AN EVALUATION OF THE DETERMINATION OF TOTAL ORGANIC CHLORINE (TOC1)
IN WATER BY ADSORPTION ONTO GROUND GRANULAR ACTIVATED CARBON,
PYROHYDROLYSIS, AND CHLORIDE-ION MEASUREMENT
by
Ronald C. Dressman
Earl F. McFarren
James M. Symons
Prepublication Copy
of
A paper presented at the Water Quality Technology Conference
at Kansas City, Missouri, Dec. 5 and 6, 1977
WATER SUPPLY RESESEARCH DIVISION
U.S. ENVIRONMENTAL PROTECTION AGENCY
26 W. ST. GLAIR STREET
CINCINNATI, OHIO 45268
-------
An Evaluation of the Determination of Total Organic Chlorine
(TOC1) In Water by Adsorption onto Ground Granular
Activated Carbon, Pyrohydrolysis, and
Chloride-Ion Measurement
Ronald C. Dressman, Earl F. McFarren, James M. Symons
Water Supply Research Division
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
INTRODUCTION
Total organic chlorine (TOC1) is a group parameter that is intended
to be a. measure of the chlorinated organics present in water. Because
chlorinated organics are not naturally occurring, and because many
chlorinated organics are known to be seriously detrimental to the health
of the water consumer, a TOC1 measurement has been suggested as an
extremely useful indicator of water quality.
The first serious attempts to measure TOC1 on a routine basis as an
indication of water quality was done on the Rhine River. Kuhn, Fuchs,
1 2
and Sontheimer ' collaborated at the Engler-Bunte Institute in Karlsruhe,
Federal Republic of Germany to develop a "pyrohydrolysis" technique for
the determination of the total organic chlorine value of compounds
adsorbed onto activated carbon.
Their initial work was directed at a method for determining the
TOC1 concentration adsorbed onto samples of granular activated carbon
(GAC) taken from full-scale activated carbon beds or sampler units. The
principal goal at that time was to evaluate or characterize the efficiency
and life of the GAC beds being used in the water treatment process to
remove organic contaminants, although the sampling system was used in an
attempt to determine TOC1 concentrations in water.
-------
2
Later, the granular—activated—carbon—sample method was modified for
the analysis of water samples for concentration of TOC1 (unpublished).
For this determination granular activated carbon ground to produce a
finely divided activated carbon (GGAC) is used to adsorb organo—chlorine
compounds from discrete water grab samples. The actual method of
pyrohydrolysis of the carbon is the same for both procedures.
GENERAL METHOD DESCRIPTION
Concentration
The concentration of organochlorine compounds onto the ground
granular activated carbon is accomplished by adding one gram of Filtra—
sorb 400* (or its equivalent), ground 90 seconds in a laboratory mill
and wetted with Super—Q water for 30 mm under vacuum, to 10 liters of
water. The sample is stirred for one hour and then the carbon is
flocculated by the addition of 50 m2. of aluminum sulfate solution (conc.
1 gm Al/i) and 20 m2. of a polyacrylamide floc aide (magnifloc 985 N,
Cyanamid — conc 200 mg/i). Before the addition of the floc aide, pH is
adjusted within the range of 6.5 to 7.5 with 5 N sodium hydroxide. After
settling, the carbon sludge Is recovered on a 47 mm diameter 8 pm pore
size cellulose acetate membrane filter (Millipore) and transferred to a
quartz boat for pyrohydrolysis. This procedure is performed twice on
each sample.
Pyrohydrolysis
Pyrohydrolysis is accomplished by combusting the carbon and its
charge of adsorbed compounds in the presence of steam and oxygen, and
condensing the steam containing the combustion product, HC1, for later
measurement of its chloride ion concentration.
*?4ention of trade names does not consititute endorsement by the U.S.
Environmental Protection Agency.
-------
3
The carbon sample is placed in a quartz boat and inserted into the
first of a series of two pyrohydrolysis furnaces (Figure 1). The furnace
into which the sample is placed has an initial temperature of less than
100°C. The second furnace is maintained at 1,000°C. They are served by
a coimnon, quartz pyrolysis tube that contains a plug of quartz wool in
the zone of the 1,000°C furnace. A flow of oxygen and steam is passed
through the pyrolysis tube at a rate of 200 m9 /min 02 and a condensation
rate of about 0.8 m2. water per minute. At the same time the first
furnace is heated to reach 700°C within 15 to 20 minutes.
In the first furnace the organic compounds are volatilized and the
carbon eventually reduced to an ash. As the volatilized organochiorine
compounds pass through the second furnace, set at 1,000°C, they are
combusted to HC1. This combustion in the presence of the steam is
termed pyrohydrolysis. The process takes about one hour, in which time
about 50 m of pyrohydrolyzate is collected in a beaker.
The oxygen, of course, provides for the oxidative pyrolysis of the
organic compounds to produce the HC1, and at the same time serves as a
sweep gas. The steam aids in preventing Cl from being adsorbed on the
walls of the system, and is also the solvent in which Cl is solubilized
for subsequent detection and measurement. Measurement can be by bench
titration, use of a select, chloride—ion probe, or inicrocoulometry,
choice to be discussed later.
-------
02 200 mI/mm
FIGURE 1:
SCHEMATIC OF PYROHYDROLYSIS SYSTEM
H 2 0
HEAT TO DELIVER
0.8 mI/mm
-------
5
Inorganic Chloride Interference
Whether the activated carbon has been taken from an operating
filtration system or recovered from a water sample, inorganic Cl is
adsorbed on the carbon. For water grab samples, the amount of inorganic
ion adsorbed onto the carbon varies with the concentration of the ion in
the water. According to data published by Kuhn et al. 1 at concentrations
of 50, 100, 400 and 1,000 mg/i, Filtrasorb 400 will adsorb the following
amounts of Cl respectively in mg per gm of activated carbon: 0.65,
1.03, 1.69 and 1.85.
Accounting for this contribution of Cl from the inorganic source
is very important, especially when the ratio of organic Cl to inorganic
Cl does not overwhelmingly favor the organic contribution, as is usually
the case. Kuhn et al. make this correction on water grab samples by
extracting a duplicate water sample. Instead of pyrohydrolyzing the
carbon sludge recovered from the duplicate sample, it is in turn extracted
with 200 m2. of 0.1 N NaNO 3 solution for at least 6 hours. Nitrate ion
displaces the Cl adsorbed onto the carbon, and the displaced Cl is
then measured in the nitrate solution and subtracted from the value
obtained by the original pyrohydrolysis This procedure of measuring and
subtracting the nitrate—displaced inorganic Cl is referred to as the
“difference” method. A second procedure that involves the pyrohydrolysis
of the carbon from which the inorganic Cl has been displaced or excluded
will be discussed as an alternative to the “difference” measurement, and
is referred to as the “direct” measurement of TOC1.
-------
6
For carbon taken from operating filter beds or sampler units con—
taming carbon, it is necessary to dry and split the carbon sample to
provide equal portions for a pyrohydrolysis determination and a nitrate
wash determination.
Calculation of TOC1
For water grab samples, TOC1 is calculated using the following
equation:
(1) (C 2 ,m . lOOO ) (vm n)( ) = C V 2 pg = pg/ 2 ;TOCI
where C = concentration of Cl in pyrohydrolyzate (mg/L)
V 1 = volume of water sample (2;)
V 2 = volume of pyrohydrolyzate (m2 )
METhOD EVALUATION
An evaluation of the method has been undertaken to determine if it
can indeed produce a valid measure of total organic chlorine. This
paper will describe the specific details of that evaluation as it
concerns itself with four main points, namely:
1. The completeness of the adsorption of organochiorine compounds.
2. The qualitative and quantitative accuracy of chloride ion
measurement in the pyrohydrolyzate.
3. The influence of inorganic chloride ion on the measurement.
4. The limits of precision, accuracy and sensitivity of the method.
Because the validity of the total organic chlorine measurements
takenon çlarifded river water during the evaluation was undefined
during this study, a1L measur ments are referred to as apparent total
organic chlorine (htTOCltt). All raw water was flocculated and settled
-------
7
because the question remains unresolved at present whether or not the
presence of suspended solids will bias the results by the combustion of
those solids recovered with the activated carbon sludge.
Certain modifications in the procedure will also be discussed.
Completeness of Adsorption of Organics
Information acquired to define the extent to which organics are
adsorbed during the extraction process included Total Organic Carbon
(TOC) measurements taken on two water samples before, between, and after
two successive extractions. The extractions were performed with one
grain of carbon on 10 liters of river water that had first been floc-
culated and settled to remove suspended solids. In this case no organic
floc aide was used in any of the flocculation steps. 250 m 9 of super—
natant liquor from each extraction step was passed through a separate 8
pm Millipore cellulose acetate filter. 70—m9. aliquots of the 250 m2
of filtered supernate liquor were taken in septa—sealed bottles for TOC
analysis. The three results were then averaged.
The filter membranes were washed before and after with three
successive washings of 250 m2, of Super—Q water which was analyzed for
TOC in the same manner. The TOC concentration value for the Super-Q
water was determined and subtracted from the filter wash— water value,
which was in turn subtracted from the sample value to obtain the TOC
concentration remaining after each extraction. For the first extrac-
tion, 88 and 82 percent removal of TOC was accomplished on the duplicate
samples, and 62 percent of the remainder was removed by the second
extraction. For the two extractions combined, 95 and 93 percent of TOC
was adsorbed from the two samples of raw water.
-------
8
Detector Selection
Two methods of detecting and measuring chloride ion were tried.
The first was a select chloride—ion probe(SIP), which is being used
routinely by Kuhn et al. 1 , and the second was the Dohrmann micro—
coulometric titration (MCT) system. The SIP appeared to be working well
on distilled—water controls, but when raw water samples were analyzed,
results were affected by a large and erratic positive bias that rendered
the results meaningless.
Table 1 presents tabulated data obtained by difference and direct
measurements using both the SIP and MCT system on four clarified river
water samples. Results of attempts to remove the interference encountered
in the use of the SIP, by treatment with hydrogen peroxide (}1202) and
copper sulfate (CuSO 4 ), are also given.
Table 1. “TOCi” in ig/9. for Clarified Ohio River Watera__
Select—Ion PrObe Interference
Method
by “Difference” “Direct”
Detector MCT SIP MCT SIP
Treatment
None
CuSO
11202 11202
1 25 108 120 26 13
2 27 120 137 53 10
3 b 36 170 144 1 6 9
32 1.92 163 52 9
a. Cl concentration 35 mgI ; suspended solids
lation and sedimentation.
b. 20 ig/t TOC1 dosed as PCP subtracted
None
11202
58
42
26
58
44
22
95
72
37
95
71
40
removed
by
floccu—
CuSO 4
-------
9
Two drops of 11202 were added to each 50—m9. pyrohydrolyzate to
(5)
remove interference from sulfur by oxidizing it to sulfur dioxide
(SO 2 ). The pyrohydrolyzate was then boiled to drive off excess H 2 0 2
that also interferes with the probe. A few grains of CuSO 4 were added
to eliminate hydrogen sulfide (H 2 S) as an interference by precipitating
copper sulfide (CuS). As can be seen in Table 1, neither treatment
brought the final result down to that obtained by “direct t ’ measurement
using the MCT system.
Evidence that interferences were present is borne Out by the fact
that values can be diminished by chemical treatment designed to elim-
inate those interferences, and by the fact that they are considerably
higher than the NCT measurements both before and after chemical treat-
ment. Because these interferences could not be completely eliminated
however, use of the SIP was discontinued and the MCT system was designated
as the detector of choice, in spite of the considerable difference in
cost. Data presented below to define blank or background levels, pre-
cision, accuracy and sensitivity limits, pertain to the method using the
MCT system for detection and measurement.
Efficiency of the Pyrohydrolysis System
The efficiency of the condenser is considered to be excellent in
view of the quantitative recovery of TOC1 dosed into the controls in the
form of pentachiorophenol (PCP). Also no chloride ion could be detected
in condensate collected after a pyrohydrolysis run, indicating that
there was no residual chloride ion left in the system. This was
-------
10
also borne out by the quantitative recovery of the TOC1 dosed into
controls in the absence of inorganic chloride ion. Thus, the operation
of the pyrohydrolysis system, in particular the condenser section, was
not found to measurably affect the precision or accuracy of the method.
Inorganic Chloride—Ion Correction
The accuracy of the method of correcting the final result for
inorganic chloride ion by subtracting out the contribution of inorganic
chloride ion adsorbed onto the activated carbon, as determined by
extracting a duplicate sample and measuring the amount of chloride ion
washed off the carbon into a 0.1 N sodium nitrate solution, was
carefully investigated.
In a test similar to that performed by Kuhn et al. to verify the
displacement of chloride ion by nitrate ion, 200—m2 aliquots of Super—Q
water previously dosed with sodium chloride to contain a chloride ion
concentration of 50 mg/L were extracted 24 hours with 2 gms of activated
carbon. The carbon was recovered by filtration and stirred for 24 hours
in 200 mR. of a 0.1 N sodium nitrate solution. The carbon was again
recovered by filtration and then pyrohydrolyzed to determine if all
chloride ion had been eliminated by the nitrate wash.
The chloride ion concentration in the pyrohydrolyzate of the
nitrate washed carbon was at approximately the same level as experimentally
determined for blank carbon (see below). Tinder the conditions of the
test, therefore, the displacement of chloride ion by nitrate ion was
essentially complete.
-------
11
A second series of tests were run to determine if the displacement
of chloride ion was sufficiently reproducible to prove a precise and
accurate correction under the conditions of analysis. On two separate
occasions six ten—liter Super—Q water samples were dosed to a concentra-
tion of 50 mg/9 , using sodium chloride, and to 20 pg/2 total organic
chlorine using pentachiorophenol (PCP). Three such samples on each
occasion were used to make a pyrohydrolysis determination on the activated
carbon, and three were used to obtain an inorganic chloride ion correction
by washing the carbon with the nitrate solution. As a check on the
value determined by this “difference” method, the nitrate washed carbon
was also analyzed to provide what is referred to as a “direct” measure-
ment.
What is reported in Table 2 are the results of obtaining the
difference measurement for each set of samples by subtracting the chloride
ion contribution determined for each of the three nitrate wash test
samples, from each of the three pyrohydrolysis samples. Nine pieces of
data were thus generated for each set of three pyrohydrolysis samples
analyzed by the “difference” method. The range of values, from —9.5 to
38 in one set and 39 to 54 in the other set, illustrate the variation
that can be expected by the difference measurement.
Table 2. TOC1a j ug/ 2 by “Difference”bMeasurement
Using Super—Q Distilled Water
Sample Number of Number of Range
Set Samples Determinations Mean a Limits
iig/ pg/Z pg/2 ,
1 3 9 20.5 ± 17.09 —95 to 38
2 3 9 47.1 ± 4.90 39 to 54
a. 20 pg/P TOC1 dosed as PCP
b. 50 mg/2 Cl dosed as NaC1
-------
12
As can be seen in Table 2, the “difference” method results exhibit
quite poor precision and accuracy, while the direct measurement exhibits
(Table 3) excellent precision and accuracy under the conditions of this
test. The data for direct measurement also includes four determinations
made independently of the above study.
Table 3. TOC1a in ugIL by “Direct” Megsurement
Using Super Q Distilled Water
Sample Number of Range
Set Determinations Mean Limits
pg/2. ugf2
1
3
19.5
+
0.50
19
to
20
2
3
18.5
±
1.50
17
to
20
3
4
22.6
1.18
21
to
23.5
1, 2 & 3
10
20.5
±
2.17
17
to
23.5
a. 20 pg/9 TOC1 dosed as PCP
b. 50 mg/9 Cl dosed as NaC1
As expected, the explantion for the lack of precision and accuracy
for measurement by difference (Table 2), at least at the low levels of
TOC1 tested, appears to be that the adsorption of inorganic chloride ion
is not sufficiently consistent or reproducible to avoid introducing
significant error of a random nature when subs tracted out by the analysis
of a “duplicate” sample. The inconsistency may in part, at least, be
due to the fact that duplicates are difficult to obtain with certainty.
On the other hand, as shown in Table 3, the displacement of inorganic
chloride ion by nitrate ion during washing apparently is essentially
complete every time, and at least for the case of PCP, does not remove
organic chlorine, thus accounting for the excellent precision and
accuracy by direct measurement.
-------
13
A third series of tests was conducted on raw Ohio river water that
was flocculated and settled to remove suspended solids. Allof these
samples were extracted one time with one gram of activated carbon rather
than the recommended two—extraction procedure to save time.
A comparison of data (Table 4) obtained by difference and direct
measurements on all 13 samples (including six dosed with 20 to 60 ig/9
Table 4. ,,TOC1,ta in jig/i for Ohio River waterb
Suspended Solids Removed by Flocculation and Sedimentation
Sample Numbers By Difference Direct
1 25 13
2 27 10
3 9 c
4 32 c 9 c
5 —1 12
6 2 17
7 2 c 6 c
8 17 d 7 d
9 4 19
10 —4 13
11 6 12
12 3 c
13 4 d
a — See Table 5 for summary of precision and accuracy
b — ci concentration 35 mg/Re
c — 20 i ig/2 TOC1 dosed as PCP subtracted
d — 60 pg/5 TOC1 dosed as PCP subtracted
-------
14
TOC1 as PCP) shows that, as it had been for distilled water controls,
direct measurement provided superior precision and accuracy on these raw
water samples. Those samples that were dosed and the dosages are
Indicated by the footnotes c and d.
Table 5 presents the mean, standard deviation, and range limits of
the values for dosed and undosed samples in Table 4 taken separately,
and for all samples combined, for both the difference and direct measure—
inents. Although the combined average values are quite similar for the
samples by both measurements, the precision of the results by direct
measurement is by far the better.
Table 5. Precision and Accuracy by Difference vs. Direct
“TOCi” Measurement on Raw Ohio River Water
Range Limits, pg/2 Mean g/9.
by Dif 1. Direct by Diff. Direct
9.5
7.05
+
4.04
14.6
3•9 ’
3.21
12.3
5.90
4.30
11.7
±1
3.74
•
3.77
The values for “TOC1” by direct measurement after subtracting the
dose are for all practical purposes the same as those obtained on
undosed samples. Thus for PCP at least, TOCI remains undiminished by
the nitrate wash, and adsorption onto the activated carbon was essentially
complete, even for one extraction, rather than the recommended two.
c , g/9
by Diff. Direct
Number and
Type of
Samples
4 Dosedb 2 to 17 6 to 15
5 Undosed —4 to 6 12 to 19
9 CombInedC —4 to 17 6 to 19
13 Combinedd —4 to 36 6 to 19
a — Cl concentration 35 mg/2
b — 20 to 60 pg/9, TOC1 dosed as PCP subtracted
c — pyrohydrolyzate untreated chemically
d — includes pyrohydrolyzates chemically treated
6.5
1.4
4.1
11.8
-------
15
Aside from providing better precision and accuracy, direct measure-
ment provides another benefit. In terms of sample treatment, the net
effect of making a direct measurement of TOC1 on nitrate—washed carbon
is to halve the amount of work that needs to be done. Two successive
extractions on one ten—liter sample would be all that is required. The
duplicate sample to obtain the inorganic chloride—ion correction is
eliminated. This not only represents a great savings in time, but also
eases the logistics of sampling.
Background Values .
The activated carbon used for extraction was pyrohydrolyzed in 2 gm
amounts to simulate the quantity required for a two step extraction, and
the pyrohydrolyzate analyzed for its Cl concentration by direct injection
of 5 Z of the pyrohydrolyzate into a microcoulometric titration cell.
The blank contribution could not be lowered by washing the carbon for 5
days with “Super—Q” water in a continuous extractor.
A second background value, termed the method blank, was established
by extracting 10 liters of Super—Q distilled water with two grams of
activated carbon and recovering the carbon by the usual procedure of
flocculation, settling and filtration. When working with distilled
water, samples are permitted to settle overnight. Pyrohydrolysis of the
carbon sludge and subsequent measurement of chloride ion by MCT revealed
the method—blank “TOCl” value to range from 8 to 10 pg/i for five
determinations. The increase over the carbon blank is not unexpected in
view of the amount of sample handling involved, the size of the glass-
ware employed, and the addition of various reagents to the sample, Thus
10 pg/i was routinely deducted as the background correction for each
analysis for total organic chlorine.
-------
16
Chloride Ion Exclusion
Before proceeding with an evaluation of the method, a modification
was studied. In an effort to simplify the direct method further by
reducing the 6 hour nitrate wash time, investigation of the addition of
nitrate ion to the sample before extraction was carried out, the purpose
being to exclude or reduce the adsorption of inorganic chloride ion
without interferring with TOC1 adsorption. By preventing chloride ion
from being adsorbed, only that chloride ion in the pore water of the
carbon sludge would remain to be washed out. This wash should be able
to be accomplished with a much more dilute nitrate solution than the
recommended 0.1 N, and In much less time, with the advantage being that
a more dilute solution of nitrate ion and the shorter time would not be
as likely to displace (desorb) adsorbed organochiorine compounds during
this step.
Sodium nitrate was added to the Super—Q water test samples to a
concentration of 500 mg/9 . nitrate ion, and sodium chloride was added to
a concentration of 100 mg/ 2, chloride ion as a severe test. The samples
were extracted once with 2 grams of activated carbon in a simplified
test to establish a background equivalent for the amount of carbon
sludge that would be recovered in two successive extractions. Several
blanks containing no sodium chloride were also run.
Figure 2 presents data for four different concentrations of nitrate
wash solutions, various sludge wash times, and various numbers of washes.
Also included is the data for blanks and several samples dosed with PCP.
-------
75ppb (NO 3 WASH)
mg/I A
N03 Wash 0 I 500 1000 200C 8500 (O.IN)
Solution
.
I 0
‘ 3 io .- __ . 0 _____ p ______
0
5.
0 0 24 3 1 0.70.5 3 1 1 1O. 1 242424241
I I I I I J I 11111 I liii I
LENGTH OF WASH PERIOD (HOURS)
0 = BLANK, NO CI IN SAMPLE; • 1 WASH, 50ppm Cl-IN SAMPLE
o = 2 WASHES, 50 ppm Cr; A = DOSED 20-6Oppb TOCI(PCP), 5Oppm cr
FIGURE 2: EFFECTS OF VARYING STRENGTH OF
WASH SOLUTION AND LENGTH OF WASH TIME.
-------
18
These data show that neither the increased concentration of nitrate
ion in the wash solution beyond 500 mg/9, nor the length of time the
carbon sludge is washed, nor whether it is washed once or twice produces
any significant difference in results. All of the results are randomly
scattered throughout the range of 7 to 15 ig/9 . “Toci” and represent a
sample background value. Since the values for the blanks (samples
containing no chloride ion) fall in the same range, the average of all
of the results can be taken as a reasonable measure of the background
level of chloride ion that will be measured in all samples containing 0
to 100 mg/P.. Cl . That background calculates to be 9.9 iigf 9.. ± 2.24 igIZ
f or a “2—gram”, activated—carbon—sludge, pyrohydrolysis sample. This
value is essentially the same as the 10 pg/P.. determined for the method
blank described previously.
Adsorption Efficiency
Recovery of TOC1 dosed as PCI ’ under conditions of the above test is
essentially the same as that previously reported for direct measurement.
This essentially complete recovery indicates that for compounds as
amenable to adsorption on activated carbon such as PCI’, at least, adding
nitrate ion to the sample does not exclude those compounds from being
adsorbed.
TOC measurements on clarified river water to which sodium nitrate
had been added to a concentration of 500 mg/P.. nitrate ion after extraction
with carbon, indicate that the presence of the nitrate ion has no adverse
influence on the completeness of adsorption of organics.
-------
19
Precision, Accuracy and Sensitivity
All of the tests to date indicate that the reliable lower limit of
sensitivity of this method should be set at no less than 10 pg/ Q above
the background value of 10 ig/JL The reliability of the number, of
course, increases with increasing TOC1. Further, the data on recovery
of TOC1 dosed as PCP show that, at least to 100 mg/9 . Cl, results are
just as reliable throughout the entire range of inorganic Cl concen-
trations. An examination of 113 cities for chloride ion concentration
revealed that 96 percent of them had concentrations below 101 mg/P..
Thus the method should have wide applicability in this country.
LIMITATIONS OF THE METHOD
Several critical aspects of this method still must be evaluated
before its validity as a measure of TOC1 can be established. First of
all, more information must be obtained to define the efficiency of the
adsorption of compounds onto activated carbon from the lowest to the
highest molecular weight compounds. The method must be demonstrated to
be all inclusive before it can be considered to be a true measure of
total organic chlorine.
Secondly, the contribution of other halides to the “TOCl” measure-
ment must be delineated. Under the conditions of pyrohydrolysis it has
been established, for example, that 70 to 80 per cent of the bromine in
bromoform and dibromochioromethane can be converted to hydrogen bromide.
Hydrogen bromide is a titratable species and is measured by the MCT
system along with hydrogen chloride. Br also interferes with
-------
20
measurements made with the chloride ion probe. It is estimated that
when using the MCT system to make a “TOC1” measurement on finished
drinking water containing various trihalomethanes, as much as 25 percent
of the measured value can be attributed to the presence of organo bromine
from the trihalomethanes. In those circumstances it would be desirable
to isolate the measurement of each halogen. An ion chromatograph is a
promising tool and should be investigated.
Finally, if the method is to be applied to water containing
significant suspended solids, it must be determined to what extent
both chloride ion and organochlorine compounds occluded in the solids
will influence the “TOCi” measurement.
Until the above limitations and uncertanties are resolved, the
method can most accurately be described as a semiquantitative measure-
ment of carbon adsorbable organohalides measured as chloride (CAOX as
Cl).
SUMMARY OF MET}IOD
A summary of the procedures now being employed In this method for
carbon adsorbable organohalides measured as chloride (CAOX as Cl) is as
follows:
1) Extract 10 liters of water, treated to be free of suspended
solids and to which has been added 685 mg of sodium nitrate per liter of
sample, for 1 hr with I gram of granular activated carbon previously
powdered to 90 sec. in a lab mill, nd wetted with Super—Q water for 30
mm under vacuum. Stir at 200—500 rpm.
-------
21
2) Settle the carbon by adding 50 inL of aluminum sulfate solution
(1 gin of aluminuui/9,) and sufficient 5 N sodium hydroxide to maintain ph
in the range of 6.5 to 7.5, and then add 20 m of polyacrylamide floc
aide (200 ing/2). Stirring rate should be reduced below 100 ppm during
floc formation (about 10 mm).
3) Allow to stand until floc has settled completely. (Time
required will depend on the nature of the sample. Anywhere from 30 mm.
to overnight may be required). Siphon off supernatant liquor to leave
about 200 m2. behind with the sludge. Quantitatively transfer sludge to
beaker. Return the supernatant liquor to the extraction vessel and
repeat steps 1 through 3.
4) Allow the sludge in the beaker to settle for 5 or 10 minutes.
Pour supernatant liquor from beaker through 8—pm pore diameter cellulose
acetate filter and wash sludge onto filter paper for vacuum filtration.
Return filtered sludge cake to same beaker, now containing 200 m2. of
685 mg/ 2. sodium nitrate (500 mg/2 . NO 3 ) wash solution, by washing sludge
from filter membrane with small amount of Super—Q water. Add a stir-
ring bar and stir moderately for 1 hour on stir jack (be sure sludge
cake has been dispersed). Settle and ref ilter sludge as before. Repeat
step four on sludge recovered from second extraction.
5) Combine sludge cakes from filter membrane into quartz boat and
pyrohydrolyze as outlined above.
6) Analyze pyrohydrolyzate by MCT by injecting one to 20 p2.
directly into the titration cell. The size of the injection for
-------
22
quantitation will be determined by the approximate concentration of
titratable halide ion in the pyrohydrolyzate as estimated by screening
injections. Three injections of sample should be made to determine an
average value for nanograms of halide ion titrated. If a Dohrmann C—300
microcoulometer is used, nanograms of halide ion titrated as Cl will
be displayed as a net integrated signal. If a C—200 is used, a strip
chart recorder must be employed and peak areas calculated. In either
case, a corrected value for nanograms of halide ion injected and titrated
as Cl must be obtained from a standard curve constructed by plotting
known amounts of Cl injected into the cell vs response.
Nanograms injected divided by ‘iP . injected is equivalent to mg/2 of
halide ion as Cl in the pyrohydrolyzate. Use this value and equation
(1) to calculate CAOX as Cl in pg/t, and then subtract 10 pg/9. back-
ground value to obtain final result. Note that when using equation (1)
to calculate pg/2. CAOX as Cl, C = the concentration of titratable halide
ion in the pyrohydrolyzate.
-------
23
App end ix
Equipment and Supplies*
1. Furnace, Split tube Lindberg 55035 (2)
(Fisher Scientific pp 384)
2. Wool, Quartz, Coleman ‘u 20 gms
(Curtin—Natheson)
3. Combustion Tube, Quartz, 74 cm long (Joint to Joint
x 18 mm I.D. mm. x 25 mm 0.D. max S 24/40
4. Boat, Quartz, 15 mm O.D. x 14 mm I.D. x 12 cm long (2) ea.
5. Adapter, T—joint 5 24/40 (cut back)
items 3, 4, 5 Paxton Woods Glass Shop — James Leue
7500 Brill Road
Cincinnati, Ohio 45243
(513) 561—8199
6. Condenser 5 24/40
7. Heating Mantle
8. Flask, round bottom, 1 liter,’ two neck 5 24/40
9. Adapter 5 24/40 with hose connection
10. Stopper 5 24/40 (2)
11. Flask, vacuum, 100 m . (4)
12. Filtering Apparatus, all glass, Millipore, xx1504700
13. Filters, Nillipore, 8 mm, 47 mm dia., SCWRPO4700
14. Mill, Analytical, Tekmar model A—105
15. Balance, Analytical
16. Stirring motor and blade
17. Jars, Cylindrical, Pyrex No. 6945, 13.2 liters (case of 4)
(Sargent—Welch Scientific Co.) — cover plates recommended
18. Titration Cell, Dohrman T—300S (w. cell enclosure)
-------
24
19. Microcoulometer, Dohrman C—300 digital; or C—200B with strip—
chart recorder
20. Meter, pH
21. Sample Bottles, 5 gal carboy, (SCIENTIFIC PRODUCTS, B7570—5A)
22. Sample Bottles, 1 gal
23. Stir Jacks
Reagents and Chemicals
1. Oxygen, extra dry
2. Carbon, granular, activated, Calgon Corporation, Filtrasorb 400
3. Flocculant, granular, ultra high molecular weight, non—ionic
polymer, Magnifloc 985N — 200 mg/9..
(Cyanamid—Industrial Chem & Plastics Div., Wayne, NJ 07470)
4. Water, Millipore “Super—Q” or equivalent
5. Aluminum Sulfate — 1 gm of aluininum/t
6. Sodium nitrate
7. Sodium Hydroxide — 5N
*Nention of brand name does not imply endoresement by the U.S. Environmental
Protection Agency.
-------
25
REFERENCES
1. Kuhn, W., Sontheimer, H., Several Investigations on Activated
Charcoal for the Deterndnation of Organic Chioro—Compounds.
Vom Wasser, 15, 65 (1973).
2. Kuhn, W., Sontheimer, H., Analytic Determination of Chlorinated
Organic Compounds with Temperature — Programmed Pyrohydrolysis.
Vom Wasser, 1 f1 1 (1975).
* U.S. 98 p 1p8IN1wIGomc0 1918— 757-140/6642
------- |