Oft'''
                   United States
                   Environmental Protection
                   Agency
Health Effects                ^  *   *C.
Research Laboratory           - ", _  • ~
Research Triangle Park NC 27711   ^ f-( (-^
                   Research and Development
EPA/600/S1 -85/020  Jan. 1986
v>EPA         Project  Summary
                   Evaluation  of  Methods for the
                   Isolation  or Concentration  of
                   Organic Substances  from
                   Water  Using XAD-4
                   Quaternary Resin

                   Shaaban Ben-Poorat, David C. Kennedy, and Carol H. Byington
                    As part of a continuing program to
                   develop methods for the preparation of
                   drinking water concentrates for biologi-
                   cal testing, a synthetic resin (Amberlite
                   XAD-4 quaternary) was evaluated as an
                   adsorbent for the concentration/isola-
                   tion of 22 specific organic solutes at ugl
                   L levels. The adsorption and desorption
                   processes were first developed and
                   tested on a laboratory scale and then
                   adapted for the processing of large vol-
                   umes of water (pilot scale). Studies de-
                   termining the effect of  humic sub-
                   stances  and inorganic salts  (sodium
                   bicarbonate at 70 ppm, sodium sulfate
                   at 120 ppm, and calcium chloride at 47
                   ppm) on the adsorption/desorption of
                   the model compounds were also per-
                   formed in an effort to simulate drinking
                   water conditions.
                    Because the  method developed
                   under this contract was ultimately in-
                   tended to produce drinking water con-
                   centrates for biological testing, the ef-
                   fect of a 2 ppm chlorine residual on the
                   generation of chlorinated organic com-
                   pounds was studied.
                    In  general, the XAD-4 quaternary
                   resin in the hydroxide form was effi-
                   cient in recovering the majority of the
                   model compounds. Glucose, glycine,
                   stearic acid and  furfural were among
                   the model compounds that could not
                   be concentrated by the resin. Caffeine
                   which could be concentrated effectively
                   in the bench-scale experiment was not
                   concentrated in the pilot scale runs.
  Mass balances developed for each
type of the compounds indicated that
accountability was generally higher in
bench/scale experiments as a drop of 10
to 20% in accountability was observed
in the three pilot-plant studies. In the
bench-scale studies, flow rates of either
150, or 48 BV/hr were used. In all pilot-
plant studies a flow rate of 103 BV/hr
was utilized. More importantly  the
amount of model compounds in  the
bench-scale studies was  3  mg for
10 cm3 resin and 1500 mg for 200 cm3
resin in the pilot-scale runs.
  A statistical evaluation of the pilot-
scale studies suggested that the pres-
ence of humic substances affected the
concentration of the model com-
pounds, particularly trimesic acid.
While the effect of salts appeared to be
significant on 1-chlorododecane and 5-
chlorouracil, this was not conclusive
because of differences observed be-
tween the bench-scale and pilot-plant
studies.
  This Project Summary was devel-
oped by EPA's Health Effects Research
Laboratory, Research Triangle Park, NC,
to announce key findings of the re-
search project that is fully documented
in a separate report of the same title
(see Project Report ordering informa-
tion at back).

Introduction
  The field of separation science  has
made great strides in recent years in de-

-------
veloping techniques for isolating, sepa-
rating and concentrating organic spe-
cies in drinking water. One impetus for
these advances has been the search for
sensitive and accurate analytical meth-
ods for  trace organic contaminants. A
second  has been the need for effective
concentration and  isolation techniques
for preparing concentrates for health ef-
fects testing since  many of the biologi-
cal tests are not sufficiently sensitive to
evaluate drinking  water directly. The
objective of this project was directed to-
wards the latter application.
  Within the realm of analytical separa-
tion systems, by far the most fruitful ap-
proach has been in the use of solid sor-
bent  techniques.  Although other
approaches have been studied (reverse
osmosis, solvent extraction, foam sepa-
ration, etc.),  none is so versatile and of-
fers so  much  potential  for selectivity,
concentration and  field-use as adsorp-
tion techniques. This project dealt only
with the investigation of adsorbents for
the isolation of organic substances for
toxicological testing. Furthermore, it
was limited to the investigation of
newly developed  synthetic  sorbents
such as the polymeric XAD-4 quater-
nary anion exchange resin rather than
traditional activated carbons.
  Our goal for  this project was to de-
velop a  system for sampling  500 liters
(or more) of drinking water which might
contain  1 to 50 ppb concentration of or-
ganic compounds. Mass balances for
each compound have been determined
to reveal  the unrecovered amount of
each compound. The mass balance de-
terminations were performed in an at-
tempt to determine  whether recovery
losses were the result of volatilization,
adsorption,  and/or chemical transfor-
mation.
  Synthetic solvents  are known to con-
tain artifacts in the resin that could be
eluted during desorption of the organic
compounds concentrated on  the resin.
Therefore, separate experiments using
XAD-4  quaternary resin  (OH~ form)
were  also performed to evaluate the
presence of artifacts, either those aris-
ing from the interaction of chlorine with
the resin or those from the resin itself.

Experimental Procedures

Preparation of Model Com-
pound Test Solutions
  Test solutions  of the model com-
pounds  selected by  the U.S. Environ-
mental  Protection  Agency (EPA) to be
used in the bench-scale and pilot-scale
studies were prepared by diluting the
required volume(s)  of  stock solution
with organic free water containing an
inorganic salt matrix. The salt matrix
consisted of 77 ppm  NaHC03, 120 ppm
Ca2S04 and 47 ppm CaCI2-2H20. Table 1
lists the final concentration at which
each model compound was tested. Dur-
ing the experiments, it was  noted  that
there was some precipitation of salt in
the reservoir prior to passing the water
through the column. The pH was, there-
fore, adjusted from approximately 8.5
to 7.0 with 1NHCI to correct the prob-
lem.
Bench-Scale Column and Resin
Preparation
  The  apparatus used during  bench-
scale studies for the  isolation/concen-
tration of organics from water is shown
in Figure 1. The adsorption/desorption
column was 37 cm  long x 1 cm I.D. In
                              most bench-scale studies, this column
                              was filled with 13 cm of XAD-4 quater-
                              nary resin (OH~ form) of 40 to 80 mesh.
                              This was equivalent to approximately
                              10 cm3 of wet resin.
                                The resin clean-up procedure used
                              during this project utilized Soxhlet ex-
                              traction with the same solvents used for
                              elution. This assured that any impurities
                              would be eluted prior to the actual ad-
                              sorbent  studies. One other solvent,
                              methylene chloride, was also used for
                              resin clean-up, but only during the resin
                              blank study. The XAD-4 quaternary
                              resin was stored in the chloride form
                              underwater rather than methanol, since
                              it is not stable when stored  under
                              methanol.
                                The resin is basically an XAD-4
                              macroreticular crosslinked polystyrene
                              into which a trimethylamine group was
                              introduced.  The resin was converted
                              from the chloride form to the hydroxide
                              form before use in the resin sorption
                              experiments.
Table 1.
Summary of Percent Recovery and Percent Total Mass Balance of All Model
Compounds in Bench-Scale Studies8
                             Cone.
                                              Recovery*
                                                         Total %
                                                      Mass Balance*
anthraquinone                  500
biphenyl                       596
bis(2-ethylhexyl)-                668
  phthalate
caffeine                         50.7
chloroform                      48.7
1-chlorododecane                502
5-chlorouracil                 5,000
2,6-di-tert-cutyl-                 696
  4-methyl phenol
2,4'-dichlorobi-                 404
  phenyl
2,4-dichlorophenol             500, 5000
furfural                         52.4
glycine                      19,700
glucose                      19,800
humic acid1                   2,000
isophorone                     500
methyl isobutyl ketone            50.5
phenanthrene                     1.0
quinaldic acid                 500, 5000
quinoline                        48.4
stearic acid                      50
2,5,2',5'-tetrachloro-             172
  biphenyl
trimesic acid                  500, 5000
                                   101
                                    71.7
                                    55.2

                                    31.6
                                    69.9
                                    42.6
                                    91.1d-e
                                    62.2

                                    66.7

                                    28.8,  66.1"
                                    56.4
                                      0
                                      0
                                    79.9
                                    77.2
                                   102
                                   112
                                    96.2,  117d
                                    86.7*
                                    60.1
                                    51.9

                                    33.6,  53.0"
101
 71.9
 77.5

 46.4
 83.6
 60.0
 91.1d-e
 63.7

 68.8

 28.8"
 75.5
 87.7s
 88.9s
 79.9
 77.4
102
112
 96.2, 117d
 86.7s
 81.5
 67.7

 34.5,  53.0d
"Studies performed with salts added, without humic acid, and at a flow rate of 150 BV/hr unless
 otherwise indicated.
b% Recovery—Represents % of model compound recovered in solvent eluants.  Values repre-
            sent average of 3 determinations.
cTotal % Mass Balance—Represents total % recovered from solvent eluants + column efflu-
                     ent + reservoir rinse.
dFlow rate was 48 BV/hr.
"Value is average of two determinations.
'Study performed in the presence of salts but the absence of other model compounds.

-------
Pilot-Plant Design
  The design of the pilot plant scale-up
was based upon  breakthrough studies
with quinaldic acid. During scale-up
from bench-scale to pilot plant it was
decided to keep two factors (residence
time and through-put) as close to con-
stant as possible. The first factor con-
trols the rate of  adsorption while the
(A) Pure Inert Gas Pressure Source

IB) Cap

(C) 2-Liter Reservoir

(D) PTFE Stopcock

(£) 24/40

IF) 1.0 cm - I.D. x 37 cm Long Glass Tube
   Packed with 13 cm Resin

(Gl Silanized Glasswool Plug

Figure 1.    Apparatus for extracting organic
           solutes from water.
through-put controls the capacity of the
resin. Experimental conditions were as
follows: column dimensions 16" by 1"
i.d., residence time 0.57 min, Bed Vol-
ume (BV) = 200 cm3, flow rate = 103
BV/hr.

Column Adsorption/Desorption
Scheme
  Prior to  passing the pH 7 test solu-
tions through the  columns (1L for
bench-scale; 500L for each  pilot-plant
run) the XAD-4 quaternary resin was
changed from the chloride to hydroxide
form by passing  75 ml of 0.1 N NaOH
through the column, followed by
organic-free water until the pH of the
effluent water was neutral (approxi-
mately 100 ml).
  In a typical bench-scale study, five
columns were set up: one column for
control of reagents and glassware; one
column for the control of the resin
blank; and triplicate columns for model
compound studies. Two flow rates were
studied in the bench-scale studies. The
initial flow rate was 150 BV/hr. A lower
flow rate of approximately 48 BV/hr was
also evaluated. Flow rates were main-
tained  by  nitrogen cylinder pressure.
Following passage of the test solution
through the column the  reservoir was
washed with 25 ml of organic-free water
and drained through the column.
  The desorption  was then  accom-
plished with a series of solvents for the
elution of neutral, basic and acid com-
pounds. The following steps were used
to desorb the different classes of model
compounds:
  1)  Six 25 ml of distilled diethyl ether.
     The first 25 ml portion of distilled
     ether was added to the resin
     column, and the column was agi-
     tated to free the resin. The column
     was then allowed to stand for 5 to
     10 minutes before draining.  The
     second through the sixth 25 ml
     portion of ether eluants were used
     for desorption  by  gravity flow
     without agitation and  were com-
     bined with the first aliquot.
  2) Two 25 ml  methanol. This  was
     used for better desorption of caf-
     feine and removal of any remain-
     ing water from Step 1.
  3) Two 10 ml ether. Used for removal
     of residual methanol from Step 2
     which could interfere in the methy-
     lation of acidic compounds.
  4)  Two 25 ml 0.1 N HCI/ether. Used for
     removal of acidic compounds.
  5)  Two  25 ml 0.1 N HCI/methanol.
     Used for removal of acidic com-
     pounds.
  6)  Two 25 ml  saturated  HCI/
     methanol. Used for removal of hu-
     mic substances.

Resin Blank Artifacts/Effect of a
2 ppm Chlorine Residual Blank
Experiment with Chlorine
  This experiment was designed to pro-
vide  information regarding the effect of
chlorine on the production of new chlo-
rinated compounds. Duplicate blank ex-
periments for both pure water and pure
water spiked with 2 ppm chlorine resid-
ual were performed by the bench-scale
resin separation/concentration proce-
dure. This was done for comparative
purposes, since only compounds identi-
fied in the  extract of the chlorine spiked
water which  are not present in the nor-
mal extract from pure blank water were
considered as true artifacts  caused by
the chlorine.  In addition, a reagent blank
was  also concentrated and analyzed.

Clean Water Blank Experiment
  This experiment was performed to
identify any  artifacts that might arise
from the resin in the course of normal
resin experiments.  Four replicate resin
blanks were  run by the normal column
adsorption/desorption scheme.

Results and Discussion

General
  The mass  balance studies required
that  analyses be made of the aqueous
effluent and  column reservoir, as well
as the organic eluant from the resin. In
order to analyze the aqueous solutions
it was usually necessary to extract the
model compounds using an organic
solvent such as methylene chloride.
This  is particularly difficult  for highly
water soluble substances like quinaldic
acid, trimesic acid, glycine and glucose
and hampers mass balance determina-
tions.

Bench-Scale Study
  Resin studies using XAD-4 quaternary
resin (OH~ form) were performed sev-
eral times  at 150 BV/hr at various con-
centrations of the model compounds.
Table 1 summarizes the experimental
results of the bench-scale experiments.
Corrections  for extraction  efficiency
were made for the analysis of aqueous
effluent portions of certain model com-
pounds. Twenty-one of the 22 model
compounds were evaluated during this

-------
phase (Table 1). It is important to note
that quinoline, glycine and  glucose
were studied at mg/L levels during this
phase of the project because of analyti-
cal considerations.
  Most hydrophobic neutral  com-
pounds showed  strong  concentration
by the resin according to the adsorption
mechanism (Van der Waa'ls attraction)
characteristic of non-ionic polymer
resins. In other words, the anionic ex-
change  capacity of the XAD-4 quater-
nary resin did not seem to affect (either
favorably or unfavorably) its capacity to
absorb neutral organic compounds.
  One hydrophobic neutral compound
which showed low  recovery in the elu-
ant concentrate was 1-chlorododecane.
The reason for the low recovery of this
aliphatic hydrocarbon is not clear but
one possibility is that it adsorbs onto
particulates and passes through the
resin in a "micelle" form.
  While recoveries of 2,2',5,5'-tetrachlo-
robiphenyl and bis(2-ethylhexyl)phthalate
were both good, a reasonable amount of
each compound was also recovered from
the water. It is possible that the flow rate
for  aqueous samples affected the  effi-
ciency with which  the compounds are
sorbed by the resin. Note the improved
recoveries for the acidic  compounds in
Table 1 that were tested at 48 BV/hr.
  Acids, in general, have posed prob-
lems due to their tendency to form salts.
The XAD-4 quaternary resin has the ca-
pacity to adsorb acids by two mecha-
nisms, i.e., Van der Waals adsorption
and anion exchange. Successful elution
with diethyl ether suggests a Van der
Waals type adsorption while the neces-
sity of  utilizing acidified organic sol-
vents is characteristic of an anion ex-
change  mechanism.

Pilot-Plant Studies
  The experimental work involved in
pilot-plant studies 1, 2, and 3 was basi-
cally the same as the  resin blank pilot
study. Twenty-one of the 22 model com-
pounds were tested in all three studies,
but the  inorganic salt  matrix was only
present in studies 2 and 3. Humic acid
was only present in study 3. This order
permitted one to evaluate the effect of
salts and humic acid on the recovery of
the  model solutes.
  The desorption steps were performed
in the columns three times with 150 ml
of solvents as in the resin blank pilot-
plant study. The resin was shaken each
time, then  allowed to stand for 20-30
minutes before it was drained. Internal
standards  were added  immediately
prior to analysis. Table 2 provides the
results of the three runs.
  A statistical evaluation of the data
using a paired t-test indicated that the
lower recoveries observed for study 3
were statistically significant from stud-
ies 1 and 2 (p<0.01). This was also true
when the data for trimesic acid was
omitted. No significant difference was
observed between studies 1 and 2.

Blank Experiment with
Chlorine
  This experiment was designed to pro-
vide information regarding the effect of
chlorine on the production of new chlo-
rinated compounds. Data from this ex-
periment indicated that no new chlori-
nated compounds  were identified  as
having originated from  the addition of
chlorine to the blank water.
Clean Water Blank Experiment
  This  experiment was performed to
identify any artifacts that might arise
from the resin in the course of normal
resin experiments. Based on the results
of these tests it was necessary to clean
the XAD-4 quaternary  resin by batch
process with saturated HCI/methanol
prior to Soxhlet  cleaning by solvents.
Artifacts such as benzoic  acid  were
found in the resin  blank experiments
but the  mass of  each artifact (per day
weight of resin) was considered negligi-
ble.
Conclusions
  The procedure used in this study for
the isolation/concentration of organic
compounds  in drinking water demon-
strated that the XAD-4 quaternary resin
(OH~), is  effective for most neutral,
acidic and semi-volatile compounds.
Because of the quaternary function of
this resin, it was possible to concentrate
several acidic compounds without any
change in the pH of the sampling water.
This is an  advantage over normal XAD
resins for  on-site compositing or grab
sampling since the problem of continu-
ous acid addition and  subsequent pH
monitoring is avoided.
  Many classes  of the organic  com-
pounds adsorbed by this resin  can  be
desorbed  by solvents such as  ethyl
ether or acidic  methanol  and  ethyl
ether. The acidic solvents can be con-
centrated to remove inorganic acid, but
some residual inorganic acid always re-
mained in the concentrated eluants. The
residual acid, or possible trace of water
in the concentrated eluants, caused an-
alytical variances  when methylation
was  attempted.  Therefore, quinaldic
acid, trimesic acid and 5-chlorouracil
were analyzed by HPLC rather than GC-
FID.
  Humic substances were concentrated
more than 50-fold on the XAD-4 quater-
nary  resin, but a  saturated HCI/
methanol solution was required for the
desorption. Percent recovery of humic
acid, methyl isobutyl ketone, quinoline,
phenanthrene, anthraquinone,  humic
acid, caffeine  and 5-chlorouracil were
markedly higher in the bench-scale ex-
periments. While the flow rate or linear
velocity appeared to  affect adsorption
for the  acidic substances, differences
observed here are probably a result  of
loading capacity since the pilot-plant
studies  utilized a lower flow rate. The
higher percent recoveries for the bench-
scale  studies  were determined to be
statistically significant using a paired t-
test with a confidence interval of 95 per-
cent.

Recommendations
  Future research on the isolation/con-
centration  of organic compounds  in
water may  need to minimize analytical
problems by use of radiolabeled com-
pounds.
  Based on the results of using several
polymeric adsorbents, it appears that
the XAD-4 quaternary resin has the best
potential for field sampling of natural
water. This resin exhibited low blank
contamination, and for the most part,
was found to concentrate most neutral,
acidic and semi-volatile compounds.
  Breakthrough experiments should be
considered for specific compounds  to
evaluate the effect of mass transfer due
to velocity changes. Breakthrough may
not occur for large volumes  of water
samples with trace organic contami-
nants due to the high capacity of the
XAD-4 resin, but high concentrations of
specific compounds may breakthrough.
  Future research on the XAD-4 quater-
nary resin for concentration of trace or-
ganics may concern:
  1)  Several different column dimen-
     sions to further evaluate the effect
     of flow rate for field sampling
     column size.
  2)  Utilization of aquatic humic mate-
     rial to  more closely simulate drink-
     ing water conditions.
  3)  Experiments with known organic
    compounds in  actual drinking
    water  samples  to evaluate the
     resin efficiency due to suspended
     matter, and soluble inorganic
    salts.

-------
Table 2.   Summary of Mass Balance and % Recoveries of All Model Compounds in All Pilot Studies (500L)

anthraquinone
biphenyl
BEHP
caffeine
chloroform
1-chlorododecane
5-chlorouracil
2,6-di-tert-butyl-
4-methyl phenol
2,4' -dichlorobi-
phenyl
2,4-dichlorophenol*
furfural
glycine
humic acid
isophorone
MIBK
phenanthrene
quinaldic acid HPLC
quinoline
stearic acid
2,5,2' ,5' -tetra-
chlorobiphenyl
trimesic acid
Amount
Spiked
(v-g)
25,000
25,000
25,000
25,000
25,000
2,500
25,000
25,000

25,000

25,000
25,000
25,000
1,000,000
25,000
25,000
500
25,000
25,000
25,000
2,500

25,000
Pilot Plant
Study No. 1
%
Recovery
50.9
62.6
22.2
0.9
34.8
34.1
32.9
52.4

54.8

40.6
49.8
0
N.A.
66.4
34.5
51.0
109.0
31.0
<1
47.0

75.6
Total %
Mass Balance
53.2
68.4
34.8
25.1
59.5
34.4
48.0
64.8

60.4

59.0
53.6
80.7
N.A.
77.2
34.5
52.6
109.6
31.0
<1
48.8

75.6
Pilot Plant
Study No. 2
% Total %
Recovery Mass Balance
40.2
51.8
22.8
0.8
42.4
87.0
43.9
42.3

35.3

34.6
55.2
N.A.
N.A.
52.2
28.1
55.4
117.2
21.4
<1
43.9

93.1
44.4
60.4
39.8
23.7
113.8
87.2
83.2
48.0

40.4

53.4
57.6
N.A.
N.A.
64.4
28.1
56.6
120.8
21.4
<1
44.8

93.1
Pilot Plant
Study No. 3
%
Recovery
39.5
40.3
15.1
0.6
36
41.8
41.7
28.0

37.8

26.2
50.2
N.A.
40.7
55.2
25.6
44.0
95.9
21.7
<1
28.7

22.3
Total %
Mass Balance
45.6
52.4
22.9
28.7
79.2
42.0
65.6
45.2

47.2

66.5
50.8
N.A.
85.7
67.6
25.6
44.6
96.4
21.7
32.5
28.7

22.3
'These totals are sums of (xg found in neutrals analysis for 2,4 DCP and acids analysis for 2,4 DCP (all GC cap FID results).
N.A.—Not analyzed.
% Recovery—Represents % recovered from solvent eluants.
Mass Balance—Represents total % recovered from solvent eluants + column effluent + reservoir rinse.
   Shaaban Ben-Poorat, David C.  Kennedy,  and Carol H. Byington  are with
    Envirodyne Engineers, Inc., St. Louis, MO 63141.
   H. Paul Ringhand is the EPA Project Officer (see below).
   The complete  report, entitled "Evaluation  of Methods for the Isolation or
    Concentration of Organic Substances from Water—Using XAD-4 Quarternary
    Resin. "(Order No. PB 86-101 84 7/A S; Cost: $ 16.95. 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:
           Health Effects Research Laboratory
           U.S. Environmental Protection Agency
           Research Triangle Park, NC 27711
                                                                                   U. S. GOVERNMENT PRINTING OFFICE: 1986/646-116/20764

-------
United States
Environmental Protection
Agency
Official Business
Penalty for Private Use $300

EPA/600/S1 -85/020
Center for Environmental Research
Information
Cincinnati OH 45268
"A-^ITVJ US.PGS
,;•;       ;.y?r.Tv j  u ^
i   v,;»r •i'?^   ;..'..MC 1
         0000329   PS

         U S  ENVIR PROTECTION AGENCY
         REGION  5  LIBRARY
         230  S DEARBORN STREET
         CHICAGO             IL   6060*

-------