United States
Environmental Protection
Agency
Environmental Monitoring and SupporT
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S4-84-003 Jan. 1984
Project Summary
Direct Determination of Total
Organic Chlorine in Water
Without Preconcentration
Judith A. Gebhardt
The goal of this research effort was to
develop instrumentation for the direct
determination of total chlorinated or-
ganic compounds in aqueous samples
without preconcentration. Two general
approaches were investigated. The first
involved isolation of the chlorinated
organics from the sample matrix by
flash evaporation. A catalytic conver-
sion step was to produce free chlorine
which was to be detected and quantified
fluorometrically. This approach did not
prove successful. An alternate proce-
dure was investigated which used a
piezoelectric crystal as the detection
device. After evaluating a number of
crystal coatings, Amine 220 was found
to have appropriate characteristics.
This material, a gas chromatographic
stationary phase, has a high affinity for
hydrochloric acid (HCI) and almost none
for sodium chloride (NaCI), which is a
major interference in total organic chlor-
ide (TOCI) or total organic halogen
(TOX) determinations. In this procedure,
the sample was delivered to the detector
system by a direct injection of the
aqueous sample through a quartz tube
which is heated to approximately
6000°C. Use of oxygen as the carrier
gas created the oxidizing environment in
which the chlorinated organics are con-
verted to HCI.
The piezoelectric crystal coated with
Amine 220 was found to have an absolute
detection limit of approximately 100 pg
of HCI. Assuming a minimum of 50
percent conversion efficiency, approxi-
mately three parts per billion (weight/
volume) of chloroform should be detec-
table in a 100 pL aqueous sample.
Difficulties were encountered in remov-
ing water from the system before it
reached the detector system. Large
amounts of water in contact with the
Amine 220-coated piezoelectric crystal
saturated the detector. Several modifi-
cations in the sample delivery system
were investigated to remove the water.
None were entirely successful, and
additional work is necessary before the
system is ready for further field evalu-
ations.
This Project Summary was developed
by EPA's Environmental Monitoring and
Support Laboratory, Cincinnati, OH, to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Description of the Approaches
The first system investigated involved:
(a) flash evaporation to separate chlori-
nated organics from salts; (b) oxidation of
the chlorinated organics to free chlorine
in a catalytic combustion furnace; (c)
derivatization of the free chlorine with
syringealdazine; and (d) fluorometric
determination.
Two flash evaporators were evaluated
in approach 1. The first was a simple
horizontal, quartztube operated at400°C
at pressures varying from atmospheric
down to 100 millitorr. Flash evaporation
of a 250 ppm solution of NaCI resulted in
a high carryover of chloride ion. In an
attempt to minimize chloride ion carry-
over, a second flask evaporator was
designed incorporating a gravity separator
and glass beads to facilitate the removal
-------
of chloride vapor and particulate. This
unit was evaluated at 400°C under
ambient and subambient pressures using
mixtures of Na36CI, 14CHCI3, and 14C-
PCB's. The experiment demonstrated that
the subambient pressures required for
good PCB recovery resulted in poor
recovery for CHCIa.
Ambient pressures resulted in good
recoveries for CHCIa but poor recovery for
PCB's. Under the conditions studied, the
carryover 36CI was unacceptably high.
Syringealdazine reacts with hypo-
chlorus acid to form a purple chromogen
that absorbs in the visible region with a
maxima at 530 nm. The detection limit
using absorption maxima was too high to
be useful for this application. Since the
unoxidized reagent exhibits strong fluo-
rescence at 530 nm at a pH of 8 or greater
and at a wavelength at 420 nm at a pH of
4, it was anticipated that the oxidation
product would also exhibit strong fluo-
rescence properties. Itwasfound that the
oxidation product is unstable at a pH of
less than 6 or greater than 8. After
extracting and stabilizing the oxidation
product, no evidence of fluorescence for
the colored product was found.
The second system investigated invol-
ved: (a) introduction of the sample by
direct injection; (b) conversion of the
organochlorine compounds to HCI by
non-catalytic oxidation; (c) removal of
water from the reaction stream; and (d)
detection of the HCI using a piezoelectric
crystal coated with a reagent that prefer-
ably adsorbs acid halides.
The frequency at which a piezoelectric
crystal oscillates is proportional to its
mass. A piezoelectric crystal detector was
assembled where the flow from the
combustion/absorber train flowed across
two piezoelectric crystals with nominal
frequencies of 9.0 MHz. One crystal was
coated with a chemical reagent that
reversibly sorbs HCI while the other
remained uncoated. A differential fre-
quency counter monitored the changing
frequency between the two crystals re-
sulting in a signal proportional to the
increased mass of the coated crystal as
HCI was sorbed.
Ten different crystal coatings were
evaluated in order to determine which
coating is best suited for detecting HCI.
The test was conducted by introducing
1-mL injections of reagent water and
reagent water containing 500 ppm of HCI
into a heated flask containing sulfuric
acid. The resulting volatilized anhydrous
HCI was swept into the detector by a
nitrogen carrier gas. Table 1 presents the
response of the crystal for each coating
evaluated compared to the response to
the same volume of reagent water.
Quadrol, 4-aminoacetanilide, triethyl-
enetetramine, and L-ornithene showed
little or no response to HCI. The response
of triphenylamine and triethanolamine
were irreversible; that is, HCI was ad-
sorbed but not released. After five repli-
cate injections, the crystal coated with
trimethylamine-HCI showed no further
response. The substrate had become
either degraded or saturated. Three coat-
ing materials, tri-n-octylamine, tri-n-
hexylamine, and Amine 220, gave a good
reversible response to HCI injected.
Amine 220 resulted in the largest change
in the crystal frequency and showed good
adherence to the crystal. This substrate
was selected for further testing.
The 9 MHzcoated with Amine 220 was
exposed to several levels of 862 to
determine if sulfur-containing compounds
would cause interferences in the detec-
tion of HCI. The results of this experiment
are presented in Table 2. Repeated injec-
tions of aqueous solutions of HCI contain-
ing as much as 500 ppm NaCI (weight/
volume) showed no effect from the
presence of the inorganic salt.
Because of the sensitivity and speci-
ficity of the crystal with the Amine 220
coating, it was chosen for use in this
system and used in all other studies.
Amine 220 is a commercial product
available through Pfaltz and Bauer,
Stamford, Connecticut. Itschemical name
is 2-(8-heptadecenyl)-2-imidazoline-1 -
ethanol.
The next task was to interface the
detector to the furnace in which HCI (or
HBr or HI) was to be generated from
halogenated organics in the aqueous
sample. Initial investigations involved
injection of aqueous solutions into a
heated quartz tube coupled to the detector
eel I by simple ground glass fittings. Using
this configuration, injections of laboratory
air produced large responses. It was
determined that this was due to "thermal
shock"; that is, rapid expansion of the
injection in the heated quartz tube. Mov-
ing the detector cell further away to
minimize this effect resulted in condensa-
tion of water on the crystal. Heating the
crystal to prevent condensation caused
the Amine 220 to bleed from the crystal
which changed the response characteris-
tics and caused a serious drift in the
baseline. It was necessary, therefore, to
introduce a component in the system
between the oven and the detector which
would remove water from the carrier gas
stream.
Table 1. Results of Coating Selection Experiments
Coating Material
Detector Response"
1 mL H2O
Detector Response'
1 mL H20 + 500 fjg HCI
Quadrol
Triphenylamine
Triethanolamine
4-Aminoacetanilide
Triethanolamine-HCI
Tri-n-octylamine
Tri-n -hexylamine
Triethylenetetramine
L-Ornithine-HCI
Amine-220
9
3
JO
17
3
1
3
3
18
18
15
3
40
19
6
280
110
3
21
960
'Change in frequency (Hz/.
Table 2. Results of Exposure of Amine 220-Coated Crystal to SOt
Amount SOi
Injected ppm (v/v)
Detector Response*
5 mL
% Frequency (Hz)
5 mL + S02
100 II.3 pg)
1.0(0.013 pg)
0.1 (0.001'3 pg)
360
360
360
368
364
363
* This represents an injection ofSmL of laboratory air. No attempt was made to remove chlorinated
compounds from the sample. Response is presumed to be due to background levels of HCI in the
laboratory.
-------
The first approach investigated was the
use of a trap containing a desiccating
material between the oven and the detec-
tor. Absorbance of water by calcium
carbide resulted in the formation of
acetylene, which seemed to strip the
coating from the crystal; therefore, this
approach was abandoned.
The applicability of phosphorous pen-
toxide was evaluated next. Although
water was effectively removed by this
technique, HCI wasalsotrapped. Heating
the trap to 95°C did not result in the
release of HCI, and this technique was
also abandoned. Anhydrous magnesium
perchlorate was investigated next. Water
and HCI were both trapped and HCI was
released slowly when the trap was heated
to 195°C. The response of the detector
using this material was not reproducible.
Water and HCI were both irreversibly
absorbed by calcium sulfate trap.
Since trapping water on a solid material
was not successful, the applicability of a
concentrated sulfuric acid scrubber was
investigated. The first scrubber configura-
tion which was evaluated was construc-
ted from a vertical glass tube packed with
glass beads coated with sulfuric acid. The
breakthrough capacity of the scrubber
was evaluated by placing a tube contain-
ing calcium sulfate downstream of the
scrubber and repeatedly injecting 0.1 mL
of distilled deionized water into the
system. The glass beads were coated
with fresh H2SO.4 after each injection.
After 20 injections totaling 2 g of water,
the calcium sulfate tube showed no
weight gain. This demonstrates that the
efficiency of the scrubber for water was
greater than 99.9 percent. Injections of
aqueous solutions of HCI yielded a mini-
mal response. When the output of the
detector was displayed on a chart re-
corder, a low broad peak was observed.
This suggests that the HCI was being
absorbed by the scrubber and gradually
bled off by the action of the carrier gas.
On the assumption that this problem
was caused by the large dead volume of
the scrubber, another configuration was
designed and tested. This scrubber was
constructed from a micro volume imping-
er filled with sulfuric acid. A significant
detector response was observed with the
injection of 0.1 ml of distilled, deionized
water indicating that the capacity of the
scrubber had been exceeded. Use of a
larger volume scrubber produced the
same results. Therefore, a modification of
the original system mustbeevaluatedfor
incorporation into the instrument.
Finally, the efficiency of the conversion
of chlorinated organics to HCI was investi-
gated. Using an oven temperature of
600°C and an oxygen flow rate of 60
mL/min, aqueous solutions of HCI and
CHCIa were injected into the system.
Chloroform solutions were prepared at
one-third the concentration of HCI solu-
tions so that injections of equal volumes
should produce the same response. Over
the three orders of magnitude that the
system was evaluated, the response to
chloroform was consistently 10 percent
of that expected based on analysis of HCI.
To improve the sensitivity of the instru-
ment, the efficiency of the conversion
step must be improved.
Conclusions
Funding constraints made it impossible
to evaluate additional parameters and
assemble a working unit.
At the time work was terminated, the
coated piezoelectric crystal detector had
been demonstrated to be sensitive enough
to fulfill the original grant objectives.
Coated with Amine 220, the crystal is
capable of detecting approximately 100
pg of HCI. The detector was also demon-
strated as being insensitive to as much as
500 ppm (weight/volume) of sodium
chloride and as much as 1.3 fjg SO*.
With this detector system, sample pre-
concentration is not required and the
analysis from start to finish is only about
five minutes. Additional development of
this system may provide USEPA with a
significant improvement in TOX method-
ology and an instrument which provides
rapid, reliable data on the levels of
chlorinated organics in aqueous samples.
-------
Judith A. Gebhardt is with Gulf South Research Institute, New Orleans. LA
70186.
Stephen Billets is the EPA Project Officer (see below).
The complete report, entitled "Direct Determination of Total Organic Chlorine in
Water Without Preconcentration," (Order No. PB 84-129 121; Cost: $8.50,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
*U.S. GOVERNMENT PRINTING OFFICE: 1984-759-015/7293
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use S300
------- |