EPA-670/4-74-009
                                       November  1974
THE DETERMINATION OF VOLATILE  ORGANIC COMPOUNDS AT THE

      yg/1  LEVEL IN WATER BY GAS  CHROMATOGRAPHY
                           By

      Thomas  A.  Bellar and James  J.  Lichtenberg

           Methods Development  and Quality

            Assurance Research  Laboratory




              Program Element No.  1BA027
        NATIONAL ENVIRONMENTAL  RESEARCH CENTER
          OFFICE OF RESEARCH AND  DEVELOPMENT
         U.S.  ENVIRONMENTAL PROTECTION AGENCY
                 CINCINNATI, OHIO   45268
          For vale by the Superintendent of Documents, U.S. Government
                 Printing Office, Washington, D.C. 20402

-------
                               REVIEW NOTICE




     The National Environmental Research Center--Cincinnati has reviewed




this report and approved its publication.  Mention of trade names or com-




mercial products does not constitute endorsement or recommendation for use.
                                      11

-------
                                  FOREWORD

     Man and his environment must be protected from the adverse effects of
pesticides, radiation, noise and other forms of pollution, and the unwise
management of solid waste.  Efforts to protect the environment require a
focus that recognizes the interplay between the components of our physical
environment--air, water, and land.  The National Environmental Research
Centers provide this multidisciplinary focus through programs engaged in

     •    studies on the effects of environmental contaminants on man and
          the biosphere, and

     •    a search for ways to prevent contamination and to recycle
          valuable resources.

     This report describes the development of a new method for quantitative
recovery and determination of volatile organic compounds in water and waste
water.  The method provides for the determination of compounds such as
organohalide solvents, e.g., chloroform, and others having low water solu-
bility at levels that heretofore could not be measured.  Using this method,
many organic compounds can be detected at 0.5 yg/1 and below.  The develop-
ment of this method has opened up a whole new area for research.  Using
the method, toxicologists will be able to measure and study the potential
chronic toxic or carcinogenic effects of low concentrations of selected
organic compounds in water.
                                              A. W. Breidenbach, Ph.D.
                                              Director
                                              National Environmental
                                              Research Center, Cincinnati
                                     111

-------
                                  ABSTRACT




     A quantitative analytical method for the concentration, isolation, and




determination of volatile hydrocarbon and chlorinated hydrocarbon solvents




in water is presented.  An inert gas is bubbled through the sample to




transfer volatile compounds from the aqueous phase to the gaseous phase.




These compounds are then concentrated on a porous polymer trap under non-




cryogenic conditions and determined by gas chromatography using a flame




ionization or microcoulometric detector.  Details of the design, fabrication,




and use of the apparatus are described.




     The method is applicable to organic compounds that are less than 2%




soluble in water and that boil below 150°C.  Application of the method to




the determination of a variety of aliphatic and aromatic hydrocarbons has




been demonstrated on several types of water including sewage treatment plant




effluents.  The lower limit of detection is 0.5 to 1 yg/1 for many compounds.




The method is useful over a concentration range of approximately 1 pg/1 to




2500 ug/1.
                                     IV

-------
                                CONCLUSIONS
     The method presented here provides for the gas chromatographic
determination of volatile organic compounds that are less than 2% soluble
in water and that boil below 150°C.   It has been found applicable to the
determination of substituted and unsubstituted aliphatic and aromatic
hydrocarbons.  Selective determination of organohalogen compounds is
possible using a microcoulometric detector.  The method is equally applicable
to finished waters,  natural surface  and ground waters, and sewage treatment
plant  and industrial effluents.  The useful concentration range for
quantitative analysis is approximately 1 yg/1 to 2500 yg/1.   The lower limit
of detection for many compounds is 0.5 to 1 yg/1.

-------
                                INTRODUCTION

     Recent legislation (1-2) requires analytical methods for the deter-
mination of hydrocarbon and chlorinated organic solvents in waste water.
In some cases a minimum detectable limit of 1 yg/1 is required for specific
compounds.  It is the responsibility of the Methods Development and Quality
Assurance Research Laboratory to evaluate existing methods and when neces-
sary to develop new methods to meet such needs.  Determination of these
substances at the 1 yg/1 level has been difficult.  Commonly used techniques
such as direct aqueous-injection gas chromatography, liquid/liquid extract-
ion, and head gas analysis have proven inadequate.  Direct aqueous-injection
gas chromatography (3-4) although generally useful for analysis of industrial
effluents, provides an approximate limit of detection of only 1000 yg/1.
Liquid/liquid extraction methods using low (5) or high (6) boiling organic
solvents followed by gas chromatographic analyses have been investigated.
These methods have provided erratic or low extraction efficiencies for
volatile compounds.  In addition, large solvent responses and solvent im-
purities can cause serious chromatographic interferences.  Distillation
techniques (7) have been employed in which a small quantity of sample dis-
tillate is collected and analyzed by direct aqueous injection gas chromato-
graphy.  Detection limits of approximately 1 yg/1 have been reported for
water soluble volatiles using this method.  Poor recoveries render the method
useless for water insoluble components.  Head gas techniques (8) have been
routinely employed for a number of years.  With this method the sample is
sealed in a partially filled container.  Each volatile organic compound es-
tablishes an equilibrium between the gaseous and aqueous phase.  At low con-
centrations the ratio of the concentration in the gaseous phase to the con-
centration in the aqueous phase is a constant  (partition coefficient) and is
unique for each organic compound.  By analyzing the gaseous phase and apply-
ing the appropriate partition coefficient the concentration can be calculated
for each organic originally present in the aqueous phase.

     Of the techniques mentioned above, the head gas method has the greatest
potential for meeting the needs set forth in the Federal Register.  For this
method to be effective, the following must be accomplished.

     1.   Transfer of 95% to 100% of the organics contained in the aqueous
          phase into the gaseous phase.

     2.   Quantitative injection of all of the organics contained in the
          gaseous phase into a gas chromatography.

If  these  can be done, 5.0 ml of aqueous sample would be sufficient to provide
a method  sensitive to 1.0 yg/1; this sample size is based upon the average
limit of  detection for direct aqueous injection techniques.

-------
     Since, in a static system, it is impossible to alter the partition co-
efficient to favor the gaseous phase 100%, it was decided to study a dynamic
system where a sweep (purge) gas is bubbled through the sample until the
volatile organics are quantitatively transferred to the gaseous phase.  The
organics that are quantitatively transferred to the gaseous phase could then
be concentrated for gas chromatographic analysis with the use of a non-
cryogenic trapping technique developed a number of years ago for ambient
air (9) and dilute emission analysis (10).  In this manner an analysis per-
formed upon the gas phase would have a direct relationship to the aqueous
phase concentration.

                                EXPERIMENTAL

Apparatus

     Several pieces of commercial equipment are modified and assembled to
meet the needs of the method.  A list of parts required and instructions
for assembling the purging device, trap, and desorbers are given in the
Appendix.

Purging Device - This specially designed glassware (Figure 1) allows finely
devided gas bubbles to pass through the sample.  Volatile, water insoluble
compounds boiling less than 150°C are transferred from the aqueous phase to
the gaseous phase in a quantitative manner using less than 300 ml of purge
gas.

Trap - The trap (Figure 2) is a short section of stainless steel tubing
packed with an adsorptive material such as gas chromatographic grade porous
polymers, silica gel, or molecular sieve.   Volatile materials are trans-
ported directly from the purging device into the trap by the purge gas.  The
adsorbant retards the flow of the purged compounds while the purge gas is
vented.  The properties of the adsorptive material are chosen to meet the
needs of the particular analysis.  The following criteria must be met:

          The volume of the purge gas passing through the adsorbant packed
     in the trap can approach but not reach the retention volume of the com-
     pound to be trapped.

          The retained compounds must not be irreversibly sorbed by the
     trap.  [Silica gel irreversibly adsorbs some aromatics above Cg (9).]

          No chemical reactions or rearrangements may take place as the
     sample is being concentrated, stored, or desorbed.  [Silica gel causes
     externally bonded olefins to rearrange to the cis- and trans-2 olefins
     (9)-]

          The adsorptive material must be thermally stable.  Chromosorb 103
     and Tenax GC have been found to perform satisfactorily.  [Divinyl
     benzene crossliked porous polymers out-gas extraneous compounds causing
     serious interferences during most gas chromatographic analyses (10).]

Desorbers - The desorbers (Figures 3 and 4) are used to transfer the contents

                                      3

-------
      OPTIONAL
      FOAM
      TRAP
EXIT 1/4
IN. O.D.

M4 mm O.D
                    INLET  1/4
                bT"~IN. O.D.
1/4 IN. O.D. EXIT
  10 mm GLASS FRIT
  MEDIUM POROSITY
           INLET
   6 mm O.D. RUBBER
   SEPTUM
       10 mm O.D.
       «- INLET
       1/4 IN. O.D.
     FIGURE 1. PURGING DEVICE

-------
  BODY ASSEMBLY
CONNECTING NUTS
  STEM ASSEMBLY
                          TRAP VENT
                     GLASS WOOL PLUG  -
                     50-60 MESH
                     POROUS POLYMER
                     BEADS
cm
18
17
16
15
14
13
12
11

 9
 8
 7
                     5% DEXSIL-300 ON  — 5
                     60-80 MESH       ~~ 4
                     CHROMOSORB-G   _Z
                     GLASS WOOL_PLJJG_ _i

                     	TRAP INLET
 3
 2
 1
 0
              FIGURE 2. TRAP

-------
INERT GAS SUPPLY
@ 60ft
     FLOW CONTROLLER
         TEFLON GASKET
                          1/8 IN. O.D.
                          TEFLON TUBING
TRAP BACKFLUSH
FLOW FITTING
      MODIFIED BODY
      ASSEMBLY

     *-CARRIER GAS
       INLET

     -LIQUID INLET OF
     GAS CHROMATOGRAPH

     GAS CHROMATOGRAPH
     COLUMN
          FIGURE 3. DESORBER #1

-------
THERMOCOUPLE WIRE
    HEATER WIRE
                               BODY ASSEMBLY
   0 TO 110V
                                     REMOVABLE GC NEEDLE
   DRILL 0.128 IN. HOLE IN HEX AREA
   OF FITTING. SILVER SOLDER
   REDUCER IN PLACE

ASBESTOS INSULATION
                           r	INERT GAS
                           \   SUPPLY @ 60tt

                              V
                 1/8 IN. O.D.
                 TEFLON TUBING
          TRAP BACKFLUSH
           FLOW FITTING
      FLOW CONTROLLER
                 FIGURE 4 DESORBER #2

-------
of the trap to the gas chromatograph for analysis.  This is done with the use
of an auxilliary carrier flow control system that backflushes the trap at
elevated temperatures directly onto the gas chromatographic column.  Desorber
#1 is used exclusively with a Perkin-Elmer 900 series gas chromatograph and
desorber #2 can be used as a universal desorber for all gas chromatographs
with a septum-type liquid inlet system.

Gas Chromatograph  - A Perkin-Elmer 900 Gas Chromatograph was equipped with
dual-flame ionization detectors and a microcoulometric detector (halide
mode).

     Column 1 consisted of dual, stainless-steel, 180 cm (6 ft) long x
2.67 mm (0.105 in) I.D. columns, packed with Chromosorb-101 (60/80 mesh).
The carrier gas was nitrogen at 50 ml/min.  The oven temperature was iso-
thermal 190°C or programmed from 120°C to 225°C at 10°/min.

     Column 2 consisted of dual, stainless-steel, 91 cm (3 ft) x 1.65 mm
(0.065 in) I.D. columns packed with 4% SE-30 on Chromosorb-P(NAW)  (60/80
mesh).  The carrier gas was nitrogen at 50 ml/min.  The oven temperature was
programmed from 60°C to 230°C at 10°/min.

     The gas chromatograph-mass spectrometer system consisted of a Varian
Aerograph Series 1400 with a Finnigan 1015C Quadrupole Mass Spectrometer
controlled by a Systems Industries 150 Data Acquisition System.  The column
was glass, 240 cm  (8 ft) long x 2 mm  (0.078 in) I.D. and packed with Chromo-
sorb-101  (50/60 mesh).  The carrier gas was helium at 30 ml/min.  The initial
oven temperature was held at 125°C for 3 min, and then programmed at 4°C/min
to 220°C.

Reagents

     Organic-free Water, water free of interfering organics, was prepared
by passing distilled water through a Millipore Super-Q water treatment
system.

     Standard stock solutions were prepared by injecting 1 to 5 yl of the
compound to be determined into a 1-liter volumetric flask partially filled
with organic-free water.  The mixture was then diluted to volume with
organic-free water to give concentrations between 1 and 7 mg/1.  Dilutions
were then made from the stock solution by pipetting a known quantity of
stock solution into a partially filled volumetric flask and diluting to
volume with organic-free water.  [For low level work (1 to 10 ug/1), a 1:10
dilution of the stock solution was prepared and secondary dilutions were
prepared from this solution as required.]

Procedure

Purging and Trapping - Samples were purged and trapped as follows:  With
nitrogen gas flowing through the purging device  (Figure 1) at 20 ml/min, the
trap inlet (Figure 2) was attached  (finger-tight) directly to the purging
device exit using a compression fitting.  The trap vent was inserted into the
exit end of the trap.  Five milliliters of sample was injected into the

                                      8

-------
purging device and purged for the specified time (11 min).  The trap was
then removed from the purging device and the vent plug was removed and re-
placed with a cap to seal the trap inlet.

     Trap Conditioning.  Newly packed traps were conditioned at approximately
200°C with a nitrogen flow of 20 ml/minute for 16 to 24 hr with one of the
desorbers and vented to the room.  Each day before use, traps were placed
into the desorber and conditioned at 130°C for approximately 10 min while
being backflushed with nitrogen at 20 ml/min.

Desorption and Analysis - Two desorbers were designed for desorption and
analysis.  One was designed exclusively for use with the Perkin-Elmer model
gas chromatograph and the other was designed for use with any gas chromato-
graph.

     Desorber #1 (Figure 3).   The gas chromatographic oven was cooled to
below 30°C with the oven door open.  The "plug" was removed from the de-
sorber and the cap was removed from the trap.  The trap was then inserted
into the desorber and locked into place.  The trap backflush flow fitting
was then locked into place on the trap exit flow fitting and backflushed with
nitrogen at 20 ml/min for 3 min between 125° and 130°C.  The trap backflush
flow fitting was removed (trap still locked in place), the oven lid closed,
and the oven rapidly heated to its normal or initial operating temperature.
Gas chromatographic analysis  was carried out under these conditions.

     After analysis, the trap was removed by:  1) inserting the trap vent
into the trap exit fitting (to vent inlet system),  2) removing the trap,
3) resealing GC inlet system with "the plug", 4) removing the trap vent, and
5) resealing the trap inlet with the cap.

     Desorber #2 (Figure 4).   The gas chromatographic oven was cooled to
below 30°C.  The needle was inserted into the liquid inlet system on the gas
chromatograph.  The trap was  then inserted into the desorber and locked into
place.   The trap backflush flow fitting was locked into the trap exit flow
fitting and backflushed with nitrogen at 20 ml/min for 3 min between 125°
and 130°C.  After desorption and sample transfer was completed, the needle
was removed from the liquid inlet system, the oven lid closed, and the oven
repidly heated to the normal  or initial operating temperature.  Gas chromato-
graphic analyses were performed under these conditions.  After sample trans-
fer the trap was removed and sealed for future use.

Investigation of Method Parameters

     Initial studies were carried out to determine the volume of purge
gas needed for quantitatively extracting selected volatile materials from
a water sample.  The purging device (Figure 1) was charged with 5.0 ml of
an aqueous solution containing methylene chloride,  chloroform, benzene,
and 2-butanone, each in excess of 10 mg/1.  As this solution was being
purged with nitrogen, 3 yl aliquots were periodically withdrawn for
analysis by direct aqueous injection.   Analyses were performed on the aqueous
mixture until the concentrations of the dosed materials were reduced to or
below the limit of detection, approximately 100 Mg/1.  This experiment was

-------
initially performed with a purge gas flow rate of 20 ml/min nitrogen.
The flow rate was reduced 65% to 13 ml/min and the experiment repeated.
The percent of the dosed compounds remaining in the aqueous phase with
respect to the purge volume are listed in Table I.
          TABLE I.  PURGING OF SELECTED COMPOUNDS FROM WATER
Nitrogen
Flow rate
ml/min
20
20
20
20
13
13
13
13




purge gas
, volume,
ml
0
20
100
300
0
6.5
85
143




Percent remaining in aqueous phase
Methylene
chloride
100
60
0
0
100
67
30
6

2

40.1
Chloroform
100
55
0
0
100
94
29
0
Solubility
1
Boiling
61.3
Benzene
100
46
3
0
100
71
6
0
in water,
0.08
point, °C
80.1
2-Butanone
100
95
96
80
100
100
86
74
%
35

79.6
     Those trap saturation volumes reported in Table II and designated by
footnote b_ were obtained by Bellar and Sigsby (9) for a dry air system.  To
determine what effect, if any, water which is inherent to the system reported
here, would have on the saturation volumes, we redetermined them using water
saturated nitrogen as the purge gas.  Little or no change was observed.  The
saturation volumes for several organochlorine compounds, not previously re-
ported, were also determined under this condition.

     The purging and trapping system was tested with selected industrial
solvents over a wide range of concentrations.  Ideally, the response for
each compound would be linear over the entire concentration range.  With the
use of the standard solutions and operating parameters described above, the
data listed in Tables III through VI were obtained.  The peak height of each
compound was measured and divided by the concentration to give the slope
between zero/zero and each data point collected.  Response curves for four
common organic solvents are shown in Figure 5.  The standard deviations from
the mean slope are also listed in the Tables.
                                     10

-------
                     TABLE II.  TRAP SATURATION VOLUMES,
                                     ml
Compound
Methane
Ethane
Propane
n- Butane
n-Pentane
n-Hexane
n-Alkanes
C7 t0 C15
Benzene
Toulene
Methylene
chloride
Chloroform
Aldehydes
C? § above
Phenols
Naphthylene
Chlorobenzene
o-Dichlorobenzene
1, 2, 4-Tri chloro-
benzene
Silica
gel
layer
<5a
<25a
>50a
>500a
>500a
>500a
d
>500a
>500a
d
d
Nonquanti-
tative
d
d
d
d
d
Pora Pak
Q
<5a
<5a
<50S
500b
>500b
>500b
>500b
>500C
>500°
>500a
>500a
>500a
d
d
d
Chromo-
sorb
<5a
<5a
500b
>500b
>500b
>500b
>500C
>500C
>500a
>500a
>500S
>500C
>500C
>500C
Tenax
GC
d
d
d
d
d
>500b
>500b
>500b
>500b
>500C
>500C
d
d
d
>500C
>500C
>500C
Retention
index
100
200
300
400
500
600
700 -
	
--

--
_» —
--
--
	
--
--






1500










 Values  reported by Bellar and Sigsby (9), 1970.

 Values  determined using water-saturated nitrogen as purge gas are same as
 those reported under dry conditions [Bellar and Sigsby (9), 1970].

"Determined for this  study.

 Not  determined.
                                     11

-------
TABLE III.  SYSTEM RESPONSE TO METHYLENE CHLORIDE
Concentration,
yg/i
5.2
10.4
20.8
52.0
104.0
260.0
520.0


TABLE IV.
Concentration,
Pg/1
6.2
12.4
24.8
62.0
124.0
310
620


Slope
0,0 to data point
78.5
76.9
78.5
77.5
76.3
72.9
85.7
78.0 Mean
3.88 Standard
SYSTEM RESPONSE TO CHLOROFORM
Slope
0,0 to data point
32.3
29.7
28.1
26.8
26.8
25.4
29.9
28.4 Mean
2 . 35 Standard
Dilution
1/100
2/100
5/100
10/100
20/100
50/100
Stock solution

deviation

Dilution
1/100
2/100
5/100
10/100
20/100
50/100
Stock solution

deviation
                           12

-------
TABLE V.  SYSTEM RESPONSE TO BENZENE
Concentration,
yg/i
3.5
7.0
14.0
35.0
70.0
175.
350.


Slope
0,0 to data point
220.6
219.4
214.9
215.8
207.5
196.0
232.0
215.2 Mean
11.4 Standard
Dilution
1/100
2/100
5/100
10/100
20/100
50/100
Stock solution

deviation
TABLE VI. SYSTEM RESPONSE TO TOLUENE
Concentration,
yg/i
3.5
7.0
14.0
35.0
70.0
175.0
350.0


Slope
0,0 to data point
120.
120.
116.
115
111.
105
124
116. Mean
6.27 Standard
Dilution
1/100
2/100
5/100
10/100
20/100
50/100
Stock solution

deviation
                  13

-------
  80
  70
  60
g 50
LU
X
  4°
LU
V)


8 30
CO
LU
DC
  20
   10
   0
                               BENZENE
                         METHYLENE
                         CHLORIDE
                             CHLOROFORM
                           I
        I
    0
100
500
600
                200    300    400
                CONCENTRATION >qg/l

FIGURE 5 RESPONSE CURVES FOR SELECTED ORGANIC
         COMPOUNDS    14

-------
     To determine the effect of variation in the physical properties of in-
dividual compounds on the efficiency of the system, a homologous series of
n-alkanes was tested.  The test mixture consisted of n-CL to n-C   in
organic-free water.  This mixture was analyzed according to the prescribed
procedure using a Tenax trap.  Tenax was used as the adsorbent because it
has a higher thermal stability than Chromosorb 103 and could be operated at
the temperatures required for desorbing the higher molecular weight alkanes.
To determine the purge volume required for quantitative transfer of hydro-
carbons over the wide boiling range, successive fractions were collected at
ambient temperature  (19.5°C) and analyzed by flame ionization (FID) gas
chromatography using an SE-30 column (see Table VII).  The test was repeated
at an elevated purging temperature  (65°C) (Table VIII).
            TABLE VII.  PURGING EFFICIENCY AT 19.5°C, % RECOVERY
           	Compound and boiling point	

Purge     n-C     n-C     n-C      n-C      n-C     n-C      n-Ci?    n~c-\l
  -           o       o       /        o        y       _L i.       j.»j       J. j
volume,
ml N-    (36°C)  (69°C)  (98°C)  (126°C)  (150°C)  (196°C)  (234°C)  (270°C)
0-60 100 100 98 90
60-120 2 6
120-240 3
240-360 1
360-480
480-600
600-720
720-840
840-960
960-1080
1080-1200
76 60
12 15
8 9
3 4
1 2
2
2
1
1
1
2
44
17
13
6
3
5
3
2
2
2
2
2
13
27
14
8
7
5
5
4
4
7
 Not 100% purged using 1560 ml N.
                                      15

-------
            TABLE VIII.   PURGING EFFICIENCY AT 65°C,  % RECOVERY
Compound and boiling point
Purge n-C
volume,
ml N2 (36°C)
0-60 100
60-120
120-240
240-360
360-480
480-600
600-720
720-840
n-C n~^7 n~^R n~^q n-C

(69°C) (98°C) (126°C) (150°C) (196°C)
100 100 100 97 76
3 10
6
4
2
1
1

r a

(234°C)
66
12
6
6
4
3
2
1
c a

(270°C)
27
24
15
11
7
6
6
4
aNot 100% purged.
                                     16

-------
     When the method was applied to a sample from a local sewage plant that
serves a diverse industrial area, the complicated FID gas chromatogram shown
in Figure 6 resulted.  The sample was analyzed again using the microcoulo-
metric detector which gave the chromatogram shown in Figure 7.  The com-
pounds identified in the chromatograms were confirmed by GC-MS.

                           RESULTS AND DISCUSSION

     The data in Table 1 show that it is possible to purge the water in-
soluble (<2% soluble) compounds from 5 ml of water using less than 150 ml
of nitrogen.  Decreasing the purge gas flow rate by 65% showed that a slight
increase in the volume of purge gas is needed for quantitative transfer.
Water-soluble materials whose partition coefficients do not favor the gaseous
phase are only qualitatively transferred regardless of the purge volume.

Trapping

     From the data reported by Bellar and Sigsby (9) and other data
exhibited in Table II, it can be seen that organics contained in small
volumes of water saturated nitrogen can be concentrated.  It is apparent
from these data that compounds with a retention index greater than 500 can
be quantitatively purged and trapped.  Retention indices given in the litera-
ture on porous polymers (11-13) make it possible to predict trap saturation
volumes for a wide variety of organic compounds.  Since most hydrocarbons
and substituted hydrocarbons commonly present in waste waters have retention
indices >500, porous polymers were used in developing this method.

     Water has a retention index of less than 300 and is not quantitatively
trapped by porous polymers.  Therefore, gas chromatographic columns and
detectors adversely affected by water can be used with a minimum of inter-
ference.

     The statistical data generated in Tables III through VI reflect an
accumulation of errors for the entire method.  Considering the number of
manipulations involved and that gas chromatographic errors are generally
+3%,  it appears that this is, indeed, a useful method.  Further study of
these data indicates that the majority of the errors are due to the volu-
metric dilution procedure.   The larger the pipet used to withdraw aliquots
of the stock solution, the larger the error.  A buret may be a more suitable
device for delivering volatile solutes.

     From the data in Table I and for the compounds studied, we estimated
that purging transferred at least 99% of the volatile, water-insoluble com-
pounds from the aqueous phase to the gaseous phase.  The data in Tables III
through VI and some unreported duplicate data show that the purging effi-
ciency is identical from 2,500 yg/1 to at least 6 yg/1.  Therefore, the com-
pounds studied can be quantitatively determined over that concentration
range.

     From the data listed in Table VII it is also apparent that the alkanes
up to Cg can be quantitatively purged using less than 500 ml of purge gas.

                                     17

-------
18

-------
               I /6/r e'9 S3N3ZN3eOHCTIHOia
1/6 "-g'f 3N3HAH13 OHCT!HOVH13i -'2'ZVl
1/6 ^8 6 3N3nAH13OaCTIHOIHl-3Vl

     l/B/rg'8 3NVH13ObOlHOIBl-l'L'L

 'si waodoacriHO
 I       ^

3SNOdS3H

    19
                                 CM
                                                          CM
                                                         ' CM
                                                          o
                                                          CM
                                                         . co
                                               LU

                                               _l
                                               LL
                                               LL
                                               LLJ
                                               I-
                                               z

                                               _l
                                               Q_
                                                                z
                                                         •CM  JO  LU
                                                                m
                                                         -co
                                                         -CD
                                                         - CM
                                               LU
                                               O
                                               LU
                                               C/)
                                                                LLJ
                                                                DC
                                                                13
                                                                g
                                                                u_

-------
Purge volumes exceeding 1.5 1 failed to transfer 100% of the Cjj through
C}5 alkanes.  Raising the temperature of the purging device and sample
(Table VIII) extended the useful range of the method to GH hydrocarbons.
If a water sample contains volatiles over the entire boiling range re-
presented by these data, it may be necessary to trap two fractions in-
order to perform a complete quantitative analyses on the sample.  This is
apparent from the data in Table II; they show that compounds with a re-
tention index of less than 600 will saturate the trap and be vented before
the high boiling materials are quantitatively purged.

Sample preservation

     Due to the volatility of the organic materials detected by this
method, common sample preservation techniques are inadequate (3,4).  The
simplicity of the trap and purging device makes it possible, however, to
collect, purge, and trap the sample at the sampling site.  The trap and con-
tents can then be sealed and shipped to the laboratory for analysis, and the
need for sample preservation is, thus, eliminated.

Application of the method

     From the experimental data reported in this paper, it is apparent
that this method has great potential for the analysis of trace volatile
organics contained in a wide variety of water sources.  For quantitative
determinations, the method is limited to organic compounds that are less
than 2% soluble in water and that boil below 200°C.  Significant qualitative
enhancement of compounds whose boiling points exceed 200°C can be expected
when the sample is heated.  The method is useful from 1 to 2,500 yg/1 with
the use of most gas chromatographs.  At concentrations exceeding 2,500 yg/1,
chromatograph column flooding and nonlinear detector responses generally
occur.  Since direct aqueous injection techniques are useful down to
1,000 yg/1 the two methods can be used together to perform analyses over a
wide range of concentrations.  For water soluble compounds, the distillation
technique should provide the supplemental methodology needed to analyze
most industrial effluents and natural waters.

     A wide variety of waste water samples have been analyzed using the
method described.  The chromatograms  (Figures 6 and 7) show the results of
one such analysis.  Qualitative identifications were made using desorber #2
and the Finnigan GC-MS system.  The quantitative analyses were obtained using
Desorber #1 with a microcoulometric detector.  Only one of the peaks in the
FID chromatogram have been identified.  At the sensitivity ranges shown,
only the chlorobenzenes are likely to appear on the FID chromatogram.  The
method worked well except for the following:  One sample collected from a
sewage treatment plant foamed excessively and caused liquid water to be
transported from the purging device into the trap.  Decreasing the sample
size from 5 to 3 ml or using the foam trap eliminated this problem.  Liquid
water entering the trap causes nonquantitative trapping and severe gas
chromatographic interferences.

     When only water-insoluble materials were detected in the sample, it
was found that the purged water could be withdrawn with a syringe and the

                                      20

-------
purging device could be recharged for successive analyses without cleaning.
When gross amounts of water soluble organics were present, a sufficient
quantity of these compounds was collected in the trap for qualitative
identification.  When this occurred it was necessary to dry the purging
device in an oven at 110°C before an interference-free successive analysis
could be performed.   Other researchers (14,15) have reported on similar
methods for the analysis of aqueous samples; their work has been primarily
qualitative.  This present work has shown that the method can be used for
the quantitative measurement of a wide variety of water-insoluble compounds
whose boiling points are less than 150°C.  By slightly modifying the method,
materials that boil at approximately 200°C can also be quantitatively
measured.  Qualitative sample concentration occurs for a wide variety of
other materials for which quantitative measurements could possibly be made
if recovery factors were experimentally determined.  Vinyl chloride is one
compound of considerable current interest that can be determined by this
method.  Analytical conditions for this specific application are under in-
vestigation.
                                     21

-------
                                 REFERENCES

1.   "National Pollutant Discharge Elimination System, Proposed Forms
     and Guidelines for Information from Owners and Operators of Point
     Sources," Federal Register, 38_ (75, Part II), 9783, Apr. 19, 1973.

2.   "Ocean Dumping Criteria," Federal Register, ^38 (94, Part II), 12872,
     May 16, 1973.

3.   Sugar, J.W., Conway,  R.A., "Gas-Liquid Chromatographic Techniques for
     Petrochemical Wastewater Analysis," J.W.P.C.F., 40, 1622 (1968).

4.   ASTM D 2908-70T, "Tentative Recommended Practice for Measuring Volatile
     Organic Matter in Water by Aqueous-Injection Gas Chromatography,"
     Annual Book of ASTM Standards, Part 23, Water; Atmospheric Analysis
     (1973).

5.   "Methods for Organic Pesticides in Water and Wastewater," U.S.
     Environmental Protection Agency,  National Environmental Research Center,
     Cincinnati, Ohio (1971).

6.   Dudenbostel, B.F.,  "Method for Obtaining GC/MS Data of Volatile Organics
     in Water Samples," Internal Report, U.S. Environmental Protection
     Agency, Region II,  Edison, New Jersey, May 14, 1973.

7.   "Procedure for Water Soluble Volatile Organic Solvent in Effluents and
     Streams," Organic Laboratory, Chemical Services Branch, Region IV, U.S.
     Environmental Protection Agency,  Athens, Georgia, Aug. 1973.

8.   "Chlorinated Organics and Hydrocarbons in Water by Vapor Phase
     Partitioning and Gas Chromatographic Analysis," Method No. QA-466, Dow
     Chemical, Louisiana Division, Plaquemine, Louisiana, Jan. 1972.

9.   Bellar, T.A., Sigsby, J.E., "Non-Cryogenic Trapping Techniques for Gas
     Chromatography," Internal Report, Reprints available from:  John E.
     Sigsby, Jr., Division of Chemistry and Physics, U.S. Environmental
     Protection Agency,  Research Triangle Park, N.C.  (1970).

10.  Bellar, T.A., Sigsby, J.E., "The Analysis of Light Aromatic Carbonyls,
     Phenols, and Methyl Naphthylenes in Automotive Emissions by Gas
     Chromatography," Internal Report, Reprints available from:  John E.
     Sigsby, Jr., Division of Chemistry and Physics, U.S. Environmental
     Protection Agency, Research Triangle Park, N.C.  (1970).
                                     22

-------
11.   Chromosorb Century Series Bulletin, Johns-Manvilie, Celite Division,
     Greenwood Plaza, Denver, Colorado, Nov. 1970.

12.   Tenax-GC Bulletin No. 24, Applied Science Laboratories, Inc., P.O.
     Box 440, State College, Pa.

13.   Hollis,  O.L., and Hayes, W.V., "Water Analysis by Gas Chromatography
     Using Porous Polymer Columns," J. Gas Chromatography, _4, 235  (1966).

14.   Zlatkis, A. and Liebich, H.M., "Profile of Volatile Metabolites in
     Human Urine," Clin.  Chem., l]_, 592 (1971).

15.   Novak, J., Zluticky, J., Kubelka, V., Mostecky, J., "Analysis of
     Organic Constituents Present in Drinking Water," J. Chromatog., 76,
     45 (1973).
                                     23

-------
                                  APPENDIX

                 LIST OF PARTS AND INSTRUCTIONS FOR ASSEMBLY

                   OF PURGING DEVICE, TRAP, AND DESORBERS

PURGING DEVICE

     The purging device (Figure 1) is constructed from glass tubing.  The
glass frit installed at the base of the sample reservoir allows finely
divided gas bubbles to pass through the aqueous sample while containing the
sample above the frit.  The sample reservoir is designed to provide maximum
bubble contact time and efficient turbulent mixing.  Gaseous volumes above
the sample reservoir are kept to a minimum to provide efficient transfer
characteristics and yet allow sufficient space for most foams to disperse.
Inlet and exit ports are constructed from 1/4-in. O.D. medium-or heavy-
wall tubing so that leak-free removable connections can be made using
"finger tight" compression fittings containing plastic ferrules.  The
optional foam trap is used to control occasional samples that foam ex-
cessively.

Parts for purging device:

     Borosilicate Glass tubing - 6 mm O.D. Standard Wall, 14 mm O.D.
          Standard Wall, and 1/4 in O.D. Medium Wall

     10 mm glass frit, medium porosity

     6 mm gas chromatographic half-hole septum or septum plugs

     5 ml syringe

     Syringe valve-Hamilton (1 ELI 2 way)

     17 cm x 20 gauge syringe needle

     Stainless steel reducing union  1/4 in to 1/8 in with Teflon or nylon
          ferrules to attach trap to purging device exit.

TRAP

     The trap is assembled and packed with the appropriate adsorptive
material according to Figure 2.  The body assembly acts as a seal for the
exit end of the trap.  The modified  stem assembly is used to attach the trap
to the desorption device.  The cap is used to seal the inlet end of the
trap when it is not in use  (finger-tight).

                                     24

-------
Parts for trap:

     Stainless steel tubing 3.175 mm  (0.125 in) O.D. x 2.67 mm  (0.105 in)
          I.D. x 21.6 cm long - (trap body)

     Body assembly Swagelok quick-connect  (QC4-B-200) (the trap vent)

     Stem assembly Swagelok quick-connect  (QC4-S-2QO) modified  (Drill
          through with a #30 (0.128 in) drill to allow trap body to pass
          through entire fitting)

     Swagelok cap (200-C with Teflon or nylon ferrules)

DESORBERS

     Desorber #1 (Figure 3) is attached directly onto the gas chromatograph
liquid inlet system after removing the septum nut, the septum, and the
internal injector parts.  The modified body assembly is screwed onto the
inlet system using the Teflon gasket as a seal.  A Swagelok plug (200-P)
is attached to one of the stem assemblies.  These assembled parts called
simply "the plug" are used to seal the desorber whenever the trap is re-
moved to maintain the flow of carrier gas through the gas chromatographic
column.  The flow controller, Teflon tubing and stem assembly are used to
provide the trap backflush flow.  This entire assembly is also used to
provide gas flow to operate the purging device.

Parts for Desorber #1:

     Brooks Model 8744 flow controller with #1 taper needle

     Teflon tubing 3.175 mm (0.125 in) O.D. x 1.65 mm (0.065 in) I.D.
          x 1.5  m (5 ft) long

     Stem assembly - 3 each Swagelok  (QC4-S-200)

     Body assembly Swagelok (QC4-B-2PF) modified with pipe threads drilled
          out using a 25/32 in drill.  The fitting is rethreaded using a
          7/16-20 bottoming tap.  (The check ball and spring located
          within the body assembly are removed and discarded.)

     Teflon gasket approximately 6.35 mm (1/4 in) thick x 19.8 mm
          (25/32 in) O.D. x 3.97 mm (5/32 in) I.D.

     Swagelok plug (200-P)

     Desorber #2 (Figure 4) may be attached to any gas chromatograph by
piercing the GC  liquid inlet septum with the needle.  The desorber is
assembled according to Figure 4 with internal volumes and dead volume areas
held to a minimum.   The heat source is concentrated near the base of the
desorber so that the internal seals of the body assembly do not become dam-
aged by heat.  The use of a detachable needle assembly from a micro syringe
                                     25

-------
makes it easy to replace plugged or dulled needles.  The flow controller,
Teflon tubing, and stem assembly are used to provide the trap backflush
flow.  This entire assembly is also used to provide gas flow to operate
the purging device.

Parts for Desorber #2:

     Brooks Model 8744 flow controller with #1 taper needle

     Teflon tubing 3.175 mm (0.125 in) O.D. x 1.65 mm  (0.065 in) I.D.
          x 1.5 m (5 ft) long

     Stem assembly Swagelok quick-connect  (QC-S-200)

     Body assembly Swagelok quick-connect  (316-QC4-B-400) modified
          (The check ball and spring are removed and discarded.)  A
          0.128 in hole is drilled through the hex area of the fitting
          and the Swagelok reducing adaptor is silver soldered in place.
          Be sure to remove plastic seals located inside the body assembly
          fitting before silver soldering.

     Stainless steel tubing 6.35 mm (1/4 in) O.D. x 4.763 (3/16 in)  I.D.
          x 13 cm long

     Cap - Swagelok (316-400-C)

     Reducer - Swagelok (316-100-R-2)

     Thermocouple, compatible with pyrometer on gas chromatograph

     Heater tape, useful up to 250°C

     Asbestos tape

     Heat resistent fiber glass tape

     Micro syringe needle - 26 gauge x 51 mm (2 in) and detachable
          syringe needle assembly from micro syringe  (Glenco)

     Variable transformer, 0 to 140 volts.
                                      26

-------
                               ACKNOWLEDGMENT




     The authors wish to thank Mr. James W. Eichelberger and Dr. Lawrence




E. Harris for providing the GC-MS analyses.
                                     27

-------
                                   TECHNICAL REPORT DATA
                            (Please read lastructions-on the reverse before completing)
1. REPORT NO.
 EPA-670/4-74-009
                             2.
              3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
 THE DETERMINATION OF VOLATILE ORGANIC  COMPOUNDS AT THE
 ug/1 LEVEL  IN WATER BY GAS CHROMATOGRAPHY
                                                           5. REPORT DATE
                                                             November 1974; Issuing Date
              6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

 Thomas A.  Bellar and James J. Lichtenberg
              8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 National  Environmental Research Center
 Office of Research and Development
 U.S. Environmental Protection Agency
 Cincinnati,  Ohio  45268
               10. PROGRAM ELEMENT NO.
                1BA027; ROAP 09ABZ;  Task 13
               11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
 Same as above
                                                           13. TYPE OF REPORT AND PERIOD COVERED
                                                           14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
 See also EPA-670/4-74-008, "The Occurrence of Organohalides  in Chlorinated Drinking
 Waters"
16. ABSTRACT
      A quantitative analytical method for the concentration,  isolation, and
 determination of volatile hydrocarbon and chlorinated hydrocarbon solvents in water
 is presented.  An inert gas is bubbled through the sample  to  transfer volatile  com-
 pounds from  the aqueous phase to  the  gaseous phase.  These compounds are then con-
 centrated  on a porous polymer trap  under noncryogenic conditions and determined by
 gas chromatography using a flame  ionization or microcoulometric detector.  Details
 of the design, fabrication, and use of the apparatus are described.  The method is
 applicable to organic compounds that  are less than 2% soluble in water and that
 boil below 150°C.  Application of the method to the determination of a variety  of
 aliphatic  and aromatic hydrocarbons has been demonstrated  on  several types of water
 including  sewage treatment plant  effluents.  The lower  limit  of detection is 0.5  to
 1 yg/1 for many compounds.  The method is useful over a concentration range of
 approximately 1 yg/1 to 2500 ug/1.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                            c.  COS AT I Field/Group
  *Hydrocarbons
  *0rganic solvents
  Gas  chromatography
  *0rganic compounds
  Water analysis
  Aliphatic hydrocarbons
  Aromatic hydrocarbons
   *n-Hydrocarbons
    Porous polymer
    Industrial effluents
    Natural waters
    Organohalogen compounds
    Analytical techniques
13B
18. DISTRIBUTION STATEMENT
                                              19. SECURITY CLASS (ThisReport)
                                                   UNCLASSIFIED
                            21. NO. OF PAGES

                                       32
           RELEASE TO PUBLIC
 20. SECURITY CLASS (Thispage)
      UNCLASSIFIED
                                                                         22. PRICE
EPA Form 2220-1 (9-73)
28
                                                     U.S. GOVERNMENT PRIHTING OFFICE: 197Jt-657-587/53l9 Region No. 5-1

-------