CD A U-S- Environmental Protection Agency Industrial Environmental Research
t • f\ Office of Research and Development Laboratory
Research Triangle Park, North Carolina 27711
EPA-600/7-78-054
March 1978
CHARACTERIZATION OF SORBENT
RESINS FOR USE IN ENVIRONMENTAL
SAMPLING
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-054
March 1978
CHARACTERIZATION OF SORBENT
RESINS FOR USE IN ENVIRONMENTAL
SAMPLING
by
R. F. Gallant, J. W. King. P. L. Levins *
and J. F. Piecewicz
Arthur D. Little. Inc.
Acorn Park
Cambridge, Massachusetts 02140
Contract No. 68-02-2150
Task 10601
Program Element No. EHB537
EPA Task Officer: Larry D. Johnson
Industrial Environmental .Research Laboratory
Office of Energy, Minerals and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington. O.C. 20460
-------
ii
-------
TABLE OF CONTENTS
Page
LIST OF FIGURES v
LIST OF TABLES viii
ACKNOWLEDGEMENT x
SUMMARY .' . xi
I. INTRODUCTION 1
II. APPROACH 4
A. Frontal and Elution Analysis Methods ... 4
B. Relation of Chromatographic Data to Sorbent-
Based Collection Devices 7
III. EXPERIMENTAL TECHNIQUE AND APPARATUS 10
A. Experimental Apparatus 10
B. Choice of Adsorbates and Adsorbents .... 14
C. Discussion of Experimental Conditions ... 18
IV. RESULTS AND DISCUSSION 21
A. Specific Retention (Elution).Volume Data . . 21
T
1. Comparison of Experimental Vg Values
with Other Chromatographic Literature
Values 26
2. Correlation of V^ with Adsorbate
Physical Properties 28
3. Flow Rate Effects on Vg 59
B. Adsorption Coefficients 69
C. Heats of Adsorption 76
D. Frontal Analysis 79
E. Adsorption Isotherms 93
V. CONCLUSIONS AND RECOMMENDATIONS 103
continued.
iii
-------
TABLE OF CONTENTS (continued)
Page
VI. REFERENCES 105
APPENDIX A 109
APPENDIX B 133
Iv
-------
LIST OF FIGURES
Figure No. Page
1 Relationship Between Frontal Breakthrough Curve
and Elution Peak 5
2 Relationship Between Initial Retention Volume
(Vj), v£ and Levels of Frontal Analysis
Breakthrough 8
T
3 Gas Chromatographic Apparatus Used to Determine V
and Adsorption Isotherms 11
4 Chemical Structure of Sorbent Resins 16
5 Log V|° vs Adsorbate Boiling Point for Tenax-GC . . 30
6 Log V|° vs Adsorbate Boiling Point for XAD-2 .... 31
7 Log V|° vs Boiling Point for Individual Adsorbate
Groups on XAD-2 33
8 Log v|° vs Boiling Point for Individual Adsorbate
GroSps on Tenax-GC 34
9 Log V|° vs Boiling Point of Adsorbate n-Alkanes
on Tenax-GC ..... 36
10 Log v|° vs Boiling Point of Adsorbate Aromatic
Hydrocarbons on Tenax-GC 37
11 Log v|° vs Boiling Point of Adsorbate Halogenated
Hydrocarbons on Tenax-GC 38
12 Log v|° vs Boiling Point of Adsorbate Ketones on
Tenax-GC 39
13 Log V20 vs Boiling Point of Adsorbate Amines on
Tenlx-GC 40
14 Log v2° vs Boiling Point of Adsorbate Aliphatic
Alcohols on Tenax-GC 41
15 Log: V|° vs Boiling Point of Adsorbate Phenols on
Tenax-GC 42
16 Log. V|° vs Boiling Point of Adsorbate Aliphatic
Acids on Tenax-GC 43
17 Log V*° vs Boiling Point of Adsorbate n-Alkanes
on XAD-2 44
continued....
-------
LIST OF FIGURES (continued)
Figure No. Page
18 Log V^0 vs Boiling Point of Adsorbate Aromatic
Hydrocarbons on XAD-2 ........ . ..... 45
19 Log V|° vs Boiling Point of Adsorbate Halogenated
Hydrocarbons on XAD-2 .... .......... 46
20 Log v|° vs Boiling Point of Adsorbate Ketones on
XAD-2 .......... . ........... 47
21 Log v|° vs Boiling Point of Adsorbate Amines on
XAD^2 ...... . ............... 48
22 Log v|° vs Boiling Point of Adsorbate Aliphatic
Alcohols on XAD-2 ................ 49
23 Log V|° vs Boiling Point of Adsorbate Phenols on
XAD-2 ...................... 50
24 Log V|° vs Boiling Point of Adsorbate Aliphatic
AciSs on XAD-2 ................. 51
25 Specific Retention Volume (20°C) vs Adsorbate
Total Polarizability for Tenax-GC ........ 57
26 Dependence on Electronic Polarizability ot for
Chromatographic Characteristics Deduced with
Small Samples for the Porous Polymer Chromosorb
102 ...................... 58
27 V vs Flow Rate - Ethylbenzene .......... 60
28 V^ vs Flow Rate - n-Hexylamine .......... 61
O
29 Incremental Surface Area Distribution (Desorption) :
XAD-2 . . .................... 63
30 Incremental Surface Area Distribution (Desorption):
Tenax-GC ....... . ............ 64
rp
31 V:; vs Flow Rate - Pentanoic Acid ......... 67
o
32 Frontal and Elution Chromatograms for n-Hexane at
Approximately 90°C on XAD-2 ........... 80
33 Frontal and Elution Chromatograms for n-Octane at
94.7°C on XAD-2 ................. 82
continued. . . .
vi
-------
LIST OF FIGURES (continued)
Figure No. Page
34 Frontal Chromatograms for n-Octane at
87.8°C on XAD-2 83
35 Frontal Chromatograms for n-Octane at 87.8°C
on XAD-2 84
36 Adsorption Isotherm of Toluene on XAD-2 Derived
from Frontal Analysis Results 96
37 Comparison of Adsorption Isotherms for Different
Types of Adsorbates 97
38 Comparison of Sorption Isotherms Generated by
Different Chromatographic Techniques at High
Challenge Concentration 98
39 Comparison of Sorption Isotherms Generated by
Different Chromatographic Techniques for Low
Challenge Concentrations 99
40 Adsorption Isotherm Dependence on Flow Rate of
Carrier Gas 101
vii
-------
LIST OF TABLES
Table No. Page
1 Adsorbate Group and Classification 17
2 Resin Capacity for n-Butylamine at 93.28C
on XAD-2 and a Challenge Concentration of
130 ppm (v/v) 20
3 Specific Retention Volumes (Vg) for Adsorbate
Vapors on Sorbent Resins (20°C) 22
4 Comparison of Vg Values with Previously
Reported Breakthrough Volumes on Tenax-GC ... 25
rji
5 Comparison of Log V^ Values for Selected
Sorbates . . . . 27
6 Comparison of Log v| Values with Values Deter-
mined by Janak (8) at 20°C on Tenax-GC .... 29
20
7 Linear Regression Parameters for Log V- vs
Boiling Point (°C) Plots of Adsorbate Classes
on Tenax-GC Resin 52
20
8 Linear Regression Parameters for Log Vg vs
Boiling Point (°C) Plots of Adsorbate Classes
on XAD-2 Resin 53
9 Total Polarizability Values (a) for Adsorbates. . 55
10 Adsorption Coefficients, K^, for Adsorbate
Vapors on Sorbent Resins (20°C) 70
11 Equilibrium Sorption Capacities, q,,, for n-
Alkanes and Aromatics on Sorbent Resins at
20°C and 1 ppm (v/v) Challenge Concentrations . 73
12 Weight Capacities of Adsorbates on XAD-2
Calculated from Adsorption Coefficients .... 74
13 Comparison of Differential Heats of Adsorption,
AHA, for Adsorbates on jtorbent Resins with
Heat of Liquefaction, A&L 77
14 Comparison of Specific Retention Volumes Deter-
mined via Elution Chromatography with Those
Determined by Frontal Analysis for n-Butylamine
at 93.2°C on XAD-2 86
viii
-------
LIST OF TABLES (continued)
Table No. Page
T
15 Comparison of Vg and Weight Capacities for
Sorbates on XAD-2 from Elution and Frontal
Analysis (Gow-Mac) 87
T
16 Comparison of Vg and Weight Capacities for
n-Octane on XAD-2 from Elution and Frontal
Analysis 89
T
17 Comparison of V and Weight Capacities for
Sorbates on XAD-2 from Elution and Frontal
Analysis (Varian) 92
18
T T
Comparison of the Vg (adsorption)/Vg (desorption)
with Weight Capacity (adsorptionJ/Weight
Capacity (desorption) Ratio 94
ix
-------
ACKNOWLEDGEMENT
This report has been submitted in partial fulfillment of the re-
quirements on EPA Contract No. 68-02-2150, Technical Directive 10601.
The authors wish to acknowledge the support and encouragement of
Drs. Larry Johnson, Project Officer and Raymond Merrill, Assistant
Project Officer, and thank them for their suggestions and review of the
work. Thanks are also given to Dr. Charles Lochinuller, Duke University,
for his encouragement and review of the work.
-------
SUMMARY
The contents of this technical report pertain to the use of
chromatographic techniques to characterize resins which are used to
trap vapors in environmental sampling schemes. Two chromatographic
techniques are described, frontal and elution analysis, which have been
applied to characterize sorbent cartridges packed with Tenax-GC and
XAD-2 sorbents. These are synthetic polymeric resins commonly used as
sampling media.
Three diverse adsorbate groups, consisting of eight distinct
chemical classes, were studied as potential pollutants. Elution
T
analysis of these vapors yielded specific retention volumes, V , which
o
can be directly related to the breakthrough characteristics of the
sorbent resins under a diversity of sampling conditions. Adsorption
T
coefficients, K., derivable from V , yield the weight capacity of the
A g
sorbent at challenge concentrations in the Henry's Law region.
Frontal analysis results confirm the elution data for sorbate
uptake of resins. A slight flow rate dependence for sorbate uptake is
noted for XAD-2. Specific retention volume data extrapolated to
ambient conditions correlate well with adsorbate boiling point and mole-
cular polarizability. These correlations allow breakthrough and weight
capacity to be estimated for a variety of adsorbate types.
A definite specificity for non-polar adsorbates is exhibited by
both resins. Tenax-GC and XAD-2 are approximately equivalent in their
capacity to trap sorbate vapors at a given challenge concentration.
This interaction is due predominantly to non-specific dispersion
forces between adsorbent and adsorbate yielding Type I, Langmuir
isotherms. Good agreement is observed between chromatographically-
predicted breakthrough volumes and actual breakthrough in sampling
experiments.
xi
-------
I. INTRODUCTION
In the past decade, adsorbent-filled cartridges have found increased
use in the sampling of volatile and relatively non-volatile organic com-
pounds . Among the media to which this method has been applied are ambient
air and stack sampling, the characterization of water for organic com-
pounds, and the sampling of industrial worker atmospheres for potential
carcinogenic vapors. The adsorbent loaded device is generally "challenged"
by the medium of interest by inducing a well-defined flow across the
sorbent bed, either by the application of a sampling pump or superimposi-
tion of vacuum. Hence, by knowing the flow rate and sampling time pre-
cisely, the concentration of the species of interest can be determined.
Sorbent modules are frequently employed as one of a number of col-
lection devices or stages in a multi-purpose sampling device, such as
the EPA-SASS train (57). The SASS train sorbent trap is primarily
designed to capture organic species that have sufficient volatility
to pass through particulate filters upstream from the sorbent bed.
For several reasons, care must be taken in designing experiments
and interpreting results with sorbent traps. Very volatile gases are
retained poorly by most sorbent resins currently used in sampling
devices. Other species will "break through" the trap if the sampled
volume exceeds the volume or weight capacity of the sorbent.
The availability of data which describe the quantitative relation-
ship between sorbent, chemical species and sampling volumes allows the
sampling conditions to be specified so that reliable results may be ob-
tained. The studies described in this report were designed to obtain
those data.
One of the more common methods of characterizing adsorbents is the
use of gas chromatography. Several reviews (1,2) attest to the
-------
popularity of this technique for thermodynamic and kinetic characteri-
zation of solid surfaces. As will be shown later, the retention time
(volume) in a gas chromatography experiment is directly related to the
breakthrough volume observed for an organic adsorbate in a sorbent
sampler. Thus, tabulations of chromatographic retention data have in-
trinsic value to the chemist or engineer designing a sampling experi-
ment involving sorbent resins. The data allow an estimate to be made as
to the suitability of a particular adsorbent for the source to be
sampled, the time required until breakthrough has occurred, and the
amount of sorbent required to collect a sufficient amount of analyte
for analytical or biological testing.
Characteristic data may be obtained by both elution analysis and
frontal analysis methods. In the elution method a small quantity of
sorbate is introduced to the sorbent in a short time. In the frontal
method the sorbent is continuously challenged with a steady state con-
centration of sorbate. Research employing the gas chromatography methods
for the above purposes has already been reported by several investiga-
tors (3,4,5,6). These include: studies to screen sorbent media for
their appropriateness in sampling (7); the effect of other agents, such
as water vapor, on the breakthrough and retention volume (8); and in-
T
vestigations relating the specific retention volume, V , to the break-
B
through volume exhibited by the sorbent device for a particular vapor at
a specified challenge concentration (9).
In an earlier report, it was shown that chromatographic retention
volume data could be correlated with frontal analysis results generated
by a sorbent trap exposure apparatus (10). This report expands upon the
previously reported studies by examining the following factors:
1. The retention characteristics exhibited by two specific sorbent
resins, XAD-2 and Tenax-GC, for a large variety of compound
types, each type representing a distinct sorbate class.
-------
2. The relationship between the equilibrium adsorption isotherm
and the retention volume results obtained in the low surface
coverage region (i.e., Henry's Law region), and its applica-
tion to sorbent sampling device design.
3. The advantages and disadvantages of several chromatographic
•T
based methods for determining V , breakthrough curves (adsorp-
.5
tion and desorption branches), adsorption isotherms, and
weight capacity of the sorbent trap.
4. The effect of flow rate on retention volume, particularly at
face velocities similar to those corresponding to actual
sampling conditions.
T
5. The correlation of elution volume (V ) data with sorbate
g
physical properties to aid in the prediction of breakthrough
volumes of other organic species.
6. The relationship between elution and frontal chromatographic
T
approaches; an elution volume V value corresponds to the 50
O
breakthrough volume on a frontal breakthrough curve.
-------
II. APPROACH
A. Frontal and Elution Analysis Methods
Several chromatographic methods have been used in this study to
produce the data on specific retention volumes, adsorption isotherms,
etc. The theoretical relationships unifying these approaches are given
in detail in Appendix A.
The underlying principle of the experimental approach is that the
T
specific retention volume, V , for an analyte, on a sorbent is related
O
in a simple manner to the equilibrium adsorption coefficient, K , so
A.
long as the experiments are carried out at low analyte concentrations
(the Henry's Law region). Under these conditions
VT = K*A°
g As
T
where V =» specific retention volume
&
K. = equilibrium adsorption coefficient
A° = adsorbent specific surface area
s
The experiments were conducted using both elution and frontal
analysis techniques. The primary difference between frontal and elution
analysis is in the method of sample introduction. In elution analysis,
a small quantity of adsorbate is injected onto the sorbent cartridge in
a short time. For the frontal analysis case, the sample introduction
time is long and continuous. The mode of sample introduction in no way
affects the appearance of the frontal boundary curve or the elution
peak maxima. Both are determined by the equation given above and theo-
retically should appear as pictured in Figure 1.
-------
Volume (Time)
FIGURE 1 RELATIONSHIP BETWEEN FRONTAL BREAKTHROUGH
CURVE AND ELUTION PEAK
-------
In the frontal analysis technique, a gas containing a specified con-
centration of sorbate is continuously fed into the sorbent cartridge.
The experimenter waits for the appearance of the boundary profile and
continues to monitor the "breakthrough" of sorbate until it equilibrates
with the cartridge for the challenge concentration specified. The
equilibration stage is signaled by the onset of a concentration vs.
time "plateau" which can be continued for as long as one wishes in a
"steady state" condition. If the breakthrough curve is sharp, or a
symmetrical sigmoid profile, then the equilibrium sorption capacity, q,
can be computed.
T
The relationship of the specific retention (elution) volume (V ) to
O
measurable parameters in a gas chromatographic experiment has been
derived in many standard treatises on gas chromatography (12) . It is
T
important to realize that V is the fundamental retention constant in
O
gas chromatography and reflects the effect of flow rate, pressure drop,
temperature, column void volume, and stationary phase weight (volume or
surface area) on the retention of an injected solute. Knowledge of the
T
value of V [frequently for convention corrected to 0°C (273°K)] allows
O
one to estimate the retention volume of a solute at another temperature
T
or for a different column length. Thus, V determined from eonventional
&
gas chromatographic columns can aid in the design of sorbent sampling
modules.
The specific retention volume is also directly relatable to funda-
mental phase distribution constants, such as the partition coefficient,
K^, or K , the equilibrium adsorption coefficient. Thus if certain
physical characteristics (such as A°) of the stationary phase are known,
s
then K. values can be obtained which allow calculation of sorbate dis-
tribution for sorption systems larger than analytical scale devices.
Further details on specific calculation procedures and explanation of
other terms used in the study are given in Appendix A.
-------
B. Relation of Chromatographic Data to Sorbent-Based Collection
Devices
The specific retention volume allows one to determine whether an
organic compound is retained by a given weight of adsorbent at a speci-
T
fied flow rate and sampling temperature and time. If V is greater
6
than the volume of gas passed through a sorbent during any specified
time period, then the sorbate will be retained by the trap. This state-
T
ment must be qualified somewhat since V actually corresponds to 50%
O
breakthrough in an elution chromatography experiment.
This concept is illustrated in Figure 2, where the actual break-
through of the sorbent begins to occur after V volumes of gas have
passed over the sorbent bed. The difference between this V, value and
T
V can be considerable, however, there are several reasons why V is not
routinely determined. For one, the value of V is dependent upon the
dispersion of the chromatographic peak in the sorbent bed, thus V is
flow rate dependent, shows a dependence on the packing structure of the
sorbent column, and is difficult to precisely locate on the chromatogram.
The specific retention volume VT, on the other hand, is easily located
O
and is the only point on the chromatographic band corresponding to true
T
thermodynamic equilibrium. Provided that it is understood that V
o
represents the 50% breakthrough volume on a corresponding breakthrough
curve, a safety factor can be built into any calculation of breakthrough
T
volume to account for the disparity between V and V .
Specific retention volume data in this report have been obtained at
higher temperatures than the normal operation of the sorbent trap module
T
of the SASS sampling train. Thus to obtain V data applicable to SASS
O
train conditions requires extrapolation to a typical ambient temperature
T
value (20°C). This is normally done by plotting the log V vs 1/T
relationship and extrapolating to the desired temperature. The values of
T
V thus obtained by extrapolation should be highly accurate if good linear
t>
regression coefficients have been obtained in fitting the experimental
-------
c
0
N
C
E
N
T
R
A
T
I
O
N
Volume (Time)
FIGURE 2 RELATIONSHIP BETWEEN INITIAL RETENTION VOLUME (V,),
VTAND LEVELS OF FRONTAL ANALYSIS BREAKTHROUGH
-------
data. A number of investigators (26,27) have employed this approach in
characterizing sorbentmedia for environmental sampling.
T
A logical concern is the validity of extrapolating log V vs 1/T
data to temperatures considerably lower than those upon which the regres-
T
sion equation is fitted. Naturally, V data should be taken as close as
6
possible to the temperature desired but this may not be possible due to
the limited volatility of the adsorbate.
A study of the temperature dependence of the adsorption isotherm is
of value since q values can be obtained ,at constant c, challenge con-
centration, for a number of values of c. Unfortunately such experiments
require the gathering of considerable experimental data to generate
isosteres (plots of log c vs 1/T for constant c values).
The primary purpose in generating isotherms in these studies was to
T
confirm the validity of V data in accurately predicting breakthrough
G
from sorbent cartridges. Note, however, that the isotherm data presented
here can be of value in other sampling and analysis contexts. For
example, knowledge of the desorption characteristics could be of value
in the thermal desorption of sorbent cartridges for chemical analyses.
-------
III. EXPERIMENTAL TECHNIQUE AND APPARATUS
A. Experimental Apparatus
Figure 3 is a schematic of the type of apparatus used to determine
T
elution volumes, V , and adsorption isotherms on sorbent cartridges.
O
A commercially available gas chromatograph was modified to include a
column inlet pressure gauge. The temperature of the column oven was
read with a thermocouple/potentiometer. Samples were injected via micro-
liter syringe for elution analysis while frontal analysis steady state
sorbent vapor challenge concentrations were generated using a syringe-
drive pump.
All concentrations cited in this report are given in ppm calculated
on a volume/volume basis.
The sorbent cartridges were proportionately scaled down from the
typical cross section of an SASS train sorbent resin canister. Stainless
steel tubing 9 cm long, 0.45 to 0.51 cm I.D., and 0.64 cm O.D. was used
to contain the resin. The internal volume of tubing was found to hold
0.40 g of 35/60 mesh Tenax-GC packing and 0.53-0.73 g of XAD-2 resin,
depending upon the I.D. of the tubing.
The sorbent cartridge was connected to two 0.64 cm x 0.16 cm (1/4 in
x 1/16 in) reducing unions drilled out to minimize dead volume. The
resin was retained in the trap by stainless steel frits at the end of
the tubing. Connections to the chromatograph were made with 0.16 cm
tubing.
The weight of the resin in the cartridge was determined by difference
after packing. Resins were changed frequently during the course of this
study to allow the collection of data on fresh sorbent. However, our
studies showed little variation in specific retention volume with use.
10
-------
Thermocouple
Potentiometer
Gas Chromatograph
Inlet Pressure Gauge
Soap Bubble
Flowmeter
Carrier Gas
FIGURES GAS CHROMATOGRAPHIC APPARATUS USED TO DETERMINE V AND
ADSORPTION ISOTHERMS 9
-------
This result is also complemented by the results obtained by Pellizzari
(28) on the constancy of collection efficiency with repeated cycling of
the resin.
The two gas chromatographs employed in this study were a Varian
Model 1200, a single column instrument employing flame ionization detec-
tion, and a Gow-Mac 550, dual column instrument employing thermal con-
ductivity detection. The latter instrument was outfitted with two Porter
VCD 1000 flow controllers and a special Gow-Mac 10-454 thermal conducti-
vity diffusion cell. The Gow-Mac instrument was employed to measure
chromatographic fronts and peaks up to flow rates of one liter per
minute, equivalent to the linear velocity in the SASS train.
The column head pressure on the Gow-Mac unit was read with a U-tube,
mercury-filled manometer, the pressure reading being taken by puncturing
the injection septum with a Becton-Dickinson No. 22 gauge needle connected
via plastic tubing to the manometer. The thermal conductivity cell was
operated at 200 milliamperes input current and at full sensitivity,
due to the limited sensitivity of this type of detector and the high
carrier gas flow rates. Frontal chromatograms obtained on this instru-
ment were recorded with a Hewlett-Packard Model 17501A recorder.
The Varian 1200 was fitted with a Wallace & Tiernan pressure gauge
(0-34 psig) to measure the column inlet pressure. Flow control was
T
provided by a Brooks Model 8743 flow controller. Most of the V data
o
were collected at maximum electrometer sensitivity and recorded on a
Linear Instruments Model 355 potentiometric recorder.
The column temperatures for both gas chromatographs were measured
with the aid of a Rubicon potentiometer. Iron-constantan thermocouples
(No. 20) were placed in contact with the sorbent cartridges and
connected to the potentiometer. To offset the effect of a "line" room
temperature EMF generating junction, a second thermocouple was connected
12
-------
in series with the oven thermocouple at room temperature. The other end
of the thermocouple was immersed in an ice bath whose temperature was
read with a Fisher-ASTM calibrated thermometer. The EMF corresponding
to the temperature of the bath (0.2-0.4°C) was then added to the origi-
nally measured EMF. The injector and detector temperature were usually
kept 50°C higher than the highest boiling point of the injected solutes.
During the frontal analysis experiments on the thermal conductivity gas
chromatograph, the injection port heater was frequently turned off since
the solute was injected directly onto the column in the oven proper.
Total gas flow rates were measured using a soap bubble flow meter
(100 mL or 250 mL capacity) with bubble transient times being recorded
with a Meyhan Model 228 stopwatch (resolution - 0.1 sec). The infusion
rates for the syringe pump were also measured by the above technique
using a 10.0 mL burette. The slow flow rates (0.05 to 0.5 mL/min) of
gas sample infused into the sorbent column by the syringe required the
use of smaller capacity soap bubble flow meters for these measurements.
Sample introduction technique varied considerably depending on
whether elution or frontal analysis was being performed. The technique
used in the elution analysis studies consisted of withdrawing a small
amount (<1 yl) of liquid sorbate in a 10 yl syringe, expelling the liquid
and pumping the syringe 50 times. This would generate a reproducible
dilute sorbate vapor concentration. With these low concentration
samples the experiments were able to be conducted in the Henry's Law
region and symmetrical peaks were obtained. For solid sorbates (i.e.,
phenols) samples were obtained from the headspace in a sealed vial.
Typically, a 2.5 mL Hamilton gas-tight syringe was used for these studies,
Frontal analysis chromatograms were generated using a Harvard Apparatus
Model 944 infusion/withdrawal pump.
A typical laboratory procedure for introducing the vapor contained
in a finite gaseous volume into the sorbent bed consisted of withdrawing
13
-------
a known concentration of vapor into a gas-tight syringe from a gas-
tight sampling bag. The contents of the syringe (Precision Scientific
Series A-2, 5 cc Pressure Lok) were then metered in through the injec-
tion port of the gas chromatograph using the infusion syringe pump.
An 8-inch, 20-gauge hypodermic needle was employed to allow direct in-
jection of the gas samples onto the sorbent cartridge.
A 5-liter sampling bag from Calibrated Instruments, Inc., was parti-
ally filled with a precisely measured quantity of He, the latter being
measured by a Singer Dry Ga's Meter, with appropriate corrections for
temperature and pressure. The required' amount of liquid sample was
added via syringe into the bag through a septum to give a particular
vapor concentration. To assure that the saturation vapor pressure of
the organic solute was not exceeded, vapor pressures were calculated
by use of the Antoine equation and the saturation concentration of the
vapor determined in the sampling bag* All concentrations were kept
below this limit.
Determination of the appropriate chromatographic conditions (infu-
sion rate, flow rate, attenuation, etc.) for the solute was done using
elution chromatography as a preliminary scanning technique. In
addition, running elution chromatograms also allowed a comparison
T
between V at the level of the plateau concentration and that correspon-
5
ding to the maximum of the elution peak. Three to four replicate
frontal chromatograms were run per compound. The amount of gas
delivered via the syringe drive was measured 4-5 times using a 10.0 niL
soap bubble flow meter. Typical infusion rates of 0.3-0.6 mL/min were
used.
B. Choice of Adsorbates and Adsorbents
As noted in the introduction, two adsorbents have been investigated
in this study, Tenax-GC and XAD-2. These sorbents differ in several of
their physical properties. Some key properties of these resins are
14
-------
listed below.
Sorbent
XAD-2
Tenax-GC
Mesh Size
20 - 50
35 - 60
Bulk
Density
(g/cc)
0.38
0.14
BET
Surface Area
(m2/g)
364
23.5
Pore
Volume
(cc/g)
0.854
0.053
The surface area of XAD-2 is about 15 times larger than Tenax-GC.
The pore size distribution of both resins as determined by mercury
porosimetry allows both sorbents to be 'classified as microporous in
character (29). XAD-2 does contain a small fraction of its total pores
in the range of 170-325A.
The XAD-2 used in these studies was cleaned according to the EPA
Level 1 procedures by serial extraction with water, methanol and
methylene chloride. Tenax-GC, obtained from Applied Science Laboratories,
Inc., State College, Pa., was used as received. The polymers differ in
chemical structure, XAD-2 being a cross-linked styrene-divinylbenzene
polymer while Tenax-GC is poly (p-2,6-diphenyl-phenylene oxide). The
chemical structures of both porous polymers are given in Figure 4.
Tenax-GC is stable at temperatures up to 400°C. XAD-2 darkens at
temperatures above 150°C, hence the taking of retention data was limited
to temperatures below 190°C. A complete summary of the physical pro-
perties of the two resins was given in the earlier report (10).
Tenax-GC and XAD-2 are both classified as weak Type III adsorbants
according to the classification scheme of Kiselev (30). The basis for
this classification is that localized negative charges are available
for interaction with the adsorbate molecules. For XAD-2, the aromatic
ir electrons are the negative charged moiety, while for Tenax-GC, both
aromatic TT electrons and the ether oxygen lone pair electrons are
available for nonspecific and specific interaction with the adsorbate.
15
-------
CH2 - CH - CH2 - CH - CHa - CH
i t
CH2 - CH -
t
- CH - CH2 - CH
i
XAD-2
Tenax-GC
Figure 4. Chemical Structure of Sorbent Resins
16
-------
In order to measure the specificity of the resin for a particular
class of adsorbates and to provide a data base of VT values for general
use, adsorbates were chosen to represent three of four sorbate classes*
specified by Kiselev (Table 1). The adsorbate groups chosen for this
study and their classification were as follows:
Table 1
Adsorbate Group Group Classification
n-Alkanes A
Aromatic Hydrocarbons B
Halogenated Hydrocarbons B
Ketones B
Tertiary Amines B
Primary Amines D
Secondary Amines D
Phenols D
Aliphatic Alcohols D
Aliphatic Acids D
The n-alkanes represent Group A, while aromatic hydrocarbons, ketones,
and tertiary amines fall into a Group B category according to Kiselev.
Group D sorbates include such chemical moieties as primary and secon-
dary amines, alcohols and acids, all compounds capable of specific inter-
actions with Type III adsorbents.
In general, the adsorbent and adsorbate physical properties, parti-
cularly vapor pressure of the sorbate, determined the temperature range
*Kiselev's category C, which includes for example organo-metallics, was
omitted.
17
-------
in which the experiments were run. Such factors as elution time and
sharpness of the boundary or peak determined the upper and lower
T
temperature over which V data could be recorded.
O
C. Discussion of Experimental Conditions
In order to obtain symmetrical elution peaks corresponding to
sorbate vapor concentrations in the Henry's Law region, great care had
to be taken to obtain as small a sample as possible. Unfortunately,
even with the dilute vapor samples obtained by consecutively pumping
the vapor space of the syringe, some skewing of the chromatogram was
observed. Other investigators (9,31) have also noted this phenomena.
In general, the small amount of asymmetry does not have a significant
T
effect on the V data.
g
Elution experiments run on the thermal conductivity gas chromato-
graph required maximum detector sensitivity and extremely small samples
(sometimes resulting in only 1% full-scale recorder response) to
approach the Henry's Law region. This approach was extremely chal-
lenging for the high flow rate experiments done on that unit since the
sample was further diluted by the high volume of carrier gas passing
through the cell.
T
Several experiments were conducted to see if the V obtained bv
g
sequential vapor space pumping of the injection syringe compared well
with a method using dilute solutions of the solute in CSj. The latter
technique had been employed in the earlier studies (10). Some improve-
ment was noted, but the improved precision obtained was in part due
to the improved experimental apparatus used to determine V^. p0r
T 6
example, the relative standard deviation in V for n-octane at 135°C
D
was 0.0199 compared to 0.112 previously recorded for this solute. A
similar trend was also apparent at 109°C (0.0331 compared to 0.112 for
the previous chromatographic apparatus) for n-octane.
18
-------
T
In addition, it was noted that the V values for n-octane were some
g
what higher than those obtained earlier (10). To determine whether the
T
presence of €82 had affected V , an n-octane sample was spiked with a
O
finite volume of CS2. The specific retention volume obtained was lower
than that found for a vapor sample introduced by the successive pumping
technique.
The reproducibility of the frontal analysis curves can best be
ascertained by noting the precision associated by integrating the
curves representing vapor uptake in front of the boundary curve.
Table 2 is an example of some resin sorbent capacities calculated for
n-butylamine at a challenge concentration of 130 ppm (v/v) on XAD-2. The
precision of the technique is excellent as judged from reproducibility
shown in Table 2.
T
The temperature dependence of V was obtained via a linear regres-
rp O
sion fit of log V vs 1/T plots. Cursory examination of the correla-
g c
tion coefficients tabulated in Appendix B for these data indicates that
T
g
T
very acceptable correlation was obtained and that extrapolated v
values may be used with confidence.
19
-------
Table 2
Resin Capacity for n-Butylamlne at 93.2°C on XAD-2
and a Challenge Concentration of 130 ppm (v/v)
Run Resin Sorption Capacity
! 0.249 mg/g of resin
2 0.238 mg/g of resin
3 0.221 mg/g of resin
4 0.242 mg/g of resin
Average 0.238 rag/g of resin
Standard Deviation 0.012
20
-------
IV. RESULTS AND DISCUSSION
A. Specific Retention (Elution) Volume Data
T
Experimentally determined V data for each temperature are pre-
sented in Appendix B. These are entered for each adsorbate class on the
T
specified resin. The listed V represent the average of these experi-
•Jl g
mentally determined V values for each compound at the specified tempera-
o
ture. For each sorbate, the regression constants (slope, intercept)
and the resultant correlation coefficient are tabulated. These data
T
should easily permit the extrapolation or interpolation to obtain V
O
at any desired temperature.
T
The regression equations for the log V vs 1/T plots have been
•j S c
used to tabulate V values at 20°C for all the sorbates investigated.
O
These are presented in Table 3 for both XAD-2 and Tenax-GC resins.
Examination of the data in Table 3 shows that, for most sorbates, the
T
V values on XAD-2 and Tenax-GC are comparable, within an order of
g
magnitude. The quantitative differences between the values may, of
course, be very significant in the design of sampling equipment.
T
For the 39 adsorbates tested on both resins, 25 of the V fs on
O
Tenax-GC are greater than those on XAD-2. Three sorbate classes, the
T
n-alkanes, halogenated hydrocarbons, and ketones, showed greater V 's
6
on Tenax-GC than XAD-2 for all the individual members of those classes.
A decided preference is shown for XAD-2 by the aliphatic alcohols (four
out of six) and by the four phenolic solutes investigated.
It is interesting to note that two of the lower members of a homo-
logous series (the aliphatic alcohols and aliphatic acids) have Vg's
that are greater on XAD-2 than Tenax-GC. An opposite trend is noted
for the sorbates in these series having a carbon number of three or
T
higher. For the three aliphatic amines examined, the Vg's are greater
21
-------
Table 3
Specific Retention Volumes (Vg)* for Adsorbate Vapors
on
Adsorbate
n-Hexane
n-Octane
n-Decane
n-Dodecane
Benzene
Toluene
p-Xylene
Ethylbenzene
n-Propylbenzene
1 , 2-Dichloroe thane
Fluorobenzene
1,1, 2- trichloroethylene
Ch lo rob en ze ne
Bromobenzene
1 , 4-Dichlorobenzene
2-Butanone
2-Heptanone
4-Heptanone
Cyclohexanone
3-Methyl-2-butanone
3, 3-Dimethyl-2-butanone
2 , 6-Dime thyl-4-hep tanone
Ace tophenone
n-Butylamine
n-Amylamine
n-Hexylamine
Benzylamine
Di-n-butylamine
Tri-n-butylamine
Sorbent Resins (20 °C)
Tenax-GC
2.58 x 101*
1.89 x 105
3.08 x 105
2.19 x 108
6.09 x 10^
7.88 x 105
3.81 x 105
8.36 x 105
1.53 x 106
2.32 x 104
8.82 x 104
8.82 x 10**
2.36 x 10^
8.41 x 106
1.73 x 107
2.21 x Wk
5.55 x 10 6
3.22 x 106
1.36 x 106
6.46 x 10^
—
—
1.23 x 107
2.67 x I0k
1.96 x 105
7.35 x 105
1.58 x 106
1.91 x 106
4.85 x 105
XAD-2
4.14 x 103
6.45 x lO4
5.06 x 105
5.24 x 10^
2.58 x 105
9.05 x 105
5.64 x 105
4.61 x 106
1.96 x Wk
3.13 x 10^
3.06 x 104
2.43 x 105
6.39 x 105
2.33 x 106
4.39 x 103
1.49 x 106
1.52 x 106
3.66 x 105
2.53 x 104
8.59 x lO4
1.61 x 107
7.70 x 106
1.80 x 10^
1.29 x 105
4.80 x 105
7.87 x 106
6.90 x 106
—
continued...
22
-------
Table 3 (continued)
Specific Retention Volumes (Vg)* for Adsorbate Vapors
on Sorbent Resins
In units of jnL/g.
Adsorbate Tenax-GC
Ethanol 9.08 x 102
n-Propanol 5.71 x 10 3
n-Butanol 4.34 x 104
2-Butanol 1.86 x 10*
2-Methyl-2-propanol 7.08 x 102
2-Methyl-l-propanol 2.88 x 10*
Phenol 2.47 x 106
o-Cresol 1.00 x 107
p-Cresol 1.40 x 107
m-Cresol 1.18 x 107
Acetic Acid 3.20 x 103 7.07 x 103
Propionic Acid 1-73 x 10* 4.00 x 10*
• A -j 1 04 x 105 7.74 x lO4
n-Butanoic Acid i.u* x ±u
.... 5 53 x 105 2.89 x 105
n-Pentanoic Acid J-JJ x ±u
23
-------
on Tenax-GC than on XAD-2.
The design of sampling experiments based on these data is relatively
simple. The sample volume divided by the weight of the sorbent resin
T
should be less than V . To allow for some margin of safety, the gas
g
volume sampled per gram of resin should be not more than one-half of
T
the relevant V value. This is important particularly in cases where
&
the elution peak profile is extremely broad.
For example, two solutes, 2,6-dimethyl-4-heptanone and 3,3-dimethyl-
4-heptanone, were difficult to chromatograph on Tenax-GC due not only
to their short retention times, but resultant skewed peak profile.
These observations are consistent with literature reports (32) of
anomolous peak broadening of methyl-branched ketones on porous polymer
packing, which in part may be due to the restricted diffusion of the
sorbate in the pores of the polymeric substrate.
T
The V values obtained in this study have been compared (Table 4)
O
with breakthrough volumes reported by Pellizzari (28) in his study,
which included five of the sorbates included in this report. The break-
through volumes obtained by Pellizzari are listed in the last column on
Table 4. For several sorbates—methyl ethyl ketone, toluene, and
1,1,2-trichloroethylene—the agreement is excellent. In all the com-
T
parisons listed, the V and breakthrough volume are never different by
g T""
more than an order of magnitude. In all but one case the V is greater
o
than the cartridge breakthrough volumes. This confirms earlier state-
T
ments that V will probably be greater than actual breakthrough volumes,
rp g
since V is determined at the actual peak maxima and not VT. However.
8 I
the VT still allows for a reliable estimate to be made of the sample
O
parameters to minimize loss of collected sample.
24
-------
Table 4
..T
Comparison of V Values with Previously Reported
O
Breakthrough Volumes on Tenax-GC
Adsorbate
2-Butanone
Acetophenone
Toluene
1,1, 2-Trichloroethy lene
Chlorobenzene
1,1, 2-Tr ichloroethylene
Chlorobenzene
T
V
g
(This Study)
2.21 x 10^
1.23 x 10?
7.88 x 105
8.82 x 101*
2.36 x 106
8.82 x 101*
2.36 x 106
Breakthrough Volume
(Pellizari, Ref. 28)
2.0 xio" (1)
1.34 xlO6 (2)
8.23 x 105 (3)
4.84 xlO4 (U)
3.27 xlO5 (5)
5.16x10" (6)
3. 26 xlO5 <7)
Notes; All values in mL/g.
(1) Table 3, p. 15, Ref. 2fi, figures at sampling rate
of 4 L/min.
(2) Table 15, p. 34, Ref. 28, volume required to elute 1/2
of the adsorbed vapor @ 25°C.
(3) Table 18, p. 38, Ref. 28, breakthrough for field
sampling study, assumed cartridge contains ^2.15 g
based on geometric considerations, sampling time =
150 mins, 1,770 L - volume air sampled.
(4) Table 19, p. 39, Ref. 28, breakthrough study, assumed
cartridge length * weight of resin, so 1.5 cm I.D. x
3 cm in length cartridge should contain 1.075 g Tenax-GC
breakthrough volume represents 50% pt.
(5) Same as (4)
(6) Table 20, p.46, Ref. 28, breakthrough volume
(7) Same as (6)
25
-------
T
1. Comparison of Experimentally Values with Other Chromatofiraphic
Literature Values
T
In order to determine the accuracy of the V data presented in
&
this report, an extensive review of the literature was conducted for
published retention volume data on the two substrates, XAD-2 and
Tenax-GC. Although a reference compendium of the chromatographic appli-
cations of porous polymers is available, relatively little temperature
T
dependent V data exists.
O
One particular reference appropriate for comparison purposes is
the work of Butler and Burke (9) who used 6.40 to 30.7 cm long, 1.6 mm
and 3.2 mm O.D. sorbent packed tubes filled with Tenax-GC and Chromosorb
102, an XAD-2 analog. Table 5 compares the results obtained by Butler
and Burke with the data obtained in this study. The agreement is
excellent for benzene on both resins and for methyl ethyl ketone and
t-butanol on Tenax-GC.
T
V values for the ketone and alcohol on XAD-2 are in serious
g
disagreement. There is no apparent explanation for the difference
T
between the V values for methyl ethyl ketone. The value for t-butanol
&
reported by Burke et al. seems inconsistent with the other trends
T
observed in their data (note how similar V is for both the Tenax-GC
g
and Chromosorbs for MEK and benzene).
rp
Burke1s data was obtained by extrapolation of V vs. l/T x 103
g c
plots up to 200°C. Burke notes that the correlation coefficient for
this relationship for the Tenax-GC/t-butanol combination was poor due
to the sample size being in the non-linear region of the sorption iso-
therm. Still the values obtained in this study agree well with those
of Burke.
26
-------
Table 5
Comparison of Log V Values for Selected Sorbates
20
Adsorbent/Adsorbate
Literature Results
Tenax-GC
Chromosorb 101
Chromosorb 102
This Study
Tenax-GC
XAD-2
t-Butanol
2.867
2.843
4.763
2.850
3.203
e g
Methly Ethyl Ketone
4.612
4.762
4.833
4.345
3.643
Benzene
4.921
4.749
4.954
4.785
4.719
to
-------
T
Additional confirmation of the accuracy of reported V 's are pro-
O
vided by a comparison of the results reported by Jnnak and coworkers (8)
for solutes: ethanol, propanol and benzene on Tenax-GC. Table 6 shows
that the agreement is excellent for these three sorbates. It should be
T
noted that one log V value for n-propanol is at 25°C (since listed
O
regression coefficients gave values both at 25°C and 20°C that were in
conflict with the values reported in the text).
T
2. Correlation of V with Adsorbate Physical Properties
g
T
The determination of a large number of V values for potential
ft
pollutants is at best a laborious task, hence a correlation scheme which
T
would permit V to be ascertained from the physical properties of the
O
adsorbate would be of immense value. In the previous work (10), it was
T
noted that a possible correlation might exist between log V of the
O
sorbate and its boiling point at atmospheric pressure. This assumed
T
relationship was based on a correlation between log V and boiling point
&
for a limited number of adsorbates, and on literature data (33). The cor-
T
relations obtained showed a linear dependence between log V and boiling
rp g
point; thus making predictions of V and therefore sorbate breakthrough
O
times and capacities on adsorbates readily available for a large number
of compounds. However, it was also noted that some sorbates deviated from
this relationship, particularly on XAD-2.
T
Figures 5 and 6 depict the relationship between log V and at-
g
mospheric boiling points (34,35) for all sorbates in all chemical
T
classes chromatographed on Tenax-GC and XAD-2, respectively. The V
g
values are for 20°C. The linear regression coefficients for all of the
data are quite low—0.800 for Tenax-GC and 0.849 for XAD-2 resin. The
dashed lines represent a typical SASS train sampling limit (4-hour
sampling time at 4 scfm) for resin charges of 130 g (XAD-2) and 29 g
(Tenax-GC) in the SASS train sorbent trap.
28
-------
Table 6
Comparison of Log Vg Values with Values
Determined by Janak (8) at 20°C on Tenax-GC
-Laboratory-
Adsorbate
Ethanol
n-Propanol
Benzene
Janak
3.061
3.761*
4.667
This Study
2.958
3.757
4.785
Value at 25°C.
29
-------
8.0
7.0
6.0
5.0
Log V2°
4.0
OJ
O
3.0
2.0
1.0
SASS
Limit
•
*r
(
A
^
A
• •
)
^
A
D
.^
J
A
A
X""
A
b ~
A
f~\
L
^
•
— o-
o
,4*
x"^
^^
3
a
a£
>^-
^
m
^
•A
0
t^
•
^
A
w
0 n-Alkanes
O Aromatics
• Halogenated Hydrocarbons
D Ketones
A Amines
A Aliphatic Alcohols
• Phenols
O Aliphatic Acids
50
TOO 150
Boiling Point (°C)
200
20
FIGURES LOG V VS. ADSORBATE BOILING POINT FOR TENAX-GC
g
-------
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
n
O
D
O
SASS Limit
O
D
• n-alkanes
O aromatics
• halogenated hydrocarbons
D ketones
A amines
A aliphatic alcohols
O phenols
+ aliphatic acids
50
100 150
Boiling Point (°C)
200
FIGURE 6 LOG v0 VS. ADSORBATE BOILING POINT FOR XAD 2
-------
The sorbates lying above the dashed line would be adequately
retained under the sampling conditions used while those below would break
through the trap. The cutoff point seems to be for solutes having a
boiling point of 130°C on XAD-2 and 150°C on Tenax-GC. Different sam-
pling conditions and/or resins might permit the collection of the more
volatile and polar adsorbates.
Matching of the data points with the legend code in Figures 5
and 6 indicates that within a single chemical class there frequently
T
exists a better linear relationship between V and the sorbate boiling
o
point. These relationships are plotted in Figures 7 and 8 for each class
of sorbates. The parallel relationship between these classes, particu-
larly on XAD-2 resin, has been observed by a number of other researchers.
The lines have not been extrapolated beyond the range of experimentally
T
determined V . Each lii
g
for the compound class.
T
determined V . Each line represents the linear regression relationship
O
The vertical displacement of each sorbate class reflects the
specificity shown by the resin matrix for that class of adsorbates. There
is a clear preference for non-polar adsorbates (such as n-alkanes and
aromatic hydrocarbons) on both resins. The more polar, Group D adsor-
bates, such as aliphatic alcohols and acids, are retained the least.
This selectivity pattern is similar to that observed by Kiselev (36) , and
indicates that, although Tenax-GC and XAD-2 are both Type III adsorbents
with the potential for nonspecific and specific interactions, the non-
specific, dispersion forces are dominating the retentive interactions
between both resins and the sorbates.
Certain exceptions in these trends do exist however. For
example, halogenated hydrocarbons are preferentially retained on Tenax-
GC as opposed to XAD-2. Increased selectivity of n-aliphatic hydro-
T
carbons on XAD-2 resin is also found. The V for phenols on Tenax-GC
O
32
-------
u>
LogV
50
100 150
Boiling Point (°C)
200
FIGURE 7 LOG V2,0 VS. BOILING POINT FOR INDIVIDUAL ADSORBATE GROUPS ON XAD-2
-------
LogV
20
u>
4.0
3.0
2.0
1.0
50
100 150
Boiling Point (°C)
200
FIGURE 8 LOG VJj° VS. BOILING POINT FOR INDIVIDUAL ADSORBATE GROUPS ON
TENAX-GC
-------
seems to be an extension of the V* trend observed for the aromatic
hydrocarbons on this sorbent. However, the phenols on XAD-2 resin are
aligned with the V* vs boiling point trend observed for the aliphatic
Figures 9-16 and Figures 17-24 are the individual VT data vs.
g
boiling point curves for each sorbate class. Linear regression has been
applied to these data and the slope, intercept and regression coefficients
for the fitted line are included in Table 7 for Tenax-GC elution and in
Table 8 for XAD-2. These parameters were used for the lines in Figures
7 and 8.
In general, considering that the adsorbate classes do not
always consist of members of a homologous series, the correlations are
adequate for predictive purposes. On Tenax-GC the n-alkanes (Figure 9),
halogenated hydrocarbons (Figure 11), ketones (Figure 12), aliphatic
alcohol (Figure 14) , and aliphatic acids (Figure 16) all show good line-
T
arity when the log V vs. boiling points relationships are plotted. It
O
could appear that the value for toluene in Figure 10 is in error and
accounts for the regression coefficient found for this particular set of
data. The data for the amines (Figure 13) had a correlation coefficient
of 0.71 when all of the data were used. The correlation was considerably
improved when the data point corresponding to tri-n-butylamine was
omitted. This has been done for the regression line shown in the figure
(correlation coefficient 0.91). It is interesting to note that of the
amines in Figure 13, tri-n-butylamine is the only member which is not a
Group D type adsorbate. Even the phenols exhibit a regression coeffi-
cient above 0.90, which is rather remarkable considering the limited
variation in structure studies in this class of sorbates.
It is worth noting that even higher correlation coefficients
can be obtained in some cases (for the n-alkylamines and n-alcohols),
if only the members of a homologous series in the adsarbate class are
35
-------
u>
ON
9.6
8.8
8.0
7.2
6.4
5.6
4.8
4.0
Hexane,.
..Octane
Decane
Dodecane
60 80 100 120 140 160 180 200 220
Boiling Point (°C)
FIGURE 9 LOG VJj° VS. BOILING POINT OF ADSORBATE
n-ALKANES ON TENAX-GC
-------
R R
6.4
6.0
5.6
5.2
4.8
4.4
4.0
^
^
^
*Ben
^-"
zene
"
^
Tc
^
luene
*.
^
•^
Et
^,
nylber
^
zene
*^
*P
^
-Xylen
e
/-
^--*
-
^,
n-Pro
^
pylbe
^"^
izene
^
-^
^
80 90 100 110 120 130 140 150 160 170
Boiling Point (°C)
FIGURE 10 LOG V VS. BOILING POINT OF ADSORBATE
AROMATIC HYDROCARBONS ON TENAX-GC
-------
00
oo
LogV
20
7.0
6.6
6.2
5.8
5.4
5.0
4.6
Fluoro-
benzene ^
X
X
•
^
1,2-0
X
/
,1,2-T
X
richlo
X
roethy
chloroethane
/
lene
X
x
Chl
/
orobe
x
izene
/
•
/
x
X
Bror
x
noben
X
zene^
/
X
^x
x
X
X
x
^
X
1,4-D
x
chlorc
ibenzi
jne
80
100
120 140
Boiling Point (°C)
160
180
FIGURE 11 LOG V VS. BOILING POINT OF ADSORBATE
HALOGENATED HYDROCARBONS ON TENAX-GC
-------
Log V
7 *>
6.8
6.4
6.0
5.6
5.2
4.8
4.4
f
/
/ 4
/
2-Bu
/3
tanone
/
-Meth
>
/
/
yl-2-b
4-Hep
/
/
jtanor
2-Hep
tanon
/
e
tanon
e*
/
3
/
• c
/
/
ycloh
/
/
jxanc
/
'
ne
/
/
• Ac
/
etoph
jnone
80 100 120 140 160 180
Boiling Point (°C)
200
220 240
FIGURE 12 LOG V VS. BOILING POINT OF ADSORBATE
KETONES ON TENAX-GC
-------
6.8
6.4
6.0
?n 5.6
LogV20
5.2
4.8
4.4
4.0
.
X
X
^
n-A
X
n-But
nylanr
^
'
ylami
ine
X
ne
4
x
•n-He
X
Di-n-b
xylam
X
utylar
ine
X
nine<
X
'
x
X
*Be
^
nzyla
X
nine
X
*'
X
'ri-n-b
X
utylar
X
nine
X
X
•
80 100 120 140
160 180 200
Boiling Point (°C)
220 240
260
FIGURE 13 LOGV^0 VS. BOILING POINT OF ADSORBATE
AMINES ON TENAX-GC
-------
5 0
4.6
4.2
3.8
3.4
3.0
2.6
/
Ethan
/
|7
/
42-
2
-Meth
/l-1-pr
2-Butanol +
/
/
Vlethy
/
/
l-2-pr(
/
\n
jpano
opano
/
-Propa
I ^ >
/
nol
/
]/
/
Jutanc
I
60 70 80 90 100 110 120 130 140 150
Boiling Point <°C)
FIGURE 14 LOG V?° VS. BOILING POINT OF ADSORBATE
ALIPHATIC ALCOHOLS ON TENAX-GC
-------
Log V
°
7.2
7.0
6.8
6-6
6.4
6.2
X
6.0
Phenol
o-Cresol
'p-Cresol
m-Cresol
180
190 200
Boiling Point (°C)
210
FIGURE 15 LOG VJj° VS. BOILING POINT OF ADSORBATE
PHENOLS ON TENAX-GC
-------
Log V
6.0
5.6
5.2
48
4.4
4.0
3.6
X
Acetic
X
n-Propionic
7*
-Butanoic
100 110 120 130 140 150 160
Boiling Point (°C)
170
180
190
FIGURE 16 LOG V VS. BOILING POINT OF ADSORBATE
ALIPHATIC ACIDS ON TENAX-GC
-------
7.6
7.2
6.8
6.4
Logvf
6.0
5.6
5.2
4.8
>
4
/
/
Hexa
/
ne
/
y
/
Octar
/
e /
/
/
/
/
/
/
A
/
/
lecane
'
60 80 100 120 140 160 180 200 220
Boiling Point (°C)
FIGURE 17 LOG VJj° VS. BOILING POINT OF ADSORBATE
n-ALKANES ON XAD-2
-------
Log V;
20
6.8
6.4
6.0
5.6
5.2
4.8
4.4
4.0
X
<
'Ben
X
zene
^
/
X
.
X
t X
""Tol
/
jene
X
X"
X'
*Eth
P-
X
-Xyle
yl ber
'ropy
^
-\e
izene
benze
X
ne A
X
^
/
x
X
80 90 100 110 120 130 140 150 160 170
Boiling Point (°C)
FIGURE 18 LOG V VS. BOILING POINT OF ADSORBATE
AROMATIC HYDROCARBONS ON XAD-2
-------
6.4
6.0
5.6
5.2
£ LogVf
4.8
4.4
4.0
Fluoro-
benz'ene •
X"
't
>f
,1,2-T
^
richlo
X"
roethy
1,2-Dichloroethane
^
lene
X
Ch
/
orobe
^
^
izene
^
»X-
x^
^
^
^x^
1,4-D
k Broi
chlorc
^X
nober
ibenze
X'
zene
;ne
^
X'
80 100 120 140 160
BoiJing Point (°C)
FIGURE 19 LOG v|0 VS. BOILING POINT OF ADSORBATE
HALOGENATED HYDROCARBONS ON XAD-2
-------
8.0
7.2
6.4
Log V20 5.6
4.8
4.0
3.2
/
X
x
^
2-But
AX
^
anone
»:
X
3-Me
X
U-Dir
thyl-2
2,6-
4-He
X
S"
nethy
-butan
Dimet
Dtanor
>/
X
-2-but
one
nyl-4-l
ie^
X
anon*
leptan
J-
^Cyc
one <
x
s
Hepta
lohex
>
s
/
none
anone
X
X
/
S
^Ace
X
X
tophe
X
none
X
x
^
X
80
100 120 140
160 180
Boiling Point (°C)
200 220 240
FIGURE 20 LOG v|° VS. BOI LING POINT OF ADSORBATE
KETONES ON XAD-2
-------
6.8
6.4
6.0
5.6
s u,vft2
4.8
4.4
4.0
/
/
r
/+
-Amy
/
/
n-But
amine
/
/
ylami
i ^^ /
ne
/
/
D
4
/
i-n-bu
/
f
n-He
tylam
,
/
/
/
ne
/
^
/
/
• B
/
;nzyla
mine
80 100 120 140 160 180 200 220 240 260
Boiling Point (°C)
FIGURE 21 LOG V^° VS. BOILING POINT OF ADSORBATE
AMINES ON XAD-2
-------
4 4
4.2
4.0
3.8
\& 9
3.6
3.4
3.2
/
/
/
/
/+
/
/
Ethane
2-
n-F
/
/
>l
Vlethy
ropar
/
/
l-2-pr(
4
-*/
/
Dpano
2-But
/
'
anol
/
/2
/
'
-Meth
/
/
^n
/l-1-pr
/
'
Butan
opano
/
ol
I
60 70 80 90 100 110 120 130 140 150
Boiling Point (°C)
FIGURE 22 LOG V^° VS. BOILING POINT OF ADSORBATE
ALIPHATIC ALCOHOLS ON XAD-2
-------
ui Log V
20
7.2
7.0
&.8
6.6
6.4
6.2
6.0
iPhenol
o-Cresol
]m-Cresol
p-Cresol
180
190
200
Boiling Point (°C)
210
FIGURE 23 LOG V VS. BOILING POINT OF ADSORBATE
PHENOLS ON XAD-2
-------
6.0
5.6
. n-Pentanoic
5.2
4.8
n-Butanoic
n-Propionic
4.4
4.0
X
> Acetic
3.6
100 110 120 130
140 150 160
Boiling Point (°C)
170 180
190
FIGURE 24 LOG V VS. BOILING POINT OF ADSORBATE
ALIPHATIC ACIDS ON XbD-2
-------
Table 7
20
Linear Regression Parameters for log V vs Boiling Point (°C)
Plots of Adsorbate Classes on Tenax-GC.Resin
ro
Adsorbate Class
Aliphatic Hydrocarbons
Aromatic Hydrocarbons
Halogenated Hydrocarbons
Ke tones
Amines
Aliphatic Alcohols
Phenols
Aliphatic Acids
Slope
0.02591
0.01513
0.02952
0.02356
0.01449
0.04975
0.03206
0.03308
Intercept
2.348
3,786
2.280
2.687
3.597
-1.012
0.6767
-0.4162
Correlation Coefficient
0.972
0.848
0.981
0.945
0.906
0.963
0.913
1.000
-------
Table 8
u>
20
Linear Regression Parameters for log V vs Boiling Point (°C)
Adsorbate Class
Aliphatic Hydrocarbons
Aromatic Hydrocarbons
Halogenated Hydrocarbons
Ketones
Amines
Aliphatic Alcohols
Phenols
Aliphatic Acids
Plots of Adsorbate
Slope
0.02309
0.02308
0.02115
0.02765
0.02613
0.03089
0,02734
0.02263
Classes on XAD-2 Resin
Intercept
3.349
2.820
2.618
1.805
2.337
0.8567
1.706
1.251
Correlation Coefficient
0.997
0.981
0.996
0.928
0.980
0.905
0.893
0.987
-------
plotted. The literature abounds in experimental data verifying this
trend. (See references 33,55,56).
T
Similar trends and success in correlating V with boiling
O
point have been observed for the results on XAD-2 resin. Inspection of
Figures 17-24 and Table 8 indicate that for the same adsorbate classes,
T
there is even a greater correspondence to linearity when log V vs.
O
boiling point is plotted than was observed for Tenax-GC.
A number of other linear relationships have been found to hold
when such parameters as molecular weight, carbon number, and electron
T
polarizability of the sorbate are plotted vs log V (38). Again, for
O
each sorbate class, the lines are vertically displaced from each other,
but of similar slope for each sorbate class. In an attempt to find a
physicochemical adsorbate parameter that might induce coincidence of
all the sorbates on one single generalized line, total molecular polari-
zability, a, was examined. Values of a are given for a selected number
of adsorbates in Table 9. The large polarizability exhibited by the
n-alkanes can readily be contrasted with the low values for polar adsor-
bates and water, which is minimally retained by the porous polymer resin.
Note that a also increases with chain length, thus a potential relation-
T
ship could exist between log V and a.
The values of a given in Table 9 were obtained from a number
of sources and represent both experimental values (38-40) as well as
calculated molar refractions (41,42) which have been converted to total
polarizabilities via
a = 0.3964 R
where R = molar refraction in cc/mole.
54
-------
Table 9
Total Polarizability Values (a) for Adsorbates
*
Adsorbate a
n-Hexane 11.78
n-Octane 15.53
n-Decane 19.22
Benzene 10.32
Toluene 12.15
p-Xylene 14.2
Acetic Acid 5,05
Propionic Acid 6.80
n-Butanoic Acid 8.58
Ethanol 5.06
n-Propanol 6.89
n-Butanol 8.72
Benzylamine 11.6
Water 1.49
* °3
In units of A
55
-------
and
R -
'
where n = refractive index of adsorbate
mw = molecular weight of adsorbate
d = density of adsorbate
The second equation is the classical Lorentz-Lorenz equation
relating R to n Hence, values of a should be readily obtainable since
tables of n , mw, and d are readily available and tt can be experimentally
determined with a refractometer.
T
Figure 25 shows log V for the solutes listed in Table 9 vs
O fTl
a. At first the correlative value looks no better than the log V vs.
O
boiling point relationship; however, the acids, alcohols and single
atnine (benzylamine) do appear to follow a single line. The aromatic
sorbates are between the n-alkane line and the polar sorbates. This
T
suggests that •» single monotonic relationship between a and log V may not
O
be possible, but that various groups of molecules do tend to correlate
T
in terms of log V vs a.
O
Data presented by Kiselev (43) indicate the validity of the
above approach. Figure 26 shows the log t (retention time) vs a
R
relationship for various sorbates on Chromosoirb 102. Note that a linear
relation exists for the alcohols (Group D sorbates) and separately
for Groups A and B molecules. This is similar to what is observed in
T
Figure 25 and indicates that V is not totally dependent on boiling
O
point or dipole moment, but is affected by a (48). In addition, this
supports the classification of Tenax-GC and XAD-2 as Type III adsorbents
(weakly specific), although it does not preclude that porous polymers
having specific adsorbing surfaces grafted into the copolymerized matrix
56
-------
LogV
6.0
5.0
4.0
3.0
2.0
1.0
5.0
O
o
o10'0
a, (A3)
o
• n-alkanes
O aromatics
+ aliphatic acids
A alcohols
A amines
15'°
FIGURE 25 SPECIFIC RETENTION VOLUME (20°C) VS. ADSORBATE TOTAL
POLARIZABILITY FOR TENAX-GC
-------
0.6
0.2
-0.2
CD
O
-0.6
-1.0
-1.4
12
a, A
FIGURE 26 DEPENDENCE ON ELECTRONIC POLARIZABILITY a FOR
CHROMATOGRAPHIC CHARACTERISTICS DEDUCED WITH SMALL
SAMPLES FOR THE POROUS POLYMER CHROMOSORB 102
(REFERENCE 43)
58
-------
may exhibit other types of more specific adsorption activity.
T
3. Flow Rate Effects on V
g
T
The V values computed in earlier sections were determined
o
using a flow rate range of 60-150 mL/min—flow rates considerably below
the equivalent flow rate experienced in SASS train sorbent traps. The
60-150 mL/min flow rate range in the 0.5 cm I.D. sorbent columns cor-
responds to a linear velocity of 6-16 cm/sec. The SASS train sorbent
module operates at about 46 cm/sec. In order to determine whether flow
T
rate has any significant effect on V and thus the breakthrough of the
O
sorbate, several experiments were conducted with a limited number of sor-
T
bates to determine V at flow rates up to one liter/minute (106 cm/sec)..
O
These experiments were accomplished using the thermal conduc-
tivity gas chromatograph which overcame the problem of flame blow-out on
the flame ionization detector; however the larger volume of gas flowing
through the thermal conductivity cell also decreased its response to the
sorbate vapor.
T
Figures 27 and 28 show the trends in V with flow rate for
O
ethylbenzene and n-hexylamine, respectively on Tenax-GC and XAD-2. As
will be shown later, these two compounds have similar adsorption co-
efficients determined at low flow rates, hence any effect flow may have
T
on V may be attributed to a loss in effective surface area. The sym-
g x
bols in Figures 27 and 28 represent the average V value of from 3 to 8
rp g
separate determinations of V , while the brackets represent the extremes
rri O
in V . Although the precision of this data is not as high as might be
desired, due to the experimental difficulties, definite trends are ob-
servable with flow rate.
The results obtained for both adsorbates on XAD-2 seem to in-
m
dicate a gradual decrease in V up to flow rates of 400 mL/min (43 cm/sec).
O
Similar trends have been noted by Janak and coworkers for n-heptane
59
-------
1,300
1,200
1,100
1,000
900
800
(ml./g) 700
600
500
400
300
200
100
XAD-2
SASS
Velocity
rr
Tenax—GC
Value Determined in vj Experiment
• XAD-2(122.5°C)
^ Tenax-GC(128.6 °C)
100
200
300
400
500 600
ml./min.
700
800
900
FIGURE 27 VT VS. FLOWRATE - ETHYLBENZENE
9
-------
1,100
1,000
900
800
700
g (ml./g) 600
500
400
300
200
100
XAD-2
Tenax—GC
SASS
Velocity
--- Values Determined in
Experiments
XAD-2 (121.5 °C)
Tenax-GC(124.4 °C)
100
200
300
400
500 600
ml./min.
700
800
900
FIGURE 28 vj VS. FLOWRATE - n-HEXYLAMINE
-------
sorption on Porapak polymers (44), by Oberholtzer and Rogers for methane
and ethane retention by molecular sieve 4A and 5A (45), and in gel per-
meation chromatography (46,47). Between 400 and 600 mL/min (43 and 64
T
cm/sec), there is a significant decrease in V , compared to the limiting
rri §
low velocity V value (79% in the case of ethylbenzene and 82% for n-
o
hexylamine). Reductions of this order of magnitude (up to 75% decrease
T
in V ) have also been observed by Moreland and Rogers in a continuation
g
of their studies on slow mass transfer in gas-solid chromatography
(molecular sieves) (48), particularly with solutes which could penetrate
the porous sorbent structure totally or partially at low gas phase velo-
cities.
Beyond 600 mL/min (64 cm/sec), there is a leveling off in the
T
drop in V with flow rate. This trend was also observed by Moreland
o
O "Ji
and Rogers (48) and, in fact, some of their V vs. flow rate curves
O
bear a close resemblance to those obtained in this study. Similar
curves to those obtained in Figures 27 and 28 have also been obtained
for n-hexylamine on XAD-2 at 80°C and 100°C.
The results obtained for both adsorbates on Tenax-GC as depicted
in Figures 27 and 28 are quite different than those for sorption on XAD-2.
T
There is a gradual increase in V with flow rate indicating a slight
O
increase in uptake of the sorbate by the resin at higher gas velocities.
This trend is also observable at 80°C and 100°C for n-hexylamine on
Tenax-GC.
As noted previously, the trends observed on XAD-2 are probably
related to the effect of slow mass transfer of the organic solute in and
out of the porous structure of the adsorbent. A comparison of surface
area distribution as a function of pore size in XAD-2 (Figure 29) and
Tenax-GC (Figure 30) indicates the presence of some macropore structure
(170-325A) in XAD-2 that is completely missing in Tenax-GC.
62
-------
+22.5
+27.5
+32.5
+37.5
+42.5
+47.5
+52.5
+57.5
+62.5
+67.5
+72.5
+77.5
+82.5
+87.5
+92.5
+97.5
+ 105.0
+ 115.0
+ 125.0
+ 135.0
+145.0
+ 155.0
+170.0
+190.0
+210.0
+230.0
+250.0
+270.0
+290.0
+325.0
+375.0
+425.0
+475.0
+550.0
% of Maximum Surface Area ( +68.764 cc/g)
Versus Average Pore Diameter, Angstroms
20%
I I I I I I I I
40%
i i i I i i i
60%
80%
100%
FIGURE 29 INCREMENTAL SURFACE AREA DISTRIBUTION (DESORPTION):
XAD-2
63
-------
% of Maximum Surface Area ( +8.665 cc/g)
Versus Average Pore Diameter, Angstroms
c
+22.5
+27.5
+32.5
+37.5
+42.5
+47.5
+52.5
+57.5
+62.5
+67.5
+72.5
+77.5
+82.5
+87.5
+92.5
+97.5
+105.0
+ 115.0
+125.0
+135.0
+ 145.0
+ 155.0
+ 170.0
+ 190.0
+210.0
+230.0
+250.0
+270.0
+290.0
+325.0
+375.0
+425.0
+475.0
+550.0
) 20% 40% 60% 80% 1C
1 1 | j 1 1 | 1 1 1 1 1 | 1 1 1 f 1 | 1 1 1 1 1 ! 1 1 1 1 1 i 1 ! 1 1 1 1 1 ' 1 1 1 1 ' 1 1 1 1 1 I 1 1 > 1 ' ' 1 ' 1 ''' ' '
__
—
....
—
— —
—
—
—
—
-
—
—
—
—
-
—
-
-
—
—
—
1
wm
•
•
M^
mmm
^•i
••
••••MHI
10%
FIGURE 30 INCREMENTAL SURFACE AREA DISTRIBUTION IDESORPT1ON):
TENAX-GC
64
-------
Hence the reduction in V on XAD-2 is probably directly related
5
to a loss in interfacial surface area available for sorption (45). It
should be emphasized again at this point, that the loss in surface area
is kinetically based as has been shown by Janak, et al (49, 50). Hence,
T
minimization of the observed dependence of V on flow can be accomplished
6
by raising the temperature of the sampling tube or lowering of the flow
rate, both which facilitate diffusion into the porous structure of the
resin.
m
Janak's (44) results for V dependence on carrier gas flow rates
o
bear some additional comment since they are highly relevant to the
T
results presented here. Janak found extremely small drops in V over a
g
relatively modest change in flow rate for non-extracted Porapak Q and P;
T
however a 14% loss in V was observed for benzene-extracted Porapak P.
O
It may well be that the methylene chloride-extracted XAD-2 corresponds
in a similar manner to Janak1s results and that results on non-extracted
XAD-2 might show a completely different behavior, although such a treat-
ment would not be acceptable for sampling because of contamination.
Janak's paper also notes results similar to those observed in
this study for Tenax-GC. There is no apparent explanation for this pheno-
mena with the possible exception that there might be unremoved liquid in
the pores of the Tenax-GC which is retaining the sorbate. Since the
sample of Tenax-GC used in this study has undergone no extraction, its
post-polymerization history might be suspect.
These results raise the question of what is really being mea-
sured by VT for these sorbent trap experiments. Giddings (51) has noted
that the conservation-of-mass equations normally used to relate phase dis-
tribution of the solute in chromatography depend on the "long-term ap-
proximation,", i.e., "that the location and profile of a chroma tographic
zone is approximated by a limiting mathematical form which is exact only
when the elution time is infinitely larger than the time of equilibration.
65
-------
between phases." In the transient situation we are dealing with, this
approximation is probably not valid. Where severe mobile phase mass
transfer control exists, as in this case, the elution time of the chroma-
tographic peak is no longer dictated by K., but represents the equili-
Q
bruim uptake at a definite KAA product. Hence the problem is not knowing
A b
K , which can be obtained from a low velocity experiment, but in knowing
A
what is the "effective surface area" seen by the adsorbate.
One disturbing aspect of the data presented in Figures 27 and 28
T
is the lack of agreement between the limiting value of Vg in the low flow
T
rate region and previous Vg values determined for those same compounds in
T
the earlier V^ elution studies. The latter values were obtained from
O
regression equations of log V? vs. 1/T plots and their values are plotted
O
as dotted lines on the left-hand side in Figures 27 and 28 of this report.
T
These V values had been determined at flow rates between 80-120 mL/min
O
and considerably smaller sorbate sample sizes. The matter was examined
with further studies using n-pentanoic acid.
The third sorbate, n-pentanoic acid, was studied taking precau-
tions to inject very small sample, sizes (just above the limit of detec-
tability) into the thermal conductivity gas chromatograph. The results
T
are plotted in Figure 31. Note the good agreement between the V values
&
obtained via the regression equations and those in the region of 'VlOO
mL/min. This suggests that the discrepancies noted for the two other
sorbates were related to differences in sample size between the two types
T
of experiments. For the n-pentanoic acid, the trend of V with flow rate
g
on XAD-2 is the same as previously recorded for ethylbenzene and n-hexy-
T
lamine. The drop in V 's absolute value with flow is 21% of its initial
O IT!
value which compares with 21% loss in V found for ethylbenzene and a
O
figure of 18% for n-hexylamine on the same sorbent.
For n-pentanoic acid adsorbing on Tenax-GC, there is a slight
T
upward trend in V with flow rate. This parallels similar t'rends in
O
66
-------
Value Determined in V
Experiments
XAD-2( 123.1 °C)
Tenax-GC(125.5°C)
100
200
300
400
500
ml./min.
600
700
800
900
FIGURE 31 vj VS. FLOWRATE - PENTANOIC ACID
-------
T
V found for the other two solutes used in this study. However it is
8 f£ >J
interesting to note that V approaches at high flow rates the V value
& O
obtained via regression analysis of earlier determined data. This
T
suggests that there probably is no true flow rate dependence of V on
O
Tenax-GC and that the effect noted is again one of sample size (i.e.,
the solute signal is very small and difficult to observe at high flow
rates due to the dilution effect of the carrier gas. The highest flow
rates are probably at a sample size comparable to that used with the
FID chromatograph).
T
The results comparing chromatographic V and experimental break-
through volumes on Tenax-GC presented in Table 4 seem to confirm the
T
small variation in V with flow rate depicted in Figures 27, 28 and 31.
rp g rp
If V did vary with flow appreciably, the agreement between the V 's in
6 o
Table 4 and the breakthrough volumes (determined at much higher flow
rates) probably would not be as good as observed.
In summary, there appears to be a small drop in sorption
capacity with flow rate for various sorbates on XAD-2 resin. The magni-
tude of this loss would seem to be independent of the sorbate type. For
Tenax-GC the capacity is fairly constant over a wide velocity range and
sorbate type.
The data discussed in this section present for the first time,
T
Vg values measured over such a large velocity range, invalidating a con-
clusion by DeLigny (52) that such measurements were experimentally unat-
tainable. Some comparison as to the magnitude of the velocities encoun-
tered in the study is given by comparing the linear velocity of the
carrier gas, v, as calculated by
F
c
68
-------
where: F = volumetric carrier gas flow rate
e = packing porosity
6 = cross-sectional area of the empty column
Assuming a value of 0.8 for e and a 6 value of 0.196 cm2 for
the sorbent tube, v = 106 cm/sec which is considerably higher than the
value of 18 cm/sec reported by Janak (44) in his low velocity studies
T
involving V variance.
6
B. Adsorption Coefficients
The equilibrium adsorption coefficient can readily be calculated from
the specific retention volume data given in Table 3 if the specific sur-
face area of the sorbent resin is known, using equation (23) (Appendix A).
A surface area of 364 m2/g for XAD-2 and 23.5 m2/g for Tenax-GC were
used in these calculations. The values of K. for each of the adsorbates
A
used in this study are presented in Table 10.
The values of K for every adsorbate used in this study are greater
on Tenax-GC than on XAD-2. This pronounced difference ranges from
similar values of K on both resins for the n-alkanes as a sorbate class,
to two orders of magnitude difference for such adsorbate groups as
aromatic hydrocarbons, halogenated hydrocarbons, and ketones.
Within a number of adsorbate classes there exists a homologous series
for which K increases with carbon number, for both XAD-2 and Tenax-GC
A
resins. The trend is in part due to the volatility differences in the
sorbates as well as the preferred adsorption of the more hydrophobic
species.
The weight capacity in grams of sorbate/gram of adsorbates can be
calculated from a knowledge of KA using equation (24) (Appendix A), for
sorbents of the same chemical identity but slightly different surface
69
-------
Table 10
Sorb en t
Adsorbate
n-Hexane
n-Octane
n-Decane
n-Dodecane
Benzene
Toluene
p-Xylene
Ethylbenzene
n-Propylbenzene
1,2-Dichlo roe thane
Fluorobenzene
1,1, 2- trichloroethylene
Chlorobenzene
Bromobenzene
1,4-di chlorobenzene
2-Butanone
2-Heptanone
4-Hep tanone
Cyclohexanone
3-Methyl-2-butanone
3, 3-Dimethyl-2-butanone
2 , 6-Dimethyl-4-hep tanone
Acetophenone
n-Butylamine
n-Amylamine
n-Hexylamine
Benzylamine
Di-n-butylamine
Tri-n-butylamine
Resins (20°C)
Tenax-GC
6.01 x 10~5
4.40 x lO"4
7.17 x 10"3
5.10 x 10"1
1.42 x lO-*
1.84 x 10~3
8.87 x I0~k
1.95 x 10-3
3.56 x 10~2
5.41 x 10~5
2.05 x 10~k
2.05 x I0~k
5.50 x 10~3
1.96 x 10"2
4.03 z 10-2
5.15 x 10~5
1.29 x lO'"2
7.50 x 10-3
3.17 x 10-3
1.50 x 10-4
—
—
2.86 x 10-2
6.22 x 10-5
4.57 x 10-4
1.71 x 10~3
3.68 x 10~3
4.45 x 10-3
1.13 x 10-3
XAD-2
1.12 x 10~5
3.38 x 10-lt
3.08 x 10~3
—
7.88 x 10-6
3.88 x 10~5
1.36 x 10-1*
8.48 x 10~5
6.93 x 10-1*
2.95 x 10-6
4.71 x 10~6
4.61 x 10~6
3.65 x 10"5
9.61 x 10~5
3.51 x 10-4
6.60 x 10~7
2.24 x 10-4
2.29 x 10-4
5.50 x 10-5
3.80 x 10~6
1.29 x 10~5
2.42 x 10-3
1.08 x 10-3
2.71 x 10~6
1.94 x 10~5
7.22 x 10-5
1.18 x 10-3
1.04 x 10l3
—
continued
70
-------
Table 10 (continued)
Adsorption Coefficients*, JCA> for Adsorbate Vapors on
Sorbent Resins
Adsorbate Tenax-GC XAD-2
Ethanol 2.11 x 10~6 2.63 x lO'7
n-Propanol 1.33xl0-5 1.40 x 1(T6
n-Butanol 1.01 x lO'1* 3.11 x ICT*
2-Butanol 4.33 x 10~5 3.06 x 10~6
2-Methyl-2-propanol 1.65 x 10~6 1.92 x 10" 7
2-Methyl-l-propanol 6>71 x 10-5 1.92 x 10"
S 76 v 10^ 5-53 x 10~'+
Phenol 5.76 x 1U
2 33 x ID'3 2-01 x 10"
o-Cresol 2.JJ x xu
1 91 v TO'2 2'27 X 10~
p-Cresol 3.27 x 1U
9 75 x 10-2 2-33 x 10
m-Cresol 2'75 x 1U
Acetic Acid
7 45 x 10-6 1.06 x 10-6
7-45 x 10 g
-5 6-01 x 10
/ m Y in
Propionic Acid 4.03 x 10
9 A? * 10-4 1'16 x 10
n-Butanoic Acid 2.42 x 10
1 9Q v 10-3 4'35 x 10
n-Pentanoic Acid I-29 x iu
T
* v
o 2
In units of moles/mm-m , derived from K. = .0 °
71
-------
areas. In addition, if the gas phase concentration or challenge concen-
tration is in the Henry's Law region where K is applicable, then q ,
A g
the equilibrium sorption capacity, can be computed for various challenge
concentrations, C .
For example, for a 1 ppm (v/v) challenge concentration, the q values
6
have been tabulated for the n-alkanes and aromatic hydrocarbons on
Tenax-GC and XAD-2, respectively, in Table 11. The q value for the
6
n-alkanes are larger in all cases on XAD-2. For the aromatic hydro-
carbons, q is greater on Tenax-GC than for XAD-2 in all but one case
6
(p-xylene).
It should be noted that many of the adsorption coefficient values
on Tenax-GC are sufficiently greater than on XAD-2, that the equilibrium
sorption capacity of Tenax-GC for certain vapors approaches that of
XAD-2 despite the much lower surface area of the Tanax-GC. An excellent
case in point is the data provided in Table 11. The K.'s for the
A
n-alkanes are of similar magnitude—hence the principal factor in the
greater sorption capacity of XAD-2 over Tenax-GC lies in the greater
surface area. For the aromatic series, the K.'s on Tenax-GC are an
A
order of a magnitude greater on XAD-2, thus this offsets the effect of
the XAD-2 surface area. Thus, one should consider the individual
magnitudes of the adsorption coefficients and resin surface areas in
selecting the resin for use.
It is interesting to use the tabulated K. values to calculate q
A Hg
values at typical frontal analysis challenge concentrations for compari-
son with q values computed via graphical integration of the frontal
&
analysis curves. Such data exist from some frontal analysis experiments
to be described in Part C of this section. Table 17 (p 92) from that part
contains weight capacity data (q ) for challenge concentrations between
O
0-10 ppm (v/v) as determined via frontal analysis. Table 12 lists the
calculated values of q . These values can be compared with the q in
o g
72
-------
Table 11
Equilibrium Sorption Capacities*, q , for n-Alkanes and Aromatics
on Sorbent Resins at 20°C and 1
g ...
ppm (v/v) Challenge Concentrations
Adsorbate
n-Hexane
n-Octane
n-Decane
n-Dodecane
Benzene
Toluene
p-Xylene
Ethylbenzene
n-Propylbenzene
Tenax-GC
9
8
1
1
1
3
1
3
7
.26
.98
.82
.55
.98
.03
.68
.70
.64
x 10~5
x 10-*
x 10-2
x 10-*
x ID'3
x 10-3
x 10-3
x 10-2
2
1
1
1
9
3
2
2
XAD-2
.67 x
.07 x
.21 x
.70 x
.89 x
.99 x
.49 x
.30 x
10-*
ID'2
1C'1
10-*
10-*
io-3
10-*
io-2
units of grams adsorbate/grams of sorbent
73
-------
Table 12
Weight Capacities of Adsorbates on XAD-2 Calculated From
Adsorption Coefficients
Adsorbate
n-Butylamine
3,3-Dimethyl-
2-butanone
n-Hexane
n-Octane
C(ppm) *
7.76
8.21
8.11
3.08
TC(°C)
101.6
115.0
87.9
142.6
V^(mA/g)
o
367
555
810
364
moles
A mm-m™*
4.33 x 10 ~8
6.31>x 10~8
9.88 x 10~8
3.85 x 1(T8
g-adsorbate
^8 g-adsorbent
6.79 x 10~6
1.43 x 10-5
1.91 x ID'5
3.74 x 10~6
ppm by volxnne (v/v)
-------
Table 17 for both adsorption and desorption frontal analysis traces.
rn
The qg in Table 12 was calculated by obtaining a V from the regression
equation constants given in Appendix B at the temperatures specified in
Table 17. The adsorption coefficient was calculated employing equation
(23) (Appendix A), while q was obtained at the same challenge concen-
O
tration values given for the sorbates in Table 17, by using equation
(24) (Appendix A).
Examination of Table 12 reveals that K is considerably lower than
the KA'S tabulated at 20°C (Table 10), a result commensurate with the
higher temperature of collection. Cross comparison of q in Table 12
g
with those for the corresponding solutes in Table 17 indicates agreement
within a factor of two for n-hexane and 3,3-dimethyl-2-butanone and
slightly higher disagreement for n-octane and n-butylamine . The agree-
ment however is very encouraging considering the data came from two
different sets of experiments and that the possibility still exists that
the challenge concentrations quoted may not be in Henry's Law range.
Equation (24) (Appendix A) can be rearranged to allow calculation
of the initial concentration in ppm (v/v) of the pollutant in the gas
stream flowing over the sample cartridge bed from the measured uptake
q , provided K is known and that the gas plume concentration of the
pollutant is in the Henry's Law region. It is the latter factor which
makes a knowledge of the adsorption isotherm of value. For example,
if an uptake of n-octane on Tenax-GC at 20°C is 7.50 x W~k grams
sorbate/gram of adsorbent, then C is
O
(v/v)
where K = 4.40 x 10"^ moles adsorbate/mm Hg-m2
A
' S
A
A! = 23.5 m2/g
7.50 x 10-tf grams sorbate/grams of adsorbent
75
-------
MW = 114.23 g sorbate/moles sorbate
ppm = volume/volume basis
It is unfortunate that Tenax-GC is not available in a higher surface
area form since thermodynamically, the K. values exhibited by many ad-
sorbate vapors make it inherently a superior sorbent to XAD-2.
C. Heats of Adsorption
As noted in the theory section (Appendix A), the magnitude and sign
of the differential heat of adsorption provides a measure of the inter-
action between the resin matrix and the sorbate molecule. Although
not totally germaine to the design of sorbent trap experiments AH. can
A
readily be calculated using equation (27) (Appendix A), and the slope
of the regression equation listed for the individual adsorbates in
Appendix B.
The differential heats of adsorption for the sorbates on XAD-2 and
Tenax-GC are tabulated in Table 13. In addition, for each sorbate, the
heat of liquefaction, AIL , is listed. There are some interesting trends
in this data which support intuitive concepts of preferential adsorption
T
already gained from the inspection of V data. For example, thirty out
&
of thirty-eight sorbates have larger AH on Tenax-GC than on XAD-2.
For some groups of adsorbates, such as the phenols, ketones and halo-
genated hydrocarbons, every member follows this trend. With respect to
the reverse trend, only the aliphatic acids show as a class marginally
greater AH on XAD-2 than on Tenax-GC.
The magnitude of the difference between AH. for the individual ad-
A
sorbates on Tenax-GC and XAD-2 can be marginal (0-1 Kcal/mole) for some
sorbate classes, such as n-alkanes, aliphatic alcohols, and the phenols,
or as large as 3-5 Kcal/mole for other classes, such as halogenated
hydrocarbons and the ketones. The magnitude of the AH. difference
r\
between the two resins for the ketones and halogenated hydrocarbons may
76
-------
Table 13
Adsorbates on Sorbent
Adsorbate
n-Hexane
n-Octane
n-Decane
n-Dodecane
Benzene
Toluene
p-Xylene
Ethylbenzene
n-Propylbenzene
1,2-Dichloroethane
Fluorobenzene
1,1,2-trichloroethylene
Chlorobenzene
Bromobenzene
1,4-Dichlorobenzene
2-Butanone
2-Heptanone
4-Heptanone
Cyclohexanone
3-Methyl-2-butanone
3, 3- Dime thy 1-2 -butan one
2 , 6-Dimethyl-4-heptanone
Ace tophenone
n-Butylamine
n-Amylamine
n-Hexylamine
Benzylamine
- — __- A
Resins with Heat of Liquefaction*, AH,
A TT
AHA
Tenax-GC
14.0 ,
14.9
18.4
24.3
13.4
16.9
14.0
15.8
15.7
12.6
14.8
14.7
18.3
19.3
20.0
13.4
19.7
19.0
16.8
14.5
—
Ifl 4
J.O • *-t
1? 6
\.L • U
15 0
J-J • v
16 2
_LvJ • £•
15 0
A.J • **
nn ——————
*M>ii
XAD-2
14.0
17.3
18.0
—
12.4
13.6
14.4
13.6
16.5
11.2
11.5
11.4
13.0
13.7
15.0
8.73
15.0
15.3
12.7
10.8
12 0
-L^> • W
17.4
15.8
10.4
12.5
13.6
17.2
u
AHL**
7.5
9.9
12.3
—
8.1
9.1
10.1
10.1
11.1
8.47
8.27
8.5
9.63
10.5
15.5
8.3
— —
• «•
10.8
8.82
9.25
^^
13.4
7.8
__
— •_
-._
77
-------
Table 13 (continued)
Comparison of Differential Heats of Adsorption*, AH., for
A ———
Adsorbates on Sorbent
Adsorbate
Di-n-butylamine
Tri-n-b utylamine
Ethanol
n-Propanol
n-Butanol
2-Butanol
2-Methyl-2-propanol
2-Methyl-l-propanol
Phenol
o-Cresol
p-Cresol
m-Cresol
Acetic Acid
Propionic Acid
n-Butanoic Acid
n-Pentanoic Acid
Resins with Heat of
— AH nn .
Tenax-GC
16.9
16.7
9.2
11.0
13.7
12.7
9.8
13.5
16.9
18.6
18.9
18.6
10.4
11.8
11.8
15.8
Liquefacl
XAD-2
16.7
10.0
11.4
10.7
11.4
9.7
10.2
15.8
17.0
17.0
17.0
10.5
12.4
11.9
13.0
All values are negative and expressed in units of Kcal/mole.
**
From J.D. Cox and G. Pilcher, Thermochemistry of Organic and
Organometallic Compounds," Academic Press, N.Y., 1970.
**
10.1
11.3
12.5
11.9
11.1
12.1
16.4
18.2
17.7
14.8
12.5
13.7
15.2
16.6
78
-------
be indicative of the presence of strong polarization forces between
Tenax-GC and the two Group B adsorbents. This enthalpy difference is
not observed for any of the Group A or D adsorbates.
A comparison of the AHA value for all adsorbates on XAD-2 and Tenax-
GC with A^ reveals that AHA exceeds A^ in all cases for the n-alkanes,
aromatic and halogenated hydrocarbons, ketones and amines. However for
three of Group D adsorbate classes: the phenols, alcohols and acids,
AR for a good number of the adsorbates exceeds AH. . The magnitude of
L, A
this difference is usually 0-1 Kcal/mole, indicative that there is not
much intramolecular attraction between XAD-2 and Tenax-GC and the polar,
Group D adsorbates.
The AH values within a given sorbate class tend to increase with
A
carbon number and/or molecular weight. This pattern has been observed
in other adsorption studies (53). Overall, the AH results indicate a
predominance of adsorbate-adsorbent interactions, as judged by the
general condition AH
-------
Response
Time
FIGURE 32 FRONTAL AND ELUTION CHROMATOGRAMS FOR n-HEXANE AT
APPROXIMATELY 90°C ON XAD-2
80
-------
taminant which is difficult to avoid during the sampling operation. The
air plateau corresponds to the air peak in alution chroma to graphy pro-
files using thermal conductivity detection and is useful to indicate
the dead volume of the system.
Distance B is the plateau due to ^50 ppm (v/v) of n-hexane vapor.
Termination of the infusion operation results in an immediate drop, the
distance which corresponds to A' , and since this is due to the air
eluting from the sorbent cartridge, A=A'. The desorption of the n-hexane
is the distance B1.
To confirm that these fronts corresponded to elution peaks in a com-
parable concentration range, the recorder chart was rewound, so that
elution chroma to grains would commence at the same starting time. Then
discrete syringe injections of the same concentration were made.
Inspection of Figure 32 shows that the elution peak maxima correspond
to the middle of the frontal profile. The two elution peaks in Figure 32
represent 0.5 mL and 0.25 mL injections of the same sample concentration
used in the frontal experiments. There is a slight skew in both peaks
which may indicate that the experiments are not in the Henry's Law
region of adsorption but overall the results are gratifying.
Figure 33 is for the sorbate, n-octane, at a carrier gas concentra-
tion of VLOO ppm (v/v). Again the frontal chromatogram maintains
essentially the same qualitative and quantitative features apparent in
Figure 32. Note that the elution chromatogram run on the same chart
paper at 1/2 height again shows good correspondence to the frontal
. profile.
Figures 34 and 35 are breakthrough curves for n-octane at two dif-
ferent challenge concentrations. Figure 35 was generated by using the
same sampling bag concentration as in Figures 33 and 34, but changing
81
-------
Response
100 ppm
Time
FIGURE 33 FRONTAL AND ELUTION CHROMATOGRAMS FOR
n-OCTANE AT 94.7°C ON XAD-2
82
-------
Response
Time
FIGURE 34 FRONTAL CHROMATOGRAMS FOR n-OCTANE AT 87.8°C ON XAD-2
83
-------
Response
Time
FIGURE 35 FRONTAL CHROMATOGRAMS FOR n-OCTANE AT 87.8°C ON XAD-2
84
-------
to 50 ppm (v/v) in the carrier gas by altering the syringe infusion
setting to reduce the delivery rate by one-half. Note that these curves
are somewhat more diffuse than the curves at 94.7°C due to the lower
temperature (87.8°). Since the lower challenge concentration curve
would have been diminished in response by 1/2, the attenuation was also
halved in recording it. Overlay of Figures 34 and 35 show that the
curves are practically identical, in striking confirmation of the quan-
titative reproducibility of the technique.
Some additional data on the reproducibility and comparison of the
specific retention volume determined from the mid-point of the frontal
breakthrough curve to those obtained from elution peak maxima are given
in Table 14. The average VT for n-butylamine on XA.D-2 at 93.2°C for the
O
frontal technique is 507 mL/g, while for the elution technique the value
is 489 mL/g. This is very good agreement and probably would be closer
were it not for the one anomolously high value of 534 mL/g for the
frontal technique. Certain kinetic contributions to the boundary profile
T
in frontal analysis may distort the V value at one-half the plateau con-
O
centration unless corrections are made. Thus it is difficult to always
T
obtain good agreement between V from the two experimental sources.
&
Nonetheless, the results depicted in Figures 32-35 are encouraging and
stand as evidence that the elution analysis peaks are an accurate measure-
ment of breakthrough.
Table 15 summarizes the results for the frontal analysis studies,
T
including values for q and V at 50% volumetric breakthrough in the
desorption branch of the frontal analysis curve as well as the adsorp-
tion branch. The q values were determined by integration of the frontal
analysis curves. The temperature and challenge concentration for the
various adsorbates run are also listed. Most of the challenge concentra-
tions are above 50 ppm (v/v), so that an adequate curve could be ob-
tained for integration purposes.
85
-------
Table 14
Comparison of Specific Retention Volumes
Determined via Elution Chromatography with Those
Determined by Frontal Analysis for n-Butylamine
at 93.2°C on XAD-2
Frontal Analysis
Run 1 534
Run 2 499
Run 3 481
Run 4 513
Avg. = 507
Elution Analysis
Run 1 497
Run 2 485
Run 3 486
Avg. = 489
86
-------
CO
Table 15
T
Comparison of V^ and Weight Capacities for Sorbates on XAD-2
- e ••
from Elution and Frontal Analysis (Gow-Mac)
Compound
Toluene
n-Buty lamine
Ethanol
3,3-Dimethyl-2-
butanone
n-Hexane
n-Octane
Challenge
Concentration
(ppm) *
141
138
133
72
118
60
VT Column
g Temp.
(mJl/g) (°C)
419
489
293
587
702
394
124.6
93.2
45.8
111.7
86.2
140.9
A/H
VT
g
451
505
346
650
715
448
Frontal
Isorption
Weight
Capacity
(g/g)
3.01 x 10"1*
2.38 x lO"*4
1.07 x 10~'f
2.36 x lO"1*
3.57 x lO"1*
1.67 x lO'4
Analysis
VT Weight
g Capacity
(mJl/g) (g/g)
428
502
363
639
708
410
2.98 x 10~4
2.55 x 10"14
0.955 x 10~4
2.31 x lO'1*
3.53 x 10""4
1.42 x lO"*4
ppm on volume/volume basis.
-------
Column 2 in Table 15 lists the challenge concentrations studied
for each sorbate. Since they vary from compound to compound, only de-
rivation of an isotherm from each front allows a comparison to be made
at equal concentration levels. However, n-hexane, n-butylamine, chal-
lenged the column at similar concentrations and temperatures. The order
of adsorption is n-hexane > n-butylamine, which is consistent with the
T
comparative V value determined earlier in elution experiments. The
&
low affinity of the resin for alcohols, even when the column tempera-
ture for ethanol is ^0°C less than for the other sorbates, is apparent
in column 6.
T
Columns 5 and 7 tabulate the V values computed for 50% of the
6 IT!
plateau concentration. These can be compared with the V for elution
O
peaks generated at peak maxima concentrations in approximately the same
range as 50% of the frontal plateau concentration. The elution values
are in column 3. The agreement in general is very good, ranging from
3-15% difference. This definitely confirms that the frontal technique
yields results equivalent to the elution data in the same concentration
range.
It is interesting to note that two of the three adsorbates in most
serious disagreement were at high challenge concentrations (133, 141 ppm
v/v). Another sorbate, n-butylamine, showed excellent agreement at 130
ppm (v/v) challenge concentration. This latter case is possible if the
sorption isotherm is linear over a large vapor phase concentration
range. An example of this will be illustrated shortly.
Frontal chromatograms were easily generated for n-octane at a chal-
lenge level of ^3 ppm (v/v) and were readily reproducible. In addition,
triplicate elution chromatograms were determined for the same solute.
These results with the data generated from frontal and elution analysis
at higher concentrations are presented in Table 16. Included in this
table are the n-octane results of Table 15 for direct comparison purposes.
88
-------
Table 16
Comparison of V and Weight Capacities for n-Octane on XAD-2 From
- - - - ---
Detector
Thermal
Conductivity
Flame
lonization
Flame
lonization
g
Elution and Frontal Analysis
Elution
Challenge T Weight
Concentration g Capacity
(ppm)* OaA/g) (g/g)
60 394
3.08 539
70.4 522
6.67 x 10~6
1.80 x 10
Frontal Analysis
•Ah.Sor'Dtion Desorption
T
V
g
(mA/g)
448
926
464
Weight VT
Capacity g
(g/g) (m£/g)
1.67 x 10 410
_5
1.62 x 10 570
1.86 x 10 550
Weight
Capacity
(g/g)
1.42 x W'1*
9.443 x 10~6
2.219 x ID"1*
ppm on a volume/volume basis.
-------
T
A comparison of column 3 with columns 5 and 7 (V data) indicate
rp g J
good agreement between the elution V values and the frontal V values
o o
for the two high n-octane concentrations. Comparison of these values
for n-Qctane at 3.08 ppm (v/v) challenge concentration shows some dis-
T T
agreement when comparing the frontal adsorption V value with the V
O O
from elution chromatography. Furthermore, there is some discrepancy
T
between the V values estimated for the 70.4 ppm (v/v) concentration
O
(run on the FID system) and those for the 60 ppm (v/v) concentration
(done on the thermal conductivity gas chromatograph).
Weight capacities for the n-octane sorbing on XAD-2 are also pre-
sented in Table 16. Here again there are some significant differences;
however the uptakes for n-octane at the 60 and 70 ppm (v/v) level agree
quite well including the peak maximum uptake value in column 3. The
T
most serious disagreement lies in the V and weight capacities of the
O
n-octane (^3 ppm v/v) for elution and frontal adsorption analysis
T
results. In this case the V value on the adsorption trace is quite
g
high and the skew lies in the adsorption branch. This is definitely
opposite to the behavior observed for n-octane in evaluating its iso-
therm by the elution by characteristic point method (ECP), as will be
shown later.
One reason for this discrepancy may lie in the possibility of some
irreversible adsorption at the low challenge concentration levels which
was not immediately apparent in the 50-150 ppm (v/v) challenge concen-
tration range. It is pertinent to note that five out of the six ad-
sorbates done on the thermal conductivity chromatograph at higher chal-
lenge concentration levels show reductions in the weight capacity of
sorbate desorbed when compared to the initial adsorption uptake. This
otherwise small difference is of the order which would appear signi-
ficant at the 3 ppm (v/v) challenge concentration level.
90
-------
To further examine this phenomena, additional frontal chromato-
grams were run for n-hexane, 3,3-dimethyl-2-butanone, and n-butylamine
at low challenge concentrations. These were all run on the Varian
flame ionization module at temperatures as close as possible to those
previously employed on the thermal conductivity chromatograph at higher
challenge concentrations (see Table 15). The resultant frontal chroma-
togram upon visual examination did not show significant difference
between the adsorption and desorption traces as had been observed for
the n-octane frontal analysis curve.
T
Computation of the 50% breakthrough V for both adsorption and
&
desorption portions of the frontal chromatogram and the resultant weight
T
capacities are listed in Table 17. Also included are V values obtained
g
by elution analysis of bag samples at the same concentration used in
the frontal analysis. In general, the same trends observed for n-octane
T
hold for the additional three adsorbate vapors. Values of V from elu-
ij. g
tion experiments correlate better with the V value from the desorption
O rn
trace, rather than the adsorption trace. The V and weight capacities
O
for the adsorption trace are slightly higher than those obtained from
the desorption trace for the three new adsorbents at low challenge con-
centrations. However, in no case is the discrepancy as high as that
observed with n-octane.
A comparison of the VT values with those obtained at higher chal-
lenge concentrations show that for n-butylamine and 3,3-dimethyl-2-
butanone, there is little difference in V* outside of experimental
error. This supports the earlier statement that even at 138 ppm (v/v),
n-butylamine is in the Henry's Law region of its sorption isotherm.
The same is true for the ketone and it, too, must be in the Henry's Law
region, where VT does not vary with challenge concentration (the weight
capacity will change as shown in Tables 15 and 17). For the two n-
alkanes, the VT values increase with lower challenge concentration,
indicative that the Henry's Law region has still not been totally reached
91
-------
Table 17
M
Comparison of V-£ and Weight Capacities for Sorbates on XAD-2
e
from Elution and
Frontal Analysis (Varian)
Frontal Analysis
ui ..4--: ™
Compound
n- Buty lamine
3, 3- Dime thy 1-2-
butanone
n-Hexane
n-Octane
n-Octane
Challenge
Concentra tion
(ppm) *
7.76
8.21
8.11
3.08
70.4
T Column
g Temp .
464
549
734
539
522
101.6
115.0
87.9
142.6
141.6
A A
yT Weight
g Capacity, q
(m£/g) (g/g) 8
574
629
932
926
464
1.65 x 10 5
3.21 x 10~5
2.59 x 10~5
1.62 x 10~5
1.86 x 10~5
T Weight
g Capacity, q
(m£/g) (g/g) g
467
609
799
570
550
1.47 x 10~5
2.77 x 10~5
2.54 x 10~5
9.44 x 10~6
2.22 x lO"1*
ppm on a volume/volume basis
-------
at 118 and 60 ppm (v/v) for n-hexane and n-octane, respectively,
The slight differences observed between the weight capacities
(Table 17) derived from the adsorption and desorption traces of the
chromatograms supports the earlier hypothesis based upon the results
in Table 15 that some irreversible adsorption is occurring for higher
boiling adsorbates and for those which show a greater affinity for
XAD-2 (the n-alkanes). Again the differences are of the order of a
few micrograms. The capacities are a factor of ten lower at 3 and 8
ppm (v/v) challenge concentrations compared to the higher values
recorded in Table 15. The weight capacity is definitely a function of
challenge concentration.
TO determine whether irreversible adsorption is related to the
temperature at which adsorption and desorption is taking place, the
frontal analysis could be run at a higher temperature. Under these con-
T
ditions, the irreversibility might be overcome and V and weight capa-
O
city should be identical from both traces. It is interesting to note
T
that for the adsorbates in Table 15 the ratios for the V (adsorption)/
T
V (desorption) are approximately the same as the weight capacity (ad-
O
sorption)/weight capacity (desorption) ratio (see Table 18). This
shows how VT and weight capacity are directly proportional in the Henry
g X
Law region and confirms that the V at 50% volumetric breakthrough
O
measures the uptake of the resin for a symmetrical mass transfer front.
E. Adsorption Isotherms
As noted earlier, the evaluation of the sorption isotherm for a
given adsorbate/adsorbent pair can be done by several chromatographic
methods. The advantage of having the adsorption isotherm is obvious
when q is a function of challenge concentration in that it permits
weight8capacities and breakthrough volumes to be calculated as a func-
tion of the same parameter. In addition, it sets the limits of KA'S
applicability, indicates the affinity of resin for a sorbate and allows
93
-------
Table 18
T T
Comparison of the V (adsorption)/V* (desorption)
with Weight Capacity (adsorption)/Weight Capacity (desorption) Ratio
Adsorbate V Ratio Weight Capacity Ratio
&
n-Butylamine 1.23 1.12
3,3-Dimethyl-2-
butanone 1.03 1.02
n-Hexane 1.16 1.15
n-Octane 1.62 1.71
94
-------
the saturation capacity of the sorbent to be ascertained.
To determine the type of adsorption isotherm and its dependence on
the nature of the adsorbate, frontal desorption chromatograms were
analyzed by the characteristic point method (Appendix A). All the iso-
therms for n-octane, ethanol, toluene, 3,3-dimethyl-2-butanone, n-
butylamine and n-hexane were Type I, Langmuir adsorption isotherms on
XAD-2. The departure from linearity was slight in many cases. A typi-
cal isotherm derived from frontal analysis (in triplicate) is shown in
Figure 36. The reproducibility is excellent between analyses. Type I
isotherms have also been reported for n-butanol and diethyl ether on
Chromosorb 102 (54).
Figure 37 shows the sorption isotherms for n-octane and n-butanol
at 98.6°C on XAD-2 and a flow rate of 108 mL/min. It is obvious that
the uptake of n-octane by the resin is considerably greater than for
n-butanol. Attempts to gather data for lower temperature were diffi-
cult due to the failure of the decaying tail of the peak to return to
baseline. Note that the sorption isotherm for n-octane is Type I and
that for n-butanol, the isotherm is linear for a rather large challenge
concentration range. These isotherms were determined by the elution
by characteristic point method (Appendix A).
Since the ECP method is much simpler to run than frontal analysis,
a rigorous comparison of the isotherms generated by these two techniques
for a common sorbate at a common challenge concentration range would be
of value. Isotherms for the two chroraatographic techniques have been
generated from the FID/GC results and are presented in Figures 38 and
39. For the higher challenge concentration range, curves I, II and
III represent the frontal analysis generated isotherm, the elution
curve generated isotherm without correction for kinetic band broadening
(whose major contributors tend to be eddy and longitudinal diffusion and
resistance to mass transfer from the resin bed), and the corrected
95
-------
3.20
2.80
2.40
2.00
1.0x10-4g/g 160
1.20
0.80
0.40
/
/
*
7
>
/
A
>
/
y
^
x
1
it.
Grams adsorbate
Grams adsorbent
Toluene Challenge Cor
Temperature - 124.6°
Sorbent - XAD-2
Flow Rate — 109 ml./i
• Run 1
* Run 2
•A. Run 3
*r
•4
/
r
ppm adsorbate
in gas phase
centration — 141 ppm
C
nin.
10 20 30 40 50 60 70 80 90 100 110 120 130 140
ppm
FIGURE 36 ADSORPTION ISOTHERM OF TOLUENE ON XAD-2 DERIVED FROM FRONTAL
ANALYSIS RESULTS
-------
1.8
1.6
1.4
1.2
1.0 x 10~3g/g
1.0
0.8
0.6
0.4
0.2
0.0
/°
50
100
vs.
Grams adsorbate
Grams adsorbent
Temp. - 98.6°C
Sorbent - XAD-2
Flow Rate 108.7 ml./min.
ppm adsorbate
in gas phase
150
200
ppm
FIGURE 37 COMPARISON OF ADSORPTION ISOTHERMS FOR DIFFERENT TYPES OF ADSORBATES
-------
2.8
2.4
2.0
1.0x10'4g/g 1.6
00
1.2
0.8
0.4
///
f
w
7
•I
?/
V
V
1
#
n
4
i
jfc.
/
V
A/
A
r
\
J
/
f.
/
4
/P
/"
CD
/
/
/
y
i
/
Y
/
r
/<
/
/
/
/
\
<
^
X^
Q
-------
1.0x10-5g/g
1.6
1 4
1.2
1.0
.8
.6
.4
.^
/
j,
r
y
ra/
W
w
•A
_.„..
/
7
V
/
X
/
/
X
/
/
/^
n y
^
/
^
(i)
&
/
/
>
/
/
X
^
X
t
^
/
>
x
^
/
/
4
/
£
/*
\
s
i
(ID
on)
Grams Adsorbate
Grams Adsorbent
n-Octane Challenge Cc
for Frontal Analys
Temperature - 142.6C
Q O A Frontal Analy
• • Elution Analy
• A Elution Analy
ppm adsorbate
in gas phase
mcentration — 3.08 ppm
s
>C. Sorbent - XAD - 2
iis(l)
;is(ll)
>is(lll) (Corrected)
1.0
2.0
3.0
4.0
5.0
6.0
ppm
FIGURE 39 COMPARISON OF SORPTION ISOTHERMS GENERATED BY DIFFERENT CHROMATOGRAPHIC
TECHNIQUES FOR LOW CHALLENGE CONCENTRATIONS
-------
elution curve isotherm, respectively.
The isotherm generated by frontal analysis would tend to give a
higher capacity for a given challenge concentration since there is no
correction for diffusion, etc., in the method. Thus curves I and II
should be similar, and they are. In addition, if one plots the iso-
therm for n-octane challenge concentration of 60 ppm (v/v) in Table 15,
it exactly overlays curve II which confirms the equivalency of the two
methods. Curve III of course is the "true" sorption isotherm.
The curves in Figure 3& correspond to the same sequence above
except for a lower challenge concentration range. These results allow
a comparison of the uptake by the resin by two independent isotherms.
Take, for example, curve III in Figures 38 and 39. For a challenge
concentration of 5 ppm (v/v), Figure 38 predicts an uptake of 1.2 x
10~5 g/g while Figure 39 yields a value of 1.0 x 10~5 g/g. This is
excellent agreement between the two isotherms.
Figure 40 depicts the sorption isotherm for n-octane at different
flow rates on XAD-2 resin. It is interesting that the resin uptake is
very similar at the three flow rates examined, although at a challenge
concentration level above 70 ppm (v/v) there appears to be a decrease
T
in uptake with flow rate. The results from elution analysis V trends
with flow rate would predict a decrease in uptake at 550 mL/min.
(Figures 27-28), however, this is not discernible in Figure 40 except
at the higher challenge concentration level. This points out the
T
superiority of the V measurement for sorbate uptake in the Henry's Law
O
region of the isotherm. One complementing fact obtained from Figure 40
is that slope of the n-octane isotherms as the challenge concentration
tends toward zero are nearly equal. This slope is equal to K , the
T
adsorption coefficient, hence the decrease in V with flow rate noted
6
earlier for sorbatas on XAD-2 must be due to the reduction in effective
surface area available to the adsorbate.
100
-------
l.Ox 1CT3g/g
Grams adsorbate
Grams Adsorbent
Temp. - 99.0°C
Sorbent - XAD-2
Sorbate — n-Octane
ppm adsorbate
in gas phase
O 105.7 ml./min.
281.1 ml./min.
+ 550.3 ml./min.
100
150
ppm
200
250
FIGURE 40 ADSORPTION ISOTHERM DEPENDANCE ON FLOW RATE OF CARRIER GAS
-------
In summary, the results obtained from the isotherm determination
T
tend to complement qualitatively and quantitatively the V data described
6 f£
earlier. Besides quantifying the limits of the applicability of V in
5
terms of the challenge concentration, the adsorption isotherms deter-
mined at higher temperatures are applicable for use in thermal desorp-
tion of organic solutes from resins. In addition, a study of isotherm
dependence on temperature allows the computation of breakthrough
volume at any challenge concentration.
102
-------
V. CONCLUSIONS AND RECOMMENDATIONS
The data presented in this report support the use of chromatographic
elution data to characterize breakthrough and sorption capacity of sor-
bent cartridges containing synthetic resin. Specific retention volume
data as a function of temperature, adsorbate type, and challenge con-
centration have been gathered on XAD-2 and Tenax-GC resin.
Complementary frontal analysis experiments indicate that the VT
g
values accurately measure breakthrough in relatively short sorbent
T
cartridges. The quantity of V data have permitted correlations to be
O m
developed between adsorbate physical properties and V , which permit
•£ 8
reliable estimation of V for other adsorbates without resorting to
6
experimental measurement.
Selectivity toward non-polar organic species is shown by both
resins, a trend supported by isotherm data of different sorbate types
on both resins. Experiments conducted at high gas velocities show
T
little variance in V with flow for Tenax-GC, but an apparent loss of
g
15-20% for several chemically dissimilar sorbates on XAD-2 at face
velocities greater than 64 cm/sec. This in part may be attributed to
mass transfer control of the sorbate in the microporous resin.
The experimental techniques and theory developed in the course of
this program are generally applicable to characterizing any sorbent
media. One of the principal advantages of the gas chromatographic
technique is the ability to determine the required data with small
quantities of sorbate. The gas chromatographic technique thus mini-
mizes worker exposure to test chemicals, especially important for
studies of suspected carcinogens, as well as saving time and money.
The data tabulated in this document also have application in analy-
sis techniques requiring thermal desorption of organic matter from
103
-------
sorbent media. In theory the data are also useful for scale-up designs
to enlarged industrial adsorption processes. However, in that case
parameters such as AH which are not of prime significance in this;
A
study, will be of greater importance.
Correlations have been observed between the specific retention
T
(elution) volumes, V , of adsorbates on XAD-2 and Tenax-GC and their
g
boiling points. These correlations may be used to extend the present
data to other compounds of interest. The correlations are best when
treated by specific compound classes.
Further studies in this area should include a survey of the effects
T
of common combustion gases, such as CC>2 and 1^0, on the V values. The
O
question of mutual interactions between two or more adsorbates, in the
vapor state or on the adsorbent surface, also warrant careful study
since the breakthrough and uptake characteristics may be mutually
dependent. Additional extensions which warrant consideration are
T
establishing V values on volatile.metallic and organometallic
O
species. Finally, the measurement of specific retention volume data on
other adsorbents, such as charcoal, which would retain some of the
species which are not retained on Tenax-GC or XAD-2 would be of value.
104
-------
VI. REFERENCES
" D'C" Physic°chercical Measurements Using Chromatography in
Advances in Chromatography, Vol 14", J.C. Gidding, E. Grushka,
J. Gazes, and P.R. Brown, eds., Marcel Dekker, Inc., N.Y. , 1976,
pp 87-198.
2. Kobayashi, R..H.A. Deans, and P.S. Chappelear, Physico-Chemical
Measurements by Gas Chromatographv in "Applied Thermodynamics",
American Chemical Society, Washington, B.C., 1968, pp 227-246.
3. Novak, J. , V. Vasak, and J. Janak, Anal. Chem. , 37., 660 (1965).
4. Gelbicova-Ruzickova, J. , J. Novak, and J. Janak, J. Chromatog. ,
64, 15 (1972).
5. Selucky, M. , J. Novak, and J. Janak, J. Chromatog., 28_, 285 (1967).
6. Wood, G.O., "Development of Air-Monitoring Techniques Using Solid
Sorbents", LASL Project R-059, NIOSH-1A-75-31, LA-6513-PR Progress
Report, September, 1976.
7. Snyder, A.D. , F.M. Hodgson, M.A. Kemmer, and J.R. McKendree, "Utility
of Solid Sorbents for Sampling Organic Emissions from Stationary
Sources", EPA Report 600/2-76-201, July, 1976.
8. Janak, J. , J. Ruzickova, and J. Novak, J. Chromatog., ^9_ 689 (1974).
9. Butler, L.D., and M.F. Burke, J. Chromatog. Sci. , 14, 117 (1976).
10. Adams, J. , K. Menzies, and P. Levins, "Selection and Evaluation of
Sorbent Resins for the Collection of Organic Compounds", EPA Report
600/7-77-044, April, 1977.
11. Conder, J.R. , Physical Measurement by Gas Chromatography in "Progress
in Gas Chromatography", J. H. Purnell, ed., Interscience, N.Y.,
1968, p. 228.
12. Littlewood, A.B. "Gas Chromatography", Academic Press, N.Y. , 1970,
pp. 48-50 .
13. Kiselev, A.V. and Y.I. Yashin, "Gas-Adsorption Chromatography",
Plenum Press, N.Y., 1969, p.20.
14. Ackerman, D.G. , "Chromatographic Adsorption Studies Using a Computer-
Controlled Data Acquisition System:, Ph.D. Thesis, Univ. of Arizona,
1973.
15. Steel, W.A. , Adv. Colloid Interface Sci, I, 3 (1967).
16. Dal Nogave, S. , and R.S.. Juvet, Jr., "Gas-Liquid Chromatography",
Interscience, N.Y., 1962, p. 89.
105
-------
References (continued)
17. Gregg, 'S.J., and R. Stock, Sorption Isotherms and Chromatographic
Behavior of Vapours in "Gas Chromatography 1958", D.H. Desty, ed.,
Butterworths, London, 1958.
18. Kuge, Y., and Y. Yoshikawa, Bull. Chetn. S oc., Japan, J38, 948 (1965).
19. Huber, J.F.K., and R.G. Gerritse, J. Chromatog., 58, 137 (1971).
20. Conder, J.R., and J.H. Purnell, Trans. Faraday Soc., 65, 824 (1969).
21. Cremer, E., and H.R. Huber, Agnew. Chem., 73, 461 (1961).
22. Sewell, P.A., and R. Stock, J. Chromatog., 50_, 10 (1970).
23. Kiselev, A.V., and Y.I. Yashin, "Gas-Adsorption Chromatography",
Plenum Press, N.Y., 1969, p. 114.
24. Cremer, E., and H.F. Huber, "Gas Chromatography," N. Brenner,
J.E. Cullen, and M.D. Weiss, eds., Adademic Press, N.Y., 1962,
p. 171.
25. Neumann, M.G., J. Chem. Ed., 53, 708 (1976).
26. Dear,D., A. Dillon, and A. Freedman, J. Chromatog., 137, 315 (1977).
27. Rappaport, S.M. et al, "Development of Sampling and Analytical
Methods for Carcinogens," LASL Project R-219, LA-6387-PR, June, 1976.
28. Pellizzari, E.D., "Development of Analytical Techniques for Measuring
Ambient Atmospheric Carcinogenic Vapors", EPA Report 600/2-75-076,
November, 1975.
29. Dubnin, M.M. , J. Colloid Interface Sci. , _2_3, 487 (1967).
30. Kiselev, A.V., and Y.I. Yashin, "Gas-Adsorption Chromatography",
Plenum Press, N.Y., 1969, pp. 11-16.
31. Lochmuller, C., Private Communication.
32. Ackman, R.G., J. Chromatog. Sci., 10, 506 (1972).
33. Gvosdovich, T.N., A.V. Kiselev, and Y.I. Yashin, Chromatographia, 6_,
179 (1973).
34. Lange, N.A., ed., "Handbook of Chemistry", Handbook Publishers, Inc.,
Sandusky, Ohio, 1952.
35. Weast, R.C., ed., "Handbook of Chemistry and Physics - 44th Edition",
Chemical Rubber Co., Cleveland, Ohio, 1964.
106
-------
References (continued)
36. Kiseley, A.V. , and Y.I. Yashin, "Gas-Adsorption Chromatography",
Plenum Press, N.Y. , 1969, p. 60.
37. Gearhart, H.L. and M.F. Burke, J. Chromatog. Sci., 3JU 411 (1973)
38. Frontasev, V.P., and L.S. Shraiber, Russ. J. Phys. Chem., 43,
229 (1969).
39. LeFevre, R.J.W., and K.D. Steel, Chem. Ind., 670 (1961).
40. Applequist, J., J.R. Carl, and K. Fung, J. Amer. Chem. Soc., 94.
2952 (1972).
41. Batsanov, S.S. "Refractometry and Chemical Structure",
D. Van Nostrand, Princeton, J.H., 1969.
42. Taterskii, V.M., V.A. Bonderskii, and S.S. Yavovoi, "Rules and
Methods for Calculating the Physico-Chemical Properties of Paraf-
finic Hydrocarbons", Pergamon Press, N.Y., 1961.
43. Kiselev, A.V., and Y.I. Yashin, "Gas-Adsorption Chromatography",
Plenum Press, N.Y., 1969, p. 58.
44. Rakshieva, N.R., S. Wicar, J. Novak and J. Janak, J. Chromatog.,
9^, 59 (1975)
45. Oberholtzer, J.E. and L.B. Rogers, Anal. Chem., 41, 1590 (1969).
46. Little, J.N., and W.J. Pauplis, Separation Sci., k, 513 (1969).
47. Kelley, R.N. , and F.W. Billmeyer, Jr., Anal. Chem., _42_, 399 (1970)
48. Moreland, A.K. , and L.B. Rogers, Separation Sci., 6_, 1 (1971).
49. Rakshieva, N.R. , J. Novak, S. Wicar and J. Janak, J. Chromatog.,
91, 51 (1974).
50. Gaha, O.K., J. Novak and J. Janak, J. Chromatog., 84, 7 (1973).
51. Giddings, J.C., Gas Chromatography 1964 (A. Goldup, ed.) Elsevier,
Amsterdam, 1965, p. 3.
52. DeLigny, C.L., J. Chromatog. , 35_, 50 (1968)
53. Hartkopf, A., and B.L. Karger, Accts. Chem. Res., 6_, 209 (1973).
54. Kiselev, A.V., and Y.I. Yashin, "Gas-Adsorption Chromatography",
Plenum Press, N.Y., 1969, p. 118.
55. Gvosdovich, T.N., A.V. Kiselev, and Y.I. Yashin, Chromatographia,
2, 234 (169).
107
-------
References (continued)
56. Sakodinsky, K., Chromatographia, JL, 483 (1968).
57. Hamersma, J.W. , S.L. Reynolds, and R. F. Maddalone, "IERL-RTP
Procedures Manual: Level 1 Environmental Assessment", EPA Report
600/2-76-160a, June 1976.
108
-------
APPENDIX A
Theoretical Considerations
Several chromatographic methods have been employed in this study
to produce specific retention volume data, adsorption isotherms, etc.
Each method, however, comes from one central theoretical derivation.
The relationship between a chromatographic elution profile (whether
it be frontal, elution, etc.) and a distribution coefficient, such as
the adsorption coefficient, K , has been derived by Conder (11) as:
(1)
3
where j = pressure correction = -r-
Y = mole fraction of the adsorbate in the gas phase
q = concentration of the adsorbate in the adsorbent in moles/cm2
(better known as the surface excess, T)
c = concentration of the adsorbate in the gas phase in moles/cm3
A°= specific surface area of the adsorbent in m /g
s
P = column outlet pressure
P = column inlet pressure.
-4
In the studies reported here, y is always smaller than 10 , hence
equation (1) reduces to
As
/T s
109
-------
If the chromatographic experiment is being conducted in the low surface
coverage region, Henry's Law applies and(-r^) is a constant such as K*
\ /
so equation (2) becomes
m JU O
VT = K* A (3)
g As
This equation is strictly applicable to the case of elution chromato-
graphy using an infinitely small sample (adsorbate) size. This situa-
tion frequently is approached in many of the environmental sampling
T
cases. It is significant that in this Henry's Law region, V is
g
independent of challenge concentration, c.
If -r^- is not constant, then V is input dependent, which means
T
that V is dependent on c. To find the amount of sorbate captured by
6
the sampling resin requires integration of equation (2) as
dc (4)
This integration can be performed on an entire elution peak or frontal
analysis profile, or if either profile approximates equilibrium, con-
secutive integrations may be performed for various values of Cg to
yield q.
Equation (4) is equally applicable to elution analysis or frontal
analysis for those cases where -r^ is either constant or not constant.
However the following points must be considered:
(1) The primary difference in frontal and elution analysis is in
the method of sample introduction. In elution analysis, a small quanti-
ty of adsorbate Is injected onto the sorbent cartridge in an infinites-
imally small time, t . For the frontal analysis case, the sample injec-
tion time (t,.) is very long (continuous until interrupted) hence the seal
110
-------
difference in the two methods is in the inequality, t
(2) The mode of sample introduction in no way affects the position
of the frontal boundary curve or the elution peak maximum, since both
are defined by equation (1) and theoretically should elute as pictured
in Figure A-l. Figure A-l is quite different from some of the chromato-
grams in standard chromatography texts illustrating the relationship
between frontal and elution analysis, in that it notes that an elution
peak maximum only coincides with the 50% volumetric breakthrough point
in elution analysis if the input concentration is one-half that for the
frontal plateau concentration.
(3) Both frontal and elution methods, even in the Henry's Law
region may have kinetic effects superimposed on their profiles due to
diffusional resistance or mass transfer effects. These phenomena account
for the fact that real peaks and fronts are not straight lines and step
functions, respectively, on the chromatogram. Such kinetic effects
must be corrected for if accurate equilibrium isotherms are to be ob-
tained from chromatographic experiments. They do, however, provide
a valuable source of information with regard to maximizing the sampling
efficiency, if they can be properly deconvoluted from the equilibrium
profile. Kinetic effects are less difficult to handle for the Henry's
Law region since the peak maximum does approximate equilibrium on the
chromatographic profile.
A. Frontal Analysis
In the frontal analysis technique, a gas containing a specified con-
centration, C-, of sorbate is continuously fed into the sorbent cartridge.
The experimenter waits for the appearance of the boundary profile and
continues to monitor the "breakthrough" of sorbate until it equilibrates
with the cartridge for the challenge concentration specified. The
equilibration stage is signaled by the onset of a concentration vs.
time "plateau" which can be continued for as long as one wishes in a
111
-------
Frontal Profile
Volume (Time)
FIGURE A-1 RELATIONSHIP BETWEEN FRONTAL BREAKTHROUGH
CURVE AND ELUTION PEAK
112
-------
"steady state" condition. If the breakthrough curve is sharp, or a
symmetrical sigmoid profile, then q can be computed by
(5)
"A
T
where V = volume of carrier gas passed to elute a symmetrical break-
8 through curve at 0.5 Cg.
CB = plateau, input, or challenge concentration
W. = sorbent weight
For asymmetric boundary profiles, integration is required to obtain
:, so that
8- g
/
V dc (6)
Co
where V = volume of carrier gas passed to obtain completion of boundary
profile.
which is the cross-hatched area in-Figure A-l.
In Figure A-l, Point A marks the commencement of the frontal analysis,
while distance AB represents the total challenge concentration presented
to the sorbent bed. Distance AG is the total volume of gas containing
sorbate at concentration, CB, required to complete breakthrough. Hence,
the ratio of the cross-hatched area (representing the adsorbate taken
up by the adsorbent) to the area ABFG, times the total sorbate passed
over the sorbent in V, volume of gas, gives the capacity of the sorbent
bed at sorbate input, CR.
The later sections of this report treat the use of frontal analysis
in determining sorption isotherms. However, the basic theory presented
here in equally applicable to isotherm determination.
113
-------
B. Specific Retention (Elution) Volumes, VJ
- * - g
The relationship of the specific retention (elution) volume
to measurable parameters in a gas chromatographic experiment has been
derived in many standard treatises on gas chromatography (12). It is
important to realize the vj, is the fundamental retention constant in
gas chromatography and accounts for the effect of flow rate, pressure
drop, temperature, column void volume, and stationary phase weight
(volume or surface area) on the retention of an injected solute.
T
Knowledge of the value of V [frequently for convention corrected
O
to 0°C (273°K)] allows one to estimate the retention volume of a solute
T
at another temperature or for a different column length. Thus, V
O
determined from conventional gas chromatographic columns can aid in the
design of sorbent sampling modules.
The specific retention volume is also directly relatable to funda-
mental phase distribution constants, such as the partition coefficient,
1C or the adsorption coefficient, K.. Thus, if certain physical charac-
teristics of the stationary phase are known, such as A°, then K. can be
S A
obtained. K. then allows calculation of sorbate distribution for larger
sorption systems than analytical scale devices.
T
Specific retention volumes, V , in this study were computed according
8 rj,
to the following formula, and represent V at the temperature of the
O
column oven in the chromatograph proper:
VT _ c r - (7)
g- WA
T
where V = specific retention volume for the adsorbate at column (sorbent
^ trap) temperature
= Fl^-jll- — 1= flow rate of carrier gas at column temperature
114
-------
F = flow rate of carrier gas at ambient temperature and pressure
ci
T = column temperature
T = ambient temperature
3.
P = ambient pressure
PW = vapor pressure of water (at temperature of flow meter)
t = peak maximum retention time
t = retention time for a completely non-sorbed solute
3.
W. = adsorbent weight
P = column outlet pressure
P. = column inlet pressure.
C. Determination of Adsorption Coefficients (K,)
The relationship between elution volume for an individual adsorbate
and the equilibrium distribution of the adsorbate between the gas phase
and the sorbent resin is described by:
where C = concentration of adsorbate in gas phase in moles/length of
column
C. = concentration of adsorbate in adsorbent in moles/length of
column
v = average linear velocity of gas phase
z = distance from sorbent cartridge inlet
t = time
The three terms in the differential equation represent respectively,
115
-------
left to right, the adsorbate removed from the flowing gas stream, the
increase of the adsorbate in the gas phase, and the increase of the
adsorbate on the adsorbent.
In order to simplify this differential equation, one may integrate
this equation for values with known limits, such as 0 ^. z = 3C /3C , equation (11) becomes:
A g
dz.
dt
(12)
upon inverting both sides of the equation. Equation (12) is the differ-
ential equation describing the rate of motion of an adsorbate zone of
116
-------
concentration, C , in terms of v Integration over the column length, L,
o
and the time it takes for the peak maximum to elute, t , gives
/
r rL
dt = (1 + y dz (13)
O ' O
Or * - (14)
The velocity of the gas stream is determined by injection of adsor-
bate, a, with a K = 0, so that its residence time in the column is t ,
so
j.F L
v = -^- (15)
a
P V
r _ AVA
g "
TA W
c. = —r-5- (17)
where T = concentration of adsorbate/m2 of adsorbent (moles/m2)
P. = adsorbate partial pressure in the gas phase (mm Hg).
A
Thus
N g x A
which yields upon substitution into (14)
vt W.A RT
—* = 1 + ATTS ° 1-r^-l (19)
117
-------
or when combined with (15)
W.A RT /~r \ ,0fn
i _i_ As c / 3F } (20)
1 ~~V\3P/
Oi i\
which can be further simplified by multiplication of the equation by
V
a
a
3Vr-'. , ART (3L.\ <21>
1 fc**"\w*A/
A x A/
T
where the left hand side of the equation is simply V (refer to
O
equation 7). For the Henry's Law region of the absorption isotherm,
T
V is constant, hence
T'
. o
V = A RT
~ » "••<• I VTri I /11\
g s c I 9PAy p— 0 (22)
where A° is the limit of A as p—^-0, and the term in parentheses is
S S
known as Henry's Law adsorption coefficient, K.. A knowledge of K. at
A A
the temperature of adsorbate collection in a resin tube is valuable in
establishing the quantity of adsorbate vapor which could be collected
under conditions approaching equilibrium. This would then allow one
to design traps of sufficient capacity to capture sufficient analyte
for assay, aid in the choice of a sorbent resin, and in evaluating the
collection of the sorbent resin.
(23)
Computation of K is readily facilitated by employing equation
A
VT = A° RT K. (23)
g s c A
T
The units of K. in this case are moles/mm-m2, thus V /A° is in ml/m2,
A 8s
R = 6.3 x 101* ml-mm/mole-°K, and T in °K. Note should be made that
118
-------
* o T T
K is proportional to q, since K.A = V , thus the magnitude of V can
A A s g g
serve as a quick guide as to sorbate affinity for the resin.
The specific surface area of the resins used in this study repre-
sents an average of several values determined by Micromeritics Instru-
T
ment Corp., and literature values (10). The quantity V /A° is
g s
referred to as the retention volume/unit of surface area by Kiselev (13)
T
In essence, this form is a form of "reduced V " in which the surface
g
area dependence has been factored out of the specific retention volume.
Thus tabulations of K. can be used to compute breakthrough volumes on
resins whose surface area characteristics vary between production lots.
It is implicit in the calculation of K that the BET measured surface
£\
area is totally available to the adsorbate. Studies by Ackerman (14)
refute this and, as will be shown later, high velocities through the
adsorbent bed may limit the effective surface area available to the
solute, particularly in the micropores of the resin.
The adsorption coefficient can be used to calculate q if the chal-
lenge concentration, c, is in the Henry's Law region of the adsorption
isotherm. Figure A-2 indicates that K would be applicable to calcu-
depending on the departure of the sorption isotherm
from linearity. This is why the total adsorption isotherm can be of
value over a wider concentration range than just covered by the
Henry's Law region.
The concentration of the adsorbate in the adsorbent, q, can be
calculated as grams of adsorbate/gram of sorbent by using the following
equation
K. A°C (760 mm Hg) (MW)
= _A_s_S (24)
g
119
-------
FIGURE A-2 RELATIONSHIP BETWEEN ADSORPTION ISOTHERM AND ADSORPTION
COEFFICIENT. KA
120
-------
where C = gas phase concentration in ppm (v/v)
6
MW = molecular weight of the adsorbate
D. Thermodynamic Functions of Adsorption
Knowledge of the temperature dependence of K allows the computa-
A
tion of several thermodynamic functions of adsorption. Values of the
molar free energy (AG ), molar enthalpy of adsorption (AH ), and molar
A A
entropy of adsorption (A~3 ) may all be calculated.
A
The differential molar free energy of adsorption AG is given by
A
AG. = -RT In K. (25)
A A
The logarithmic term usually contains a standard state defined for
K. in units consistent with a unitless logarithmic term. Unfortunately
there is no universally accepted standard state of adsorption process,
although several have been suggested (15).
The computation of AG has not been pursued in this study since
it primarily reflects the trends in K and hence ultimately is pro-
T
portional to V . It lends little to the discussion of the experimental
O
results, but could be of interest to theoretical surface chemists.
The differential molar enthalpy of adsorption, AH is calculated as
AHA = 3A GA/3(l/Tc) (26)
or
3 log V
-2'302R 71I7F? (27)
c
121
-------
Thus plots of log VT vs. 1/T readily yield AHA for the temperature
range of interest.
The differential molar enthropy of adsorption, ASA> is calculated
from the Gibbs-Helmholtz equation, where
AS
AHA ~ AGA (28)
'A T
AS. will contain both the error inherent in the determination of AG.
and the slope measurement for AH and hence frequently is of limited
A
statistical significance.
The differential molar heat of adsorption is perhaps the parameter
of most significance in this study since its magnitude and sign relative
to the heat of liquefaction, AIL , of the sorbate allows a qualitative
assessment to be made of the strength of adsorbent/adsorbate inter-
— — 9 log VT
actions. Since AIL is negative, and AH is negative if ° g is
X A 9 d/Tc)
positive, then AIL < AH is indicative of a strong enthalpic interaction
between the sorbent and sorbate. If AIL -^ AH , then marginal inter-
action of the adsorbate with the resin is occurring, and the resin is
merely serving as an inert surface allowing liquefaction of the adsor-
bate.
E. Determination of Adsorption Isotherms From Chromatographic Profiles
There is a direct relationship between the shape of a chromato-
graphic profile (either in elution or frontal analysis) and the shape of
the equilibrium adsorption isotherm. Figure A-3 illustrates this
correspondence for three types of cases. The top set of graphs depict
a linear, Langmuir, and anti-Langmuir isotherm, respectively. Beneath
each type of isotherm is the resultant elution and then frontal profile,
shown in the presence of kinetic band broadening of the chromatographic
122
-------
Linear
Langmuir
Sorption Isotherm
Anti- Langmuir
NJ
Time
Time
Elution Analysis
A
Time
Time
Time
Frontal Analysis
Time
FIGURE A-3 RELATIONSHIP BETWEEN ADSORPTION ISOTHERM, ELUTION ANALYSIS, AND FRONTAL ANALYSIS
-------
zone. Thus, if the skew factor on an elution peak is pronounced, one
can immediately ascertain the type of sorption isotherm governing phase
equilibrium. The same corollary holds true for the frontal analysis
profiles shown in Figure A-3. For the linear isotherm case, the elution
peak is assumed to be a symmetrical Gaussian distribution, likewise for
the frontal case, both adsorption and desorption branches of the frontal
profile are symmetrical.
In examining a chromatographic profile for skew one should be aware
of several factors. For example, unfavorable adsorption/desorption
kinetics can contribute to a skewed profile. Operating the chromato-
graphic experiment at mobile phase flows far from the minimum in the
van Deemter equation (16) can cause excessive band broadening, particu-
larly in the high flow regime where resistance to mass transfer is
maximized. It is also possible to start at an input concentration
which corresponds to one section of an isotherm containing an inflec-
tion point. Thus a different input or challenge concentration may be
in the other portion of the isotherm, and hence give a skew in the oppo-
site direction from that encountered at the previous challenge concen-
tration.
This is illustrated in Figure A-4 by some literature data repre-
senting the work of Gregg and Stock (17). The top half of Figure A-4
represents Langmuir isotherms with the associated frontal chroma togram
for each case. For these cases, the skew in the desorption trace cor-
responds to the isotherm. For the second example pictured in the lower
half of Figure A-4, the isotherm is concave to the pressure axis in one
region of concentration and convex in the other region. The frontal
chromatogram inserts show that depending on the input pressure (concen-
tration) one of several frontal profiles could result. The same pheno-
mena can be encountered in elution analysis (18).
Several chromatographic methods exist for isotherm determination,
varying in their experimental complexity, attainment of equilibria, and
124
-------
300
(b)
100
Pressure,p
150
200
20 40
60
t
BO
100
min
Systems showing Type I isotherms, on silica gel A. (a) Sorption isotherm for (i) n-hexane, (ii) n-pentane; circles,
determined with the sorption balance; crosses, calculated from the chromatogram. (bl Chromatogram for n-hexane.
(c) Chromatogram for n-pentane, l^ time of change-over to pure nitrogen.
U
o>
I
10
) 50 100 0 20 40 0 20 40
Pressure p mm . t (min) t (min)
Cyclohexane on block-dried calcium carbonate (Type II and Type III isotherms), (a) Isotherms with
the calcium carbonate: (i) in dry form, (ii) after treating with water; circles, determined with the
sorption balance; crosses, calculated from the chromatograms. In (b) and (c) the column was
saturated to a concentration corresponding to X and in (d) and (e) to a concentration corresponding to Y.
\' time of change-over to pure nitrogen.
FIGURE A-4 COMPARISON OF ISOTHERM CURVATURE WITH FRONTAL CHROMATOGRAM
EXPERIMENTAL DATA (REFERENCE 17).
125
-------
applicable adsorbate concentration range. Huber (19) has classified such
techniques into equilibrium and nonequilibrium classes. The former class
of techniques experimentally includes complex pneumatic instrumentation
to produce concentration plateaus (frontal analysis) while the latter
techniques involve elution pulses which are simpler to produce experi-
mentally, but require extensive computational corrections in the higher
adsorbate concentration ranges. Purnell (20) has advocated an alternative
classification scheme based upon the method used to produce the chromato-
graphic profile.
Elution by characteristic point (ECP) has been amply described in
the literature (21, 22). Two basic methods exist in the ECP method for
isotherm determination; one in which a series of peak maxima specific
T
retention volumes, V , are determined, each yielding one point on the
rn g
isotherm. These V volumes will be invariant in the Henry's Law region,
g
but will show a variance in the nonlinear portion of the isotherm as the
concentration of the adsorbate in the carrier gas stream decreases from
the injected maximum concentration to zero. The second method involves
the determination of a portion of the sorption isotherm from the locus of
points on the diffuse profile of the elution pulse, implying that such a
boundary for all ideal forms of chromatography represents a true equilibrium
profile. Hence, a coincidence criterion is developed by which the diffuse
edge of a large elution profile may be regarded as a summation of discrete
peak maxima. Thus, as Kiselev (23) has noted, the proper choice of experi-
mental conditions can assure a close approach to equilibrium, and the
simplification of determining small portions of the total isotherm via a
single peak (20) is realized. Figure A-5 shows several examples of
elution by characteristic point, the isotherms derived from this method,
and varying degrees of coincidence of peak maxima with the asymmetrical
boundary profile.
In order to calculate the equilibrium adsorption isotherm, the
chromatographic profile must be corrected for kinetic band broadening.
126
-------
4 5
Q4 _
Q2 .
±
;312/3 ;SM"
_
o
E
1 -
0 100 200 300 400 500 600 700 800
p, mm Hg
1 '
25
20
0)
I 15
"" 10
0.5
90°
0 10 20 30 40 50 60 70 80
p, mm Hg
119°
90'
0 20 40 60 80 100 120 140
p, mm Hg
FIGURE A-5 EXAMPLES OF CHROMATOGRAMS FOR VARIOUS SAMPLE
SIZES AND THE RESULTING ISOTHERMS FOR CHROMOSORB 102
USED AT VARIOUS TEMPERATURES WITH WATER, DIETHYL
ETHER AND n-BUTANOL (REFERENCE 54)
127
-------
To determine the divergence from equilibrium of the chromatographic
profiles, the simple graphical correction suggested by Cremer (24) has
been used for the elution by characteristic point method mentioned earlier.
This procedure involves bisection of the chromatographic peak profile
at its peak maximum value, by erecting a perpendicular to the base line
(Figure A-6). The partitioned peak may be regarded as consisting of two
sections: one representing the change in concentration with respect to
time due to solute distribution nonequilibrium, and the other consisting
of the solute distribution nonequilibrium contribution plus the concen-
tration change due to equilibration between the phases. Subtraction of
the nonequilibrium area from the area representing both processes yields
the true equilibrium profile (dashed line in Figure A-6). This process is
depicted in Figure A-6 for n-butanol (0.05 yl liquid injection) eluting
from XAD-2 at
-------
ro
Divergence from Equilibrium
Volume
FIGURE A-6 CREMER METHOD FOR CORRECTING ELUTION PROFILE FOR KINETIC
CONTRIBUTION (0.05 jUl LIQUID INJECTION OF n-BUTANOL ON XAD-2 @ 100°C.)
-------
A = peak area
W. = sorbent weight
A
p = partial pressure of sorbate over resin
R = 6.232 x 104 mL-mm/mole ~ °K
s = chart speed
T = column temperature
h = vertical height corresponding to partial pressure
of sorbate in the gas phase
j = James-Martin compressibility factor
F = flow rate of carrier gas at column temperature
Equations (29) and (30) are directly derivable from equation (4)
with a change in integration variables (25).
An analog similar to the elution case presented above exists for
frontal analysis. Frontal analysis by characteristic point consists of
calculating the equilibrium isotherm from a single diffuse boundary pro-
file. Again it assumes that intermediate frontal analysis profiles coin-
cide with the diffuse boundary profile. The resin bed is challenged with
different concentrations of sorbate, thus several points on an adsorption
isotherm can be determined as given in equation (6) and depicted in
Figure A-7. The variable changing in this case is C- in equation (6),
hence determining the isotherm simply becomes a series of integrations at
different values of C^,.
D
Figure A-7 shows the evaluation of q for both the desorption as
O
well as the adsorption branch of a frontal analysis chromatogram. Here,
q is determined by graphical integration so the q , .is given
g j o r e Hg> adsorption &
as
_ Area of ABDF
qg, adsorption Area of ABDEF DE B
(31)
and
_ Area of A'B'DJF'
qg, desorption ~ Area of A'B'D'E'F' A'F' °B
(32)
130
-------
Adsorption
Desorption
Bf
D'
c
0
N
C
E
N
T
R
A
T
I
O
N
E A'
Volume (Time)-
FIGURE A-7 GRAPHICAL EVALUATION OF q FROM FRONTAL ANALYSIS ADSORPTION AND
DESORPTION CURVES
-------
where:
Vnv = volume of eluent required to completely adsorb
uhi ,
sorbate
V.,„,= volume of eluent required to completely desorb sorbate
Point A marks the moment that sorbate is entering the sample cart-
ridge while point B' marks the withdrawal of the challenge mixture and
beginning of the dissipation of sorbate from the cartridge. If the
experiment is performed in the Henry's Law region, then
iD'niT?'
Area ABDF = Area A'B'D'F
and
, adsorption ^g, desorption
The two methods for determining isotherms described above each offer
distinct advantages. The elution analysis technique of determining iso-
therms is experimentally simple and probably adequate for the challenge
concentration ranges encountered in many environmental sampling situa-
tions. Unfortunately the sorbate is diluted by the carrier gas, limit-
ing the concentration range over which the sorption isotherm can be
determined (this is not a problem in frontal analysis). However, another
advantage in the elution by characteristic point method is that the
kinetic contribution to distortion of the chromatographic profile can be
corrected with less difficulty than for the frontal analysis method.
132
-------
APPENDIX B
Specific Retention Volume Data
133
-------
OJ
.Table B-l
Specific Retention Volumes (ml/g)
Resin: XAD-2
Adsorbate
(ml/g)
Temperature
n-Hexane
n-Octane
n-Decane
— & —
261
740
1870
693
2280
7760
1670
4440
•V — * —
110.0
89.7
73.0
130.4
108.7
89.7
148.8
130.4
Adsorbate Class:
Sloe
Intercept
3.06580 -5.58655
3.77333 -6.51827
3.93204 -6.10014
Aliphatic
Hydrocarbons
Correlation
Coefficient
0.99995
0.99973
1.00000
n-Dodecane
7530
148.8
-------
Ui
Table B-2
Specific Retention Volumes (ml/g)
Resin:
Adsorbate
n-Hexane
n-Octane
n-Decane
n-Dodecane
Tenax-GC
vj (ml/g)
82.4
266.0
778.0
168.0
430.0
144.0
208.0
562
1850
684
2380
Temperature
(°C)
111.9
89.2
69.9
131.0
110.9
89.2
148.3
130.5
110.1
148.3
130.8
Adsorbate Class: Aliphatic
Hydrocarbons
Correlation
Slope Intercept Coefficient
3.06668 -6.04832 0.99988
3.26578 -5.86329 0.99961
4.01717 -7.21373 0.99974
5.30902 -9.76855 1.00000
-------
Table B-3
Specific Retention Volumes (ml/g)
Resin:
Adsorbate
Benzene
Toluene
P-Xylene
Ethylbenzene
n-Propylbenzene
XAD-2
v£ (ml/g)
O
164
329
858
204
428
1100
458
1060
2730
430
1030
2190
774
1880
Temperature
(°C)
129.8
110.3
90,0
150.3
129.8
110.3
150.3
130,0
110.3
150.3
130.0
111.5
150.0
131.7
Adsorbate Class: Aromatic
Hydrocarbons
Correlation
Slope Intercept Coefficient
2.71260 -4.53329 0.9960
2.97094 -4.72089 0.9970
3.14611 -4.77490 0.9998
2.96535 -4.36304 0.9981
3.60543 -5.63485 1.0000
-------
Table B-4
Specific Retention Volume (ml/g)
Resin:
Adsorbate
Benzene
Toluene
o-Xylene
Ethylbenzene
n-Propylbenzene
Tenax-GC
VT (ml/g)
&
107
313
657
109
271
889
229
552
1370
192
483
1520
372
1000
2730
Temperature
(°C)
129.9
109.8
90.7
149.9
131.0
109.8
150.9
131.0
109.3
151.1
131.0
109.0
150.8
131.0
109.9
Adsorbate Class: Aromatic Hydrocarb
Correlation
Slope Intercept Coefficient
2.93478 -5.22599 0.98509
3.69321 -6.70084 0.99983
3.05239 -4.83089 0.99836
3.45687 -5.86934 1.00000
3.42707 -5.50489 0.99790
-------
Table B-5
Specific Retention Volumes (ml/g)
00
Resin:
Adsorbate
1 ,2-Dichloroethane
Fluorobenzene
1,1,2-Trichloro-
ethylene
Chlorobenzene
Bromobenzene
1 , 4-Dichlorobenzene
XAD-2
V* (ml/g)
~8
38.4
54.2
96.9
50.9
75.0
132.0
51.2
86.6
141.3
60.6
108.6
167.8
107.7
188.3
315.4
164.9
301.3
538.4
Temperature
(°C)
161.3
146.6
132.4
161.3
146.6
132.4
161.9
145.0
130.9
192.8
174.5
161.3
192.8
174.5
161.3
192.8
174.5
161.3
Adsorbate Class: Halogenated
Hydrocarbons
Correlation
Slope Intercept Coefficient
2.45466 -4.08429 0.99072
2.52280 -4.11409 0.99523
2.49480 -4.02811 0.99994
2.84505 -4.32456 0.99994
2.98486 -4.38164 0.99872
3.28498 -4.84322 0.99803
-------
Table B-6
Specific Retention Volumes (ml/g)
Resin: Tenax-GC
Adsorbate
(ml/g)
u>
Temperature
1,2-Dichloroe thane
Fluorobenzene
1,1,2-Trichloro-
ethylene
Chlorobenzene
Bromobenzene
1 , 4-Dichlorobenzene
— 6
21.3
31.2
75.5
22.8
38.0
102.2
24.5
39.9
108.1
27.0
51.8
83.2
51.1
97.3
167.9
69.2
146.3
235.9
161.6
148.3
126.5
161.6
148.3
126.5
161.6
148.3
126.5
186.0
173.0
161.6
186.0
173.0
161.6
186.0
173.0
161.6
Adsorbate Class: Halogenated
Hydrocarbons
Sloe
Correlation
Intercept Coefficient
2.74521 -5.00108
3.23296 -6.08617
3.20783 -6.00053
3.99802 -7.26918
4.22532 -7.49334
5.36271 -7.64892
0.99791
0.99975
0.99938
0.99774
0.99967
0.99454
-------
•o
o
Table B-7
Specific Retention Volumes (ml/g)
Resin: XAD-2
Ads orb ate
(ml/g)
Temperature
2-Butanone
2-Heptanone
4-Heptanone
Cyclohexanone
3-Methyl-2-butanone
3,3-Dimethyl-2-
butanone
— &
25.3
39.5
69.4
201.7
480.4
1161.8
175.6
427.8
1034.6
185.4
399.2
822.1
41.4
78.9
146.3
67.3
133.4
278.9
174.8
153.0
133.0
175.4
153.0
133.2
175.4
153.2
133.4
175.6
153.0
133.4
175.8
153.6
133.4
176.2
153.2
133.4
Adsorbate Class: Ketones
Slope
1.90864
3.27760
3.33420
2.78674
2.35244
2.62583
Intercept
-2.86767
-5.00653
-5.19071
-3.94194
-3.62170
-4.02293
Correlation
Coefficient
0.99616
0.99990
1.00000
0.99995
0.99976
0.99883
continued....
-------
Table B-7 (continued)
Adsorbate
2,6-Dimethyl-4-
heptanone
Acetophenone
V* (ml/a)
— g
519.0
1415.5
3990.3
267.6
626.6
1501.6
3808.0
(°C)
175.8
153.0
133.4
197.0
174.6
153.0
133.4
Slot
3.80074
3.45276
Correlation
Intercept Coefficient
-5.75828
-4.92004
0.99953
0.99990
-------
Table -B-8
Specific Retention Volumes (ml/g)
Resin:
Adsorbate
2-Butanone
2-Heptanone
4-Heptanone
Cyclohexanone
3-Methyl-2-butanone
Ace tophenone
Tenax-GC
v£ (ml/g)
17.5
37.8
99.3
146.7
469.3
1902.6
126.5
406.8
1544.4
168.3
476.8
1585.9
26.9
66.6
183.6
88.5
225.8
646.5
Temperature
152.8
131.8
110.6
152.8
131.8
110.6
152.8"
131.8
110.0
153.6
131.8
110.2
152.8
131.8
.UO.O
196.6
174.3
153.6
Adsorbate Class: Ketones
Correlation
Slope Intercept Coefficient
2.92774 -5.64138 0.99917
4.31585 -7.97739 0.99969
4.14677 -7.63681 0.99997
3.66388 -6.36534 0.99988
3.17976 -6.03588 0.99987
4.01805 -6.61535 0.99906
-------
Table B-9
Specific Retention Volumes (ml/g)
Resin:
Adsorbate
n-Butylamine
n-Amylamine
n-Hexylamine
Benzylamine
Di-n- Bu ty lamlne
XAD-2
V| (ml/g)
41.8
76.0
126
276
92.4
175
298
895
183
362
799
355
597
876
447
617
1060
Te mperature
(°c)
169.7
150.7
131.3
109.7
169.7
150.7
131 .
110.1
169.7
151.0
131.0
169.7
160.8
150.3
169.7
161.2
150.3
Adsorbate Class : Amines
Correlation
Slope Intercept Coefficient
2.27928 -3.51956 0.99771
2.74263 -4.24533 0.98721
2.96738 -4.44033 1.00000
3.75608 -5.91593 0.97836
3.64200 -5.58430 0.99810
-------
Table B-10
Specific Retention Volumes (ml/g)
Resin: Tenax-GC
Adsorbate
n-Butylamine
n-Amylamine
n-Hexylamine
Benzylamine
Di-n-b utylamine
Tri-n-butylamine
1C
£ (ml/g)
~O
19.6
31.9
64.9
161
418
35.5
64.8
150
449
1360
61.2
135
313
1120
257
562
1350
100
250
647
33.3
64.3
157
643
Temperature
(°C)
170.0
150.9
130.6
110.4
91.3
170.0
150.9
130.8
110.4'
91.3
170.2
150.9
130.8
110.4
170.2
150.9
130.9
170.0
150.9
130.9
170.0
150.9
130.9
110.4
Slope
2.75520
3.28387
3.55078
3.27994
3.70514
3.65348
Adsorbate Class: Amines
Intercept
-4.97164
-5.90947
-6.24518
-4.98957
-6.35624
-6.77637
Correlation
Coefficient
0.99535
0.99637
0.99508
0.99998
0.99913
0.98674
-------
Ln
Table B-ll
Specific Retention Volumes (ml/g)
Resin: XAD-2
Adsorbate v£
O~
Ethanol
n-Propanol
n-Butanol
2-Butanol
2-Methyl-2-propanol
2-Methyl-l-propanol
(ml/g)
15.2
31.8
60.7
43.7
96.5
211
74.0
128
275
54.4
88.2
194
485
16.4
30.2
62.5
60.7
101
210
Temperature
(°C)
130.7
111.1
90.7
130.9
110.6
90.5
150.3-
130.6
110.7
149.7
130.9
110.6
90.5
130.8
111.1
90.7
150.1
130.9
110.7
Adsorbate Class: Aliphatic Alcol
Correlation
Slope Intercept Coefficient
2.18956 -4.22468 0.99408
2.48352 -4.50203 0.99903
2.34307 -3.67715 0.99593
2.49183 -4.19021 0.99449
2.12922 -4.06001 0.99990
2.22674 -3.48989 0.99632
-------
Resin: Tenax-GC
Table B-12
Specific Retention Volumes (ml/g)
Adsorbate Class: Aliphatic Alcohols
Adsorbate
Ethanol
n-Propanol
1-Butanol
2-Butanol
2-Methyl-2-
propanol
2-Methyl-l-
propanol
vj (ml/g)
— a
21.9
41.0
86.0
67.6
147
344
70.4
175
437
48.2
120
267
719
12.9
25.6
58.5
49.5
131
296
Temperature
(°C) Slope
110.1
91.4 2.01955
71.3
110.1
91.4 2.39945
71.3
130.2'
110.4 2.99335
91.3
129.9
^J'g 2.76500
71.5
111.5
91.4 2.15199
71.1
130.2
110.4 2.95273
91.6
Correlation
Intercept Coefficient
-3.93050 0.99999
-4.42767 0.99935
-5.57235 0.99968
-5.16541 0.99876
-4.49011 0.99964
-5.61303 0.99582
-------
Table B-13
Specific Retention Volumes (ml/g)
Adsorbate
Phenol
o-Cresol
p-Cresol
m-Cresol
Resin; XAD-2
Vg (ml/g)
137.9
221.2
337.4
228.5
374.7
593.7
261.6
439.2
684.9
262.1
436.1
688.3
Temperature
(°C) Slope
196.6
185.0 3.44622
173.0
196.6
185.0 3.71282
173.2
196.6
185.0 3.70588
173.0
196.6
185.0 3.71874
173.0
Adsorbate Class: Pheno
Correlation
Intercept Coefficient
-5.18973 0.99837
-5.53959 0.99918
-5.46240 0.99763
-5.49053 0.99842
-------
-p-
00
Table B-14
Specific Retention Volumes (ml/g)
Adsorbate
Phenol
o-Cresol
p-Cresol
m-Cresol
Resin: Tenax-GC
V* (ml/g)
S
68.4
131.9
269.1
103.6
206.1
463.3
118.8
231.7
545.1
116.7
235.6
525.0
Temperature
<°C)
185.2
170.3
153.8
185.2
170.3
153.8
185.2
170.3
153.2
185.2
170.3
153.8
Adsorbate Class: Pheno
Correlation
Slope Intercept Coefficient
3.70378 -6.24066 0.99956
4.05497 -6.83084 1.00000
4.12943 -6.93879 0.99977
\
4.06926 -6.80843 0.99993
-------
Table B-15
Specific Retention Volume (ml/g)
Resin; XAD-2
Adsorbate Class: Aliphatic Acids
Adsorbate
Acetic Acid
Propionic Acid
n-Butanoic Acid
n-Pentanoic Acid
Vg (ml/g)
50.0
107
238
117
262
746
148
263
681
289
441
640
Temperature
(°C) Slope
131.1
109.9 2.29043
88.0
131.4
109.9 2.70997
88.0
151.4
130.7 2.59865
109.8
151.4
142.2 2.83271
130.7
Correlation
Intercept Coefficient
-3.96313 0.99923
-4.64209 0.99867
-3.97513 0.98817
-4.20089 0.98691
-------
Table B-16
Specific Retention Volumes (ml/g)
Ui
o
Adsorb ate
Acetic Acid
Propionic Acid
n-Butanoic Acid
n-Pentanoic Acid
Resin: Tenax-GC
Vg (ml/g)
25.6
47.1
97.6
34.4
61.2
149
338
70.4
156
390
1050
127
363
1000
Temperature
(°C) Slope
131.3
110.7 2.27624
90.7
150.9
131.3 2.58210
110.7
90.7
150.9
130.6 3.01511
110.7
90.7
150.9
130.6 3.44349
108.9
Adsorbate Class : All-
Correlation
Intercept Coefficient
-4.25864 0.99523
-4.57032 0.99769
-5.26958 0.99988
-6.00272 0.99647
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-78-054
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Characterization of Sorbent Resins for Use in
5. REPORT DATE
March 1978
Environmental Sampling
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R.F.Gallant, J.W.King, P.L.Levins* and
J.F.Piecewicz
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc.
Acorn Park
Cambridge, Massachusetts 02140
10. PROGRAM ELEMENT NO.
EHB537
11. CONTRACT/GRANT NO.
68-02-2150, Task 10601
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 3/77-1/78
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES BERL-RTP task officer is Larry D.
541-2557.
Johnson, Mail Drop 62, 919/
is. ABSTRACT
repOrf- describes the use of chromatographic techniques to character-
ize resins which are used to trap vapors in environmental sampling schemes. It
describes two such techniques (frontal and elution analysis) which have been applied
to characterize sorbent cartridges packed with Tenax-GC and XAD-2 sorbents, two
synthetic polymeric resins commonly used as sampling media. Three diverse adsor-
bate groups, consisting of eight distinct chemical classes, were studied as potential
pollutants. Elution analysis of these vapors yielded specific retention volumes
which can be directly related to the breakthrough characteristics of the sorbent
resins under a diversity on sampling conditions. Adsorption coefficients, derivable
from the specific retention volumes , vield the weight capacity of the sorbent at
challenge concentrations in the Henry's Law region. Frontal analysis results con-
firm the elution data for sorbate uptake of resins. A slight flow rate dependence for
sorbate uptake is noted for XAD-2. Specific retention volume data extrapolated to
ambient conditions correlate well with adsorbate boiling point and molecular polariz-
ability. These correlations allow breakthough and weight capacity to be estimated
for a variety of adsorbate types.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Sampling
Analyzing
Polymers
Sorbents
Properties
Chromatography
Vapors
Elution
Air Pollution Control
Stationary Sources
Frontal Analysis
13B
14B
07D
11G
07A
9. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
163
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
151
------- |