapid heating.
Too rapid application of the appropriate voltage for the desired
sensitivity can also burn the bead. This resulted in the loss of one new
-72-
collector. The voltage should be increased slowly to allow the temperature
of the bead to equilibrate slowly.
Moisture in the injected sample can have a depressing effect on the
detector's response. This can be compensated by increasing the voltage to
the bead; at some level the response will remain stable at a constant,
although depressed value. This will accelerate deterioration of the bead,
however, perhaps through a combination of the effects of higher voltage and
exposure to moisture. If the collector is in good condition, the detector
will recover to its normal operating state when injections of the moisture
laden sample cease.
The response of the detector is susceptible to fluctuations in ambient
temperature. It was observed that opening and closing the cover over the
detector resulted in a transient rise in an otherwise stable offset. The
existing insulation over the oven and around the detector was then
supplemented with glass wool and the GC was subsequently operated with the
cover in place.
The HP manual recommends a collector height of 0.075" (Hewlett-
Packard, 1974), but the Operating Note, "Evaluating and Optimizing the
Performance of the Nitrogen-Phosphorous Detector," (Hewlett-Packard, 1978),
indicates i.hat the height can be anywhere from 0.015" to 0.15". HP
reported that they usually obtain best results with settings between 0.025"
and 0.050". All analyses of DIMP presented in this report were performed
with a collector height of 0.075", but some experimentation was done. With
the 2Z OV 101 HP test column installed, injections of the performance
evaluation sample (PES) were made with a setting of 0.075". The average
-V-V-V-.--'.
. - - - . •.*.*'
r v -a-.v.
* » * •_* *• * » » * » * % *
r.\Vvw\
phosphorous response (area per microliter) for 7 injections was 1370. The
setting was then lowered to between 0.050" and 0.060". The average
phosphorous response for four injections was 2600, an increase of about
1001. The neight was then lowered to between 0.025" and 0.050". A single
injection of PES dropped the offset from 91 to 14 mm. It appears from HP's
Operating Note that the optimum setting may vary from one collector to the
next. It is probably not necessary to go through the relatively cumbersome
procedure of optimizing the setting each time a new collector is installed;
unless the sensitivity is unacceptably low, recalibration with standard
solutions should b^ sufficient.
Replacement of the chimney assembly of the detector apparently
resolved certain problems. The offset had been highly erratic, eventually
decaying to zero despite voltage increases. Spiking also occurred when
there was any movement in the area around the detector. With a new
chimney, the problem was eliminated. It is conceivable that earlier
difficulties in obtaining a stable offset were attributable to
deteriorating performance of this component.
Performance Evaluation —
The 2X OV 101 te3t column was installed in order to check the overall
operation of the GC. The collector was cot changed initially. The
performance evaluation sample supplied with the test column contains
azobenzene, octadecane and malathion in 2,2,4-trimethylpentane. Evaluation
of performance is based on the ratios of azobenzene and malathion peak
areas to the octadecane peak area and the ratios of these areas to peak to
peak noise measured in millimeters. According to HP, "The response of the
(N-P detector] to the evaluation sample will vary from instrument to
instrument, from collector to collector, and even from day to day." As
long as the ratio of peak area to octadecane area is greater than or equal
to 3.5 for azobenzene and 7.5 for malathion, and the ratio of peak area to
noise is greater than or equal to 1000 for azobenzene and 2000 for
malathion, the collector is performing acceptably. It is also noted,
however, that for a given analysis, acceptable performance is that which
allows the analysis to be accomplished and that this might be quite
different from their definition of acceptable (Hewlett-Packard , 1978).
There are therefore no precisely defined criteria for acceptable
performance.
A stable offset was obtained for the evaluation and the chromatograms
resulting from injections of PES appeared reasonable; that is, the relative
retention times were correct, the peak areas were reproducible., the peaks
were well defined and the factory performance criteria based on area ratios
were met. The malathion response was consistently lover than the
azobenzene response, however. According to HP this "can almost always be
traced to a bad column," since the malathion is the most readily adsorbed
compound in the sample (Hewlett-Packard, 1978). Other observations were
consistent with this possibility. For example, successive injections
produced increasing malathion peak areas which according to Supelco can be
caused by "adsorption of components and saturation of active sites with
sample (priming the column)" (Supelco, 1983). The offset also increased
slightly after each injection which may be attributable to a response to
residua1 malathion. With a new collector in place, essentially the same
pattern was observed, although the malathion peak areas were about 2502
greater and the azobenzene areas were about 502 greater. Some
experimentation with collector height was also done as described above.
HP's Operating Note mentions that most collectors seem to have some
"steady state" offset at which they are most stable. This is usually 252
to 502 less than the checkout procedure offset of 75 to 100 mm. If the
offset is higher, it will slowly return to the steady state value,
resulting in a change of sensitivity (Hewlett-Packard, 1978). Monitoring
the offset over a 12 hour period revealed a docline to about 802 of the
initial value (88.5 mm). According to HP, the offset should be stable over
an eight hour period. In the first eight hours of observation the offset
had declined to 75 mm or 852 of the initial value. This may have been a
manifestation of the detector's tendency to approach some steady state
offset value. The significance of offset variation was tested as described
in the Results section. Between offset values of 75 and 88.5 mm, the
response for 1.0 mg/1 DIMP in acetone increased by 4.52. Although
intentional changes in the voltage applied to the bead are not necessarily
equivalent to uncontrolled factors influencing the offset, this result
suggested that offset variations may have a significant impact on the
response. In general, however, on the basis of the performance evaluation
it appeared that the detector was functioning normally.
Columns
Several different columns were tried in conjunction with the simulant
DIMP, including both solid and liquid phase types. Because of the
multiplicity of factors effecting the response, it was difficult to assign
particular causes to observed aberrations. It was clear that in many
instances, the detector was not functioning properly. On the other hand,
column and septum leaks were occasionally responsible. Despite the
uncertainties, it was possible to differentiate the performances of various
columns. These are summarized in Appendix 3. The beat results were
obtained with the 13Z Carbowax 20M on 80/100 chromosorb W-HP glass column.
The operating conditions that were finally established after investigation
of the detector and columns were reported in the Methods section.
The data collected for DIMP demonstrates the applicability of the
procedure. The gas chromatograph provided a rapid, selective, sensitive
0
technique. The standard curves prepared were reproducible and provided a
linear response with high correlation coefficients. Typical correlation
coefficients greater than 0.99 were observed. Replicate injections of DIMP
in acetone had a coefficient of variation of about 7Z for the poorest
conditions tested at low concentrations approaching the sensitivity limits
of the detector. At concentrations of 0.05, 0.10 and 20.0 mg/1 the
coefficient of variation was only 3Z. The observed variations are well
within the variations observed for this type of analysis.
DIMP was found to be stable in acetone, the elution solvent. The
stability of DIMP in acetone has important implications for the mundane
manipulations and processing of samples. The levels of DIMP in acetone
showed little change over a oeriod of 20 days for a range of 0.1 to 10.0
mg/1 at 4°C and 23-25°C under dark and natural light conditions. For
DIMP this means that samples may be collected, adsorbed and eluted in the
field, and shipped to a laboratory for subsequent quantitative analysis
-77-
.■wwc
without any special handling. Determinations in the field may be conducted
without any special precautions concerning the sample.
The adsorption and elution of the DIMP on the octylsilane bonded
silica gel cartridge was quantitative and overall recoveries in tap water
and tap water with 5000 mg/1 of salt at neutral and slightly alkaline pH
were 94.12 or greater for samples of 1.0 mg/1 DIMP and greater than 89.22
for samples of 10.0 mg/1 DIMP. There appeared to be decreased recovery for
water samples at acidic pH at the 10.0 mg/1 level. Only 86 OX and 88.82
of the DIMP was recovered under these conditions. These results must be
treated with caution since variation in detector sensitivity was
encountered during the analysis. It is also possible that the extraction
columns were overloaded for the 10.0 mg/1 samples. Relatively large
volumes of 15 aud 25 ml were passed through the adsorption columns for
these samples. During reverse osmosis trials, only 2 ml were extracted for
the feed water samples.
The addition of the extraction step increased the variability
associated with the overall measurement. The coefficient of variation for
replicate extraction and analysis was 14.62 for low concentrations found in
the product water. It should be noted that the levels approached the lower
sensitivity limit. For the higher concentrations observed in the feed
water the coefficient of variation was only 2.62 The data suggested the
overall recovery of the DIMP was sufficient for concentrations to be
expected in this study. The data also suggested that variation may be
decreased by increasing the sample size . The adsorption step could be
used to concentrate the low level samples and increase the quantity of
material analyzed by the gas chromatograph.
-78-
Once worked out, Che procedure was found to be simple, easy, rapid and
relatively inexpensive. The sample collection, extraction and gas
chromatographic procedures could all be automated and large numbers of
samples could be processed efficiently. The procedure may be interupted at
several points allowing transport to places where the analysis may be
continued more efficiently and economically. The analytical approach is
sufficiently developed to test other simulants developed in the literature
search. Preliminary trials have been conducted with DMMP and DEHP. These
compounds may be analyzed under chromatographic conditions similar to DIMP.
STABILITY IN AQUEOUS SOLUTIONS
To be of value as a simulant for the agents to be removed by reverse
osmosis the simulant should be stable in aqueous solution to facilitate
testing. Information in the literature suggested that DIMP was stable in
water. At the temperatures of 4-40°C and pH values of 5 - 10 expected to
be encountered in the aquatic environment, the hydrolysis of DIMP would be
8 low (Bel'skii et al., 1975). Studies were conducted in our laboratories
to determine the stability of DIMP under conditions closer to field
conditions. DIMP was added to tap water and tap water with 5000 mg/1 NaCl
at pH 5.0, 7.0 and 9.0. The latter conditions with salt were intended to
approximate brackish waters. Each test was conducted over the range of
concentrations of DIMP expected to be encountered in the test system (0.10,
1.00, and 10.0 mg/1). For each concentration of DIMP and aquatic
conditions, the simulant appeared to be stable for a 14 day period. While
the stability of DIMP was expected, the data collected covered the
conditions to be used in subsequent experiments. In addition, the DIMP
-79-
stability trials in aqueous solutions served to establish workable
experimental protocols for future studies.
REMOVAL OF DIMP BY REVERSE OSMOSIS
Trials were conducted in the reverse osmosis test stand supplied by
Ft. Belvoir Research and Development Center at the JHU field station at
Edgewood Area. DIMP was added to dechlorinated tap water at about 20 mg/1
and the reverse osmosis unit was operated as a closed loop system. The
temperature range over the course of the test was 23-33°C and the pH was
7.1. DIMP was rejected at better than 992 for the conditions tested. The
level of DIMP was reduced from 15 mg/1 in the feed water to 0.04-0.08 mg/1
in the product water. The addition of 5000 mg/1 of salt had little
demonstrable effect on the rejection of DIMP in the reverse osmosis test
system . Similar rejections were found for the trials conducted by the
CRDC group with their reverse osnmosis test stand and assayed in our
laboratory. The average level of DIMP in the feed water was 22.6 mg/1 and
the average concentration in the product water was 0.12 mg/1 in tap water
trials. The average level of DIMP in the feed water was 22.0 mg/1 and the
average level in the product water was 0.08 mg/1 in the brackish water
trial. Again slightly better removal was observed for the brackish water.
Unfortunately no data is available at this time for the
anticholinesterase agents from CRDC to permit a comparison with the
rejection rates observed for DIMP. The suitability of DIMP as a simulant
therefore cannot be judged. with respect to removal by reverse osmosis.
Some data for the removal of anticholinesterase agents by reverse osmosis
was reported by Lindsten (1978). Removals of 99.9 and 99.1 were reported
CONCLUSIONS
1. OIMP and DMMP were chosen as possible simulants for acetyl choline
esterase inhibitors for testing reverse osmosis membranes.
c2. The system developed for the extraction and analysis of DIMP was
sensitive enough to permit assay down to the 0.02 mg/1 range presently
established by the Surgeon General for acetyl choline esterase inhibitors.
3; The extraction, elution, and assay steps provide a method that may be
adapted to other simulants and lower levels of detection.
JU DIMP was stable in acetone for at least 20 days and allows for a
flexible sample processing schedule.
^5,- DIMP was stable in water and will not require special consideration for
development of membrane test protocols.
6. The levels of DIMP in the product water was 0.04 to 0.09 mg/1 in tap
water and 0.03 to 0.06 in the brackish water trials. This corresponded ‘to
about 99.71 and 99. 8Z rejection, respectively. The salt rejection for the
brackish water trials was approximately 99Z. Similar DIMP rejection was
observed for trials conducted by CRDC.
Jl*,. A comparison between removal of the DIMP and the acetyl choline
esterase inhibitors by reverse osmosis cannot be made. Information on the
rejection of nerve agents by the membrane used in this study is not yet
available.
-82-
RECOMMENDATIONS
1. The current project has set up general operational methods and
procedures that allow for the assay and application of simulants for the
testing of reverse osmosis membranes and reverse osmosis systems. A good
deal of time and effort was directed toward the development of an
analytical methodology and procedures that would be applicable to a broad
range of possible simulants and have the necessary sensitivity and
specificity. The studies should be continued and expanded to evaluate
additional simulants and other membranes.
2. The ultimate test for the simulants would be a thorough comparison of
the simulant with the agent under conditions as close to "in use"
conditions as possible. Unfortunately, open testing of the nerve agents is
not possible but after preliminary comparative testing under controlled
conditions, the functional simulants can be tested in the field. This
would allow a thorough evaluation of the effects of different water quality
parameters on the rejection rates under realistic conditions.
3. The analytical method used in this study employed a gas chromatograph
with a nitrogen-phosphorous detector. The gas chromatograph performed
well, but considerable time was spent with the detector. The advantages of
the gas chromatographic method for analysis warrants further work to
evaluate other detectors.
LITERATURE CITED
Bagley, F.D. et al. 1977. Simulant Review and Selection DPG Document
No. DPG-1R-T-125A U.S. Army Dugway Proving Ground, Dugway, Utah 84022.
Banks, W. , Sharpies, F. 1964. Arthur D. Little Research Institute, Final
Report to Office of Saline Water.
Bel'skii, U.E., et al. 1975. Kinetics of Dialkyl Methylphosphonate
Hydrolysis Izu. Akad. Maus SSSR, Ser Kuin, 72:78155.
Blais, P. 1977. Polyamide Membranes. In: Reverse Osmosis and Synthetic
Membranes. Edited by S. Sourirajan, National Research Council, Canada
Publications.
Breton, E.J., Jr. 1957. Water and Ion Flow Through Imperfect Osmotic
Membranes. Office of Saline Water Research and Development Progress Report
No. 16 PB 161341.
Broadus, J. 1982. DIMP and DMMP in Water. Chemical Analysis Method,
Rocky Mountain Arsenal, Commerce City, CO 80022.
Chian, E.S., et al. 1975. Removal of Pesticides by Reverse Osmosis.
Environmental Science and Technology, 9(1): 52.
Coon, P.A., et al. 1982 (Draft) Simulant Users Handbook. Chemical
Systems Laboratory Special Publication, Chemical Systems Laboratory,
Aberdeen Proving Ground, Maryland 21010
Fasamo, R. eit al^. 1982. Analytical Methods Development for Dimethyl
Methylphosphonate Diisopropyl Methylphosphonate and Trimethylphosphate.
Report DRXTH-TE-Cr. Final Task Report prepared for U.S. Army Toxic and
Thazorcloric Material Agency, Aberdeen Proving Ground, Maryland. Arthur D.
Little Inc., Cambridge, Massachusetts.
Hewlett Packard, Inc. 1974. Gas Chromatograph Instrument Manual Series
5830A, Avondale, PA.
Hewlett Packard, Inc. 1978. Operating Note: Evaluating and Optimizing
the Performance of the Nitrogen-Phosphorous Detector.
Lindsten, Don C. 1972. Memorandum Report, 600 Gallon Per Hour Reverse
Osmosis Water Purification Unit. Project Officer U.S. Array Material
Development and Readiness Command.
Lonsdale, H.K. e£ a_l . 1965. J. of Appl. Polymer Sci. 9:1341
Nusbaum, I. 1981. Membrane Process-rDesign and Application. In:
Proceedings Twenty-Third Annual Public Water Supply Engineer Conference.
Champaign, Illinois.
-84-
Pusch, W. 1977. Determination of Transport Parameters of Synthetic
Membrane by Hyperfiltration Experiments.
Sherwood, T.K. £t a_l . 1967. Desalination b> Reverse Osmosis 9 & EC
Fundamentals 6:1.
Sourirajan, S. 1970. Reverse Osmosis. Academic Press, New York.
Spiegler, K.S., and Lavial, A.D.K. 1980. Principles of Desalination.
Academic Press, New York.
Standard Methods for the Examination of Water and Wastewater. 1980.
APHA-AWWA-WPCF.
Supelco, Inc. 1983. Troubleshooting Guide, Bellefonte, PA.
BIBLIOGRAPHY
Anderson, J.E., e£ £l . 1972. Factors Influencing Reverse Osmosis
Rejection of Organic Solutes From Aqueous Solutions. The Journal of
Physical Chemistry. 76(26) :4006 .
Ford, A., £t £l . 19/4. Removal of f2 Virus from Water by Army Water
Purification Units. NTIS AD-A-005557.
Gregg, S.J. 1961. The Surface Chemistry of Solids. The Whiterfriars Press
Ltd., London, England.
Lacey, R.E. 1972. Membrane Separation Processes. Chemical Engr.,
79:56-74.
Lindsten, D.C. and Schmitt. 1976. Decontamination of Water Containing
Chemical Warfare Agents. Report 2125, U.S. Army Mobility Equipment
Research and Development Center, Fort Belvoir, Virginia.
Loeb, S. 1966. High Flux Cellulose Acetate Membranes. In: Merten (ed)
Desalination by Reverse Osmosis, The MIT Press, Boston, Massachusetts.
Lonsdale, H.K. 1982. The Growth of Membrane Technology. J. of Membr. Sci.-
10:81-181.
Merten, U., and Bray, D.T. 1966. Reverse Osmosis for Water Reclamation.
Third Inter. Conf. on Water Pollut. Res., Munich, Germany.
Michaels, A.S. and Porter, M.C. 1971. Membrane Ultrafiltration. Chem.
Tech. 57.
Outterson, G.C. and Prociv T.M. 1980. Ed. Proceeding of Toxic Substance
Control: Decontamination Symposium Sponsored by Chemical Systems
Laboratory, Aberdeen Proving Ground, Maryland 21010.
Porter, M.C. 1975. Selecting the Right Membrane. Chem. Eng. Prog. 71:55,
Reid, C.E. 1966. Principles of Reverse Osmosis. In: Merten (ed.),
Desalination by Reverse Osmosis. The MIT Press, Boston, Massachusetts, pp.
1-15.
Reid, C.E. and Breton, E.J. 1959. Water and Ion Flow Across Cellulosic
Membranes. J. Appl. Polymer Sci . 1:133.
Riley, R.L. ejt £l . 1971. Preparation Morphology and Transport of
Composite Reverse Osmosis Membranes for Seawater Desalination. Office of
Saline Water Symposium on Membrane Transport.
Saltonstall, C.W., Jr. 1576. Practical Aspects of Sea Water Desalination
by Reverse Osmesis. Principles and Desalination 18:315-320.
Sliger, H.B. and Quinn, R. 1976. Application of Membrane Processes
Desalination 19.
Tang, T.L. Don, et a_l_. 1981. Application of Membrane Technology. Ind.
Water Eng., 18-26.
-87-
RECORD NUMSER
1 ENTRY
2 TYPE
'V
2 CONTAINS AG CN T
M COMMON NA “E
I TECH NAME
S MOLEC WT
TS F CR HU LA
21 LCGTEN VPR PRESS HMHG
22 X21 TEMP CEGC
2M LlC DENS GM/CC
25 X22-2M OEMS TEMP OEGC
76 VAPCR DENS ATM
27 X26 TEMP OEGC
28 BOILING POINT DEG C
23 X 28 PRESS HMHG
II ULN
22 MELTING pt oecc
2M SURFACE TEJISICN DYNES/CM
25 X3M TEMP OEGC
28 LCGTEN CEHTISTCKC VISCOSITY
23 X38 TEMP CEGC
MO LOGTEN VOLATILITY M0/METER3
Ml X 40 TEMP OEGC
M2 MOLEC DIFFUS CO EF F CM2/SEC
M3 X M2 TEMP DEGC
MM LOGTEN HEAT QF VAPORIZATION K CA
M7 REFRACTIVE INCEX
Me XM7 TEHP DEGC
MB SCL IN MATER
67 LOGTEN ETV KCAL/KG
ZB SPEC HEAT CP KCAL/CKG 0EGC3
7M F LA CH POINT CEGC
73 HILCE3RAN3 SCL PARA
= 20
= C CM PC UNO
: NO
= 2 ISC2-MZ7 HCXYETHYLOCTHE R CS 3 ME
= diethyl glycol oi methyl ether
= 13M.17
= C6HM03
: . MS E
= 25
= . 25 S M
= 25
= M «S 300
= 25
= 162
= 760
= 102C201
= -68
= 23.50
= 2 S
= .003
= 25
= M .324
= 25
— . 06 1
= 25
\J M3 = 1 .888
= 1 .4037
= 20
= HISCI2LE
= 2. ISO
= .5003
= 70
= 3 ,0
R EC CRC NUM8ER
3
1 ENTRY
2 TYPE
? CONTAINS AGENT
4 COMMON NAPE
S 1)4
= C CM POUND
: IIC
= diethyl sul n te or cis
E HOLEC MT
i
3 FORMULA
3 LC3TEN VPR PRESS M.MHG
22 X 21 TEMP CEQC
24 LI3 DENS CM/CC
25 X 23-2 4 DENS TEMP CEGC
28 BOILING POINT DEGC
23 X28 PRESS MMHG
”1 ULN
34 SURFACE TENSION OYNES/CM
IE X 34 TEMP DEGC
*8 L05TEN CEN7IST0KE VISCOSITY
T Xie TEMP CEGC
«0 LOGTEN VOLATILITY MG/ME TER3
<a X4G TEMP DEGC
42 MCLEC OIF FUS CCEF* CK2/SEC
44 LOGTEN HEAT OF VAPORIZATION XCAL/tC
E7 LOGTEN ETV KCAL/KG
TE HILCEERANO SOL PARA
: 135.13
= C4H1002S
= .533
; 25
= 1.0789
= 25
= 123
= 730
: 2 0 2SC
= 22.70
- j r
- -.108
: 25
: 4 .405
= 25
= .05 4
s 1 .832
= 2.053
= 8.3
a CUCTE SOURCE
: HC3
Appendix 1
RECORD NUH9ER 6
X E..IRY
7 TYLC
3 CONTAINS AGENT
4 COMMON NA HE
5 MCLEC UT
%
33 FORMULA
21 LOGTEN VPR PRESS MMHG
22 X21 TEMP SEGC
21 LIB, CEN3 GM/CC
35 X23-24 DENS TEMP CECC
38 3CILING POINT OEGC
31 ULN
r«» SURFACE TENSION CYNES/CM
75 X34 TEMP OEGC
78 LOGTEN CE NT IS TO XE VISCOSITY
79 X 28 TEMP CEGC
40 LOGTEN VOLATILITY MG/KETER3
«1 X 40 TEMP CEGC
42 MCLEC OIFFUS CCEFF CM2/SEC
43 X42 TEMP CEGC
= 67
= CCMPC'JMC
= NC
= ETHYL DIMETHYL »H CS -V IT L
= 1.32.11
= C4H11C3P
= .531-
= 25
= 1 .0040
= 25
= 124
=10 2PC2
= 31.20
= 25
= - .237
= 25
= 4 .4 03
= 25
= .057
= 25
44 LOGTEN HEAT CF VAPORIZATION KCAL/WJ = 1.350
37 LOGTEN ETV KCAL/KG
2.4 37
Appendix 1
/
V V *»' V VJ W ■ >»'. «v S'. . - . . .
J EC CSC NUM3ER 1
1 ENTRY
2 TYPE
2 CONTAINS AO ENT
4 CCHMCN NAME
5 TECH NAME
6 HCLEC WT
i
33 FORMULA
21 LOOTEN VP R PRESS HMHG
24 LIQ OEMS CH/CC
25 X23-24 DEL'S TEMP CEGC
28 3 OX LINO POINT OEOC
23 X28 PRESS PMHS
n VLN
24 SURFACE TENSION OYNES/CM
25 X34 TEMP OE SC
35 LOOTEN CENTXPOI EE VISCOSITY
\
37 X3S TEMP OE GC
28 LOOTEN CENTISTCKZ VISCOSITY
23 X 38 T CMP SE GC
*0 LOOTEN VOLATILITY MG/METER3
41 X 40 TEMP OEOC
42 HCLEC OIKFUS CCEFF CM 2/ SEC
43 X 42 TEMP OEOC
44 LOOTEN HEAT OF VAPORIZATION KCAl/IC
45 LOOTEN HEAT OF COMBUSTION K CAL/ KG
47 REFRACTIVE INCEX
48 X 47 TEMP OE SC
67 LOGTEN ETV KCAL/KC
S8 SPEC HEAT CP KCAL/CKG 0E0C3
72 LOGTEN SATN VAPCR CCNC MG/M3
74 FLASH POINT OEOC
"5 HILSE2RANS SOL PARA
: 13
= C CM PC UNO
= NO
= SIS
r S IS C2-E TH YL HE XY L3 PHf.^ PH CN AT s.
■t
- 306.42
s C 1SH3S0P
= -3.134
= .33 0 0
= 25
= 230
= 7S0
= 4Y2I10 2PH0
= 23.50
= 25
= .785
= 25
= .734
= 25
= 1.025
= 25
= .038
= 25
= 1.633
: 3.888
= 1.4416
= 25
= 2.513
= .4000
= 1 .0 25
= 165
= 6.2
Appendix 1
RES OR C NUMBER 1
1 ENTRY
2 type
I CONTAINS AS ENT
a COMMON NAME
E TECH NAME
S NOLEC HT
n FORMULA
21 LCOTEN vpr press mmhg
22 X21 TEMP CE SC
-a L IS CtNS GM/CC
28 3CILING POINT OEGC
23 X28 PRESS HMHG
II ULN
IN SURFACE T Cl SI ON OYNES/CH
IS X 14 TEMP OEGC
IS LCGTEN CENTIPCISE VISCOSITY
17 X26 TENP OEGC
■*8 LCGTEN CENTISTCKE VISCOSITY
19 Xie TEMP OEGC
ao LOCTEN VOUTILITY MG/METER3
ax xaa temp oegc
42 MCLEC OIFFUS CCE TT CM2/SEC
ai X42 TEMP OEGC
aa LOGTEN HEAT OF VAPORIZATION KCAUKJ
a? REFRACTIVE INDEX
aa xa7 TEMP OEGC
a9 SCL IN WATER
£7 LCGTEN ETV KCAL/KG
G8 SPEC HEAT CP XCAL/CKG DEGC3
69 X6i TEMP OEGC
71 AUTOIGNITION TEMP OCOC
7% FLASH POINT OEGC
= 33
Z CCMPCUND
= NO
= diethyl phcsphona te cr OEHP
r diethyl phosphite cr diethyl hyorcsen p
- 13S.11
: caHiici?
= .SOS
= 25
= 1.0578 ,
= 183 •
= 760
= 20 2PHC
5 3 0.35
= 2 5
= .068
: 25
= .os:
5 2 S
= a. 373
= 25
= .069
= 25
= 1.820
= 1 .4373
• : 2 D
S SOL
= 2.170
s • S3S3
= 55
= 224
= 32
75 HIL0E3RANC SOL PARA
8.1
Appendix 1
RECORD NUMBER 4
1 ENTRY
2 TYPE
3 CONTAINS A3 ENT
4 COMMON NAME
5 MCLEC WT
33 FORMULA
21 LOOTEN VPS PRESS MM HG
22 X21 TEMP OEGC
24 LIQ DENS GM/CC
23 X23-24 SENS TEMP DE SC
28 BOILING POINT CESC
21 ULN
24 SURFACE TENSION OYNES/CM
23 X34 TEMP CEGC
25 LOOTEN CENTIPCISr VISCOSITY
37 X3S TEMP DE GC
28 LCGTEN CENT IS "*0 ME VISCOSITY
33 X38 TEMP CEGC
40 LOOTEN VOLATILITY MG/METER3
41 X 40 TEMP DE GC
<C MCLEC OIFFUS CCEFF CM2/SEC
43 X 42 TEMP DE GC
44 LCGTEN HEAT OF VA P0R17A TI ON K CAL/ M3
47 REFRACTIVE INCE X
48 X47 TEMP DE SC
73 HILDEERANC SOL PARA
42
CCM POUND
NO
DIETHYL S.EBACATE Cfl D ES
2S8.3E
C14H26C4
-3.232
25
. 3597
2S
307
20V8V02
3 2.30
25
.738
25
.732
25
.832
25
.043
25
1.814
1.4 363
23
7.8
Appendix 1
RECORD NUM8ER s
1 ENTRY
2 TYPE
I CONTAINS AGENT
4 COMMON NAME
E TECH NAME
C MCLEC WT
23 FORMULA
21 LOGTEN VP R PRESS MM Hfl
72 X21 TEMP OEGC
04 LIS DENS CK/CC
25 X23-24 OENS TEMP CEGC
26 VAPOR OENS ATM
28 SOILING POINT 0C5C
29 X28 PRESS MMHS
21 ULN
32 MELTING PT CEGC
34 SURFACE TENSION DYNES/CM
IS X34 TEMP CEGC
36 LOGTEN CENTIPOISE VISCOSITY
37 X 36 TEMP DEGC
38 LOGTEN CENTISTCKE VISCOSITY
39 X 38 TEMP CEGC
40 LOGTEN VOLATILITY MG/fCTERI
41 X40 TEMP OEGC
42 MCLEC OIFFUS LCCFF CM2/SCC
43 X42 TEMP OEGC
44 LOGTEN HEAT OF VAPCRI2ATI0N KCAL/IC
43 LOGTEN HEAT OF COMBUSTION K CAL/ KG
4 1
COM POUND
NO
DIETHYL PKTKALATE OR CEP
ETHYL PHTHALATE
2 22.22
C12H14C4
-3.180
25
1.1 230
25
7 .£530
2 36
760
2 OVR 8V02
-40
36.10
25
1.349
2 5
1.300
25
1.556
25
.049
25
i.7eo
2.762
*
■■ n i
Appendix 1
i •
9
47 REFRACTIVE INDEX
48 X47 TEMP OE SC
49 SCI. IN UATER
68 SPEC HEAT CP KCAL/CKG DEOC3
71 A OT CI3NIT ION TEMP OESC
71 LCCTEN SATM VAPCR CCNC HG/HI
73 X72 TEMP 0E9C
74 FLASH POINT DEOC
7* HH.DE3RAN3 SOL PARA
77 CCST UNITS
78 COST 5U0TC
81 QUOTE SOURCE
= 1.5062
: 25
VINSOL
= .4500
= 457
= .897
: 25
= 152
: 1.0
s POUND
s .50
t NATHEISCN-C0LE-3EU.
APPENDIX 2
DETAILED CHEMICAL PROPERTIES AND REACTIVITY
OF CHEMICAL AGENT SIMULANT COMPILED BY
ENVIRONMENTAL TECHNICAL DIVISION CRDC
IDENTIFIER
CAS REG NO
1445-75-6
. ;HE MICA L FORMULA:
>YNONYMS: Fhosphonic acid, methyl-, diisopropyl ester; phosphonic acid, methyl-, bis (1-rnethylethy
■ster; diisopropyl methylphosphonate; methanephosphonic acid, diisopropyl ester.
3ISCRIPTORS: DIMP belongs to a group of compounds known as organophosphates.
■; CHEMICAL AND PHYSICAL PROPERTIES:
Property
Value (Ref)
Property
Value (Ref)
]
s
/~
•lolecular weight
180 (1)
Specific gravity
0.98 g/ml (1)
iOiling point
174°C (1)
Solubility
0.1 - 0.2* (1)
^ lash point
71°C(2)
*
xamx
DIMP
Diisopropyl methylphosphonate
MILITARY APPLICATION: DIMP is used as a simulant for the G-agents. The compound has spectral
haracteristics similar to those of the G-agents, and is therefore used in general remote detection.
N DUS TRIAL APPLICATION: D1MP has no industrial application.
TORAGE, SHIPPING, AND HANDLING: DIMP is classified as a combustible liquid as defined in the US
epartment of Transportation 49 CFR 173.115 (b).3 The compound is not specifically listed as a
a azardous waste under the Resource Conservation and Recovery Act (RCRA), (40 CFR 261.33), and its
* igh flash point (71 °C) does not qualify it as a hazardous waste on the basis of ignitabiiity, as defined in
V 0 CFR 261.21.
gy £>jyv/J TZctf (?/</ 6/0
L ’ ChAAe/jt <zf/£/r\ flGivT ^
Appendix 2
!
i
3
s
i
TOXICOLOGY: Acute toxicity of DlMP.
ROUTE
SPECIES
DOSE
EFFECTS/REMARKS (Ref)
Intravenous
Rabbit
224 mm^/kg undiluted
caused local irritation (5)
Percutaneous
Rabbit
> 200 mm^/kg
undiluted
no irritation at the site of
application (5)
Ocular (eye)
Rabbit
0.25 mm^/eye
undiluted
inflammation, rmld to severe, neg in
24 hours, lacrimation, edema - slignt,
neg in 24 - 48 hrs. (5)
Intraperitoneal'
Mice
> 250 mg/kg undiluted
ld5q <»
Inhalation
(total exposure)
Mice
Ct = 24,811 mg
min/rrr (t=43 min)
0/10 died in 14 days after exposuie in
a 386 liter chamber. The average
chamber concentration was 5 77
mg/r.v*. No toxic signs. (5)
Subcutaneous
Rat
> 200 mg/kg
undiluted
LD50 (5)
Oral
Duck
Bird
Mammal
Cattle
1490 mg/kg
1000 mg/kg
503 mg/kg
750 mg/kg
LD5Q (6)
LD50 (6)
LDen (6)
LD50 (7)
Carcinogenicity: An extensive search of the literature did not present any data on the carcinogenity of
Dl MP.
Mutagenicity: Hart® has reported that specially purified samples of DlMP proved to be non-mutagenic
when administered to mice, rats, and dogs.
Teratogenicity: Hart® also reported that no teratogenic effects were observed in rats given dietary
levels of 80, 250, or 750 ppm on days 6 through 15 of gestation. He observed that the compound
produced no teratogenic effects it rats when dietary levels of 300-3000 ppm was given on days 6 through
15 of gestation. Dietary incorporation of DlMP at 300-3000 ppm produced no dose-related reproductive
response in the rat over 3 successive generations with 2 matings per generation.
Health Hazards: DlMP imposes a minor acute inhalation hazard. The compound is slightly irritating to
the eyes, nose, skin, and respiratory tract. Presently, no threshold limit value (TLV) has been
established for DlMP.
Plant Data: No, data was found on this subject.
Appendix 2
CHEMICAL REACTIVITY:9
Alkali and Alkaline Earth Metals: An exothermic reaction may occur upon mixing DIMP with alkali and
alkaline earth metals.
Azo Compounds: Azo compounds may react with DIMP to produce hazardous conditions.
. Caustics: The hydrolysis of 01MP under alkaline conditions yields isopropyl alcohol and metal salt of
TrieThylphosphonic acid.
Epoxides: The reaction between D1MP and epoxides may produce hazardous conditions.
Mineral Acids: Excessive strong mineral acids can cause decomposition of DIMP to yield primarily
alcohol and methylphosphonic acid.10
Organic Peroxides: There is very little available information on the reaction of DIMP with organic
peroxides. The reaction between the organic peroxides and DIMP may produce hazaidous conditions.
Oxidizing Agents: The exhaustive oxidation of DIMP can yield toxic and corrosive fumes of oxides of
phosphorus, sulfur, nitrogen, and heat.
Oxidizing Mineral Acids: Excessive oxidizing acids can decompose DIMP to yield heat and toxic fumes
of nitrogen oxides, sulfur oxides and phosphorus oxides.
Reducing Agents: For information on the reducing agenu, see alkali and alkaline earth metals above.
Water Reactives: The water reactive materials may react with DIMP to produce highly unstable
mixtures, heat and toxic and/or flammable gases.
ENVIRONMENTAL FATE: Organophosphorus compounds such, as DIMP are subject to biological and
chemical degradation upon entering the natural environment.11'3' The ultimate degradation product is
orthephosphoric acid (H3PO4) or orthophosphate salts. Chemical degradation occurs primarily through
hydrolysis. The hydrolytic behavior of phosphate diesters such as DIMP is similar to that of the
corresponding phosphate triesters1 11,51 while the hydrolytic behavior of the phosphonate inonoester
parallels tliat of the equivalent phosphate diester.10 Under alkaline conditions, DIMP hydrolyzes much
more rapidly to produce the monoisopropyl "ester than the monoester does to produce methylphosphonic
ucid. DIMP and the inonoester hydrolyze at aoproximatcly the same rates under acid conditions.
Isopropyl rnethyiphosphonate is very stable under neutral conditioa*. The primary products anticipated
upon ciipiplete hydrolysis of DIMP are isopropyl alcohol (flammable, low boiling liquid),
methylphesphcnic acid, and various amounts of isopropyl rnethyiphosphonate, depending on
environmental conditions and the length of time. The rate of chemical hydrolysis of isopropyl
rnethyiphosphonate to produce methylphosphonic acid and ultim? :,y phosphoric acid or its salts may be
very slow, especially under alkaline conditions. However, hydre’ysis rate can be greatly accelerated by
the presence of microorganisms, enzymes, and other factors the environment, DLvlP's hydrolysis
product, methyl phosphonic acid, is a very stable compound. e compound can be recrystallized from
fairly strong hydrochloric acid or heated in boiling «udiui:. hydroxide for several hours without
change.12 Methyl phosphonic acid and other phosph«*e derivatives are susceptible to further
degradation by photolysis with sunlight and ultraviolet radiation to yield phosphonic acid derivatives.
Phosphpnat.es may be assimulated *.n<J subsequently serve as a sole source of phosphorus for aquatic
plants.11'01 Methyl phosphonic acid is also very slowly oxidized by ozone to orthophosphonic acid,
carbon dioxide, and water. DIMP14 and its hydrolysis products are water soluble and isopropyl alcohol is
nighly volatile. These factors would facilitate their disposition in the environment.
Appendix 2
CONCLUSIONS: DIMP is an irritant of the eyes, nose, skin, and repiratory tract, and prlonged
inhalation and skin contact should be avoided. Since it is not known whether DIMP is a carcinogen,
personnel should take extra precautionary measures, and wear protective clothing, rubber gloves, and an
approved respirator. It is suggesteu that tne data gaps concerning the phytotoxicity and carcinogenicy
of DIMP be investigated in the future.
REFERENCES:
1. Rosenblatt, David H. et. al., Problem Definition Studies on Potential Environmental Pollutants:
Physical, Chemical, Toxicological, and Biological Properties of 16 Substances, US Army Medical
Research and Development Command, Forrestal Building, 1975.
2. Allen, Craig R.f The Relationship Between Oxygen Index and tne Flashing Propensity of
Explosively Disseminated Liquids, October 1977.
3. Code oi Federal Regulations, Vo]. 49, Parts 100-177, US Government Printing Office, Washington,
DC, 1981.
4. Code of Federal Regulations, Vol. 40, Parts i 90-399, US Government Printing Office, Washington,
DC, 1981.
5. Jacobson, Keith, H., The Acute Toxicity of Some Intermediates in GB Manufacture, Chemical
Corps Medical Laboratories Special Report, February 1953.
6. Aulerich, R. 3., Coleman, T. H., Polin, D., Ringer, R. K., Howell, K. S., Toxicology Study of
Diisopropyl Methylphosphonate and Dicyclopcntadiene in Mallard Ducks, BoDwhite Quail and Mink,
Michigan State University East Lansing Department of Poultry Science, DAMD17-76-C-6G54, April
76 - June 79.
7. Palmer, 3. S. ct. al., Toxicologic Evaluation and Fate of Diisopropyl Methylphosphonate (DIMP)
and Dicyf'lopen.tadiene (DCPD) in Cattle, Science and Education Administration College Station
TX Veterinary Toxicology and Entomology Research Laboratory, March 77-Sep 79.
8. Hart, E. P.., Mammalian Toxicolcgical Evaluation of DIMP and DCPD, Government Reports
Announcements and Index (GRA 1), Issue 15, i960.
9. A Method for Determining the Compatibility of Hazardous Wastes, April i 980.
10. Keary, Leonard, Canadian Journal of Chemistry, Vol 43, pg 2637, 1965.
11. Griffith, ‘E. 1., et. a!., Environmental Phosphorus Handoook, John Wiley and Sons, New York,
(a) pg 250; (bj pg. 259; (c) pg 242, 1973.
r2. Corbridge, D. E. C., Phosphorus, An Outline of its Chemistry, Biochemistry and Technology;
Elsevier Scientific Publishing Co., New York, pg. 204, 1978.
i,
13. Libby, R. A., Inorganic Chemistry, Vol 10, No. 2, pg. 386, 1971.
14. Coon, Phillip, Research Division, Chemical Research and Development Center, Aberdeen Proving
Ground (APG), MD, July 1983.
b u
Appendix 2
IDENTIFIER
DMMP
Dimethyl methylphosphonate
75b-73-t>
CHEMICAL FORMULA: (CH30)2P(0)Cri3
SYNONYMS: Methanephosphonic acid, diinethyl ester; dimethyl methane-phosplioruitc;
dimethoxy mcthylphosphinc
DISCR1PTORS: DMMP belongs to a group of stable organophosphorus esters known as the dialkyl aJkyl-
phosphonates. It is classified as a diester of meli.yipncsphonic acid.
CHEMICAL AND PHYSICAL PROPERTIES: DMMP is a clear, colorless, mobile liquid witn a very rnild
characteristic odor. The compound is miscible with water, alcohols, esters and aromatic solvents, but
immiscible in aliphatic hydrocarbons. Selected chemical and physical parameters are listed below:
Property
Value (Ref)
✓
Property
Molecular weight
124.1 (1)
Flash point
Boiling point
X .
181°C, 54 mm Hg (4)
Viscosity
(Centistokes)
Melting' point
below -50°C (2)
Vapor pressure
Specific gravity
1.174, 20°C(i)
Volatility
Vapor specific gravity
*4.3 (3)
Solubility
note: ‘estimated values
/ Value (Rei)
1Q4.4°C; Open
Cup
(Cleveland) (1)
1.81, 25°C (1)
*0.61 mm Hg,
20^C (3)
4100 mg/m^ (3)
miscible
MILITARY APPLICATION: DMMP is extensively used as the simulant for simulating non-persistent
chemical agents. The Gornpound is a volatile agent simulant, and is used in vehicle
penctration/vulnerabiiity studies, protective mask filter element quality assurance tests, freon
decontamination tests, chemical units, and teams decontamination capabilities studies (with K123
thickener), aircraft spray tank dissemination tests (with K125 thickener), and the shelter vulnerability
tests.
INDUSTRIAL APPLICATION: DMMP is used quite extensively in industry as a flame retardant additive
and viscosity depressant in resins, such as unsaturated polyesters and epoxies. It is also used in heavy
.nctal extraction, solvent separation, preignition additive for gasoline, as an anlifoam agent, plasticizer
and stabilizer, textile conditioner and antistatic agent, and as an additive in solvents and low
temperature hydraulic iluids.
STORAGE, SHIPPING, AND HANDLING: DMMP is classified as a combustible liquid as defined in the
US Department of Transportation (DOT)'-49 CFR 173.113 (b)/ and all storage, shipping, and handling
procedures must be in accordance with the regulations therein. DMMP is not specifically listed as a
tazardous substance, and its flash point (1Q4.4.°C) and oral toxicity value (130 mg/kg) do not qualify it as
hazardous waste as defined in 40 CFR 261. 21. 6 The compound has been reported in the Environmental
. rotection Agency (EPA) Toxic Substances Control Act (T5CA) inventory since i*80/
1
Appendix 2
TOXICOLOGY: Acute toxicity of DM UP.
ROUTE
SPECIES
DOSE
EFFECTS/RF.MARKS/(Ref)
Oral
Rat
150 mg/kg
(in corn oil)
LD^q (H)
Rat
>4640 ing/kg
ld50 (9)
intragastric
Rat
■>3000 rng/kg
o
Q
intraperitoneal
White Leghorn
Hen
50 mg/kg
lowest dose that produced visible
detectable ataxia, produced no
delayed neurotoxic activity (8)
Intraperitoneal
Mouse
250 J*l/kg
0/10 died in 24 hr, 2/10 died in 7 days,
weakness ataxia, prostration (10)
Inhalation (total
exposure)
Mouse
3900 mg min/m^
(Time = 10) (nominal
concn. 77 ppm)
0/10 died in 10 days (Benesh machine)
(ID
Percutaneous
Rabbit
> 4740 mg/kg
LD50 (9)
Subcutaneous
Mouse
50 mg/kg
300 mg/kg
100 mg/kg
0/2 died in 10 days (12)
0/2 died in 10 days (12)
0/2 died in 10 days (12)
Eye Irritation
Rabbits
4740 mg/kg
nonirritant (9)
Skin Effects
Human
240 1/kg
no effect (11)
Carcinogenicity:
Little*^ reported
that DMMP had no effect in asays which were indicative of
dioxyribonucleic acid (DMA) damage or measured neoplastic transformation.
Mutagenicity: According to Little, DMMP produced no mutagenic responses in the Ames Salmonella
mutagenicity assay:. - •
Teratogenicity: DMMP is currently being tested for teratogenic acitivity in a study conducted urder the
auspices of the National Toxicology Program, National Institute of Health. No official data has been
released on the teratogenicity of DUMP.
Health Hazards: DMMP causes irritation of the eyes, skin, and respiratory tract. According to
Dunnik,14 tlie compound was toxic to the reproductive system of male rats. The author reported that
with increasing doses of DMMP, the number of pregnancies decreased, the mean litter size decreased,
and the percent of resorptions increased. The male rats showed some weight decrease, and at high doses
DMMP showed an increase in the number of abnormal sperms. 4 Presently, no Threshold Limit Value
(TLV) has been established for DMMP in humans.
Plant Data: Libby, 5 indicated in a recent publication that phosphonates will undergo photolytic
reactions with sunlight to produce orthophosphates. The orthophosphates tend to serve as a sole
phosphorus source to aquatic plants.
Appendix 2
CHEMICAL REACTIVITY:16
Alkali and Alkaline Earth Metals: When DMMP is mixed with these metals, an exothermic reaction may
occur.
A 7.0 Com pounds: DMMP may react with azo compounds to produce hazardous conditions. However,
Tittle imormation is available on these conditions.
Caustics: DMMP is slowly hydrolyzed under alkaline conditions to produce an alkali salt of
methylphosphonic acid and methyl alcohol.18
Non-oxidizing Mineral Acids: The non-oxidizing mineral acids can hydrolyze DMMP to highly flammable
'methyl alcohol (flash point: li°C) and methylphosphonic acid, a fairly strong acid (first ionization
constant (pkj) 2.3 20 - 25°C. 17
Organic Peroxides: Mixing DMMP with organic peroxides may create hazardous conditions; however,
little information is available.
Oxidizing Agents: Exhaustive oxidation of DMMP can yield toxic and corrosive fumes of oxides of
phosphorus and other toxic compounds.
Oxidizing Mineral Acids: Excess oxidizing mineral acids can decompose DMMP to yield toxic fumes
such as nitrogen oxides, sulfur oxides, and phosphorus oxides.
Reducing Agents: Dialkyl alkylphosphonates in general are resistant to reducing agents. Materials such
as sodium or aluminum amalgam have little effect. Stronger reducing agents do react, but information
is scant. An exothermic reaction may occur especially if the DMMP contains some water.
Water Reactives: DMMP can react with water reactive materials to yield heat along with toxic and/or
flammable gases.
ENVIRONMENTAL FATE: "Having entered the natural environment, organophosphorus compounds are
degraded by biological and/or chemical reactions to orthophosphate, the ultimate degradation
product."1® Hydrolysis is the primary chemical procedure for degrading organophosphorus compounds
entering the environment. Phosphonate diesters (such as DMMP) are similar to phosphate triesters in
their hydrolytic behavior.1® The lower molecular weight dialkyl alkylphosphonates are moderately
resistant to hydrolysis; however, hydrolysis will occur both under acidic, and less rapidly, under alkaline
conditions. The primary products anticipated upon exhaustive hydrolysis are methyl alcohol and
methylphosphonic acid or its salts (alkaline hydrolysis). Methyl alcohol is very volatile (boiling points
64.5°C), and completely miscible with water; therefore, it would have little tendency to accumulate in
the environment. The lower alkyl phosphonic acids such as methyl phosphonic acid are hygroscopic
(absorbs water from the atmosphere) white crystalline solids. Methyl phosphonic acid melts at
105°C. ,19 Methyl phosphonic acid is a fairly strong acid (pkj: 2.3 at 20 to 25°C), and it would tend
to form water soluble salts in an alkaline environment, i nerefore, both the free methyl phosphonic acid
and its salts would be washed away over a period of time. Dialkyl alkylphosphonates in general are
resistant to reaction with oxygen and oxidizing agents.
CONCLUSIONS: DMMP has been reported as an irritant ol lire eyes, skin, and possibly a runspeaiic
irritant of the upper respiratory tract. In addition, the compound causes sterility of the reproductive
system in male rats. The compound Iras been reported to produce no mutagenic responses in the ames
Salmonella Assays, and nad no effect in assays which were indicative of L)N/\ damage or ineusured
neoplastic transformation. However, personnel should avoid contact by wearing protective doming,
rubber gloves, and an approved respirator since the compound produces sterility in male rats, and the
teratogenic effects are not available.
REFERENCES:
1. Mobile Chemical Company, Product information bulietin, DialkyJ Alkylplrosphonates, Industrial
Chemicals Division, page 3.
2. Toxicology Laboratory Report T-4 1 25, Stauffer Chemical Company, Western Research Center,
Westport Connecticut, 06880.
3. Lyman, W. J.; et. al., eds.. Handoook of Chemical Property Estimation Methods, New York:
McGraw-Hill Book Company, 19X2.
4.. Tomlinson, G. J., and A. H. Samuel, Literature Survey of Physical and Chemical Properties ol
Agents VX, GD, HD, and HL, Vol 1, Final Report, July 19X0. Chemical Systems Laboratory
Contractor Report, ARCSL-CR 80051 (batteile).
5. Code of Federal Regulations, Vol 49, Parts 100-177, US Government Printing Office, ‘Washington,
DC, 1981.
6.. Code of Federal Regulations, Vol 40, Parts 190-399, US Government Printing Office, Washington,
DC, 1981.
7. Lewis, Richard J., and Rodger L. Tajtken, Registry of Toxic Effects of Chemical Substances, US
Department of Health and Human Services, February, 19X2.
8. Hollingshaus, J. G.; et., al., Delayed Toxicity and Delayed Neurotoxicity of Phosphorothioate and
Phosphorothioate Esters, Journal of Toxicol. Environ. Health 8: 619-627, 1981.
9. Morey, H. W. Jr., "Toxicology Data on Fryol DMMP," Letter, Stauffer Chemical Company,
Specialty Division, Westport, Connecticut, 6 Aug 1980.
10. Jones, Jr., H. W., et. al. The Relationship of Cholinesterase Inhibiting Activity to the Toxicity of
Some Organic Phosphorus Compounds, Medical Division Reports no. 134, p. 11, April 1948.
11. Ceiling, E. M. K. et al., (Compiled by HD Young) Division 9, National Defense Researcn
Committee Office of Scientific Research and Development, OSRD No. 4176, Status Report on
Toxicity and Vesicant Test of Compounds Referred to the University of Chicago Toxicity
Laboratory Aug 1, 1944, Oct 3, 1944. Unclassified Report.
4
12. The University of Chicago Toxicity Laboratory, Informal Monthly Progress Report on Toxicity and
Irritancy of Chemical Agents, Informal Report No. N.S. 1, p. 39, april 15, 1945.
13. Little, Arthur D., Evaluation of Dimethyl Methylphosphonate and Exo-Tetrahydrodi-
(Cyclopentadiene) in a battery of in Vitro Short-Term Assays, Air Force Aerospace Medical
Research Laboratory, 1983.
14. Dunnick, June, Personal Communication, NIEH5, North Carolina, 1982.
15. Libby, Robert A., The Photolysis of Two Diphosphonates, Inorganic Chemistry, Vol 10, No. 2, 1971.
16. A Method for Determining the Compatibility of Hazardous Wastes, EPA-600/2-80-076, April 1980.
17. Van Wazer, J. R., Phosphorus and Its Compounds, Vol 1, Intcrscicnce, New York, 1958.
18. Griffith, £. Jr., et al., Environmental Phoshporus Handbook, John Wiley and Sons, New York, 1973.
19. Corbridge, D. E. C.; Phosphorus, An Outline of the Chemistry, Biochemistry and Technology;
Elsevier Scientific; New York; 1978.
APPENDIX 3
PERFORMANCE OF SELECTED COLUMNS
FOR GAS CHROMATOGRAPHIC ANALYSIS
Performance of Selected Columns for Gas Chromatographic
Analysis
1. 15% DEGS on 80/100 chromosorfc WAW, 6' x 0.125" O.D., 0.085" I.D.,
stainless steel; Date: 4/29/77; max. temp. 200 C.
Prior use: unknown.
Most recent period of use: from ? to 9/23/83.
This column was replaced after difficulties had been encountered in
obtaining a stable offset. Looking at the chromatograms in retrospect, it
may be that the collector was the actual source of the problem, although
deterioration of the column cannot be ruled out entirely. The column was
operated at oven temperatures from 150 to 175°C.
2. HP test column: 2% OV 101 .m 100/120 chromosorb W HP, 4' x 2 mm I.D.,
1/4" O.D., glass; no date; max. temp. 350°C.
Prior use: Received with the N-P detector and used during initial checkout
Most recent period of use: from 11/3/83 to 11/10/83.
This column was conditioned at 250°C overnight. The peak areas for DIMP
were not reproducible. When operated with an oven temperature of 150 C,
peaks showed unacceptable tailing. This was eliminated by raising the
"temperature to 175°C, but at this temperature, the separation between the
negative acetone peak and the DIMP peak was insufficient. Integration of
the DIMP peak therefore started before the response had returned to
baseline. Reduction of the slope sensitivity lead to good positioning of
the end integration mark, but did not ameliorate the peck separation
problem at the start.
3. 10Z Carbowax 20M on 80/100 chromosorb W HP, 6' x 0.125" O.D., 0.085"
I.D., stainless steel; Date: A/29/77; max. temp. 225°C.
Prior use: not known precisely. Reasonable results were obtained in the
past with this column.
Most recent period of use: 9/28/83 to 11/1/83 (from 9/29 to 10/12, the GC
was not used because the oven heating element had burned out).
Initially, the system seemed to perform acceptably with this column.
However, subsequent difficulty in obtaining a stable offset resurfaced. In
addition, injection of acetone resulted in dramatic drops in the offset.
There was no substantial reason to doubt the integrity of the collector and
a test comparing demoisturized acetone to untreated acetone yielded no
indication that the solvent was the source of the problem. Column bleed
with no other changes in the system was a possible explanation. Subsequent
installation of solid phase columns resulted in a quite stable offset,
which suggests that this column was not in good condition. It was operated
at 150°C throughout this installation period.
4. 10Z Carbowax 20M on 80/100 chromosorb W-HP, 2 m x 0.25" O.D., 2 mm I.D.,
glees; Date: 11/8/83; max. temp. 225°C.
Prior use: none.
Because the Carbowax column described above had performed adequately in the
past, a new column was employed. Use of the glass column allowed on column
injection. Operating with a column temperature of 165°C and carrier gas
flow of 30 ml/min, the DIMP retention time was about 1.8 min. Peaks were
reproducible and well formed. Satisfactory performance of this column
resulted in its use for. all gas chromatographic analysis of DIMP samples.
5. 80/100 Porapak QS, 6' x 0.125" O.D., stainless steel; Date: 10/24/83;
mak. temp. 250°C.
Prior use: none
Most recent period of use: 11/1/83 to 11/2/83
This solid phase column was tried since it would eliminate the possibility
of column bleed. A stable offset was obtained. With injections of 0.2 and
10 mg/1 DIMP in acetone, the acetone peak spread from 3 to 6 minutes and no
other peaks were observed after a 20 minute wait. Rather than spend time
experimenting with various temperatures a second solid phase column
containing Tenax was installed.
6. 80/100 Tenax GC, 6' x 0.125 O.D., stainless steel; Date: 10/24/83, max.
temp. 375°C.
Prior use: none
Most recent period of use: 11/2/83 to 11/3/83
The column was conditioned at 300°C. The acetone peak appeared at about
0.75 minutes with an oven temperature of 200°C. No DIMP peak was evident,
but after 20 minutes and oven temperature increases to 290°C, a low bump
did appear which may have represented the DIMP.
8-85