United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S2-80-193 Mar 1981
Project Summary
Evaluation of Solid
Sorbents for Water Sampling
J. C. Harris, M. J Cohen, Z. A. Grosser, and M. J. Hayes
Based on the results of screening a
variety of sorbent/solute pairs, it
appears that a combination of two
dissimilar resins is the most promising
approach to a general purpose sorbent
sampling system. In particular, a
combination of XAD-2 and XE-347
resins offers the possibility of collection
and recovery of both non-polar and
polar organic species. This XAD-
2/XE-347 system was, for the most
part, found to be comparable with
solvent extraction in regards to re-
covery; polar species such as phenols
and pyrrole were recovered in higher
yield from the cartridge.
This Project Summary was devel-
oped by EPA's industrial Environmen-
tal Research Laboratory, Research
Triangle Park, NC to announce key
findings of the research project which
is fully documented in a separate
report of the same title (see Project
Report ordering information at back).
Introduction
This report describes and presents
results of a systematic laboratory in-
vestigation of the applicability of macro-
reticular resins for general and com-
pound specific sampling of organics,
based on small scale chromatographic
experiments."The objective was to
develop screening procedures and a
data base that would allow selection of
the resin(s) best suited to a particular
sampling application and selection of a
sampling module size and configuration
that gives quantitative collection of
contaminants from the aqueous stream
of interest.
The Effect of Flow Rate
(Linear Velocity)
One important consideration in the
design and utilization of a sorbent cart-
ridge is the rate at which the sample
passes through the column. The advan-
tage of using a high flow rate is that the
sampling time is reduced, for a given
volume of sample. High flow rate,
however, may have significant disad-
vantages. A higher flow rate usually
produces a higher pressure drop across
the column. (This sometimes requires
use of higher pressure tubing, such as
steel, instead of Teflon). More impor-
tantly, high flow rate affects retention.
Flow rate is usually expressed in
terms of volume of solution per unit time
(e.g., mL/min.). The critical parameter,
however, is not flow rate per se, but
rather the linear velocity of the solution
through the column, because a mea-
surement of linear velocity takes into
account the diameter of the column and
thus the mechanics of mass transfer,
while a measurement of volumetric
flow rate does not. Conclusions drawn
from observations of a small-scale
system can be extrapolated or scaled up,
using linear velocity, to a larger scale
system.
The effects of linear velocity on
adsorption were studied by generating
breakthrough curves at various linear
velocities. Figure 1 shows breakthrough
curves up to the 50% breakthrough
point for a particular system at six
different linear velocities. Apparent
linear velocity (face velocity) was calcu-
lated as: L = F/TTr2, where F is the flow
-------
50
40
Flow Rate.
O 4.0 mL/min, 20.0 cm/min
A 2.0 m/L/min, 10.0 cm/min
O ; 0 mi/mm, 5.0 cm/min
O 0.75 mL/mm, 3.8 cm/min
^7 0.50 mL/min, 2.5 cm/min
QO 25 mL/min, 1.3 cm/min
30
tti
20
10
100
200
300 400
Volume ml/cc
500
600
700
Figure 1.
Normalized breakthrough curves at various flow rates (10 ppm
acetophenone on XAD-2 6.2 cm x 0.5 cm ID column)
rate (mL/min.) and r is the radius of the
column (cm).
At low linear velocities (e.g., 4.0
cm/min.) the breakthrough curve is
virtually independent of this parameter.
This is especially evident in that the
value of V5o is unchanged with respect
to flow rate, and is consistent with the
fact that, melutionchromatography, the
retention volume of a specific compound
is also independent of flow rate (at low
flow rates). The breakthrough curve, as
the linear velocity is increased, shifts
closer to the vertical axis: this means
that the solute breaks through sooner.
The effect of linear velocity on the shape
of the breakthrough curve is an effect
which had not been expected. Figure 1
illustrates that at low linear velocities.
the breakthrough curve has a distinctive
concave appearance while at increased
linear velocity, the curve becomes
distorted. Since uptake is proportional
to the area above the curve, it can be
seen that uptake efficiency is greater
where the curve is concave than where
it is convex. Collection efficiency at the
1/2 V50 point is almost 100% at low
linear velocities but considerably lower
at high linear velocities.
It should be noted that once an appro-
priate linear velocity criterion has been
determined for a sorbent/analyte com-
bination in a chromatographic experi-
ment, it can be used in design of a large-
•scale sampling system. A sorbent cart-
ridge of at least 3.7 cm diameter would
be required, for example, assuming a
4.0 cm/mm, velocity criterion and
hour, 10-liter water sampling constrain
Sorbent/Solute Interactions
A systematic study was performed t
determine the retention behavior c
several sorbents when challenged wit
specific model organic compounds a
solutes. Eiecause of the demonstrate
success of this resin for collection c
non-polar compounds,XAD-2 resin wa
evaluated using polar compounds o
XAD-2. These data demonstrate tha
XAD-2 has a significantly lower affinit
for polar compounds than non-pola
compounds (Table 1). Its affinity fo
phenol is so low that in order to sample
10 L sample with 97% collection effi
ciency (i.e., at 1/2 V50), a resin bed c
267 cc would be necessary.
Table 1 also shows V50 data for th<
polar model compounds on XAD-7, ,
more polar, acrylic ester resin that ha
previously been used for collection o
these species. However, V50 value
within the same range as the XAD-:
values indicate that neither XAD-2 o
XAD-7 is satisfactory for the sampling o
polar species. The carbonaceous resin
XE-340, XE-347 and XE-348 showei
high affinity for the polar compound
tested in contrast to XAD-7. For ex
ample, no breakthrough occurred afte
16 hours at 10 mg/L phenol (1601
mL/cc)
Comparisons among the XE-34(
series resins were made using benz^
alcohol as the analyte Table 1 present
V50 data for some polar compounds oi
the Ambersorb resins. XE-347 ha
slightly greater affinity for benzyl alcohc
than XE-348 and significantly greate
affinity than XE-340.
The V50 values for polar analytes oi
XE series resins are at least 5-10 time
higher (allowing for differences n
challenge concentration) than those fo
the XAD series. The XE series resins an
moderately polar, indicating fairly stron<
dipole-dipole interactions in the adsorp
tion process. Recoveries of simila
compounds are generally lower fron
XE-347 than from XAD-2 resin. A pos
sible explanation for this is the carbona
ceous surface, which may, like carbon
show extremely strong solute/sorben
bonding.
The ion exchange resins imtiall'
showed promise for polar organics, witl
IRA-93 and IRA-904 showing substan
tially higher capacity for phenol ani
benzoic acid in distilled deionized watJ
than did XAD-2 or XAD-7. Howeve?
-------
Table 1. Vso Data for Polar Compounds on Various Resins
V5o mL/cc
Compound
XAD-2* XAD-73 XE-347" XE-348" XE-349*
Phenol
Benzole Acid
Benzyl Alcohol
75
101
99
115
84
c
584
870
623
c
c
582
c
c
161
^Concentration: 10 mg/L
^Concentration: 100 mg/L.
cNot determined.
breakthrough on these resins was
virtually instantaneous when the aque-
ous solution contained as much as 250
ppm of sodium chloride.
Design of the Full Scale
Sampling Cartridge (Scale-Up)
No single resin could be used to
adequately sample a wide range of
chemical classes. XAD-2 has high
affinity for non-polar species but fairly
low affinity for polar compounds. The
best resms for adsorption of polar
species were found to be the Ambersorb
XE-34 series, with XE-347 exceeding
the volumetric capacity of the others. A
two-stage cartridge was designed using
kthis information. Samples entering the
cartridge would encounter the XAD-2
' first. Then any polar materials breaking
through the XAD-2 would be adsorbed
by the XE-347 stage
Stainless steel was chosen for con-
struction of the cartridge because it is
chemically inert to organic species, easy
to work with, and rugged enough for
field work. The inside diameter of the
cartridge was determined by the rela-
tionship between the desired sampling
rate and the linear velocity criterion
established in this work as a constraint
for efficient collection.
The minimum length of the resin bed
was determined by the retention char-
acteristics of the resins used To obtain
satisfactory uptake of the compounds
(95% efficiency), it was necessary to use
sufficient resin to ensure that the 10Lof
sample present added no more than 1 /2
Vso for any solute of interest. It has been
generally accepted that an adsorbent
bed should have a length-to-diameter
ratio considerably greater than one to
avoid channeling and back eddy effects.
The requirements for a sampling cartridge
are not as severe as those for a chroma-
tographic column. For this reason the
length of the cartridge was set at 15 cm,
7.5 cm for each of the two bedsfapproxi-
Knately 2.1 length-to-diameter ratio)
The cartridge was thus somewhat "over-
designed" compared to the criteria for
direct scale-up from the laboratory
results.
Use of Full Scale Cartridge
Adsorption of Model
Compounds
The first test of the full scale cartridge
involved verification of predicted break-
through behavior of a mixture of seven
model compounds A 10 L sample,
containing the model analytes at a
concentration of 10 ppm, was passed
through the sorbent cartridge m 4
hours, at a 4 cm/mm linear flow rate.
Hourly samples of the effluent of the
cartridge were taken to verify that no
breakthrough occurred.
Table 2 shows the recovery of seven
model compounds from each stage of
the sorbent cartridge. The non-polar
compounds such as tnchloropropane
and naphthalene were efficiently ad-
sorbed by the XAD-2 stage and the polar
compounds were trapped by the XE-347
stage.
Overall Recovery
Overall recovery of seven model
compounds was determined using the
full-scale sorbent cartridge and this was
compared to recoveries generated by
solvent extraction using methylene
chloride Table 3 shows data averages
for seven and eight determinations for
the sorbent cartridge.
At first glance the recovery of
dodecane, a strongly hydrophobic com-
pound, appears to be surprisingly low
for both methods. However, this is
rather frequently reported for aliphatic
hydrocarbons and appears to be related
to the difficulty of dissolving these
species in the water sample. In many
cases the dodecane is not in solution at
all, but exists rather as micelles and is
therefore not available to be extracted
Effect of Solvent on Desorption
from XE-347 Resin
Originally, it was thought that
methanol, being more polar, would be a
better solvent than methylene chloride
for recovery of compounds adsorbed on
a polar resin (XE-347) However, it was
found that the recovery of phenol using
methylene chloride was 90% of the
Table 2. % Recovery of Model Compounds by Batch Extraction
From
Compound XAD-2 Resin XE-347 Resin
Total
1 , 2. 3-trichloropropane
Naphthalene
Diethylphthalate
Acridine
Phenol
Dodecane
Pyrrole
78
84
89
97
47
17
0
0
0
0
0
27
0
42
78
84
89
97
74
17
42
Table 3. % Recovery of Model Compounds by Two Methods
Sorbent Cartridge Solvent Extraction
Recov-
Compound ery %
1 , 2, 3-trichloropropane
Naphthalene
Diethylphthalate
Acridine
Phenol
Dodecane
Pyrrole
76%
80
88
99
76
18
57
SD*
18
18
19
15
19
13
24
% RDS**
24
23
22
16
26
70
43
Recov-
ery %
75
92
101
89
17
41
46
SD
12
12
13
21
7
4
9
% RDS
15%
13
13
23
42
9
20
* Standard Deviation.
** Relative Standard Deviation.
-------
recovery obtained using methanol. For
this reason, use of the methylene
chloride as the solvent for both resins
was favored.
Effect of Method of Desorption
A comparison was made between the
recoveries obtained using batch extrac-
tion versus continuous extraction of
compounds adsorbed in the sorbent
trap A substantial increase in the
recovery of pyrrole was obtained using
the continuous extraction method. No
significant change was noted in the
recovery of the other compounds. The
use of a continuous extractor for subse-
quent samples was favored because of
these results and the ease of operation.
Use of Sorbent Cartridge on
Environmental Sample
The full scale sorbent cartridge was
used to analyze the organic content of
10 L of coal preheater scrubber water
obtained from a steel manufacturing
firm. This same sample was also sub-
jected to a Level 1 solvent extraction*
with methylene chloride.
for a particular organic analyte; of prin-
cipal advantage is the ability to deter-
mine the required data upon small
quantities of analyte, thus minimizing
worker exposure to test chemicals as
well as saving time and money
Uptake is greatly affected by the flow
rate through the cartridge. Decreased
collection efficiency at high sample flow
rates is presumed to be due to mass
transfer effects. Therefore, design cri-
teria for full scale sampling systems
include a limitation on the linear (face)
velocity of the sample through the cart-
ridge. The flow rate must be such that
the face velocity (sample flow rate in
mL/min. divided by the cross-sectional
area of the resin bed) is -4 cm/min. to
ensure optimum equilibrium uptake.
This constraint combines with V50 to
determine the size and shape of the
resin bed required for any particular
sampling application.
A combination of two dissimilar resins
is the most promising approach to a
general purpose sorbent sampling sys-
tem based on the results of screening a
variety of sorbent/solute pairs. In par-
ticular, the XAD-2/XE-347 system was t
found to be comparable with solvent
extraction in regards to recovery of
model compounds. Low molecular
weight and polar organic species such
as phenol and pyrrole were recovered in
higher yield from the cartridge than by
solvent extraction. High levels of sus-
pended particulate in the aqueous
sample will require a slight modification
to the sorbent cartridge and/or incorpo-
ration of a filter upstream of the sorbent
bed.
The mixed bed XAD-2/XE-347 sor-
bent cartridge is recommended for use
for general and for compound-specific
sampling of organics from water at
levels of 10 100 ppm and below. If an
aqueous stream is suspected to contain
more than 100 ppm of organics or if its
organic loading is unknown, break-
through may occur before 10 L total
sample is collected. Finally, if more than
1 g of organic material is removed from
the 3.8 cm by 15 cm XAD-2/XE-347
cartridge, the possibility that the weight
capacity limitation could lead to prema-
ture breakthrough should be considered.
Conclusions and
Recommendations
The chromatographic approach to
characterization of sorbent/solute be-
havior allows rapid, simple quantitative
evaluation of resins for water sampling
applications. The volumetric break-
through capacity, V5o, as determined in
a small scale frontal chromatography
experiment, has been shown to be a
fundamental property of the sorbent/
solute system. At low (< 10 ppm) solute
concentrations, uptake of solute by the
sorbent is governed by volumetric ca-
pacity, rather than by weight capacity, of
the resin. Collection efficiency is inde-
pendent of concentration; a 10 ppb
solution can be expected to break through
a given sorbent cartridge just as rapidly
as a 10 ppm or higher solution.
Values of V50 determined in small
scale laboratory experiments were
valuable in setting design parameters
for a full-scale sampling cartridge. The
use of such screening studies is highly
recommended whenever it is desired to
determine the optimum sampling system
*Lentzen et al , "IERL-RTP Procedures Manual
Level 1 Environmental Assessment" (Second Edi-
tion), EPA-600/7-78-201 (NTIS PB 293795),
October 1978
J C. Harris, M. J. Cohen, Z. A. Grosser, and M. J. Hayes are with Arthur D. Little,
Inc., Cambridge, MA 02140.
Larry D. Johnson is the EPA Project Officer (see below).
The complete report, entitled "Evaluation of Solid Sorbents for Water Sampling,"
(Order No PB 81 -106 585; Cost. $9.50, 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:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
* US GOVERNMENT PRINTING OFFICE 1961 -757-0 U/7042
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
------- |