PB87-228516
Soil Gas Sensing  for Detection  and
Mapping of Volatile Organics
Nevada Univ.,  Las Vegas
Prepared  for

Environmental  Monitoring Systems  Lab
Las Vegas,  NV
Aug 87
                     U.S. DEPARTMENT OF COMMERCE
                   National Technical Information Service

-------
                                                       PB67-22B510
                                                EPA/600/8-87/036
                                                August 1987
               SOIL GAS SENSING FOR DETECTION AND

                  MAPPING OF VOLATILE ORGANICS
                               by
Dale A. Devltt, Roy B. Evans, William A. Jury and Thomas H. Starks
                  Environmental Research Center
                 University of Nevada, Las Vegas
                    Las Vegas, Nevada   89154

                  Bart Eklund and Alex Gholson
                       Radian Corporation
                          Austin, Texas
                      Cooperative Agreement
                          CR 812189-01
              J. Jeffrey van Ee, Technical Monitor
              Advanced Monitoring Systems Division
           Environmental Monitoring Systems Laboratory
               Office of Research and Development
                    Las Vegas, Nevada   89114
           Environmental Monitoring Systems Laboratory
               Office of Research and Development
              U.S. Environmental Protection Agency
                    Las Vegas, Nevada   89111

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing/
  REPORT NO.
   EPA/600/8-87/036
                             2.
              RECIPIENT'S-ACCESSION NO.
              PB87-22B5167AS
4. TITLE AND SUBTITLE
 SOIL GAS SENSING FOR DETECTION AND MAPPING OF VOLATILE
 ORGANICS
             5. REPORT DATE
                 August 1987
             6. PERFORMING ORGANIZATION CODE
'. AUTHOR(S)
Dale A. Devitt, Roy B. Evans,  William A. Jury, Thomas H.
Starks, Bart Eklund and Alex Gholson	
                                                           8. PERFORMING ORGANIZATION REPORT NO.
 I. PERFORMING ORGANIZATION NAME AND ADDRESS
 University of Nevada, Las Vegas, NV  89154
 Radian Corporation, Austin  ,  TX  78766
             10. PROGRAM ELEMENT NO.

               D1Q9	
             11. CONTRACT/GRANT NO.

               Co-op.  Agree. 812189-01
12. SPONSORING AGENCY NAME AND ADDRESS
 Environmental Monitoring Systems Laboratory-Las  Vegas
 Office of Research  and Development
 U.S. Environmental  Protection Agency
 Las Vegas, NV  89193-3478
             13. TYPE OF REPORT AND PERIOD COVERED
               Project Report/Summary
             14. SPONSORING AGENCY CODE

               EPA/600/07
15. SUPPLEMENTARY NOTES
 16. ABSTRACT
 The sensing of soil  gas is a relatively new approach that shows great promise in
 assisting scientists in the detection and mapping  of volatile organics  in  the subsur-
 face.  This document is an attempt at compiling all pertinent information  on the
 current state of  the art of soil gas sensing as it relates to the detection of subsur-
 face organic contaminants.  It is hoped that such  a document will better assist all
 those individuals who are faced with assessing  the extent of the contamination of our
 soil ground-water systems.  Soil gas monitoring has been shown to be a  cost effective
 means of delineating the size and movement of organic contaminants in the  subsurface.
 It has also been  shown to provide immediate information of the lateral  extent of soil
 and ground-water  contamination and to minimize  and more accurately predict the number
 and location of conventional monitoring wells that must be drilled.  As such, this
 document addresses five important areas related to soil gas monitoring; 1) site
 specific parameter considerations, 2) transport and retention of organics  in soil and
 ground-water, 3)  sampling methods, 4) analytical methods and 5) statistical treatment
 of soil organic vapor measurements.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.IDENTIFIERS/OPEN ENDED TERMS
                           C. COSATI I icId/Group
18. DISTRIBUTION STATEMENT
 RELEASE  TO  PUBLIC
19. SECURITY CLASS (ThisReport/
UNCLASSIFIED
21. NO. OF PAGES

     281
                                              2O. SECURITY CLASS (Thispage)
                                               UNCLASSIFIED
                                                                         22. PRICE
 EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION is OBSOLETE

-------
                           NOTICE
     The  information  in this document has  been funded  wholly or
in part by the United States Environmental Protection Agency
under  cooperative  agreement  number CR  812189-01  to  the
Environmental Research Center, University of Nevada,  Las Vegas.
It has been subject to the  Agency's peer and administrative
review, and it has been approved  for publication as  an  EPA
document.

     Mention of  trade names or  commercial products  does  not
constitute endoresement or recommendation for use.
                              ii

-------
                           ABSTRACT
     The  sensing of soil gas is a relatively new approach that
shows great promise in  assisting scientists  in the detection and
mapping of volatile organlcs in the subsurface.  This  document is
an attempt at compiling all pertinent information on  the current
state  of  the art  of  soil  gas  sensing as  it  relates  to the
detection of subsurface organic  contaminants.  It is  hoped that
such a document will  better assist all those individuals who are
faced with assessing  the  extent  of  contamination of  our  soil
ground-water systems.

     Soil  gas monitoring has been shown to  be  a  cost  effective
means of delineating  the size and  movement of  organic
contaminants in  the  subsurface.  It  has  also been  shown to
provide immediate information of  the  lateral  extent  of  soil and
ground-water contamination  and  to minimize  and more accurately
predict the number and  location of conventional monitoring wells
that  must  be drilled.   As  such,  this document addresses five
Important areas related to soil gas monitoring;  1) site  specific
parameter considerations, 2) transport and retention  of organics
in soil and ground-water, 3)  sampling methods, 4)   analytical
methods  and 5)  statistical  treatment of  soil organic vapor
measurements.

     The  document begins by  outlining  many of the  parameters
(water solubility, microbial  influence,  ground-water  flow, etc.)
that  must be considered  by  the  scientist  before utilizing soil
gas  sensors in  a  field  monitoring program.   Next,  the complex
soil, air,  water  and hydrocarbon system  is addressed with an
overview  of the important processes involved in the transport and
fate  of organic  contaminants  in  the  soil.   Additional sections
address the correct sampling  and analytical methodologies for
monitoring volatile organics  in  the  subsurface, covering such
sampling methods  as headspace,  ground probe, flux  chamber and
passive sampling  techniques.   Analytical  methodologies  covered
outline the most  appropriate  methods to utilize for a  given
contaminant monitoring  program.  A statistical treatment  of soil
organic vapor measurements is also included  to ensure that soil
gas  monitoring  programs  address the  requirement  for  data
precision.  The  statistical  section also gives  greater insight
into  understanding the  spatial  patterns of  soil organic vapor
measurements.
                               iii

-------
     Finally,  case studies are  included  to give the unfamiliar
reader examples of the  design,  procedures and  results of soil
gas monitoring programs  that have been successful  in  delineating
the size and lateral extent  of a  subsurface organic  contaminants.
                                iv

-------
                            CONTENTS


Notice	11
Abstract	ill
Figures	vi
Tables	xi
   1.   Introduction 	    1
            Soil gas sensing for detecting and mapping
            volatile organics 	    1
   2.   Site Specific Parameter Considerations 	  19
            Chemical and physical properties of the
            organic compound.	  19
            Properties of the unsaturated zone	59
            Hydrogeologic properties	I  79
            Characteristics of the spill	82
            Miscellaneous 	  82
   3.   Transport and Retention of Dissolved and
       Immiscible Organic Chemicals in Soil and
       Ground-Water	89
            Processes governing transport of organic
            chemicals through soil. .  .	92
            Movement of hydrocarbon vapor through
            soil  . . '	109
   U.   Measurement Methodologies  	 125
            Sampling methods  	 125
            Sampling design and sampling quality
            assurance techniques	157
   5.   Analytical Methodologies 	 168
            Selecting the proper methodology	168
   6.   Statistical Treatment of Soil Organic
       Vapor Measurements	199
            Components of variance analysis 	 200
            Interpolation and concentration
            contouring	•	208
   7.   Case Studies	212
            Hydrocarbon plume detection at
            Stovepipe Wells, California 	 213
            Study of ground-water contamination from
            industrial sources at Pittman, Nevada  	 236
   8.   Summary and Conclusions	256
            Utilization of soil-vapor measurements	256
References
   Chapter 1   	  17
   Chapter 2	85
   Chapter 3	116
   Chapter H	162
   Chapter 5	19^
   Chapter 6	211
   Chapter 7	255
Appendices
   Chapter 3	119
Subject Index  	 267

-------
                            FIGURES

Number                                                     Page
 1.1   Relationship between number of volatile compounds
       and organic priority pollutants in ground-water
       in the vicinity of hazardous waste disposal sites
       (M - 113 sites)	   2

 1.2   Typical behavior in porous soil following a sudden,
       high volume spill  	   5

 1.3   Behavior of product after spill has stabilized ...   6

 1.4   Organic gas concentration distribution in the
       vadose zone expected from diffusion	   9

 1.5   Chloroform and carbon tetrachloride depth
       distribution 	  10

 1.6   Soil-gas vertical profile at a site in northern
       California	13

 2.1   Plot of log solubility vs log vapor pressure
       illustrating the tendency for compounds in a
       homologous series to lie on a 45° diagonal of
       constant Henry's law constant	  46

 2.2   Relationship between the water solubility of a compound
       (mg/liter) and its partition coefficient between soil
       organic C and soil solution (Koc)	48

 2.3   Diagram showing an economical and safe way to contain
       chlorinated hydrocarbons (TCE) compounds 	  50

 2.4   Thickness of center of oil lens versus time where k
       values are permeabilities	56

 2.5   Teztural triangle, showing the percentages
       of clay, silt,  and sand in the basic soil
       textural classes 	  62

 2.6   Possible migration of product to outcrop
       followed by second cycle contamination 	  63
                               vi

-------
2.7   Benzene uptake by soil as a function of the
      relative vapor concentration 	  65

2.8   Cross-sectional view of soil pore area	66

2.9   Comparison of diffusion coefficient moisture
      curves for various median pore sizes 	  67

2.10  Relative permeability graph where swater is the
      percent saturation 	  69

2.11  Biodegradation rate based on oxygen recharge ....  74

2.12  Hypothetical ground-water system 	  81

2.13  Diagram showing how oil on a water table can
      be trapped in a cone of depression created by
      drawdown of a pumping well	84

3.1   Oil migration pattern (A,B)	91

3.2   Generalized shapes of spreading cones at
      immobile saturation (A, B, C). . .	93

3.3   Infiltration of kerosene into a porous medium. ...  95

3.4   Oil retention capacity as a function of time ....  96

3.5   Subsurface redistribution of a surface spill ....  98

3.6   Relation between thickness of oil layer and
      spreading time	100

3.7   .Solubility of hydrocarbons in water	106

3.8   Comparison of actual and idealized
      concentration depth profiles below a waste
      spill  in ground water	108

3.9   Steady state vapor concentration profiles between
      groundwater and the soil surface, for a compound
      undergoing first order degradation 	 115

4.1   Soil  core sample sleeve	131

4.2   Ground-probe design used by Russell and
      and Appleyard	133

4.3   Ground-probe design used by Neglia and
      Favretto   	134
                               vii

-------
4.4   Ground-probe design used by Tackkett 	 136

4.5   Ground-probe design used by Thornburn, et al . . . . 137

4.6   Ground-probe design of Lovell, et al	 139

4.7   Ground-probe design used by Swallow and Gschwend . . 140

4.8   Ground-probe design used by Swallow and Gschwend . . 141

4.9   Ground-probe design used by Walther, et al	142

4.10  Ground-probe design used by LaBrecque, et al . . . . 143

4.11  Ground-probe driver and extractor used by
      LaBrecque, et al	144

4.12  Sampling manifold and pump used by
      LaBrecque, et al .	145

4.13  Ground-probe design used by Radian Corporation . . . 147

4.14  Ground-probe design used by Crow, et. al	148

4.15  Surface-flux chamber and peripheral
      equipment	150

4.16  Curie-point wire accumulator device  	 153

4.17  Build-up and attenuation of volatiles  from
      gasoline through a sand column and through
      undisturbed wet clay soil	154

4.18  Hypothetical diffusion pattern and measured
      surface flux anomaly	155

6.1   Density curves for normal and skewed
      distribution 	 202

6.2   Meaningless contours 	 210

7.1   Location of study area	214

7.2   Comparison of the assumed location of  gasoline
      plume from USGS wells at the beginning of field
      work (May 1980) with the plume extent  shown by
      subsequent Lockeed wells 	 215

7.3   Diagram of Lockheed-EMSCO SOV probe tip and
      shaft	217

                               viii

-------
7.4   SOV probe driver and extractor	218

7.5   SOV sampling manifold	219

7.6   Map of total organic vapor in ppm as benzene
      from SOV sampling, August 1981, Stovepipe
      Wells, California	222

7.7   Map of ethane/propane from SOV sampling,
      August 1984, Stovepipe Wells, California 	 223

7.8   Map of butane from SOV sampling, August 1984,
      Stovepipe Wells, California	224

7.9   Map of pentane from SOV sampling, August 1984,
      Stovepipe Wells, California	  .  . 225

7.10  Map of benzene from SOV sampling, August 1984,
      Stovepipe Wells, California.  . 	 226

7.11  Map of isooctane from SOV sampling, August
      1984, Stovepipe Wells, California.  ; 	 227

7.12  Summary  of  drilling results,  August 1984,
      Stovepipe Wells, California	 229

7.13  Levels of volatile organlcs  in well SP5 as  a
      function of  depth, August 5,  1984,
      Stovepipe Wells, California	230

7.14  Cross-section of isooctane levels  (ppm) across  the
      contaminant  plume, August 1984, Stovepipe
      Wells, California	231

7.15  Cross-section of benzene  levels  (ppm) across  the
      contaminant  plume, August 1984, Stovepipe
      Wells, California	232

7.16  Cross-section of  pentane  levels  (ppm) across  the
      contaminant  plume, August 1984, Stovepipe
      Wells, California	233

7.17  Cross-section of  butane  levels  (ppm) across the
      contaminant  plume, August 1984, Stovepipe
      Wells,  California	234

7.18  Cross-section of  ethane/propane  levels  (ppm)  across
      the contaminant  plume,  August 1984, Stovepipe
      Wells,  California	235
                                ix

-------
7.19  General location map	237

7.20  Hydrogeologic cross-section with the locations
      of sampling boreholes along the contaminant
      plume	238

7.21  Ground-water quality based on total dissolved
      solids	239

7.22  Isocontour projection of benzene concentrations
      (ppm) in ground-water	240

7.23  Hydrogeologic cross section of the transect	212

7.24  Locations of monitoring wells along the Pittman
      Lateral	213

7.25  Ground-water concentrations of chloroform,
      benzene, and chlorobenzene	216

7.26  Probe locations in the area of the
      chloroform/carbon tetrachloride	247

7.27  Probe locations in the area of the benzene/
      chlorobenzene contaminant plume	247

7.28  Chloroform concentrations at 4-foot depth as a
      function of distance across plume	249

7.29  Soil-gas chloroform concentrations at 4-foot depth
      as a function of ground-water chloroform
      concentration	251

7.30  Chloroform and carbon tetrachloride depth
      distribution.  Coefficient of determination	254

8.1    Flowchart of soil-gas surveys	258

8.2    Flowchart of soil-gas measurements 	 260

-------
                            TABLES

Number                                                     Page
 1.1    Most frequently reported substances at 546
       NPL sites	    3

 1.2   Air/water concentration ratios for some common
       industrial solvents at 23°C. . 	    7

 1.3   Soil gas monitoring	   12

 1.4   Soil gas monitoring	12

 1.5   Analytical methods for soil gas sampling	12

 2.1   Site specific parameter considerations 	   20

 2.2   Physical and chemical properties of various
       organic compounds.	   21

 2.3   Chlorinated hydrocarbons 	   19

 2.4   Liquid group classification of various organic
       compounds	52

 2.5   Dielectric constants, densities and water
       solubilities of various halogenated and
       nonhalogenated solvents	57

 2.6   Intrinsic permeability of soils permeated
       by  water and organic  liquids	58

 2.7   Summary of organism growth  in various
       substrates	72

 2.8   Fate of organic compounds applied  to a sandy
       soil	73

 2.9   Average utilization of  substrates  in aerobic
       acetate-grown biofilm column  after acclimation  ...   76

 2.10  Average utilization of  substrates  in
       methanogentic acetate-grown biofilm column
       after acclimation	77
                                xi

-------
2.11  Blodegradation of the components of gasoline  ....  78

3.1   Residual oil void fraction, So ...........  94

3.2   Oil lens thickness above ground-water ........  99

3*3   Koc and T^/2 values for various miscible
      organic chemicals along with an estimate of
      the travel time required to migrate L - 1000 m
      through ground water using eq. 8 with Jw-1md~1,
      * - 0.5, pb - 1.5 gcm~3, roc - 0.005 ........ 104
3.4   Saturated vapor density, water solubility and
      Henry's constant KH for various volatile and
      semivolatile organic chemicals ........... 112

3.5   Time to diffuse L - 1  m through a soil with
      *-0.5, a-0.3, Dvair-4300 cm2d~ foe- .005
      using eg 6 and (A17) . . .  ............. 113

4.1   Criteria for selecting soil sampling equipment . . . 128

5.1   Description of selected portable analyzers ..... 171

5.2   Instrument response to selected organic
      compounds ...................... 173

5.3   Description of selected portable gas
      chromatographs .......  .. .......... 182

5.4   Summary of suggested calibration and quality
      control requirements for analytical systems ..... 1 87

6.1   Sample standard deviations  for raw and
      transformed chloroform measurements ordered
      by size of sample mean ............... 206

6.2   Chloroform concentrations (ppbv) measured on
      gases drawn after each of a series of purges
      of the same probe .................. 206

6.3   Confidence intervals for o2 based non s2 as a
      function of degrees of freedom and assuming a
      normal distribution for data  ............ 207

7.1   Background concentrations as  ppmv benzene ...... 221

7.2   Concentrations of chloroform  in ground water
      samples collected from wells  along the Pittman
      lateral (micrograms/liter )  ......... .... 244
                              xii

-------
7.3   Concentrations of benzene and chlorobenzene
      In ground-water samples collected from wells
      along the Pittman lateral (micrograms/llter) .... 215
7.i»   Observed chloroform concentrations over
      the chloroform contaminant plume 	 248

7.5   Chloroform and carbon tetrachloride concentrations
      in soil gas as functions of depth	252
                               xiii

-------
                         CHAPTER 1

                       INTRODUCTION




SOIL  GAS SENSING FOR  DETECTING AND MAPPING VOLATILE ORGANICS
     Interest in the measurement of  concentrations  of  volatile
organic  compounds  in the pore-space gases of  soil  was
stimulated by  enactment of Superfund (the  Comprehensive
Environmental Response,  Compensation,  and  Liability  Act, or
CERCLA) and by  the November 1984  reauthorization of RCRA  (the
Resource  Conservation and  Recovery Act of 1976)  which directed
the EPA to promulgate standards  for underground storage tanks
to include provisions for leak  detection.   The applications
discussed  in this report are principally appropriate for the
Superfund situation where  contamination  of  the  subsurface has
occurred and must be  assessed before taking remedial action:
removal  and  biological treatment of the contaminated  soil
and ground water or  both.  In this case, the usual objective in
measuring organic gases in soil is to  map the lateral extent of
soil and ground-water contamination or both while  at  the  same
time minimizing the  number of conventional  monitoring  wells
which  must be  drilled.  Soil gas concentration serves as a
surrogate for actual  measurements  of the  concentrations of the
compounds of interest  in ground water.   Maps of soil gas
concentrations  can  be used to site  ground-water  monitoring
wells more  efficiently.

     Volatile  compounds  are  components  in the soil  and
ground-water contamination at many, if not  most,  Superfund
sites.   Figure  1.1 shows  the relationship between the number of
volatile compounds and the  number of organic  priority
pollutants found  in  a survey of  ground-water  monitoring  data
from 113 Superfund  sites (Plumb,  1985).  A  regression  line
through  these  data shows a linear relationship with good
correlation between the total number of volatile compounds and
the total  number  of  organic priority pollutants detected per
site.   Table 1.1  lists the 25 compounds  most  frequently
reported at Superfund  sites;  15  of  these 25 are volatile
organic  solvents.   In addition  to  whatever toxicity  these
volatiles  may themselves  possess, they may serve as  tracers for
other,  non-volatile components provided that their gases arrive
near the  surface in measureable concentrations.

                               1

-------
88-
                                     •  All available data averaged

                                     A   One or more sites considered an outlier
                                     w   and excluded from average
                                         (8 of 121 eltes excluded)
        2   4    8    8   10  12  14   16  18  20  22   24
   TOTAL NUMBER OF VOLATILE COMPOUNDS DETECTED PER SITE
    Figure 1.1.  Relationship between number of volatile compounds
                 and organic priority pollutants  in ground-water  in
                 the vicinity of hazardous waste disposal sites  (N-
                 113 sites) (Plumb, 1985).

-------
TABLE 1.1. MOST FREQUENTLY REPORTED SUBSTANCES AT 5U6 NPL SITES
Rank
1
2
3
4
5
6
7
8
9
10
1 1
12
13
11
15
16
17
18
19
20
21
22
23
21
25
Substance
Triehloroethylene
Lead
Toluene
Benzene
Polychlorinated biphenyls (PCBs)
Chloroform
Tetrachloroethylene
Phenol
Arsenic
Cadmium
Chromium
1,1, 1 -Tr ichl or oe thane
Zinc and compounds
Ethylbenzene
Xylene
Methylen-e chloride
Trans-1 ,2-Dichloroethylene
Mercury
Copper and compounds
Cyanides (soluble salts)
Vinyl chloride
1 ,2-Dichloroethane
Chlorobenzene
1 , 1 -Dichloroethane
Carbon tetrachloride
Percent
of Sites
33
30
28
26
22
20
16
15
15
15
15
U
11
13
13
12
11
10
9
8'
8
8
8-
8
7
  From:  The Hazardous Waste Consultant, 1985.

-------
    Although  the processes governing the  movement of organics
in the subsurface are discussed in detail  in Chapter 2, a  brief
discussion of the process will illustrate the  logic behind soil
organic  gas measurement.   Figure  1.2 shows  a  subsurface
cross-section immediately following a sudden,  high-volume spill
of an organic  fluid.   The  spilled fluid has  moved vertically
downward through  the  unsaturated zone to  form  a  lens  on  the
water table  (which has a specific gravity less than water)  and
has left  behind  a  column  of  soil contaminated  by  residual
product.  The organic fluid  lens will immediately  begin to move
under the force of gravity and to spread put on the water  table.
Under the  conditions shown here,  the lens will  continue
spreading until  it  eventually disappears, for  all  practical
purposes.  However, most petroleum fuels  are a mixture of many
compounds with  a  wide spectrum of properties  such  as molecular
weight and water solubility.  Some fraction of the compounds,
generally quite  small, will dissolve in ground  water and will
move downgradient  with the  ground-water gradient  flow.   Figure
1.3 shows the  situation after  the spill  has stabilized days,
weeks, or months  after the occurance.   The column of  soil
contaminated  by residual product during  the original spill
remains.   The  lens of  organic fluid has spread  downgradient and
is much  thinner.   The dissolved fraction of  the spilled fluid
is migrating  with the  ground-water^ flow.  Gases   from  the
spreading organic fluid lens have begun to move  upward  through
the soil column above  the path of  the  spreading organic  plume.
When the  floating organic lens has  disappeared,  the volatile
component dissolved  in ground water will evolve from the  ground
water into the gases of the soil  pore  spaces.

     In  the situation where there exists a  floating lens of
free organic fluid,  the initial  concentration  of  organic gases
above the  lens can be estimated  from its  gas pressure.   In the
situation where there  exists a small  concentration of  dissolved
organic  volatiles in ground water,  the initial  concentrations
of volatile organics  in the pore space gases immediately above
the  water  table can  be estimated  from Henry's Law, as discussed
 at length  in Chapter 2.   Table  1.2  shows  air:water
 concentration ratios measured  for some common industrial
 solvents at room temperature (Thompson,  198U).  From Table 1.2,
 we would expect  a ground-water  concentration of 26yg/l of TCE
 to yield a pore-space gas  concentration of   10yg/l  in the
 unsaturated zone  immediately  above the  water  table.   These
 values  must be regarded as  rough approximations,  but they
 illustrate the point that  there  are well-understood physical
 principles  relating  the concentrations of dissolved volatile
 organics in ground water  to the  pore space  gas concentrations
 of these same volatiles.

     The migration  of the  organic gases upward through the
 vadose  zone is a  complicated process, as discussed  in  Chapter

-------
                                                               GROUND  SURFACE
                                 SPILL SITE
                                                          I;;.: PRODUCT  MIGRATING
                                                          '•'•DOWNWARD  AND
                                                           |:i ACCUMULATING ON
                                                           :•: WATER TABLE
                                                     xv*T*a3i*^—•• — ••• ••• ••••••• i MT,
                                                     ?VVVV-I-rc~AVl L L A R Y "f O N E I V
WATER  TABLE -_--•_--
Figure 1.2.  Typical  behavior  in  porous  soil  following a  sudden, high volume
             •pill (NYDEC.  1984).

-------
0.
                                              o o O O O O O O O O O O O
         Figure 1.3.   Behavior of product  after  spill  has  stabilized  (NYDEC,  1984).

-------
 TABLE  1.2.   AIR/WATER CONCENTRATION RATIOS FOR SOME COMMON
                     INDUSTRIAL SOLVENTS AT 23°C
        Compound                         Air: Water Ratio
                                         of air:yg/1 of water)
1,1  dichloroethylene (DCE)                     1:1

1,2  transdichloroethylene                      1:3

methylenechloride                              1:12

1,1,1 trichloroethane (TCA)                    1:15

trichloroethylene (TCE)                        1:26

carbontetrachloride                            1:1

tetrachloroethylene (PCE)                      1:17

chloroform                                     1:9

F-113                                          *»:1


(Marrin and Thompson, 1984)

-------
2,  but as a first-order approximation,  it can be regarded  as  a
diffusion process,  represented by

                     3C/at - D0[32C/3y2]

where

    C  - soil gas  concentration
    t  - time
    D0  - diffusion coefficient
    y  - vertical  distance above contaminant (or water table)

    Confining  discussion  to  "conservative"  gases  (where no
chemical or biochemical  processes add to or subtract from the
pore space gases) and  to a vertical cross-section in which the
subsurTace  is homogeneous  with  uniform  porosity  and
permeability in the unsaturated zone, it  is possible to outline
the qualitative nature  of  the  expected  vertical profiles of
organic gas concentrations  in soil gas.   In the relatively
simple situation where the evolution of the  organic gases  from
the  dissolved components  in ground  water has  reached  a
quasi-steady state,  this equation  indicates that  soil gas
concentrations linearly decrease from the initial concentration
immediately above the water table predicted  by  Henry's Law (or
Table 1.2) to zero  concentration  at the soil surface.  Figure
1.1 shows this  situation schematically.   Figure  1.5  is  a
vertical profile of carbon tetrachloride  and chloroform above  a
ground-water  contaminant plume In Nevada (Kerfoot and Barrows,
1986).   The  profile exhibits a  linear decrease of concentration
as the surface is approached.  These experimental data, showing
soil  gas concentrations which  Increase linearly with  depth,
support the predictions of  the steady  state model.  The
subsurface  vertical cross-section  is rarely homogeneous,
however; more often, a succession of strata is encountered in
which there  are varying permeabilities, porosities, soil  types,
and moisture  contents.  These varying properties influence the
upward  movement  of organic  gases and varying vertical gas
concentration gradients.

     Chemical and biochemical  processes  should also affect the
vertical gas  concentration profiles.  Figure 1.6 contrasts the
profiles of  PCE and benzene  reported from a  site in northern
California  (Marrin and Thompson,  1984).   The  same rapid
disappearance of  benzene vapor with distance above the water
table was reported  in  Nevada  associated  with the  vertical
profile of carbon  tetrachloride and chloroform of Figure 1.5
(Kerfoot and  Barrows, 1986).   Apparently, different  processes
affect the  movement  of  halocarbon  gases  such as
tetrachloroethylene, dichloroethane and chloroform than  affect
non-ha 1ogenated  hydrocarbons such  as  benzene.   Several
explanations  for this differing  behavior have  been suggested,

                                8

-------
                                                 GROUND SURFACE
a.
ui
o
                      Soil Concentration
                                                           C  max
     IWATER TABLE
      Figure 1.4.  Organic  gas concentration distribution in  the

                  vadose zone expected from diffusion.

-------
         COMPOUND  CONCENTRATION

                      (ppbv)
          0       100      200      300
Ill
0
u.
cc
CO
0?
-I 0
m~
X
H
0.
UJ
O

u
1 -

2 -
3 -

4 -

5 -
6 -

N
t O v
v \
^ O
\N%C ARSON TETRACHLORIDE
\\
• VN>
\ X\
CHLOROFORM • *>
\\
1 '< \
               W«t«r  at  12-1/2  f««t
Figure 1.5.  Chloroform  and carbon tetrachloride  depth
          distribution.  Coefficient of determination (r^ -
          .99)  Kerfoot  and Barrows, 1986).
                        10

-------
including biological  degradation, adsorption by organic matter
and clays, and water solubility.   Whatever the reason for the
differing behavior of hydrocarbons and  halocarbons,  the
halocarbons  generally appear  to  behave  "conservatively" and
are, therefore, better tracers for soil-gas investigations than
the hydrocarbons (Marrin and Thompson,  1984).

     The basic  approach  in  a soil-gas  investigation  at a
particular site is  now apparent.   The vertical profiles  of
organic gases present in  the soil pore spaces are  measured and
plotted for several locations at the site.  Selection of tracer
gases  for  the site  is  aided  when prior  information  on
contaminant concentrations  in ground water  is available*  Based
on the vertical profiles, the particular organic soil  gases
present,  and the  sampling and  analytical  methodologies
available, one or more tracer  gases  are  selected.   A  sampling
depth  is  also selected,  based  on the  measured  vertical
profiles, which is  expected to produce soil gas concentrations
well above the minimum concentrations detectable with  the
analytical techniques at  hand.  This  is shown schematically on
Figure  1.4.   When  this constant sampling  depth is used,  soil
gas  samples are  collected  and measured  across the  site
preferably on a regular grid  pattern.  These  values are  then
plotted on a map  and  are  contoured either  by hand or with a
computer algorithm.  The desired  result  is  a  contour  plot  of
soil gas concentrations at a  constant depth  across the site;
the investigator hopes  that this plot  is  related in a  more  or
less linear  way to  contaminant concentrations in  ground water
or in the buried waste  stratum of interest.

     Field measurements of soil gas have usually  taken one of
three basic approaches:  (1) measurement of emission fluxes  at
the surface; (2) measurement  of  pore space gas concentrations
at some depth in gas samples collected  with  subsurface probes
or with  shallow,  temporary  wells;  or  (3)  use  of  passive
collection devices buried at relatively shallow depths (Reid,
1985;  Marrin and  Thompson,  1984;  Bisque, 1984; Manos,  1985;
Kerfoot and  Barrows,   1986; Spittler, 1985).  These basic
methods  are listed in Table  1.3.  The  first two approaches
involve collection of gas samples  in the  field for subsequent
analysis;  the third   involves  adsorption  of the gas onto a
collection medium such  as activated charcoal:  the adsorbed gas
is later purged from  the  activated charcoal and Is  analyzed in
the laboratory.  With surface collectors and  flux chambers and
also with driven probes and temporary  wells, gas samples can be
collected with any  of the techniques listed in  Table  1.4.
Figure 8.2   is a  flow chart  which describes the process  of
collecting and analyzing a  soil gas  sample with driven probes
and  temporary wells.   Collection techniques  include  drawing
samples with a gas  chromatograph  syringe directly from the
sampling train,collection  of gases by  adsorption onto activated

                               11

-------
          TABLE 1.3.   SOIL GAS MONITORING
                  Sampling Methods
            o    Driven probes
            o    Shallow "wells"
            o    Petrex tubes
            o    Collectors on surface
            o    Flux chambers
          TABLE 1.4.  SOIL GAS MONITORING
                Collection Methods
          o    GC syringe
          o    Charcoal cartridge
          o    Stainless steel cannisters
          o    Tedlar or teflon bags
TABLE 1.5.  ANALYTICAL METHODS FOR SOIL GAS SAMPLING
         Method                          Comment
 Draeger tubes
 Organic vapor analyzer
 Field GC
 Lab GC
very crude
over 0.5 ppmv
generally over 10 ppbv
can be less than 1 ppbv
                           12

-------
                         \    X
          Observed BenxeneX     \

               concentrations
                                        N Linear Function
      possible behavior  "* —
                        Observed PCE concentrations
16
    r
    0
          60            100


Soil Concentration..
180
200
                                                                         .030
                                                                 260
       Figure  1.6.  Soil-gas vertical  profile  at a site in  northern
                    California (Marrin and Thompson, 1984).
                                         13

-------
charcoal or Tenax from an airstream  pumped  from  the probe or
well,  collection in stainless steel  cannisters,  and collection
in Tedlar or Teflon bags.

     Table  1.5  lists  analytical methods which have  been used in
soil  gas  investigations.   These include:   (1)  field
measurements with either a portable  organic gas analyzer (OVA)
or with  a transportable  gas  chromatograph,  (2)  laboratory  gas
chromatographic analysis of gas samples (via evacuated flasks
or cannisters or teflon or  teddar  bags)  or  charcoal  or  Tenax
cartridges  collected  in field,   or  (3) laboratory  gas
chromatograph/mass spectrometry  (GC/MS) analysis of  field
samples.  These methods  are discussed in detail in Chapter 4.
Another method which has been  used in field measurements
utilizes colorimetric  devices  usually called  Draeger  tubes.
Table 1.5  also  lists  the  usual  minimum  detectable
concentrations associated with  these various methods  in
practice.   Draeger tubes are the  least accurate and,  usually,
the least sensitive  of the methods listed in Table 1.5.  OVA's
are also relatively insensitive  and should  not  be expected to
produce usable  information  where the total non-methane
hydrocarbon concentrations  are   less  than  0.5  ppmv.
Transportable gas  chromatographs  have been used  in  field
applications where individual gas species  concentrations were
greater  than  10 ppbv (Barker, 1980).

      Laboratory-based gas chromatography can reliably  and
economically measure individual gas  species concentrations down
to  1  ppbv;  absolute detection limits are much smaller but such
low concentrations are generally of  little interest to  soil  gas
surveys.  The sensitivity and discriminative abilities of
laboratory-based GC/MS measurements can  extend  down  to
concentrations  considerably  less than  1  ppbv; however,  the
substantial expense is rarely  Justifiable in  soil  gas
investigations.

     The range  of gas measurement sensitivities listed in Table
1.5 means  that soil  gas  measurements can  detect very  small
concentrations of tracer  gases.  This sensitivity together with
the specificity of  the  gas  analyses  gives soil  gas
investigations some  important  advantages  over other indirect
tools for ground-water investigations such  as resistivity  and
other electrical geophysical  techniques.   The electrical
techniques  depend on the presence of substantial quantities of
dissolved  ionic  solids in  ground  water  to  create detectable
differences  in aquifer conductivity between the  contaminant
plume and the surrounding unaffected  areas.   Detection  and
mapping contaminant plumes with electrical  methods requires a
significant  conductivity  contrast between contaminant plume and
background  as they are  observed  from the  surface.  A  rule of
thumb,  sometimes  quoted, suggests  that a conductivity contrast

                               14

-------
of 1.5:1 or better is usually  needed with a "shallow" aquifer
(water table  <  30 feet)  to  satisfactorily  map a contaminant
plume  with electrical  methods.  This  would require total
dissolved solids concentrations in the plume to be at least 1.5
times the background TDS  values.  For deeply burled aquifers.
the necessary contrast  is higher because  of the  screening
presence  of the  overburden.   Not all Superfund sites have
conductive contaminant  plumes.  Even in cases where such plumes
exist, the contaminants  of  interest are usually organlcs which
add little or nothing  to  ground-water conductivity.  Because of
these limitations, electrical  techniques can be expected to
"underdefinew contaminant  plumes;  plume outlines identified by
electrical methods can  be expected to be less extensive than
the actual area of interest.   Plume outlines defined with the
most  sensitive  soil  gas measurement techniques may tend to
"overdefine"  ground-water  contaminant  plumes because  the
organic gases rising from the  ground-water plume will tend to
spread laterally as they  rise.

     Soil-gas  measurements  have  also been suggested as a means
of detecting  leaks from  underground storage  tanks;  in  this
application the focus  is  on  detecting leaking tanks before they
become serious environmental  problems.  The  EPA has estimated
that the total  number of underground storage tanks in the U.S.
is in the neighborhood of 3,000,000 to 5,000,000 and that as
many as 100,000  are  leaking today while another 350,000 may
leak within 5 years (Jeyapalan,  et al., 1986;  Hazardous Waste
Report, 1984; Predpall.  et  al.,  1981).  Because measurement of
volatile organic vapors in soil  pore gas is  probably the most
sensitive known  technique for  detecting material leaked from
underground storage  tanks,  the 1984  RCRA  amendments  have
provoked considerable interest in such methods.  The technical
objectives of  leak detection are distinct from those of site
assessment..   The goal in site  assessment is  to use soil-gas
measurements as a surrogate for soil and ground-water sampling
to map the lateral extent of contamination.  The goal in leak
detection is to provide an alarm when tank leak rates exceed a
given  value  (current suggested maximum leak rates are 0.05
gallons/hr).   To do this  with soil-gas monitoring requires a
known relationship between  leak rates or volumes and soil-gas
concentrations.  As will  be demonstrated, this  relationship is
highly  site-specific.   Other  problems associated with leak
detection by  soil  gas monitoring must be solved to make  the
technique workable.  Among these problems are the following:

     o  New leakage  must be detected in environments where
        soil  already  contains  residual vapors from previous
        spills and leaks.
                                15

-------
     o   No  "action levels" exist to aid in deciding what soil
        gas concentrations  should trigger remedial  action.

     o   The relationships between ground-water  concentrations
        of volatile organics and the resulting  soil  pore gas
        concentrations of  those organics are very  complex, and
        predictive  models may require more  data  about  a
        particular site than are likely to be available.

     o   Currently available measurement methods with adequate
        sensitivity may  be  too expensive  for use as  leak
        detectors.

While the problem of detecting leaks from underground storage
tanks requires a different technology than  does Superfund site
assessment, the  following  discussions of  gas vapor migration
and current measurement technology should serve as a background
for the  development of leak  detection methods.
                                16

-------
                       REFERENCES
1.   Barker,  Nicholas J. and Richard C. Leveson.   "A Portable
    Photoionlzation GC for Direct Air Analysis," American
    Laboratory, December 1980.

2.   Bisque, Ramon E.  "Migration Rates  of  Volatiles  from
    Buried  Hydrocarbon  Sources Through Soil Media,
    "Proceedings  of the  NWWA/API Conference  on Petroleum
    Hydrocarbons and Organic Chemicals in  Ground Hater  -
    Prevention, Detection, and Restoration,  National Water
    Well Association,  Houston,  Texas, November 5-7,  1984.

3.   Hazardous  Waste Report--Trends  &  Analysis.  The  Next
    Regulatory Battle:   Leaking Underground Storage  Tanks.
    May  1984.

4.   Kerfoot, H. B. and  L.  J.  Barrows.  "Soil-Gas  Measurement
    for  Detection of Subsurface  Organic  Contamination,"
    report, Contract  68-03-3215,  Environmental  Monitoring
    Systems  Laboratory,  Las Vegas, Las  Vegas, Nevada,  February
    1986.

5.   Kerfoot, H. B. and  Cynthia  Mayer.  "The Use  of Industrial
    Hygiene Samplers for Soil-Gas  Surveying," Ground-Water
    Monitoring Review, Fall 1986, vol.  6, no. 4.

6.   La Brecque,  D.  J.,  S. L. Pierett and  A.  T. Walker.
    Lockheed Engineering and Management Services,  Inc.,  and  J.
    W. Hess, Desert Research  Institute (EPA Report in  Draft),
    198U.

7.   Manos,  Charles G.,  Kenneth  R.  Williams, W.  David  Balfour
    and  Shelly J.  Williamson.  "Effects  of  Clay Mineral-
    Organic Matter Complexes  on Gaseous Hydrocarbon Emissions
    from Soils, "Proceedings of the  Second NWWA/API  Conference
    on Petroleum Hydrocarbons and  Organic Chemicals in Ground
    Water -  Prevention,  Detection,  and Restoration,  National
    Water  Well Association, Houston, Texas, November 13-15,
    1985.

8.   Marrin, D.  L. and G.  M.  Thompson.   "Investigation  of
    Volatile  Contaminants in  Unsaturated  Zone Above TCE
    Polluted  Ground-water,"  EPA Project  Report,  Project  CR
                              17

-------
    811018-01-0,  Robert  S.  Kerr Environmental  Research
    Laboratory, Ada, Oklahoma, 1984.

 9.  New York State Department of  Environmental Conservation,
    Division  of  Water,  Bureau  of Water  Resources.
    "Recommended Practices  for Underground Storage of
    Petroleum," Albany, New  York, May 1984.

10.  Plumb, R.  H. Jr.   "Disposal Site  Monitoring Data:
    Observations and  Strategy Implications,"  Second  Annual
    Can ad1 an - A m e r 1 can  Conference on  Hydrogeology -
    Groundwater: A Soluble  Dilemma;  National Water Well
    Association,  Banff, Alberta, June 25-29, 1985.

11.  Spittler, T. M. and W. Scott Clifford.  "A New Method  for
    Detection  of  Organic Vapors in the Vadose Zone," National
    Water  Well Conference  on Characterization and Monitoring
    of Organic Vapors in the Vadose Zone,  Denver, Colorado,
    November 1985.
                               18

-------
                         CHAPTER 2


           SITE SPECIFIC PARAMETER CONSIDERATIONS
    Because of  the  heterogeneous nature of soil and parent
material found in the vadose zone, the movement  of  organic
compounds  in both  the  liquid and  vapor phase are  often
difficult to predict with any degree of certainty  (varying in
both space  and time)..  However, if the variability of site
specific parameters is  properly  recognized,  correct
interpretations can be  made.  This  chapter  covers the
significance of the parameters listed in Table 2.1.   Special
attention  will be given to the parameter influence on soil
gas monitoring and on  contaminant plume detection  in the
saturated zone.

CHEMICAL AND PHYSICAL  PROPERTIES OF THE ORGANIC COMPOUND

   (1) Vapor pressure:  The pressure  of  the vapor of  a
      liquid confined such that the vapor  collects above
      it is referred  to as  the  vapor pressure.  Thus, at
      spill sites,organic  compounds with  high vapor
      pressures would be  expected to  be  present to  some
      degree  in the  vapor  phase of soil  pores.   Highly
      volatile  fuels such as  gasoline  are  known to
      evaporate relatively  fast even in  the  subsoil,
      forming an  envelope of hydrocarbon  vapors  around
      the core  of the spill (Schwille,  1975).   If an
      organic compound  has an exceedingly low vapor
      pressure (e.g., pesticides), it would not be  a good
      candidate for  soil vapor monitoring.  A list of  the
      vapor pressures of  many organic compounds  can be
      found in Table 2.2,  compiled  by Mackay and Shlu
      (1981).

  (2)  Water solubility:   The extent  to which an organic
      compound (solute)  dissolves  in  a solvent (water),
      is referred to  as  the water  solubility  of the
      compound.  Organic compounds  with high  water
      solubility  would be expected to partition primarily
      into  the  liquid water  phase.   The  rate  at which
      these compounds would  move through  the unsaturated
      zone  would  therefore be  controlled  to  a great
      extent  by the  unsaturated  hydraulic  conductivity of


                             19

-------
       TABLE 2.1.   SITE SPECIFIC PARAMETER CONSIDERATIONS
    Chemical and Physical Properties of the Organic Compound

     1)   Vapor pressure
     2)   Water solubility
     3)   Henry's law constant
     4)   Concentration
     5)   Organic carbon distribution coefficient (K  )
     7)   Density                            oc
     8)   Viscosity
     8)   Dielectric constant
     9)   Boiling point
    10)   Molecular weight
B.  Properties of the Unsaturated Zone.

     1)  Air filled porosity
     2)  Volumetric water content
     3)  Soil organic matter
     H)  Soil texture
     5)  Vapor pressure of water in the soil pores
     6)  Shape and size of pores
     7)  Depth of unsaturated zone
     8)  Retention
     9)  Temperature and temperature gradients
    10)  Microbial influence
    Hydrogeologic Properties

     1)  Ground water flow  (direction, velocity, gradient)
     2)  Water table oscillations
     3)  Lithology of the aquifer
D.  Characteristics of the Spill
E.  Miscellaneous

     1)  Rainfall
     2)  Background water quality
     3)  Barometric pressure and wind
     H)  Proximity to rivers, lakes and pumping wells
                                20

-------

Compound
Data at 25°C
Methane

Ethane

Propane

n-Butane
I ao butane
2,2-Olmthyl-
propane
Data at 25°C
n-Pentane




laopentane


n-Hexane






TABLE 2.2.
MW «p,°C
for gaaeoua alkanea
16.04 -162.5

30.7 -163.3

44.11 -189.7

58.13 -138.4
58.13 -159.6
72.15 - 16.6

for liquid alkanea
72.15 -129.7




72.15 -159.9


86.17 - 95






PHYSICAL AND CHEMICAL PROPERTIES OF VARIOUS^ ORGANIC
bp,°C Vapor Solubility
preaaure S,g/e>3
p. kPa*

-164 27260 24.1

- 68.6 3990 60.4

- 42.1 941 62.4

- 0.5 243 61.4
- 11.7 357 48.9
9.5 172 33.2


36.1 68.4 38.5
39.5
40.0
40.4
47.6
27.9 92.6 47.8
48.0
49.6
68.95 20.2 9.5
9.47
9.52
12.3
12.4
16.2
18.3
COMPOUNDS
Henry* a
conatant kPa
calc exptl

67.4

50.6

71.6

95.9
120
373


128
125
123
122.2
103.7
140
139
134.7
190
191
190
147
144
110
98.9

law
recoM

67.4+
2.0~
50.6+
i.r
71.6+
2.4~
95.9+
120+
373+
11.2

125+10




138+5


170+25






(continued)

-------
TABLE 2.2 (continued)
Compound MH «p,°C bp,°C Vapor
pressure
p. kPa*
2-Methyl- 86.17 -153.7 60.3 28.2
pentane

3-Methyl- 86.17 63.3 25.3
pentane
2,2-Di»ethyl- 86.17 - 99.9 49.7 42.6
butane

2,3-Olroethyl- 86.17 -128.5 58.0 31.3
butane
N> n-Heptane 100.21 - 90.6 98.4 6.11




2-Methyl- 100.21 -118.3 93. 0 8.78
hexane
3-Wethyl- 100.21 -119 92 8.21
hexane
2,2-Oimethyl- 100.21 -123.8 79.2 14
pentane
2,3-Olmethyl- 100.21 89.8 9.18
pentane

Solubility
13.8
13.0
15.7
12.8
13.1
18.4
21.2
23.8
19.1
22.5
2.93
2.24
2.66
2.19
3.37
2.54

2.64
4.95
4.40

5.25


Henry's
constant kPa
calc exptl
175.
186
154
172
171
199
173
154
141
120
209
273
230
280
182
346

312
166
318

175


ISM
•'/•el*
recom
170+15


172+8

173+16


130+10

230+50




346+21

240+75

316+8

175+7

(continued)

-------
                                                         TABLE 2.2 (continued)
N>
CO
Compound MW «p,°C bp,°C Vspor
pressure
o. kP. *
2,4-Oimethyl- 100.21 -119.2 60.$ 13.1
pentane
3,3-Dimethyl- 100.21 -134.5 86.06 11.0
pentane
n-Octane 114.23 - 56.23 125.7 1.68




3, Methyl- 114.23 -120.5 115 2.6
heptane
2,2,4-Trl- 114.23 -107.4 99.2 6.56
me thy 1-
pentane
2,3,4-Trl- 114.23 -109.2 113.5 3.60
TO thy 1-
pnntene
2,2,5-Tri- 128.26 -105.8 124.1 2.21
mathyl-
hexane
n-Nonane 128.26 - 51 150.8 0.571


Solubility
W-i3
4.06
4.41
5.50
5.94

0.66
0.431
0.493
0.85
0.8B
0.792

2.44
2.05

1.36
2.30

1.15
0.54

0.122
0.098
0.22
Henry's
conatsnt kPs
cslc exptl
323
298
239
186

325
499
438
253
244
376

308
365

302
179

219
467

601
748
333
ISM
•'/HOI*
recoM
300+25
186+10

300+50




376+15

330+30


190+15


350+120


500+200


                                                                                                                        (continued)

-------
TABLE 2.2 (continued)
Compound MM «p,°C bp,°C
4-Wethyl- 128.26 -113.2 142.4
octane
n-Oecane 14B.28 - 29.7 174.1

Undecane 156.32 - 25.59 195.9
Dodecane 170.33 - 9.6 216.3


Tetradacane 190.38 5.86 253.7


Hexadecane 226.44 18.17 287


Data at 25°C for aolld alkanea
Octadecane 254.4 28.18 316.1

Elcosane 282.6 36.8 343
Vapor
preaaura
p. kPa *
0.903
0.175

0.0522
0.0157


0.00127

0.00124
0.0000898

0.000917
7.44xlO~6a
1.30xlO-5"
2.59xlO~5"
6.92X10-66

2.18xlO-7"
2.67xlO~6a
1.58xlO-7b
Solubility
S,g/«3
0.115
0.052
0.024
0.044
0.0034
0.0037
0.00844
0.00606
0.0022

0.00628
0.0009

0.0021
0.00608
0.0019
Henry 'a
conatant kPa
calc axptl
1010
500
10BO
185
723
7B6
317
34.7
110

3.24
22.6

0.84
1.463
2.92

0.025
0.288
law
•3/«ol"
recoil
1000+100
700+300

1855+760
750+250











                                                              (continued)

-------
to
Cn
Compound MW «P»°C bP»°C
Hexacoaana 366.7 56.4 412.2
"Extrapolated value fron liquid atate.
Calculated fro» the extrepolatad vapour praaaure with
Data at 25°C for cycloalkanea
Cyclopentane 70.14 - 93.88 49.26

Cyclohexane 84.16 6.55 80.7


Methyl. 84.16 .142.14 71.8
cyclopentane
Methyl- 98.19 -126.6 100.9
cyclohexane
l-cie-2-Oi- 112.2 - 50.1 129.7
methylcyclo-
hexane
1,4,-trena- 112.2 - 37 119.4
Dimethyl-
eye lohexane
1,1,3-Tri- 112.3 - 14.2 104.9
methylcyclo-
pentene
Vapor
preaeure
p. kP.*
7.32x10
3.55x10
a fugacity ratio

42.4

12.7


18.3

6.18

1.93


3.02


5.3


Solubility
s,g/«3
-12" 0.0017
-12b
correction.

156
160
55
57.5
66.5
42
41.8
14
16
6.0


3.84


3.73


Henry 'a
conatant kPa
calc exptl
7.7xlO-7


19.1
18.6
19.4
18.6
16.1
36.7
36.8
42.8
38.0
36.1


88.2


159


law
recoai



18.5+1.1

18.0+2.0


36.7+1.4

40+3.0

38+5.0


88.2+4.0


159.+8


                                                                                                                              (continued)

-------
                                                          TABLE 2.2 fcontinued)
NJ
Ccmpoind
P ropy Icy clo-
pentane
Pentylcyclo-
pentane
Data at 25°C
Ethene
Propane
1 -But ana
2 -Methyl -
propone
3 -Methyl -1-
butene
MM
112.2
140.26
for gaseous
28.5
42.08
56.12
56.12
70.14
«p,°C tp,°C
-117.3 103.0
- 63
alkenes
-169.2 -103.7
-185.3 - 47.4
-185.4 - 6.3
-140.4 - 6.9
-168.5 20.0
Vapor
pressure
p. kPa *
1.64
0.152
6070
1140
297
304
120
Solubility
s,g/«3
2.04
0.115
131
200
222
263
130
Henry's law
conatant kPa •'/•ola
calc exptl reco«
90.2 90.2+4.4
185 185+18
21.7 21.7+2.0
21.3 21.3+3.0
75 75+4.0
64.8 64.8+6
54.7 54.7+6
Calculated using atmospheric pressure.
Data at 25°C
1 -Pent one
2-Pentene
1- Hex one
2-Methyl-l-
pentene
for liquid alkenea
70.14 -138 30.0
70.14
84.16
84.16
-151.4 36.9
-139.8 63.4
-135.7 60.7
85
66
24.8
26.0
148
203
50
78
40.3 40.3+2.0
22.8 22.8+1.0
41.8 41.8+1.0
28.1 28.1+1.2
(continued)

-------
                                                     TABLE 2.2  (continued)
Compound
4-Methyl-l-
pentene
2-Heptene
(trana)
1 -Octane
Data
Butadiene
2-Hethyl-l,
3 -butadiene
1,4-Penta-
dlene
"Calculated
Data
Propyne
l-8utyne
1-Pentyne
MW
84.16
98.19
112.2
at 25°C for dlenea
54.09
68.13
68.13
•p,°C bp,°C Vapor
preaaure
P. kP« *
-153.6 53.9 36.1
-136.6 95.7 6.45
-101.7 121.3 2.32
-108.9 - 4.4 281
-146 34 73.3
-148.3 26 98
Solubility
s,g/«3
48
15
2.7
735
642
558
Henry's
constant kPa
calc exptl
63.2
42.3
96.4
7.46«
7.78
12.0
law
M3/*Ol*
recoil
63.2+3.5
42.3+4.0
96.4+7.1
7.46+0.2
7.78+0.12
12+0.6
using atmospheric preaaure.
at 25°C for alkynea
40
50.09
68.13
- 101 - 23.2 558
-125.7 8.1 188
- 90 40.18 57.6
3640
2870
1570
1.11"
1.91"
2.5
1.11+.04
1.91+0.07
2.5+. 05
Calculated uaing atmospheric pressure.
                                                                                                                       (continued)

-------
                                                          TABLE  2.2 (continued)
Co
00
Compound MM np,°C bp,°C Vapor
pressure
p. kPs *
Data at 25°C For •onosromstlca
Benzene 78.11 5.53 00.1 12.7







Toluene 92.13 -95 110.6 3.80






Ethyl- 106.2 -95 136.2 1.27
benzene




p-Xylene 106.2 13.2 138 1.17




Solubility
S,g/«3

1780
1755
1769
1790
1779.5
1740
1869
1770
515
517
544
534.3
500
519.5
627
152
177
131
208
161
175
185
198
157
156
200
Henry 'a
conetant kPa
cslc exptl

0.557 0.562
0.565
0.561
0.554
0.557
0.570
0.533
0.560
0.68 0.673
0.677
0.632
0.655
0.70
0.674
0.558
0.887 0.854
0.762
1.03
0.648
0.837
0.771
0.671
0.628
0.791
0.797
0.621
law
•3/«ol«
recoil

0.550+.025







0.670+.035






0.8Q+.07





0.710+.08




                                                                                                                            (continued)

-------
                                                            TABLE  2.2 (continued)
to
Conpouid MM «P,°C bp,°C Vapor
preaaure
D. kPa*
m-Xylene 106.2 - 47.9 139 1.10




o-Xylene 106.2 - 25.2 144.4 0.682




1,2,3-Trl- 120.2 - 25.4 176.1 0.202
me thy 1-
benzena
1,2,4-Tri- 120.2 - 43.8 169.4 0.271
methyl-
benzene
1,3,5-Trl- 120.2 - 44.7 164.7 0.328
me thy 1-
banzane
Propyl- 120.2 -101.6 159.2 0.449
benzene
leopropyl- 120.2 - 96.6 154.2 0.611
benzene

Solubility
S,g/m3
162
196
173
146
134
175
170.5
167
204
213
75.2


57
51.9
59
97.0
48.2

55
120
50
48.3
65.3
Henry's
constant kPa
calc exptl
0.721
0.596
0.675
0.80
0.872
0.535
0.549.
0.561
0.459
0.440
0.323


0.571
0.627
0.552
0.407
0.818

0.981
0.450
0.147
0.152
0.112
IBM
«3/«ol"
recm
0.70040.10




0.50*0.06




0.32340.02


0.590+.04


0.60+. 20


0.700*. 30

0.130+.025


                                                                                                                              (continued)

-------
TABLE 2.2 (continued)
Compound MM mp,°C bp,°C Vapor
pressure
p. kPa*
l-Ethyl-2- 120.2 - 80.8 165.2 0.330
methyl-
benzene
l-Ethyl-4- 120.2 - 62.4 162 0.393
methyl-
benzene
n-Buytyl- 134.2 - 88 183 0.137
benzene


•
w laobutyl- 134.2 - 51.4 172.8 0.248
o benzene
s-Butyl- 134.2 - 75.5 173 0.241
benzene
t-Butyl- 134.2 - 57.8 169 0.286
benzene
1,2,4,5- 134.2 - 79.2 196.8 0.0659
Tetra-
me thy 1 benzene
1-iao- 134.2 - 67.9 177.1 0.204
P ropy 1-4-
methyl benzene
Solubility
93.05


94.85


12.6
11. 8
15.4
17.7
50.0
10.1
17.6
30.9
34.0
29.5
3.48


34.15


Henry ' a
constant kPa
calc exptl
0.427


0.498


1.46
1.56
1.04
1.04
0.368
3.30
1.84
1.05
1.13
1.30
2.54


0.80


law
•'/•ol*
recoil
0. 427+. 025


0. 498+. 03


1.30+.25




3. 30+. 13
1.40+.40

1.20+0.10

2.54+0.20


0.80+.10


                                                                (continued)

-------
TABLE 2.2 (continued)
Conpound MM mp,°C bp,°C
n-Pentyl- 148.23 - 73.0 203.4
Data For poly nuclear aromatlca at 25°C
Naphthalene 126.19 80.2 218












1 -Methyl- 142.2 - 22 244.6
naphthalene




2 -Methyl- 142.2 34.6 241.1
naphthalene

1-Ethyl- 156.2 - 13.8 258.7
naphthalene
Vapor
preaaure
p. kPa *
0.0437

1.09xlO-2








1.04xlO-2
1. 16xlO-2
3.11xlO-2a
l.OBxlO-2*
8.84x10-'



7.50xlO"3a
7.17xlO-)a
9.03xlO~3a
7.24xlO~3b
9.07xlO-3a
2.51xlO-3a

Solubility
S,g/m5
10.3

34.4
31.2
31.7
33.5
31.3
30.8
1.69
30.0
22.0




28.5
30.0
25.8
29.9


25.4
24.6

10.7
10.0
Henrx'a
conatant kPa
calc axptl
0.62

0.0407 0.0489
0.0448
0.0441
0.0417
0.0446
0.0454
0.0441
0.0466
0.0635




0.0041 0.0263
0.0419
0.0487
0.0420


0.0405
0.0419



law
V/nwl"
recoil
0.60+.06

0.0430*. 004












0.0450+.004










                                                                  (continued)

-------
                                                          TARIF  7 ? {continued)
to
ro
Compound MM
2-Ethyl- 156.2
naphthalene
Biphenyl 154.21









Acenaphthene 154.21



Fluorene 166.2



Phananthrene 178.23





«P,°C bp,°C Vapor
pressure
p. kPa *
liquid 4.21xlO-3
3.24xlO~3a
71 255.9 1.30xlO-3






5.80x10-*
3.92xlO-3a
7.55xlO-3a
96.2 277.5 3.07xlO~3a
5.96x10-*°

4.02xlO-3a
116 295 8.86x10"*°

1.13xlO-2a
1.66xlO~3a
101 339 2.67X10-50





Solubility
S.9/.3
8.0

7.48
7.0
7.45
7.50
7.08
5.94
3.87



3.88
3.93
3.47

1.90
1.98


1.18
1.07
1.29
1.60
1.15
1.002
Henry's
constant kPa
calc exptl
0.0822

0.0268 0.0413
0.0286 0.0304
0.0269
0.0267
0.0283
0.0337
0.0518



0.0237 0.0148
0.0234 0.0157
0.0265

0.00775 0.0101
0.00744


0.00403 0.00398
0.00445 0.00365
0.00367
0.00297
0.00414
0.00475
IBM
•3/«ola
recoM


0.028+.002









0.024+.002



O.OOB5+.002



0.0040+.OOOB





                                                                                                                             (continued)

-------

tJ
Compound MM m»,°C

Anthracene 178.23 216.2




Pyrene 202.3 156




Fluorathene 202.3 111


1,2-Benzan- 228.3 160
thracene
bp,°C Vapor
preaaure
p. kP«*
2.27X10-50
1.59x10-*"
4.64x10-*"
340 1.44X10-60



8.32xlO-7c
3.17xlO-5"
1.44x10-*"
360 8.86xlO~7c




375 1.79x10-*"
2.54X10-*6


6.67xlO-13
(20-)
Solubility

0.075
0.073
0.041
0.046
0.030
0.148
0.135
0.132
0.175
0.171
0.260
0.265
0.206
0.236
0.014
0.01
Henry's law
constant kPa »V«ol"
calc exotl recom

0.0034 0.073 0.0060*. 003
0.0034 0.676
0.0063
0.0056
0.0085
0.00121 0.0011 0.0012+.002
0.00133
0.00136
0.00102
0.00105
0.198 0.22*. 03
0.194
0.249
0.219


                                                                                                                           (continued)

-------
CO
TABLE 1 1 ( continued)
Compound MW mp,°C bp,°C
3,4-Benzo- 252.3 175
py rene

"Extrapolated values from liquid atate.
Calculated from the extrapolated vapor preaaure with a
cCxtrapolated From aolld vapor preaaure.
Data for halogenated alkenea and alkenea at 25 °C
Chloromethane 50.5 -97.7 -24.2






Dichloro- 84.9 -95.1 39.7
methane






Trichloro- 119.4 -63.5 61.7
methane


Vapor
preaaure
o. kPa *
6.67x10



Solubility
-13 0.0012
0.0038
0.0040

Henry 'a lew
conatant kPa wP/mol*
calc exptl recom
1.4xlO-7
4.«xlO-8
4.21x10-°

fugacity ratio correction.

unleaa otherwise
570
480(20°)



499(20°)

58.40

46.53
(20°)
'48.31
(20°)
21.08
(1.5°)
25.60

32.80
(20°)

atated.
5350
7400(20°)

6270(20°)

7250(20°)

19400
13200
13200(20°)



22700(1.5°)

7900
7950
8000(20°)



0.951" 0.95+.05
0.691" 0.731
(20°) (20°)
0.817"
(20°)
0.706"
(20°)
0.256 0.272 0.26+.02

0.299 0.301
(20°)


0.079
(1.5°)
0.387 0.322 0.38*. 03
0.383
0.496
(20°)
                                                                                                                        (continued)

-------
U)
Compoind MVf mp»°C




Carbon 153.8 -22.9
tetrachloride




Chloroethana 64.9 -136.4



1,1-Oichloro- 98.97 -96.98
ethane

1 ,2-Oichloro- 98.97 -35.36
ethane



1,1,1-Tri- 133.4 -30.4
chloroethane



bp,°C Vapor
presaure
D. kP«*
20.06
(20°)
8.9
(1.5°)
76.5 15.06

12.13
(20°)
12.0
(20°)
12.27 100.7
(20°)


57.5 30.10
24.42
(20°)
83.47 10.93
8.52
8.40
8.93

74.1 16.53
13.20
(20°)
13.33
(20°)
Solubility
s,g/»3
8200(20

10300(1.5°)

1160
800
8DO(20<>)

785(20°)

5710(20")

4700(20°)

5100
5500(20°)

8700
8800
8000
8000(20°)

720
730(20°)

950(37°)

Henry's
conatant kPa
calc exptl
0.292
(20°)
0.102
(1.5°)
1.586
2.895
2.331
(20°)
2.351
(20°)
1.145
(20°)
1.391
(20°)
0.585
0.439
(20°)
0.124
0.096
0.104
0.111
(20°)
3.06
2.41
(20°)


0.283
(20°)


2.16



2.21
(20°)







0.099





3.47
(20°)


law
•Vaol"
recon




2.0^0.4

2. 30*. 2
(20^)


(.204-. 20
(20°))


0.5B+.02


0.1U.01




2.B+.04




                                                                                                                      (continued)

-------
                                                         TABLE 2.2 (continued)
CO
Compound MW "P,°C




1,1,2-Tri 133.4 -36.5
chloroethane

1,1,1,2-Tetra- 167.85 -70.2
chloroethana
1,1,2,2-Tetra- 167.85 -36
chloroethane

1,1,2,2,2- 202.3 -29
Pentachloro-
ethane
Hexachloro- 236.7
ethane

Vinylchloride 62.5 -153.8





hp,°C Vapor
preaaure
p. kPa *
12.80
(20*)
5.33
(1.5*)
113.8 4.04
3.30
(20*)
130.5 1.853

146.2 0.867

0.647(20*)
162 0.60
0.444(20*)
-
186 0.044
0.028(20*)

-13.4
344(20*)

308(20*)


Solubility
s,g/«3
480(20*)

880(1.5*)

4420
4500(20*)

1100

3000
3200

480
500
(20*)
8
50
(22*)
2700
90
(20*)

60
(10*)
Henry 'a la
conatant kPa ar
calc exptl
3.56
(20*)
0.808
(1.5*)
0.122
0.8978
(20«)
0.283

0.0485
0.0455

0.253
0.180(20*)

1.302


2.35d

70.4(20*)d

105.6(10*)d 117.6
(10*)
M
/•Ol«
recoM




0.12402


0.28+.02

0.048+. 04


0.221.04











                                                                                                                       (continued)

-------
TABLE 2.2 (continued)
Conpoind MM «p,°C
1,1-Oichloro- 96.94 -122.1
ethene

1,2-Oichloro- 96.94 -80.5
ethena (els)
1,2-Oichloro- 96.94 -50
ethena (trane)

1,1,2-Trl- 131.4 -73
chloroathene





Tetrachloro- 165.83 -19
ethena






Trichloro- 147.5 -14.7
propane
bp,°C Vapor
preaaura
10 *
P. kPa
37 79.73
66.0(20«)

60.3 27.46

47.5 43.47
34.65(20*)

87 9.87

7.86(20*)

8.0(20*)
3.27
(1.5*)
121 2.48

1.90(20*)

1.80(20*)

0.64(1.5*)

156.9 0.413

Solubility
S,gV
400
5500
(20*)
3500

6300
300
(10*)
1100
1000
1100
(20*)

1000
(1.5*)
140
400
120
(20*)
150
(20*)
130
(1.5*)
1900
(20*)
Henry 'a law
conatant kPa wi*/mol*
calc exptl recoil
13.32
1.16 15.61
(20*) (20*)
0.761

0.669


1.179
1.30
0.939 0.904
(20*) (20*)

0.430
(1.5*)
2.94 1.239 2.J+.4
1.03
2.62 2.03
(20*) (20*)
1.99
(20*)




                                                           (continued)

-------
                                                           TABLE 2.2 (continued)
u>
00
Compound MM n
-------
TABLE 2.2 fcontinued)
Compound HW mp,°C bp,°C
Bronomethane 94.94 -93.6 3.56

riuoromethane 34.03 -141.8 -78.4
"Extrapolated valiea fro* liquid atata.
^Calculated from the extrapolated vapor preaaure with a
Extrapolated from aolid vapor praaaura.
"Calculated using atmospheric preaaura.
Data for halogenated aroma tics at 25 °C
Chlorobenzene 112.56 -45.6 132






o-Oichloro- 147.01 -17.0 180.5
benzene



m-Oichloro- 147.01 -24.7 173
benzene

p-Oichloro- 147.01 53.1 174
benzene

Vapor
preaaura
p. kPa*
183.9

3536
Solubility
S.Q/.3
18040
(20-)
1770
ttnm\
Henry's
constant kPs
calc exptl
0.533(20°)d

1.95(30°)d
ISM
ar/wol*
recom



\JU-f
fugacity ratio correction.

1.581



1.590


0.196



0.20
0.307


0.0902°



471.7
500
490
503

448(30*)
488(30°)
145.2
145
152
92.7

123.2
123
120
83.1
87.2


0.377 0.382
0.356 0.314
0.363
0.354



0.198 0.193
0.198
0.190


0.366
0.366
0.367
0.160 0.240
0.152


0.35+.05
•





0.19+.01




0.36+.02


0.16*. 02

(continued)

-------
TABLE 2.2 (continued)
Compound MW



1,2,3-Tri- 181.45
chlorobenzene
1,2,4-Tri- 181.45
chlorobenzene
1,3,5-Tri- 181.45
chlorobenzene
1,2,3,4-Tetra- 215.9
chlorobenzene
1,2,3,5-Tetra- 215.9
chlorobenzena
1,2,4,5-Tetra- 215.9
chlorobenzene
Pentachloro- 250.3
benzene
Hexachloro- 284.8
benzene

mp,°C bp,°C Vapor
praaaure
D. kPa*



53 218 0.0530"
0.0280b
16.95 213.5 0.0606

63 208 0.077"

47.5 254 0.00876"
0.00521b
54.5 246 0.0186"
0.009Bb
140 243 0.0101"
0.000 72b
86 277 8.89xlO-3"
2.19xlO~3b
230 322 3.44x10"*"
1.45xlO-6
(20-)
Solubility
s,g/*3
79
76
90.6
16.6
31.5
25.03
34.57
25.03
6.59
4.31
3.50
0.595
0.560
0.0050

Henry 'a law
conatant kPa n'/nol*
calc exptl race*
0.168
0.174
0.146
0.306 0.127
0.161
0.439
0.318
0.233
0.0884
0.261
0.593 0.159
0.261
0.977
0.0050
(20-)
(continued)

-------
TABLE 2.2 (continued)
Compound MM mp,°C
2-Chloro- 126.6 -39
toluana

F luorobenzena 96.11 -41.2


a,a,a-Tri- 146.11 -29.11
fluorotoluene
Broinobenzana 157.02 -30.62








m-Olbromo- 235.92 -7
benzene
p-Oibrono- 235. 92 87.33
benzene

2-Bromoothyl- 185.07 -67.5
benzene
bp,°C Vapor
preaaura
D. kPa *
179.3 0.173
0.236(20")

85.1 10.20


102.06 4.98

156 O.S52

0.570




0.997(35")

218 0.057(35*)

219 0.0215
0.018(35°)

218 0.0326"

Solubility
s,g/«3

466
(30«)
1553
1540
<30«)
450.7

410
360

500
(20-)
446
(30«)
458
(35-)
67.47
(35«)
20.0
26.42
(35")
39.05

Henry's
constant kPa
calc exptl

0.0641
(30»)
0.631


1.61

0.211 0.247
0.241





0.342
(35-)
0.199
(35«)
0.254
0.161
(35«)


law
•'/boi*
recoM



0.631.06




0.21+.04















                                                               (continued)

-------
TARIF 7 ? (continued)
Compound MM mp,°C bp,°C
lodobenzene 204.01 -31.21 188. 3
p-Dliodobenzana 329.91 131 285
1,4-Bromo- 191.46 68 196
chlorobenzene
1-Chloro- 162.62 -2.3 258.8
naphthalene
2-Chloro- 162.62 61 256
naphthalene
Extrapolated from liquid atata.
"Calculated from the extrapolated vapor preaaure with a
cExtrapolated from aolid vapor preaaure.
Data For peatlcidea
Ltndane 290.83 112.9
Vapor
pressure
D. kPa *
0.132
0.0344
Fugacity ratio correction
8.39x10-*
4.35xlO-6
. (20«)
4.13xlO-6
(20-)
2.8DxlO-6
<20")
Solubility
S,g/m3
180
229
340
(30«)
1.4
1.86
44.88
22.4
11.7
•
7.3
7.80
Henry 'a law
constant kPa mVmol'
calc exptl recom
0.150 0.13+.02
0.118
0.147
0.355
0.0319
3.34x10-* (3.2+.2)xlO-A
3.13x10-*
                                                              (continued)

-------
TABLE 2.2 (continued)
Compound MW np,°C bp,°C Vapor
pressure
D. kPa*
(20")
1.25xlO-6
(20")
4.0xlO-3
(20-)
Aldrin 557.9 59.60 7.99xl(T7
Oleldrln 373.9 175 6.59xlO~7
3.47xlO~7
(20-)
3.87xlO-7
(20«)
2.53xlO-7
(20-)
l.OftxlO'7
(20»)
2.67xlO'8
(20")
2.37x10-°
(20-)
DDT 354.5 109 185 1.34xlO-8


2.53xlO-8
(20»)
Solubility
s,g/«3



0.2
0.017
0.25
0.20






1.2xlQ-3
(20»)
5.5xlO-3
(20-)
3.1xlO~3
(20«)
l.OxlO-3

Henry's
constsnt kPs
calc exptl



1.43xlO-3
4.09xlO-3
9.86x10"*
1.23xlO-3






3.9xlO~3
(20-)
8.46X10"3
(20«)
1.53xlO-3
(20")
8.97xlO-3

ISM
m'/mol"
recom



(2.8±1.4)xlO-3
(l.l±.2)xlO-3






(5.3+3.B)xlO-3




                                                               (continued)

-------
                                                    TABLE 2.2 (continued1
Compound
Parathlon





Methyl.
parathion



Malathion



Chlor-
ptir 1 foa

MW np.°C bp,°C Vapor Solubility Henry's IBM
pressure S,g/ra* constant kPa •'/•ol*
p. kPs* calc exptl iscom
297.27 6.1 113/ 24
(76) .0067 5.04x10-* 11.9(20°) 1.23x10-*
(76) (20°)
11.0(40°)
2.61xlO~6
(20°)
5.85xlO-7
(20°)
7.6xlO-7
(20°)
263.18 35-36 109/ 3.94xlO~* 25(20°) 4.51xl--5 (2.0±1.0)xl
-------
    water  in the  porous  medium.   Compounds with high
    water  solubility  (from  surface spills)  would of
    course have shorter downward  travel  times as
    reflected  by  classical breakthrough curve analysis.
    Pfannkuch  (1984) points  out that for oil spills the
    hydrocarbon components with differing solubilities
    will dissolve out differentially and will produce a
    simultaneous  ageing  and leaching effect  on the
    spill.  See  tables  by Mackay and Shiu (1981)  for a
    listing of the  solubilities of  various  organic
    compounds.
(3)  Henry's  law  constant (Ky):  According to Mackay and
    Shiu  (1981),  "The Henry's  law  constant  is
    conventionally expressed as  a ratio of partial
    pressure in the  vapor to the  concentration in the
    liquid."   It is thus a coefficient that reflects
    the  air-water partitioning.   Such information is
    helpful  in understanding  in  what phase an organic
    compound  would  most likely  be  found.   Thus,  an
    organic  compound with a high  vapor pressure and low
    water solubility would be expected to be favored in
    the  vapor phase and would  therefore be a relatively
    good candidate for soil vapor  monitoring.  However,
    if the  detection  of organics in ground water  is of
    primary  interest,  a false  positive might  result
    from this situation.  The  relationship between
    solubility and vapor pressure  is  plotted in Figure
    2.1  by  Mackay and Shiu (1981)  and  shows  that
    compounds within a homologous  series generally lie
    on a 45°  diagonal  of constant Henry's law constant
    and that pesticides  typically have very low vapor
    pressures.  Marln  and  Thompson  (1984) have found
    that actual field  partition  coefficients are less
    than laboratory  coefficients  and that the field
    data suggest larger partitioning into the liquid
    (ground  water)  phase.   They  theorize that this is
    probably the  result of either  lateral diffusion or
    of equilibrium conditions  that  are typically  never
    achieved in the  field between  diffusion in soil gas
    and water.

(i»)  Concentration:   The concentration refers to the
    amount  of organic  compound per  unit  amount  of
    solvent  (air/water)  in units such as g/m3, ppb/v.
    Kreamer  (1982)  concluded  that  the diffusion of a
    gas  from areas of high concentration  to  low,
    resulting only  from the existing concentration
    gradient,  was the mechanism  of greatest importance
    for  gas  transport  in the  unsaturated  zone.   The
    concentration of the organic  compound in the ground

                            45

-------
10 —

10* -
-
:, :
10° -
io-1-
0.
jf A
u? 10~ -
flC
• -3
•9 10 —
III •«
ff
Ct -4
0 10 -
io-6-
io-e-
io-7-
io"-

40 •

"*10
10
1C

V HALOOENATED HYDROCARBONS
6 POLYNUCLEAR AROMATIC8 O
• DIENE8 9k^»
* CYCLOALKANE8 O"A ®
-• MONOAROMATIC8 ^ (9 gS7 » e^7 *
0 PESTICIDES ffiigb JL
O ALKANE8 ^ V ^ X *
^ ALKENE8 o^5 ^5? * *
« ALKYNES <5O^ oT^7 **
°° .*•"*."
0 \7


*/*

*r
&
Jy ° DO
° ° 0
o
o
o
a
o




















I I I I I I I I I
f7 io~e io~8 io'4 io'3 io~2 lo'1 10° io1 10* io3
                       SOLUBILITY, mol/m*
Figure 2.1.
Plot of  solubility vs vapor pressure illustrating
the tendency for compounds in a homologous series
to lie on a 45° diagonal  of constant Henry's law
constant  (MacKay and Shiu, 1981).
                              46

-------
    water will dictate to a great  extent  the vertical
    concentration gradient  of  the vapor  in  the
    unsaturated zone.   Thus  if  the concentration of the
    organic compound in question  is low in the  ground
    water  (and the compound  is  not Insoluble  in water),
    then  most likely  the vapor  gradient will  be
    difficult  to detect,  and either a  different
    compound should be  selected as  a tracer  or  a more
    intensive sampling grid should be imposed to more
    accurately delineate the  contaminant plume.

(5)  Koc:    The  Koc for  an  organic compound  is  a
    coefficient that relates  the  partitioning  of the
    organic compound between the adsorbed phase and the
    soil  solution relative to  the organic  carbon
    fraction.  The Koc  reflects  the affinity  of an
    organic compound to  adsorb out of solution  onto
    soil   organic material.   Figure 2.2  shows  the
    correlation between  water  solubility  and Koc and
    demonstrates  that  those compounds with  low  water
    solubility often possess  higher Koc values (Wilson,
    et  al., 1981).  Although  this  value will reflect
    the potential of  an organic  compound  to  be
    attenuated  in the unsaturated zone, it  is of course
    totally  dependent on the  presence  of organic
    material.   Often the organic carbon content in the
    unsaturated zone  will  decrease  with  depth
    (influence  of vegetation)  and can  be almost
    entirely void  in subsurface  coarse material.

(6)  Density:  The density of an organic compound refers
    to  the  amount  of substance  per unit volume (g/cm3).
    [Schwille (1984) indicates that next to  solubility
    the difference in density  between contaminant and
    ground water  is the  next  most  important parameter
    in  determining the  contaminant  migration relative
    to  the aquifer.]  Hackay  (1985) states  that density
    differences  of about  1  percent are  known  to
    influence fluid movement  significantly and that
    with few exceptions the density  differences between
    organic  liquids  and water are in excess  of  1
    percent and  often 10 percent.   Such high  densities
    and  limited solubilities  of  chlorinated
    hydrocarbons  (Table 2.3) led  Byer  (1981)  to
    conclude  that this was  the primary cause for the
    widespread chlorinated hydrocarbon contamination of
    underground water sources.  Byer  (1981) also showed
    how the density of an organic  compound  such  as TCE
    can be used effectively to trap the compound in a
    recovery process (Figure  2.3).
                           47

-------
 o
 o
o
o
   8.6 "1
   8.0
   2.6  -
    8.0
                       OE     Q
                        r
                                                       JKL
    1.6
              T

              2
T

 3
                   LOO WATER SOLUBILITY 
-------
                            TABLE 2.3  CHLORINATED HYDROCARBONS
Compound
Hex achloroe thane
1,1,1,2,2-Pentachloroethane
Tetrachloroethylene
1, 1,2, 2-Tetrachlore thane
Carbontetrachloride
1 ,1 ,1 ,2-Tetrachloroethane
Chloroform
1,2, 2-Trichloroethy lene
1,1,2-Trichloroethane
1,1,1-Trichloroethane
Methylene Chloride
Cis-1 ,2-0 ichloroethy lene
Trans 1,2-0 ichloroethy lene
1,2-0 ichloroethane
1 ,1-0 ichloroethy lene
1 , W) ichloroethane
Cheaical
Foraula
C2«6
C2HC15
C2C14
C2H2C14
CC14
C2H2d4
OC13
C2HC13
C2HjCl3
CzHjCl,
CHjClj
C2H2C12
C2H2C12
C2HAC12
C2H2C12
C2H4C12
8 P oC
760an Ha.
186.3
162
121.2
146
76.7
129
61.7
87
113.5
74
40.4
60
47
83
37
57
Solubility in
lOOa H->0 (1)
0.005 ga
0.05 ga
0.015 ga
0.29 ga
0.08 ga
0.015 ga
0.8220 oa
0.1100 gi
0.4500 ga
0.4400 OB
2.000 ga
0.35 ga
0.63 ga
0.869 ga
4.00 ga
0.55 ga
Density
20/4°C
2.2091
1.6796
1.6227
1.5953
1.5940
1.59069
1.4832
1.4642
1.4397
1.3390
1.3266
1.2837
1.2565
1.2351
1.218
1.1757
^Chlorination  of siaple  hydrocarbons yields a variety  of chlorinated hydrocarbons varying in
physical properties in a regular fashion depending upon the degree of chlorination.  Usually,  as
the chlorine content of the hydrocarbon is increased there is a progressive decrease in specific
heat,  dielectric constant and solubility in  Mater.  The substitution of chlorine onto  the
hydrocarbon iaparts  an increase in  solvent  power, viscosity, nonflaaaability, cheaical
reactivity and density.  (Byer, et al.,  1981)
                                             49

-------
  EFFLUENT
                                               OVERFLOW
                                              TREATMENT
      TO

    RE-USE
,'f
.'* «•>«/»
•&3
*<$#
•sfeW
                       Nitroglycerine
                                              8EPERATION
                                                    LAYER
Figure 2.3.  Diagram showing an economical and safe way  to
            contain a chlorinated hydrocarbon (TCE) compound.
            The  so-called nitroglycerine  trap collects TCEs
            from the plant via floor drains.  Traps must  he
            concrete lined  to prevent seepage.   It  is
            economical to reprocess TCE when volume used  is
            large (Byer,  1981).
                             50

-------
    Organic compounds with specific gravities of less
    than 1.0 associated with solubilities of less  than
    1  percent are often referred to as floaters,  see
    Table 2.1 compiled by the New York State Department
    of  Environmental Conservation  (1983)-   If  the
    floater  is  also classified as  being  highly
    volatile, the compound would be a  good candidate
    for  soil  vapor monitoring as it  would  not be
    diluted   in the aquifer and potentially  could
    establish a steeper vapor concentration gradient.
    However, floaters would have difficulty moving past
    subsurface obstructions.  Thus, it would  also be
    critical that vapor concentrations be correlated
    with the  concentration of the organic floating at
    the water  table and that  a properly  screened
    monitoring well be used  to avoid sampling ' water
    deeper in the aquifer.

(7)  Viscosity:   The viscosity of a  liquid  organic
    compound is  a measure of the degree  to  which it
    will  resist  flow under a given force measured in
    dyne-seconds per cm^.   According to Schwille
    (1984),the viscosity of  the organic fluid (such as
    oil) will affect the  velocity of the flow  process.
    Mackay  (1985) adds that  it is the  combination of
    density and viscosity that will  govern the
    migration of an  immiscible organic liquid in  the
    subsurface.  Thus,  it  is the viscosity of an
    immiscible organic fluid (such as  oil) that  will
    control the lens thickness on  a  water  table.
    However, even with high  viscosity  fluids  such as
    oils, Holzer (1976) points out that  many months
    after a spill, it is often too late to determine
    whether the  spill was. caused by a catastrophic
    release or a slow steady "drip".  Figure 2.4  shows
    that after 300 days,  little difference is  noted in
    the thickness of a  spill  developed  under  two
    different permeability  values.

(8)  Dielectric constant:  The dielectric constant of  a
    medium is a  parameter  that relates  the  relationship
    between two  charges and  the distance of separation
    of  the two charges to the  force of attraction.  In
    a clay medium this constant reflects the degree to
   - which the clays will  either shrink or  swell.
    Liquids with a high  dielectric constant (Table 2.5)
    such as water would  be  expected to cause the  clays
    to  swell.   Conversely,  those liquids with  a  low
    dieletric constant would  cause  the clays to shrink
    and, therefore, increase In permeability  after
    exposure to  concentrated organic liquids.   Table

                           51

-------
liquid
Ace tal deny de
Acetic acid
Acetic anhydride
Acetone cyanohydrin
Acetyl bromide
Acetyl chloride
Acrolein
Acrylonitrile
Allyl alcohol
Allyl chloride
Ammonium hydroxide
Amyl acetate
Aniline
Benzene
Benzonitrile
s^Benzoyl chloride
Benzyl chloride
Bia(chloromethyl) ether
Bia (2-chloroisopropyl)
ether
Bia (2-chloroethoxy)
methane
Bia (2-chloroethyl)
ether
Bis (2-ethylhexy)
phthalate
Bromoform
(trtbromantethane )
Butyl acetate
Butyl am ine
Butyl benzyl phthalate
Butyric acid
Poisonous
_
-
-
X
-
-
X
-
X
X
-
-
X
-
„
-
-
X

_

_b

X

X

-
-
-
-
-
flammable Corroalve Reactive
X
X
X X
-
X X
X - X
X - X
X
X
X
X
X
_
X
. - -
X X
X
_

_ _ _

J> J> J»

- •

_

.
X
X - -
_
X
Volatile Floater
X
-
-
-
-
X
X
X
X
X X
-
X
-
X X
-
-
-
X

X

xb _b

X

X X

-
X
X -
-
-
Amenable to
biological
treatment
X
X
X
X
X
X .
X
X
X
X
X
X
X
X
X
X
X
X

x

X

X

-

-
X
X
X
X
Bio-
da gradeable
X
X
X
X
X
X
X
X
X
X
—
X
X
X
X
X
X
X



X

X

-

-
X
X
X
X
Highly toxic
to
aquatic life
-
—
-
X
—
—
X
-
—
—
™
~
—
-
-
-
-
ND

ND

ND

ND

ND

ND
-
-
-
-
(continued)

-------
I iquid Poisonous
Carbon disulfida
Carbon tetrachlorida
Chlordane
Chlorobenzena
Chlorodibromomethana
Chloroethana x
2-Chloroathyl
vinyl ether
Chloroform
2-Chlorophenol x
4-Chlorophenyl phenyl
ether xb
Chloroeulfonlc acid
Crotonaldehyde x
Cyclohexane
Diazinon
1,2-Oichlorobenzene
1 , 3-Dichlorobenzene
• L.
Dlchlorobromomethana -
1 , 1-Oichloroathane
1 ,2-Dichloroe thane
1 , 1-Dichloroathy lane
(vlnylidane chloride)
1 ,2-Oichloropropane
2 . 2-Dichloropropionic
acid
1,3-Oichloropropylene x
Dichlorovoe
Diethy lamina
Dlethyl phthalate
Dimethyl phthalate
Di-.n-octyl phthalate x

Flammable
x
a»
x
_
X

X
-
-
J>
-
X
X
-
-
_
_b
X
X

X
X

-
X
-
X
-
-
-

Amenable to
Corroeive Reactive Volatile floater biological
treatment
x
x
x
.-XX

X
X
-
_b _b xb J»
X X
X - X X
x x
• — - —
_
_
_b J> _b
- ' X -
X

X
X

X
X
X
x - x x
X -
X
X

X
-
X

X
X
X
_
X
X
X
* *
X
X
-
X
X

-
X

-
X
-
X
X
X
X

degredeable to
aquatic life
x
-
X

X
-
X
.
X
X
X
~
X
X
-
X
X

-
X

-
X
-
X
X
X
X

X
ND
ND

ND
ND
••
-
ND
—
x
ND

—
••
ND
ND
—

ND
•


—
x
ND
ND
ND
ND
(continued)

-------
Liquid
Di-rv-octyl phthalate
Diaulfoton
Dodecylbanzeneaulfonic
acid
Endrln aldehyde
Epichlorohydrin
Ethlon
Ethyl benzene
Ethylene dl bromide
Ethylenedlamina
Formic acid
Furfural
Hexachlorobutadlane
Hydrochloric acid
Hydrofluoric acid
en laophorone
*~ laoprene
Halathion
Methylena chloride
(dichloromethane )
Methyl methacrylate
Hevinphoa
Naphthenlc acid
Nitric acid
Nitrobenzene
^-nitroaodimethylamine
^-nltroaodi-n-
propy lamina
Parathion
Phosphorous oxychloride
Phosphorous trichloride
Pol/chlorinated
blphenyia (PCBS)
Prop ionic acid

Poisonous Flammable Corrosive
_ — -
X

X
_
XXX
- - -
X
X - X
X -
- - X
- X
_ — -
X
X
_
X
_ _ _

_ .
X
X - X
X
X
X
.

_b J> J>
X
X
- X

_
X

Reactive Volatile
X
-

— —
-
-
-
X
X
-
-
- -
-
- -
-
-
X
- -

X
X
-
-
-
_
X

_b xb
-
X
X X

-
-

Floater
-
-

_
-
-
-
X
-
-
-
-
-
-
-
X
X
-

-
X
-
-
-
-
-

_b
-
-
-

-
-

Amenable to
biological
treatment
X
-

X
-b
X
-
X
X
X
X
X
X .
X
X
X
X
-

-
X
-
X
X
X
-

-
X
X
X

-
X

Bio-
degradeable
X
-

X
_b
X
-
X
X
X
X
X
-
-
-
X
X
-

-
X
-
X
-
X
-

-
X
X
-

-
X

Highly toxic
to
aquatic life
ND
ND

ND
. X
-
X
-
-
-
-
-
ND
-
-
to
-
X

ND
-
X
-
-
-
M)

ND
X
-
-

X
-
(continued)

-------
Ln
Ul
TABLE 2. 4. (continued)
liquid Poisonous
Propionic anhydride
Propylene oxide
Quinoline
Styrene
Sulfuric acid
Sulfur mono-chloride
2,4, 5-Trich lorophenoxy
acetic scid with
aroinss
1,1,2,-Tetrachloroethane
Tetrachloroethylene
Tetraethyl
pyrophosphste x
Toluene
1 ,2-trana-
Dichloroethylene
1 ,2, 4-Trichlorobenzene
1,1, 1-Tr ichloroethsne
1,1,2-Trichloroethane
Trichloroethylene
Trlethylamlne
Vinyl acetate
Vinyl chloride
(chloroethylene)b
Xylene
rismmable Corrosive
x
x
.
X
X
X


X
x
-

-
X

-
^ _
-
-
-
X X
X

X
X
Reactive Volatile
x
X
-
X
-
X


X
X X
X

X
X

X
-
X
X
X
X
-

X
•• —
rioster
-
-
-
X
-
-


.
-
-

-
X

-
-
-
-

X
X

X
X
Amenable to
biological
treatment
x
x
-
x
X '
X


-
- . ,
-

.
X

X
X
X
X
X
X
X

X
X
Blo-
degrsdeable
x
x
-
x
-
-


-
-
-

-
X

X
X
X
X
x-
X
X

X
X
Highly toxic
to
aquatic life
.
• ND
-
-
-
-


-
ND
ND

NO
-

ND
ND
ND
-
ND
- •
-

ND
—
  ND
  b
Denotes chemical is in this category.
Denotes chemical is not in this category.
Denotes no data available.
Denotes assumed valuaa based on chemical groups of  similsr  type  substancee.
   (New  York  State  Department  of Environmental  Conservation, 1983)

-------
   10 -i
    6 —
c
•9
Ml

I   4
o
    2  -
                                   k  - .0026 cm/s
       10
                       30
 I
too
800
                           TIME (days)
       Figure 2.4.  Thickness of center of oil lens versus time where
                   k values are  permeabilities with  respect  to oil
                   (Holier, 1976).
                                  56

-------
       TABLE 2.5.  DIELECTRIC CONSTANTS, DENSITIES AND WATER
                    SOLUBILITIES OF VARIOUS HALOGENATED AND
Name
Water
Methanol
Ethanol
Acetone
1-Propanol
1-Butanol
b-Pentanol
Pyridine
Phenol
Dichlorome thane
1-Bromopropane
1,1, 1-Trichloroethane
9 »
Aniline
Chloroform
Bromoform
Trichloroethylene
Toluene
Benzene
Carbon tetrachloride
Cyclohexene
Hexane
Tetrachloroethylene
Dielectric
Constant
78.5
32.7
24.6
20.7
20.3
17.5
13.5
12.4
9.8
8.9
8.1
7.5
6.9
4.8
4.4
3.4
2.4
2.3
2.2
2.0
1.9
2.2
Density
(g/cm3)
1.00
0.79
0.79
0.79
0.80
0.81
0.81
0.97
1.05
1.31
1.34
1.34
1.02
1.48
2.89
1.48
0.87
0.88
1.60
0.78
0.65
1.6
Solubility
_
Miscible
Miscible
Miscible
Miscible
Miscible
Miscible
Miscible
Miscible
1.32Z
Slightly Soluble
Soluble
0.82Z
0.10Z
0.1 11
Slightly Soluble
Slightly Soluble
0.08Z
40 mg/1
—
150 mg/1
(Anderson, et al., 1984)
                                    57

-------
        TABLE 2.6   INTRINSIC  PERMEABILITY OF  SOILS  PERMEATED  BY
                      WATER AND ORGANIC LIQUIDS

                    Intrinsic  Permeability (KT9cm2)
                 Nonpolar  Solvents	Polar Solvents
Soil type
Sand
Sandy-clay
Clay
Water
145.
1.6
0.5
Kerosene
179.
158.
10.1
Xylene
151.
146.
14.0
Ethylene
glycol
135.
62.8
2.4
Isopropyl
alcohol
177.
93.2
6.4
(Anderson,  et al.,  1984)
                                   58

-------
       2.6  (Anderson,  et a1.,  1984) shows  the impact
       various solvents will  have  on the  intrinsic
       permeability of  various soil types.  Such a result
       would of course mean that in a clay soil or  horizon
       the  concentrated organic plume will  reach the
       ground water in a  shorter travel time and that the
       plume  would  expand  to  a larger  volume.  In
       addition,  the vapor  phase moving back toward the
       surface in the same area would  not be restricted to
       the same degree.

  (9)   Boiling Point:  The  boiling point of a compound is
       the temperature at which the external pressure of
       the  liquid is in equilibrium  with the saturation
       vapor  pressure of the liquid.   Thus, for  higher
       boiling points there is a general association with
       lower vapor pressures.  Again, it  is those organics
       with higher vapor pressures and  thus lower  boiling
       points that would be  better suited for soil  vapor
       monitoring.  A listing of the  boiling points for
       various organic compounds is included in table 2.2
       compiled by Mackay and Shiu (1981).

 (10)   Molecular  weight:  The  molecular  weight  of an
       organic compound  is the sum total  of the weights of
       the  atoms  that compose it (see table 2.2).   Mackay
       and  Shiu (1981)  indicate that  for liquid  alkanes
       there  is a tendency for the Henry's law constant to
       increase with increasing molecular weight as the
       solubility falls more  than  the vapor pressure.
       High molecular weight hydrocarbons (especially
       aromatics) are decomposed through biodegradation at
       a much slower rate (American Petroleum Institute,
       1972)  and  thus would persist in the unsaturated and
       saturated zones longer.

PROPERTIES OF THE UNSATURATED ZONE

   (1)  Air  filled porosity:   The air  filled porosity of a
       porous medium such as soil  is defined as the  ratio
       of the  volume of  air in the soil  pores to the  total
       volume  (volume of air, water,  and soil combined).
       It  is  thus indicative  of soil aeration and is
       inversely related  to the degree  of saturation.  The
       air-filled porosity  is an important parameter in
       estimating the  diffusion of gas  in  soil and
       unconsolidated material.  The diffusion coefficient
       of oxygen  is approximately 10,000 times lower In
       water  than in air (Letey, et  al., 196U).   Thus,
       soil  organic  vapors migrating  toward  the soil
       surface would be restricted if the water  content

                               59

-------
    increases  and  the  air-filled porosity decreases.
    Vapors moving  into  low air-filled porosity zones
    could potentially be resolubilized.  Because  of the
    restrictions  in  flow   and  the  possible
    resolubilization,  vertical  soil gas profiles would
    be poorly  established.  This parameter  would be
    expected to change  dramatically in the unsaturated
    zone as  it  is dependent on the position relative to
    a  wetting front  (rainfall)  and on changes in
    texture  (water  holding capacity).

(2)  Volumetric water  content:   The volumetric water
    content is the ratio of the volume of water in a
    porous medium  to  the total  volume.  When  the water
    fills  the  entire  pore volume  the  medium is
    saturated.  Coarse soils  have lower volumetric
    water  contents  at saturation than  do medium
    textured soils  and  medium textured soils  less  than
    clayey soils.   Under unsaturated flow conditions,
    it is the  unsaturated hydraulic conductivity and
    the hydraulic head gradient that  dictate the  change
    in volumetric  water  content with time.   As the
    volumetric  water content increases, the air  filled
    porosity  decreases, and the path for vapor flow
    becomes  restricted.  Thus, as the percent water in
    the  soil  pores  goes up,  the  possibility of a
    vertical  soil organic vapor gradient  being
    established is lessened and  so is the likelihood of
    correlating soil gas measurements with ground water
    concentrations.   Reid (1985) has  indicated  little
    success  in  soil vapor monitoring studies when the
    vadose zone contains high clay  and water  contents.
    Water content  in  subsurface layers has also been
    shown to be the decisive factor in determining the
    shape of an oil body and the distribution  after
    percolation (Scwille, 1975).

(3)  Soil organic matter:   The amount of organic  matter
    in a soil  varies  according to the vegetation,
    climate,  and  the  rate   of  decomposition.
    Agricultural soils  often possess  in excess of 2
    percent organic matter whereas  desert soils  can be
    almost entirely void of organic matter.   Organic
    material  generally has  high surface  areas and
    exchange properties  ideal for adsorption of organic
    compounds.  The Koc value reflects the impact of
    this organic material to adsorb organic compounds
    out  of  solution.   Chiou (1985) states that "the
    extent of uptake of  nonionic  organic compounds from
    water  on  a  great  variety  of  soils is  closely
    related to soil organic matter  content."   Chiou

                            60

-------
     also indicated that  "when soils are fully hydrated,
     adsorption of the organic solutes  by soil  minerals
     becomes relatively insignificant compared to the
     uptake by partitioning into soil organic  matter,
     presumably because water is preferentially  adsorbed
     by  minerals."

(4)   Soil texture:  The texture of a soil refers to the
     proportions of various particle  size groups  in  a
     soil  mass.   These  particle size  groups  are
     typically called sand, silt, and clay.  Figure 2.5
     shows the textural triangle and  the various  soil.
     textural classes.  As the clay  content increases,
     the water holding capacity increases, the  exchange
     capacity increases, and  the rate of diffusion
     decreases.   Thus,  if  a  high  clay content layer
     exists  in the subsurface (textural  discontinuity)
     or  if the entire vadose zone is comprised of clayey
     soil, it will  act as  a retarding layer to  the
     vertical flux of volatile organic  carbons.  Swallow
     and Gschwend (1983) point out  that  it is  the  rate
     of  flux through the most retarding layer that  will
     control the vertical flux.  Figure 2.6 (American
     Petrolem Institute,  1972)  shows the possible
     effects of clay layers and lenses  on the migration
     of  contaminants in the  unsaturated zone.  Indurated
     layers  such as petro calcic layers  can also alter
     the flow path of contaminants  both in the liquid
     and vapor phases.  Conversely,  gravel layers  have
     been shown to act as a conduit for organic vapors
     to  diffuse laterally from more  contaminated soils
     (Marin, et al.,  1984).

(5)   Vapor  pressure of water in the  soil pores:   The
     aqueous vapor pressure measured  in soil pores is
     for the most part considered to be vapor saturated:
     as  Hillel (1971) points out, a change in matric
     suction between 0 and  100  bars is accompanied  by  a
     vapor  pressure change of only  1.6  millibars.
     Temperature has a much  larger  impact on vapor
     pressure as a 1°C change has almost the same effect
     as  the 100-bar suction change.   Since vapors  tend
     to  move from warm to cold  areas  in a soil,   the
     vapors would therefore  tend to move  downward during
     the day and upward during the night.   At the  soil
     surface and perhaps down even several Inches, the
     vapor pressure can drop  below saturation because of
     higher gas mixing and exchange rates.   The presence
     of  electrolytes (often  concentrated near the  soil
     surface via evaporation) can also  lower the vapor
     pressure.  If the vapor  pressure in  the soil pores

                             61

-------
                        100
                  Percent  by weight Sand
Figure 2.5.  Textural triangle,  showing the percentages of clay
            (below 0.002 mm),  silt (0.002-0.05 mm) , and  sand
            (0.05-2.0 mm) in the  basic  soil  textural classes
            (American Petroleum  Institute,  1972)  (Hillel,
            1971).
                             62

-------
           Possible Migration of  Product to Outcrop
           Followed by Second Cycle  Contamination
        AREA OF
          SPILL
                                   PERMEABLE SOIL
                                              IMPERMEABLE
                                              CLAY


                     Effect of Clay  Lene  In Soil
  LAND SURFACE
 LAYER OF  SATURATION
                                         UNSATURATED  ZONE
                                                 !* CLAY LENS
        ^m
        SATURATED 2 O N E =-=

Figure 2.6.  Possible migration of product  to outcrop followed
             by second cycle contamination  (American Petroleum
             Institute, 1972).
                                63

-------
    can be reduced, it has been shown by Chiou (1985)
    to have a significant effect on the  adsorption  of
    organic  vapors.   He states that the mineral
    fraction of a dry or slightly hydrated soil  will  be
    a powerful adsorbent for organic vapors at  lower
    vapor  pressures and that this  may become the  most
    important  process  in the  uptake of  organic
    compounds  by mineral  rich unsaturated  soils.
    Figure 2.7 shows the  impact of relative humidity  on
    the adsorption of benzene and Indicates a large
    increased adsorption  as the relative  humidity drops
    below  90 percent.

    This  relative humidity-adsorption  phenomena may
    cause significant reductions in  the  amount  of
    organic vapor measured at or near  the soil surface.
    Thus,  deeper  soil  gas probes would  be advised for
    precautionary means  (precluding the  use of surface
    gas measuring devices).  Probes located  beneath a
    depth of 12  inches in almost  all situations would
    encounter a saturated aqueous vapor,  thus  avoiding
    the  increased adsorption  phenomena shown  in
    Figure 2.7.

(6)  Shape and size of  pores:  Knowledge of  the  shape
    and size of pores or pore size  distribution  in soil
    is important in any  understanding  of the tortuous
    path  vapors  must  traverse in reaching the  soil
    surface.  Figure  2.8 is a drawing  of the possible
    variation  in pore  size  an organic vapor  might
    encounter.   Note  that  some pores are totally
    blocked by interstitial  water,  and the rate  of
    diffusion  is  thereby reduced by orders of magnitude.
    It  is -also  worth noting  here  that  the total
    porosity  does  not  provide  any  indication of the
    pore size  distribution.  Nielson  and  Rogers (1982)
    using a  mathematical  model to  estimate  radon
    diffusion  in earthen  materials  calculated the
    diffusion coefficient for radon by using nine pore
    size  distributions.  Figure  2.9  shows  the impact
    various median  pore  sizes and water contents can
    have  on the  diffusion coefficient.   In  the  drier
    range,  a difference of  an  order  of magnitude  is
    observed  in  the diffusion  coefficient, and  Nielson
    and Rogers attributed this  to the  differences  in
    the median pore size.  Clayey soils  tend to have a
    more  uniform pore size  distribution than  do coarser
    soils (Hillel,  1971) whereas  the  coarse  soils tend
    to have larger  mean  pore  sizes which will  transfer
    fluids faster under saturated conditions  and vapors
    faster under unsaturated  conditions.   The diffusion

                           64

-------
     40
     3O —
E
%*
O
     10 -
                  I
                 0.2
 I
0.4
o.e
o.a
1.0
            RELATIVE VAPOR CONCENTRATION. P/P
Figure 2.7.  Benzene uptake by soil  as  a function  of the
             relative vapor  concentration, where P is  the
             equilibrium partial pressure  and P° is  the
             saturation vapor pressure of the compound  at the
             system temperature (Chiou, 1985).
                              65

-------
                            OPEN PORE AREA
SOLID GRAINS
INTERSTITIAL
      WATER
                      UNIT CROSS SECTIONAL  AREA
                                                               AIR
Figure 2.8.  Cross-sectional view  of soil  pore  area (Nielaon
             and Rogers, 1982).
                              66

-------
£   -2
010   -
VI

O
z
o
  io~  -j
 O
 O
  to'6 =
LOO-NORMAL PORE SIZE
     DISTRIBUTIONS WITH
     QEOM.STD.DEV. * 8
     POROSITY » 0.4
     DENSITY = 1.T Q/cm3
     r- MEDIAN PORE  RADIUS
          I

          0
0.2       0.4        0.6

   MOISTURE SATURATION
                                       0.8
1.0
    Figure 2.9, Conparison of diffusion coefficient moisture
               curves  for various median pore sizes (Hielson  and
               Rogers, 1982).
                               67

-------
    coefficient  must  .therefore, compensate for this
    tortuous path for vapor flow;  this is accomplished
    by  replacing  the  diffusion  coefficient with the
    effective diffusion coefficient  (see chapter  3).

(7)  Depth  of  unsaturated  zone:   Depth  of  the
    unsaturated zone  or depth to  the water table is a
    very  important site parameter in soil organic vapor
    monitoring.  Spreading of organic contaminants is
    enhanced  by a  shallow  unsaturated  zone  as the
    opportunity  time  and volume for adsorption  and
    retention  is decreased.   The shorter the vertical
    travel time is  in the unsaturated zone, the  greater
    is  the opportunity for  miscible organics to be
    dispersed  into  the  ground water.   [As  the
    unsaturated  zone  increases in size (vertically and
    laterally), the possibility of textural changes and
    distinct  horizontal  layers forming is  also
    increased.]  [Deeper sampling might be recommended
    to  avoid these  discontinuities, which would alter
    the vertical vapor gradient.]  Deeper  sampling
    correlates  to  increased cost and  to  decreased
    convenience.  In  terms  of  vapor monitoring, a
    deeper unsaturated zone means a greater distance
    over which  a  vertical  vapor gradient must  be
    established.  Thus, by the time vapors arrive at or
    near  the soil surface from deep water tables, the
    concentration  may be below detection levels and
    thus  provide little Information on  the spatial
    extent of  the  plume.  Successful  vapor correlations
    under  such conditions  would  require  the
    concentration  of the organic  in the ground water to
    be  exceedingly  high and that the unsaturated zone
    be  comprised primarily of coarse gravelly.material.

(8) Retention:   Depending  on  the solubility of the
    organic  compound,  the texture of the soil,  and the
    pore  size distribution, a certain percentage of  the
    liquid contaminant will be  retained  in  the soil
    pores.  Column studies by McKee (1972) showed  the
    specific retention  of water to  range from 10.37 to
     17.67 percent  in the soils he Investigated  whereas
     the specific retention of gasoline measured was
    only  6.10 percent.  Other studies that he conducted
     showed that even after 8MH  pore volumes of water
     had  passed  through  a  column  contaminated with
     gasoline,  the effluent still  had a gasoline taste.
     In  another column  McKee passed 3750 pore volumes of
     air and still  could not completely vaporize the
     gasoline.   The explanation for this comes directly
     from  Figure 2.10,  where Williams and Wilder (1971)

                            68

-------
     1.00
     0.80
     o.eo

     0.40  —


     0.20  —
     O. 10
     0.08
     0.08

     0.04  .
     0.02  -
      0.0
ȣ.:j.:;.:?::: CLEAN :
   I  GASOLINE
 ::•:::::: FLOW :•
           0%
        26%
60%
76%
1 00%
                               Wat»r
Figure 2.10.
Relative  permeability graph where ^water  is  the
percent saturation  of water  and kr is  the
permeability  ratio  (ratio of  observed
permeability at a given  saturation relative to
the  permiability at  100  percent saturation)
(Williams  and Wilder,  1971).
                            69

-------
    state that in region III little  flow  of  gasoline
    will take  place.   This, they state, is because:

      "The smaller  capillaries are  filled  with
      gasoline  only when  the  pressure  drop
      overcomes capillary forces.  Thus until the
      pressure drop P gaa  - P water l3 greater  than
      P capillary*  & ia impossible to  move the
      snapped  off gasoline bubbles through  the
      throats (of the pore).   As a  result,  the
      gasoline does not fill  the neck  or pore
      throat  and thus becomes extremely difficult
      to move.  The water which is used to flush
      the sand will therefore tend to flow through
      unblocked  and continuous water  filled
      channels rather  than  through the gasoline-
      blocked  channels."

    A  continuous release  of contaminant over  long
    periods  of  time,  because of  this retention
    solubility factor,  would make  it Increasingly
    difficult  to  describe the extent  and  the  cause  of
    the contamination as  pulses might soon overlap.

(9)  Temperature and temperature gradients:   Soil
    temperature  and  the  gradient  that  is  established
    within the unsaturated zone can  have  an impact  on
    the status  of  organic compounds.  If great
    temperature  gradients  exist  (surface layers),
    thermal diffusion will readily  take  place.   Hillel
    (1971) indicates that "the  effect  of  warming the
    soil  is to lower the suction  and raise the vapor
    pressure   of  soil  water.  Hence the  effect  of a
    thermal gradient is to induce flow and  distillation
    from  warmer  to  cooler  regions."  Thus,  organic
    vapors migrating from the  ground water  to  the  soil
    surface during summer months and during the daytime
    will typically have  to move  against a  temperature
    gradient  (i.e., movement  by concentration gradient)
    when they  enter the surface horizon. During winter
    months when  the  soil surface  may actually freeze,
    vapors would possibly be  unable to escape and would
    therefore concentrate  or  be  driven  to  move
    laterally.  Organic  compounds  that have   boiling
    points lower  than soil temperatures will of course
    be  highly volatile,  such as the  gaseous   alkanes
    propane   and  isobutane  which  boil at-42.1°C and
    -11.7°C (see tables by Mackay and Shiu).

    Soil  temperature can also have a large impact  on
    microbial growth.   Cullimore  (1982)  states  that

                            70

-------
     lower soil temperatures tend to increase bacterial
     migration  down  through  the  soil  and  that
     Pseudomonads will  actively degrade substrates
     (substances upon which enzymes  act)  under  aerobic
     conditions if the ground water temperature is above
     six to eight degrees  centigrade.

(10)  Microbial influence:  If conditions are  optimum
     (pH,  temperature, aeration,  nutrients,  detention
     time), the presence  of microbial populations in the
     subsurface can lead to a significant biodegradatlon
     of  organic compounds. The extent of  biodegradation
     would be  dependent  on the number and  species  of
     micro-organisms reaching  a critical  level  in
     relationship to  the degree of difficulty  in
     breaking  down the compound in question.  Table 2.7
     gives a summary of  the  growth  data for various
     micro-organisms on varying substrates in a study by
     Jamison et al. (1975).  Note  that no one organism
     thrived on all seven substrates and that  no  one
     substrate  supported  growth for  all  six
     micro-organism classifications.

     A  significant lag period is often required for the
     active microbial population  to  reach the  optimum
     density.  Wilson  (1981) showed  in his studies that
     a  three-week lag  period was  required for  the
     microbial  community  to  shift  in  favor  of
     nitrobenzene degrading organisms.  Table  2.8  shows
     the fate  of organic compounds  applied to  a  sandy
     soil  in  the experiment  by  Wilson  (1981).   The
     results show that the substituted benzenes degraded
     to a much larger  extent than  the  halogenated
     hydrocarbons.

     The rate  and extent of  degradation  is often
     controlled  by oxygen  (limiting factor)  as  the
     decomposition is primarily an  oxldatlve process.
     Figure 2.11 by Baehr and Corapcioglu (1984)  shows
     the influence of  different  levels  of  oxygen
     recharge on biodegradation rates.  Raymond  (1983)
     estimated that 3.5  pounds of oxygen were required
     to  degrade 1.0 pound of  gasoline hydrocarbons.
     Thus, the rate at which oxygen will diffuse through
     a  subsurface horizon will  control the rate  of
     replenishment and thus the rate of decomposition.

     If  the concentration of the contaminant is too low,
     it  may be below the  minimum level  required  for
     maintenance of the micro-organism. However, Bouwer
     (198U) points out that the simultaneous utilization

                            71

-------
            TABLE 2.7.  SUMMARY OT ORGANISM GROWTH IN VARIOUS SUBSTRATES

                   Nocardia   Paeudo-   Acineto-   Micro-     Flavo-    Unclassified
                               •onas     bacter    coccus   bacteriua

n-Alkanes             +          -         -         *          -

Cyclic alkanes    .               ...

Alkyl-aubstituted
   cyclicalkane       .....

Monoaethylakanes      +          -         -         +          •».

Diaethylalkanes       -          _         _         .          +

Trinethylalkanes      +          +         >         _          _            +

Aroaatics                        +         *         +                       +
+  Hydrocarbons utilized by •icro-organisa.
-  Hydrocarbons not utilized by •icro-organia

Unison et •!., 1965).
                                         72

-------
        TABLE 3.8.  FATE CF GRGANIC OQPOMS APPLIED TO A SANDY SOIL
      CoffiMund
Ccroentraticn   Volatilized   Cesium    Deg'aded  or
   applied                   effluent        not
                                        accounted for
mgttiter
,
Halogenated hydrsacarocns:
Chlcpofora


mCTuXJ TJU \JULlUcw ICBIB

1 ,2-Cionlopcethane

T_fc _i_n_mWl j«i^_at-V>*»^»**.
Tetracnicroecnene
1 ,1 ,2-Trionlcroethane
TWchloroethene

Substituted benzenes:

LfllUuUtauJalc


1 ,4-Oichlcrobenzene

1 ,2,4-Trichlcrocenzene

Toluene

Nitrobenzene

«*J 	 -j ^ rin,-,,-, _.» _
nmnoi liwjMi ™,i^™
Bis(2-chlcroethyl )ether


6.90
0.25
0.82
0.18
0.81
0.25
0.15
1.00
0.16
0.90
0.18

1.0U
0.18

0.80
0.13
3.40
0.57
0.90
0.20
0.92
0.16

0.86
0.16
% of material applied

5«±9t
61±10
5^14
94±16
49±9
50±7
101 ±29
27±16
47±13
58±14
88±18

27±7
54±21

ND5
ND
ND
ND
38±11
66±19
ND
ND

ND
ND

4l±4t
31±9
liQt If
UCAl 7
61±9
37±5
19±16
65±12
61±5
28± 1
21±13

33±9
26±15
(34±24)
37±4
49±10
46±11
39±3
2± 1.7
13±6
80±29
60±20 •

91±15
86±1

5±11t
8±14
-2±17
-39±18
-10±14
+13±11
-20±7
8±10
-8±7
1*»±15
-9±11

40±7
20±21
(12±13)
63± 4
51 ±10
54±32
61 ±20
60±8
21 ±15
20±25
40±31

9±31
14±20
IMeans ±95% confidence intervals.
{Material in effluent as determined by the purge and trap method.
5 Not determined.
(Wilson, 1981)
                                           73

-------
                             I n t • r «n • d I • t •
                                           High
             2.0
                 0.0  0.5  1.0  1.6  2.0   2.6   S.O
             Milligrams of Oxygen Per Cubic Centimeter
                    of Soil Recharged Per Year
Figure 2.11.  Biodegradation rate based on  oxygen  recharge
            (Baehr and Corapcioglu, 1984).
                          74

-------
of  several  different  substrates  is  possible
(secondary  utilization).   Table 2.9 indicates  that
secondary  utilization was possible for several  non-
chlorinated aromatic  hydrocarbons and chlorinated
benzenes in  an  aerobic  biofllm.  The halogenated
aliphatics were not transformed under aerobic
conditions but were nearly completely oxidized  in a
methanogenic  biofilm (Table 2.10).

With petroleum  products, it is the straight chain
paraffinic  hydrocarbons  that are most susceptible
to  biodegradation; the branched  chain paraffins and
cycloparaffins follow in terms  of susceptibility  to
biodegradation.  The aromatic hydrocarbons and the
non-hydrocarbon  compounds of high molecular  weight
would  be decomposed  at  the slowest rate (American
Petroleum Institute, 1972).  The following example
shows  the  differences in the oxidative products for
a  straight chained   paraffin hydrocarbon
(hexadecane)  and  a  cyclic hydrocarbon  (n  -
Dodecylbenzene)

    C16 H34 * 12.502—*-12(CH20) + H C02 + 5 H20
 hexadecane          bacterial cells

        + 11.5 02—»-5 (CH20) * 6 H20 +       
-------
    TABLE 2.9.  AVERAGE UTILIZATION OF SUBSTRATES  IN  AEROBIC
        ACETATE-GROWN BIOFILM COLUMN AFTER ACCLIMATION

        Substrate               Influent cone.*       Percent
                                     Wg/L             removal*
Primary
  acetate                           1.0 mg/L            99.7±0.3

Secondary
Chlorinated aromatics
  chlorobenzene                     9.5±2.5               91 ±3
  1 ,2-dichlorobenzene               9.6±2.M               97±1
  1,3-dichlorobenzene               9.8±1.8               71±8
  1 ,i»-dichlorobenzene              10.8±1.8               99±1
  1 ,2,4-trichlorobenzene            9.2±1.6               95±3

Nonchlorinated aromatics
  ethylbenzene                      9.1±2                 99±1
  styrene                           7.6±1.5              >99
  naphthalene                      13.8±3.5               99±1

Halogenated aliphatics
  chloroform                       28.5+1.2                2±20
  1,1,1-trichloroethane            15.9±3-3                5±27
  tetrachloroethylene               9.8±3.7                2±HO


*0ne  standard deviation of  the  mean values  is  given  (Bouwer,
 1984).
                                76

-------
TABLE 2.10.   AVERAGE UTILIZATION OF SUBSTRATES IN METHANOGENIC
        ACETATE-GROWN BIOFILM COLUMN AFTER ACCLIMATION	
        Substrate
                 Influent cone
                      ug/L
Percent
removal1
Primary
  acetate

Secondary
Chlorinated aromatics
  chlorobenzene
  1 ,2-dichlorobenzene
  1 ,3-dichlorobenzene
  1,H-dichlorobenzene
  1,2,1-trichlorobenzene
Nonchlorinated
  ethylbenzene
  styrene
  naphthalene
aromatics
                    100. mg/L
                     22.±5
                     15.±3
                     10.±3
                     10. ±3
                     11 .±3
                     12.0±H
                      7.9±2
                     28.8±7
  93±2
   0±15
   0±15
   0±15
   0±15
   0±15
   7±26
   8±26
  -2±29
Halogenated aliphatics
chloroform
carbon tetrachloride
1 , 2-dichloroethane
1,1,1 -trichloroethane
1 ,1 ,2,2-tetrachloroethane
tetrachloroethylene
br omodichl or om ethane
d i br omochl or om ethane
bromoform
1 ,2-dibromoethane

28. ±7
17. ±1
22. ±3
18. ±2
27. ±1
15.4*1
26. ±3
25. ±2
26. ±2
27. ±2

99±1
>99
-1±20
97±3
97±3
76±10
>99
>99
>99
>99
  One standard deviation of the mean values  is  given  (Bouwer,
  1984).
                                 77

-------
             TABLE 2.11.  BIODEGRADATION OF  T>€  COMPONENTS OF GASOLINE
Coaponenta of Sunoco 260
n-Propane
n-Butane
n-Pentane
n-Hexane
n-Heptane
n-Octane
Olefina-CA
Olerins-Cc
01efins-C6
laobutane
Cyclcpentano
Cyclohexane
Hethylcyclopentane
Methylcyclohaxane
2-Methylbutane
2-Methy Ipentane
3-Hethy Ipentane
2-Methylhexane
3-Hethy Ihexane
2-Methylheptane
3-*tethylheptane
4-Mathylheptane
2,2-0i»ethylbutane
2,3-OiMthylbutane
2 , 2-4>ia»thy Ipentane
2 ,4-Oi«ethy Ipentane
3, 3-0iawthy Ipentane
2, 3-Di»ethy Ipentane
2,5-Oi«ethy Ihexane
2,4-Di»ethy Ihexane
2,3-Diaethy Ihexane
3, 4-0 ine thy Ihexane
2, 2-0 iaethy Ihexane
2,2-Oi«ethylheptane
1 ,l-Dia»thylcyclopentane
1,2 and 1,3-Oiaethylcyclopentane
1,3 and l,4-0ia»thylcyclohexane
1 ,2-OJMthylcyclohexane
2 ,2, 3-TrlMthylbutane
2 ,2, 4-Tri«ethy Ipentane
2, 2, 3-Tri«ethy Ipentane
2,3,4-Triaethy Ipentane
2 , 3 , 3- Triwthy Ipentane
2, 2, 5-TrlJW thy Ipentane
1 ,2,4-TriMethylcyclopentane
Ethy Ipentane
E thy Icyclopentane
Ethy Icyclohex ane
Benzene
Ethylbenzene
Toluene
o-Xylene
•-Xylene
p-Xylene
Heavy ends
Initial
concen-
tration
Ulitre
Trace
0.63
0.55
1.36
0.37
0.34
0.11
1.04
0.51
0.11
0.17
0.12
0.41
0.05
3.29
1.72
1.30
0.74
0.66
0.35
0.46
0.15
0.28
0.86
0.42
0.53
0.04
0.48
0.53
0.46
0.54
0.09
0.05
0.09
0.12
0.12
0.02
0.16
0.03
3.47
0.17
1.89
1.97
0.51
0.03
0.08
0.11
0.06
0.41
1.36
2.22
1.62
3.28
1.03
8.97
Concentration at
192 hr.
U-litre
Control Saaple
Trace
0.37
0.25
0.78
0.20
0.18
0.18
0.30
0.16
0.13
0.05
0.06
0.18
0.04
1.34
0.83
0.56
0.53
0.37
0.15
0.31
0.08
0.16
0.37
0.19
0.31
0.02
0.24
0.33
0.22
0.29
0.08
0.04
0.06
0.04
0.10
Trace
0.05
0.02
2.40
0.10
1.22
1.35
0.35
Trace
0.05
0.04
0.06
0.45
1.61
2.67
2.18
4.29
1.31
11.80
Trace
0.37
0.06
0.15
Trace
Trace
0.12
0.14
0.07
0.12
0.04
Trace
0.14
Trace
1.31
0.73
0.47
0.36
0.48
Trace
0.10
Trace
0.09
0.36
0.15
0.25
Trace
0.21
0.22
• 0.20
0.19
Trace
Trace
Trace
Trace
Trace
Trace
Trace
Trace
1.95
Trace
0.97
1.02
1.02
Trace
0.04
Trace
Trace
0
0
0
0
0
0
1.13
Percent
biode graded
over and
above
control
0
0
70
46
49
54
0
16
18
0
0
45
10
75
0
6
7
23
0
38
45
48
25
0
9
11
45
0
20
0
19
84
75
62
25
78
0
26
62
13
54
13
16
23
0
0
31
95
100
100
100
100
100
100
87
(Jaaison,  et  al., 1975)
                                        78

-------
       products  that result from the  oxidation process  can
       sometimes be more soluble  and toxic than  the
       original compound.
HYDROGEOLOGIC PROPERTIES

   (1)  Ground-water  flow  (direction,  velocity, gradient):
       Once a contaminant reaches the ground water, if it
       is  soluble,  its  fate in terms of dispersal will
       then be controlled  to  a great extent  by  the
       direction  and velocity of the ground-water flow.
       If  the concentration of the contaminant is low and
       if  the  spill  is small, the contaminant is  quickly
       diluted by  the process of mixing and diffusion so
       that  a  plume  is difficult to delineate.  However,
       "for most ground water flow regimes, mass transfer
       is  predominantly diffusion controlled and  therefore
       Independent of flow rates.   This is  due  to the
       generally low flow velocities in natural ground
       water flow fields"  (Pfannkuch,  1984).  Thus,
       knowledge  of  the direction of flow would most  often
       be  the deciding factor in the initial decision to
       locate gas probes  and monitoring wells.  Andres
       (1984)  points out  that well  locations  have
       sometimes been  inaccurately located  when the
       direction of  ground water flow was predicted on the
       basis  of the  location of hydrologic boundaries and
       site topography.

       For contaminants  that reach the ground water but
       are immiscible in  water (such as many petroleum
       products), the contaminant will follow the water
       table  gradient.   If a steep gradient exists,  a
       greater interface  will be developed, which will
       lead to a greater opportunity for the dispersal of
       slowly dissolving constituents into the ground
       water.

  (2)  Water  Table Oscillations:  Changes in the depth of
       the water table can have a large impact on vertical
       transport  of  contaminants.  McKee (1972) observed a
       considerable  rise  in gasoline that had  filtered
       down to the  water  table as the water table rose
       over a  three  year  period.   Oscillations  in the
       water  table  could  allow hydrocarbons that float to
       move over  or  under  subsurface obstructions  that
       might  otherwise prevent their further migration.
       In  an  underground  gasoline tank leak  study In
       Montana, Reichmuth  (1984) noted that when  the water
       table was lower during the winter,  a gravel layer

                              79

-------
    was exposed  that possessed considerable  void space.
    During  the time of  gravel exposure,  Reichmuth
    suggested  that  gasoline  vapors were transported by
    horizontal  flow.  That correlations of soil organic
    vapors  with  ground-water concentrations  could be
    impaired  by  an oscillating water table  is  a
    possibility  that grows out of Marin and Thompson's
    (1984)  statement that  changes in  the water  table
    level contribute to a vertical gradient  that is not
    indicative  of steady state.

(3)  Lithology of  the  aquifer:   Once a contaminant
    enters  the confines  of an  aquifer, its further
    migration  will  be dictated to a great extent by the
    physical properties  of  the  sediments that make up
    the  aquifer  (Figure 2.12)  (American Petroleum
    Institute,  1972).  Barriers to flow (retardation of
    flow) can  occur if  lateral changes  take  place in
    either texture (unconsolidated sediments) or  rock
    formations (consolidated sediments).   In
    sedimentary  rocks,  Osgood  (1974)  states  that the
    orientation  of  the  rock and  the  primary
    depositlonal characteristics  of  the unit are as
    Important  as permeability and porosity in  dictating
    flow.   The  depositional characteristics would
    include such features  as cementation and packing.
    In more steeply dipping rock  units,  Osgood  (1974)
    suggests that the  dominant flow  direction  of the
    hydrocarbons would  be parallel  to the strike,
    downslope,  and that deviations  from  the  strike
    direction  would be  controlled  by  Jointing  and
    fracturing.   Non-uniform  characteristics  of the
    aquifer sediments would,  of  course,  cause
    non-uniform  advancement  of the contaminant  plume
    relative to the  ground surface  as  water  flow  would
    be channeled through zones of lower resistance.  If
    these sediments  also  comprised a  large  portion of
    the  vadose  zone,  then organic vapors  would  also
    follow the  path  of least resistance and  would  move
    through the  fissured and fractured  rock  according
    to the paths  that were  dictated.   Lateral flow of
    vapors  could be tremendous under  such  conditions
    and could thus negate any  hopes for correlating the
    vapor  concentrations  with the ground-water
    concentrations.  Schwille  (1984)  indicates  that
    with fissured  rocks, the gas  tracer method  would
    only be  useful if the fissured rock were  covered by
    a layer  of  porous loose  rock.
                           80

-------
oo
                                                                 GROUND-WATER FLOW
                        OLDER BEDROCK
             Figure  2.12.    Hypothetical  ground-water  system  (American Petroleum  Institute,  1972).

-------
CHARACTERISTICS OF THE SPILL

Greater  knowledge of the  history of the  spill can  often
provide the Investigator with greater insight  as  to the proper
location for  soil gas  probes  (first approximation).   Such
characteristics as the total volume lost,  the length of time
the product was spilled  (continuous  vs. one-time  spill), area
of the spill, and  the age of  the spill,  can  be very helpful in
better understanding the possible extent  of unsaturated and
saturated zone contamination.

MISCELLANEOUS

   (1) Rainfall:   Depending  on the  frequency and the
      amount of  rainfall that  occurs,   organic
      contaminants in the  unsaturated zone will be
      susceptible to  leaching.   In areas  of  high
      rainfall,  oscillations  in the ground water may
      occur which would bring  contaminants closer to the
      soil surface.  Obviously, any  input of water  will
      lead to decreased air-filled  porosities  and to
      reduced vapor movement.  Vertical concentration
      gradients  will be altered as vapors reaching the
      rainfall saturated zone will either concentrate,
      move laterally, or will be resolubilized to some
      extent.  Vapor migration (upward, lateral) and
      resolubilization can lead to a wider spread of the
      contamination area.   If the area in question is
      covered with vegetation,  then the amount of
      leaching will be  dependent not only on the amount
      of rainfall but also on  the evapotranspiration, the
      amount of water  in  storage,  and  the effective
      rooting depth.  Rainfall  will also  delay field
      measurements and make it extremely difficult to
      compare soil-gas concentrating  before and after  a
      rainfall event.

  (2) Background water  quality:  The more contaminated
      the ground water, the more difficult  it is to
      delineate the spatial extent of the particular
      contaminant  in question.  In some cases, several
      plumes  may exist, that are partially or completely
      over-lapping  and  that  represent different spills
      over a  period of  time.   Greater  instrument
      sensitivity would be required In those cases where
      the  background  contained  numerous  organic
      contaminants  at concentrations that were  orders of
      magnitude higher  than  the contaminant being
      monitored   (see  section  on  analytical
      methodologies).
                             82

-------
(3)   Barometric pressure and  wind:   Early work  by
     Buckingham (1901) showed that changes  in  barometric
     pressure  had  little influence on  soil gas transport
     in most cases,  with its greatest influence  on  the
     gases in the soil pores at or near the soil surface.
     In a study conducted by Reichmuth (1981),  gasoline
     vapors  detected in a basement down  gradient from an
     underground storage  tank  leak,  worsened during
     periods of  high  wind  and low barometric pressure.
     He concluded that such conditions were optimal  for
     maximum  earth out gassing  .  Other factors  that
     would  maximize  this  gas  exchange  would be  the
     absence of  vegetation (resistance to  wind flow) and
     the  presence  of  coarse permeable  soil.   However,
     one  would  have  to conclude that in  almost  all
     cases, -if soil  gas  probes were located several  feet
     below  the soil  surface,  the vertical  vapor
     concentrations  measured would not be  influenced to
     any extent by  this  surface phenomena.

(i»)   Proximity to  rivers, lakes, and pumping wells:   The
     presence of  rivers  and  lakes  would  mean that
     contaminants reaching  the aquifer  would  be
     intercepted and dispersed even further (see  Figure
     2.13).   Such  Interception of contaminant flow would
     mean an alteration  in the subsurface  boundaries of
     the  contaminant  plume.  Another consideration is
     the  proximity of  soil-gas probes  and  monitoring
     wells  to pumping wells  as shown in Figure  2.13
     (American Petroleum Institute, 1972).  Altering  the
     ground-water  table by creating a  cone of depression
     would cause immiscible organic compounds  (floaters)
     to  move laterally and deeper relative to the  soil
     surface.  This condition would not only  alter  the
     movement of  a plume  but also  the  distance  over
     which a vertical  soil  gas gradient would  have to be
     established.
                            83

-------
                              .V.V.V.V.V.V ORIGINAL  WATER TABLE
                                       WATER MOVEMENT
Figure 2.13.
Diagram showing how oil  on a water table  can be
trapped  in a  cone  of depression created  by
draw-down  of  a pumping well (American Petroleum
Institute,  1972).
                                84

-------
                         REFERENCES
1.  American Petroleum Institute.  The migration of petroleum
   products in  soil and  ground  water.  Pub.  No.  4149.
   Washington,  D.C.,  1972.

2.  Anderson, D.  C.  and S. C. Jones.  Fate of organic liquids
   spilled on soil.   National Conference on hazardous waste
   and environmental  energencies.  Houston, Texas, 1984.

3.  Andres, K.  G.  and  R.  Canace.   Use of the  electrical
   resistivity  technique to delineate  a  hydrocarbon  spill  in
   the coastal  plain deposits  of  New Jersey.   In Petroleum
   Hydrocarbons and Organic Chemicals in Ground Water.
   National Water Well Assoc., 1984.

*.  Baehr,  A. and M.Y. Corapcioglu.  A  predictive model for
   pollution from  gasoline  in  soils and ground water.   In
   Petroleum Hydrocarbons and Organic Chemicals  in Ground
   Water.  National  Water Well Assoc.,  1984.

5.  Bouwer,  E.  J.   Biotransformation of organic micro-
   pollutants in the subsurface.  In Petroleum Hydrocarbons
   and Organic Chemicals in  Ground Water, National Water Well
   Assoc., 1984.

6.  Buckingham,  E.   Contributions to our knowledge of the
   aeration of  soils.  U.S. Department Agri., Bureau of  Soil
   Bulletin 25.,  1904.

7.  Byer,  H. G,  W.  Blankenship and R.  Allen.   Ground  water
   contamination by chlorinated  hydrocarbons:   causes  and
   prevention.  Civil  Engineering - ASCE.,  1981 March, pp.
   54-55.

8.  Chiou, C.  T.   Soil  sorption of organic vapors  and effects
   of humidity on sorptive mechanism and  capacity.   Environ.
   Sci.  Technol., in press,  1985.

9.  Cullimore,  R.    Bugs in  the  wells, microbes  in  ground
   water.  Canadian Water Well,  1982.   Vol.  8, No.  2.  pp 24-
    25.
                             85

-------
10.   Hlllel, D
     Processes.
     288 p.
15
16
17
                  Soil and Water,  Physical Principles and
                Academic  Press.  New York and London, 1971.
11.  Holzer,  T. L.  Application of ground water flow theory to
     a subsurface oil spill.   National Water Well Assoc.,  1976.
12.  Jamison
                V. W.,  R.  L.  Raymond  and J.  0.  Hudson.
     Biodegradation of high-octane  gasoline.   In Proceedings of
     the Third International  Biodegradation Symposium.   Applied
     Science Publishers LTD,  London, 1975.
13.  Kreamer
                       In  situ Measurement of Gas Diffusion
     Characteristics in Unsaturated Porous Media by Means of
     Tracer Experiments.   Ph.D.  Dissertation.   University of
     Arizona,  1982.

 14.  Letey, J.  and L. H.  Stolzy.   Measurement of  oxygen
     diffusion rates with  the  platninum Microelectrode,  I.
     Theory and Equipment.   Hilgardia.  1964.  Vol.  35, No. 2.
     pp 545-554.
     Mackay,  D.  and W.  Y
     organic compounds.   J.
     10,  No. 4.
                          .  Shiu.  Henry's
                          Phys. Chem.  Ref.
law constants for
Data,  1981.   Vol.
     Mackay,  D. M., P. V. Roberts and J.  A. Cherry.  Transport
     of Organic contaminants in ground water.  Environ. Sci .
     Technol.,  1985.  Vol.  19, No. 5.  pp 384-392.

     Marrin,  D. L., and G. M.  Thompson.   Remote detection of
     volatile organic contaminants in ground water via  shallow
     soil-gas sampling.  In Petroleum Hydrocarbons and Organic
     Chemicals in Ground Water, National  Water Well Assoc.,
     1984.
18.  McKee,  J.  E.,
    Ground water.
    293-302.
                   F.  B.  Laverty and R. M.  Hertel.  Gasoline
                    Journal WPCF,  1972.  Vol. 44,  No. 2.
                                                          in
                                                          PP
19.   New York  State Dept.  of Environmental Conservation.
     Technology for the  storage of hazardous liquids,  a state
     of  the art review.   Albany, New York,  1983.
20.   Nielson,  K.  K.  and V. C. Rogers.  A mathematical
     radon  diffusion  in earthen materials.   U.S.
     Regulatory Commission, 1982.  NUREG/CR-2765.
                                                   model for
                                                    Nuclear
21.   Osgood, J. 0.  Hydrocarbon dispersion in ground water
     significance  and  characteristics.  Ground Water,  1971
     Vol. 12, No.  6.  pp  427-438.
                              86

-------
22.  Pfannkuch, H.  Determination of the contaminant source
    strength from  mass exchange  processes at the petroleum  -
    ground water interface  in  shallow aquifer systems.  In
    Petroleum  Hydrocarbons and Organic  Chemicals  in Ground
    Water. National  Water Well Assoc., 1984.

23.  Raymond,  R.  L.   Personal communication as  quoted for A.
    Baehr in A predictive model for pollution from gasoline in
    soils  and  ground water, 1984.   In Petroleum Hydrocarbons
    and Organic  Chemicals in Ground Water.  National Water
    Well Assoc.,  1983.

24.  Reichmuth,  D.  R.    Subsurface gasoline  migration
    perpendicular to  ground-water gradients -  a case study.
    In Petroleum  Hydrocarbons and Organic Chemicals in Ground
    Water, National  Water Well Assoc., 1984.

25.  Reid,  G.  W.,  G. Thompson and  C. Oberholtzer.  Soil  vapor
    monitoring as  a  cost effective method  of  assessing ground
    water  degradation from volatile chlorinated hydrocarbons
    in  an  alluvial   environment.    Second  Annual
    Canadian/American Conference  on Hydrogeology - Hazardous
    Waste  in Ground  Water:  A Soluble Dilemma. National  Water
    Well Assoc.,  1985.

26.  Swallow,  J.  A. and  P.  M.  Oschwend.   Volatilization of
    organic compounds from  unconfined aquifers.  Proc. of  the
    3rd National  Symposium on Aquifer  Restoration and Ground
    Water  monitoring. National Water Well Assoc., 1983.

27.  Schwille,  F.  Ground-water  pollution by mineral  oil
    products.  Ground Water Pollution Symposium, 1971.   AISH
    Publ .  No. 103. 1975.

28.  Schwille, F.   Migration of organic fluids  immiscible with
    water  in  the  unsaturated zone.  From B. Yaron,  G.  Dagan
    and  S. Goldshimd (eds.) Pollutants  in Porous Media:  The
    Unsaturated Zone between Soil Surface and Groundwater.
    Springer-Verlag, 1984.

29.  Schwille, F.    Petroleum contamination of  the subsoil -  a
    hydrological  problem.  In  P.  Hepple  (ed.)  The Joint
    problems of the  oil  and  water  industries.   Proc.
    Symposium,  The Institute  of  Petroleum,  Brighton,
    January  18-20, 1967,  pp 23~54.

30.  Williams, D. E. and D. G. Wilder.  Gasoline  pollution of  a
     ground-water reservoir  - a  case history,  1971.  Ground
    water, Vol. 9, No. 6.   pp 50-54.


                              87

-------
31.   Wilson, J. T. ,  C.  G.  Enfield, W. J. Dunlap,  R.  L. Cosby,
     D.  A. Foster  and  L.  B.  Baskin.  Transport and  fate of
     selected organic  pollutants  in a sandy soil.   J.  Environ.
     Qual..  1981.   Vol. 10, Mo.  1.  pp 501-506.
                               38

-------
                         CHAPTER  3
     TRANSPORT  AND RETENTION OF DISSOLVED AND IMMISCIBLE
          ORGANIC CHEMICALS IN SOIL AND GROUND-WATER
INTRODUCTION


   A large  variety and  quantity  of different  organic
chemicals,  some  of  which pose a health  hazard, are accidentally
or  deliberately applied to soil  where they may migrate to
ground-water.  For  example, many  different pesticides, mostly
applied  during  agricultural operations,  have appeared  in the
ground-waters  in a  number  of  different states (Pratt, et al. ,
1985).   In addition,  large numbers of  dissolved  organic
compounds  are accidentally  released  into soil from  leaking
waste disposal  sites  or storage tanks.  These compounds migrate
downward with flowing water and can enter and  contaminate
underground water supplies.  A  second large class of organic
liquids found  in soil  are  petroleum products  which are released
to  soil  by accidental leaks or  large spills and which  may be
largely  immiscible in water.   These spills may  occur from
underground  storage tanks which have either  corroded, ruptured,
or  have  faulty  connections.  Similarly, petroleum  products
might enter  the soil when a tanker truck releases its contents
in  a highway accident.

   In a complex soil,  air, water  and hydrocarbon system, an
organic chemical compound,  depending on its properties and on
the soil  conditions, may  be  found in a number of different
phases: as an immiscible  liquid, as a dissolved component of
the soil-water solution,  or  as a  gas.   In addition, the
immiscible liquid may  be flowing or Immobilized,  and the
dissolved  components may be moving freely within soil solution
or  may be  absorbed to soil  mineral surfaces or to stationary
organic matter in  the soil.   This chapter will provide an
overview of the processes  important in the  transport  and fate
of organic  contaminants in  soil,  focusing separately on
(1) petroleum mixtures which  are  transported or  retained  in the
soil  as  a  fluid largely  immiscible in water and on (2) soluble
organic  contaminants which dissolve readily  in water and
predominantly move by  convection  within flowing  solution.
                              89

-------
Movement of Liquid Oil Through Soil

    When a large quantity  of spilled oil is introduced at the
soil surface,  it will Infiltrate under the  influence of gravity
principally as an  immiscible  fluid separate  from water.  The
exact path and rate of the infiltration as well as the extent
of  lateral  movement will  depend in a complex manner on the
permeability of the soil to  water and oil,  on the water and oil
content, and on  the presence  of structural voids which
contribute substantially to  spreading of the spill.  Although
attempts are being made  to  formulate  the oil and water
transport problem mathematically (Baehr and Corapcioglu, 1981),
quantitative  values  for  the exact water and oil flow  paths and
flow  rates  in natural  soils  are for practical  purposes
unpredictable.  Nonetheless,  experimental observations, small
column  experiments,  and  model  calculations  which  use
simplifying assumptions  have produced a general picture of the
oil entry and transport process which  describes  the  main
features of a spill event.

    Qualitatively, as oil  enters the vadose zone it  displaces
air  but not water from the  pore spaces,  and infiltrates
vertically under the  influence of gravity at a rate limited by
the permeability of  the  oil-filled pore  space.  As  the  oil
passes  through a given region  of the porous medium,  it leaves
behind  a  residual  and  largely  immobilized  concentration of
insoluble oil which varies  between approximately 5 and 20
percent  of the void space  depending on the  type of oil and the
characteristics of the soil (Dietz, 1970).  Thus, if  the total
quantity of oil spilled into the soil is less  than the amount
required  to fill the  residual  pore space in  the vadose zone,
then the body of the spill will not reach the ground-water and
will remain in a pendular  volume poised  above the water table
(see  Figure 3.1A).   If,  however, there  is an  excess  of
insoluble oil, then  part  of  the flowing oil will reach the
ground-water where most of the remaining  oil will spread into a
thin  film  occupying  the  volume  Just  over   the
saturated-unsaturated  zone interface (see Figure 3.1B).

Stabilized Oil Spill Profile

    After the  transient  infiltration phase has concluded, the
insoluble oil body will  be spread over  a  relatively fixed
volume  of soil which  may or may  not extend to ground-water.
Those components  of the  oil  body which  are soluble  will
continually dissolve  into  the  soil solution,  and subsequently
may migrate with  flowing  water.   In addition,  volatile
components of the oil  body  exposed to soil air interfaces will
evaporate into the soil  air,  and subsequently may  migrate
upward  and laterally  by  vapor  diffusion.  The principles used
                              90

-------
                   \  \  \
        GROUND  SURFACE
                              OIL PHASE
                               OIL  COMPONENTS
                                DISSOLVED IN  WATER
                                             CAPILLARY FRINGE
           (A)
  Figure 3.1.  Oil migration pattern (Case No. 1)  (Schwille,
              1984).
                                       GROUND SURFACE
UNSATURATED  ZONE
OIL PHASE  (OIL  BODY)
                                      VISUAL LINE OF SATURATION

      OIL  COMPONENTS  DISSOLVED IN WATER
                                              WATER TABLE'
      SATURATED ZONE

           (B)
   Figure 3.1.  Oil  migration pattern  (Case No. 2) (Schwille,
              1984).
                                91

-------
in quantifying  these  phenomena  are  discussed  in  the section
describing transport processes.

Movement  of Dissolved Organic Chemicals Through Soil

    Unlike oil  products,  many  organic contaminants readily
dissolve in water  and  do not exist  as separate  immiscible
liquids  in  soil.   These compounds move downward through
unsaturated soil in flowing soil  solution, where their movement
is attenuated  to  varying degrees  by adsorption-desorpbtion
reactions with  stationary organic matter and  soil mineral
surfaces.   Since the  adsorption processes  are largely
reversible, these' dissolved compounds  do not immobilize in soil
and will move  as  long as water  is  flowing.   Thus, unless a
particular organic  compound is  completely degraded by soil
microorganisms or chemical reactions,  it  will  only reside
temporarily in a vadose  zone which receives a  net annual input
of water and will eventually  migrate to  ground-water.  A
quantitative description of the adsorption  and transport
processes  for  dissolved chemicals are given  in the next
section.

PROCESSES GOVERNING TRANSPORT OF  ORGANIC CHEMICALS THROUGH SOIL

Transport of Liquid Oil

    As an oil  spill enters the soil,  it may spread  laterally as
it infiltrates downward.  The extent of this lateral  spread is
highly dependent  of  the heterogeneity, permeability  and
moisture status  of the soil-water profile  and cannot be
predicted with  any  confidence for  any soil  conditions.
Virtually the only  quantitative  information  about lateral
spreading has been  obtained from small scale  laboratory oil
Infiltration experiments, usually in  Hele-Shaw  cells with
transparent  walls  (Dietz, 1970; Schwille,  1984).  In the
absence of any quantitative information,  several  qualitative
comments may be  made  about the shape of  the spill volume
relative  to the soil texture of the vadose  zone.  For the moat
part,  infiltration of  oil in homogeneous soils  of a single
texture will in  the  absence of any large  structural voids or
barriers  to  vertical movement be relatively uniform,  and the
lateral extent  of  the  oil plume  will be smaller  in a coarse
textured soil  (Figure 3.2A) than  for a fine-textured soil
(Figure 3.2B).   The  spreading of  the infiltrating  plume will
continue  with time and will produce  a cone-shaped vadose  zone
profile until the plume  reaches ground-water.   If  the soil is
heterogeneous or layered, substantial lateral  flow  can occur at
the boundary of  a  region  of lower permeability where the oil
will build up and  flow  laterally (Figure 3-2C).   In addition,
highly irregular plumes of oil may flow through fissures and
cracks in bedrock  or very impermeable soil  and may produce a

                             92

-------
                       LAND SURFACE
     A - HIGHLY PERMEABLE.  HOMOGENOUS SOIL

     B - LESS  PERMEABLE, HOMOGENOUS SOIL

     C - STRATIFIED SOIL WITH VARYING PERMEABILITY
Figure 3.2.  Generalized  shapes of spreading cooes at immobile
             saturation (American Petroleum Institute,  1972).
                                93

-------
complexly  shaped volume of spilled material in the vadose zone
(Figure 3.3).

    Since  spatial resolution of the hydrocarbon spill  volume is
generally too  imprecise in actual field studies or cleanups to
allow  any  generalizations to  be made about the extent of
spreading,  specific Idealized spill shapes  are  usually assumed
in estimating the potential for ground-water contamination.  A
major assumption made  in constructing this idealized picture is
the  concept  that a  fixed immobilized residual oil volume
fraction,  which depends only  on the type  of oil material,
remains  in  the  soil  after  a  spill has  infiltrated  and
stabilized.  Table 3.It adapted  from information given in Dietz
(1970),  gives prototype values  for the residual Immobilized oil
void fraction  SQ (fraction of total void space) remaining after
an oil  spill  has passed through a volume of soil.  In  addition,
residual void fractions of 0.05  or  less for  coarse  textured
soils  have been reported by  Pfannkuch  (1983) and Schwille
(1984).  It should be cautioned  that these  values  are  only
rough  estimates and that actual  values may differ in soils of
different texture (Schwille, 1984).  Part of the reason for the

     TABLE  3.1. RESIDUAL OIL VOID FRACTION So (ADAPTED  FROM
                         DIETZ. 1970)
Type of Oil
Light oil (gasoline)
Medium oil (diesel, light fuel)
Heavy oil (lube, heavy fuel)
Residual Void
0.10
0.15
0.20
Fraction

large  variation in reported values for  the  residual  oil void
fraction is that the drainage of  oil  under  gravity following
wetting to a high degree of saturation  is  a dynamic process
which is very rapid during the early  stages and which undergoes
a slow  but continual decrease with time for many days  after the
initial rapid drainage (Figure  3.1).   Thus,  the residual oil
content in soil after 5 days may be  considerably higher than it
is  at   100  days.  For  this reason,  the residual oil  void
fraction,  SQ , should be regarded as  an  index similar to the
water content "field capacity" which  is very  poorly defined In
fine-textured soils (Hlllel, 1971).  Nevertheless,  the  residual
oil content is a useful parameter for making rough calculations
of  the  spatial volume occupied by  a  spill.  For example, from
an estimate of the  residual void  fraction  percent,  S0, and the
total  porosity, $, one calculates  that  a spill of volume, V0,
will eventually occupy a soil volume  V$ where

                VS  - V0/*S0                               M )
                               94

-------
                  q ~  1.26 f/d
                                              mln
                                       8  cm
                                        43
                                            11 mln
                                            43
                                            = 8  cm
Figure 3.3.
Infiltration of  kerosene into a
porous medium  through a  narrow
                                       1.25 L/D.
                                       is 8 cm.
fissure  at a  rate
Capillary fringe  height
Top:     beginning  stage
Below:    after infiltration  finished.
(Schwille,  1984)
                        95

-------
 O   40 —



>•   SO —

*O
a.
CD
o
c
CD
**
CO
oc
20 —



10 —



 O •
           I

          SO
 I

• 0
180
                       Time (days)
Figure 3.4.  Oil  retention  capacity  aa a function of  time
            (Schwille, 1975).
                           96

-------
   By making  specific assumptions  about the shape of the oil
spill in the vadose zone,  the amount of oil  required  to  reach
the ground-water may be calculated.

   EXAMPLE

       A gasoline truck spills 10,000 gallons  (37.85 m3) Of
   gasoline  (S0 -..10) on  the  soil surface.   The gasoline
   infiltrates over  a surface area of  9 m2.   Calculate the
   volume  of  oil  which will  reach  the  ground-water  table
   located  at a  depth of 30 meters  assuming  no lateral
   movement in the vadose zone.  Assume that  the  soil  porosity
   * is 0.4.

   Solution

       The total  volume of  residual oil  in the  soil  after
   stabilization may be calculated by using Eq. (1).  Thus

       VS - (37.85 m3)/(o.H)(.io) - 946.25 n»3

       The total  volume  Vv occupied by  the  oil  in  the 30 m
   thick vadose  zone, assuming  no  spreading  (i.e.,  a
   cylindrical plume), is equal  to

                   Vv - AL                           .     (2)

   where  A is the surface spill area (9 m2)  and  L  is the  depth
   to ground-water (30 m).   Thus, in this case Vv -  270  m3 and
   this  volume is filled by a quantity Vv
-------
       GROUND-WATER CONTAMINATED BY  SOLUBLE COMPONENTS
       FLUID  OIL FLOATING ON WATER  TABLE
       RESIDUAL SATURATION
Figure 3.5.  Subsurface  redistribution of a surface  spill
            (American Petroleum Institute, 1972).
                                98

-------
       TABLE 3.2  OIL LENS THICKNESS ABOVE GROUND-WATER
                     (AFTER DIETZ.  1970}

Type of Sand

Extremely coarse-very coarse
Very coarse-moderately coarse
Moderately coarse-moderately fine
Moderately fine-very fine
Grain
Diameter
(mm)
.5-2
.2-. 5
.05-. 2
.015-. 05
Zone
Thickness
(cm)
2-9
9-22
22-28
28-45
water  mound may be roughly calculated  from the volume of  oil,
VQ, which reaches the  ground-water using  equation (3).

               AG - VG/$TS0                              (3)

   Estimates of the film  thickness vary considerably among
different researchers, in  part  because film  thickness,  like
residual  oil  saturation,  is  a  dynamic  quantity which
decreases slowly over time (Figure 3.6).   Values  given by
Schwille (1967), largely obtained  from model experiments, are
generally of the order of  1  cm  or  less.   Furthermore, the
effective  thickness of  the  oil lens  may be  altered by
ground-water table fluctuations  occurring during  the  time of
lateral  redistribution.   This  vertical motion can  spread a
layer  of  immobilized oil over  a much greater vertical thickness
than  if  the redistribution occurs over a motionless water
table.

   EXAMPLE  (continued from  above)

       The 27.05 m3 of  oil which enters  ground-water is
   assumed  to form a  symmetric circular  film of thickness T -
   0.01  m.  Thus, using  equation  (3)  the area of the film is
   AQ -  67625 m2.  In this  case,  the film will form  a  circle
   of radius 147 m.

        In  practice,  this estimate of lens  thickness may be too
    small if ground-water level  fluctuations are  frequent in
    the  area common to the  spill boundary.  For example, if the
    effective thickness of the film is increased to 0.1  meter,
    the  AQ  is  reduced to  6763 m2 and the  radius  of  the spill
    over  ground-water  reduces  to H6.5 m.

Transport of Dissolved Chemicals Through  Soil

    The  most important  processes governing transport of
dissolved  organic chemicals  through soil  are mass  flow or
convection of  chemicals with flowing  soil solution and
hydrodynamic dispersion, the spreading of chemicals in  soil by
movement around  solid  obstacles.  Many  dissolved  organic

                              99

-------
                                                180
                           Time (days)
Figure  3.6.  Relation between  thickness of oil  layer and
            spreading time (Schville, 1975).
                             100

-------
chemicals  do  not move freely  in solution but are  attenuated to
varying  degrees by reversible adsorption to stationary  soil
organic-matter and,  to  an  extent, to clay mineral surfaces.

   An  important  index for  describing adsorption  is  the
distribution coefficient  Kd (cm3g~1) wnich  is defined as the
ratio  of adsorbed concentration Ca (vg g~1  soil)  to dissolved
concentration  Cj, (jig cm~3  solution) at equilibrium,  or

                                                         (A)
Another  index
per  unit soil
  which is defined as the distribution coefficient
  organic carbon  fraction f
called the  organic
(em3g~1), or
                            OC
         carbon  distribution
(see Appendix)  is
coeff1ci ent,  Koc
                                                        (5)

Koc na3 been shown to  vary less between soils  than  Kd for a
given chemical (Hamaker  and Thompson, 1972).  Thus,  it
represents the adsorption  potential of a given compound.  Large
compendia of Koc values  for pesticides and  other organic
chemicals are available  in different references  (Kenaga, 1980;
Rao and Davidson, '1980;  Jury,  et  al., 1984).
   As  a further  attempt  to standardize  the adsorption
potential of a given organic chemical, measurements  have been
•ade  of the adsorption  of  compounds to octanol  (Lambert,  1968).
The octanol-water partition coefficient, Kou,  has  been measured
or calculated for a large number of organic  chemicals  (Rao and
Davidson, 1980).   Furthermore,  various  regression coefficients
have  been developed  between Koc and Kow, including the relation
       log Koc
                ow
                  - 0.18
             (6)
(used by Rao and Davidson,  1980)  for 13 pesticides  (r2  - 0.91).
Attempts have also been made to  calculate Koc  or  Kow from more
basic chemical properties  or from chemical structure  (Briggs,
1969).  For example, Kenaga  (1980)  used  the  following
regression with water solubility,  Cj, (ug cm~3).
       log Koc - 3.64 - 0.55  log  Cj,
                                             (7)
to obtain Koo values  for  a variety of organic  compounds.  He
stated that
magnitude.
the equation  was accurate within  1 .2 orders  of
    In  the  Appendix it is  shown  that in a flowing  solution, the
average effective velocity,  Vg, of  a dissolved organic chemical
which undergoes adsorption  is:
                               101

-------
        VE - Jw/(PbKd + «) - Vw/(pbKd/© + 1)              (8)

Where Jw (cm D~1)  is water flux

      Pb (8 cm"3)  is soil bulk density

      9 (cm~3) is  volumetric water content

and   Vw - Jw/0 is pore water velocity

    If a concentrated  pulse or  front  of chemical is suddenly
applied to the soil, not all of the molecules  will  move  at  the
same  velocity Vg because of dispersion.   However, Vg will
describe the average velocity of  the pulse or equivalently,  the
velocity of the center of mass of the pulse.

    EXAMPLE

        Three compounds, chloride (Koc - 0), benzene (Koc -  83)
    and n-octane (Koc « 6800) are introduced into an aquifer of
    porosity  4  -  0.5,  water flux Jw - 1 m d"1,  bulk density p&
    •  1.5  (g  cm~3), and  organic carbon fraction foc - .005.
    Calculate  the average  travel time of these compounds  to a
    well L  .-  1000  m downstream.

    SOLUTION

        When  Eq.  (5)   is used,  the Kg values of the three
    compounds  (Chloride, benezene, n-octane) are (Kd - focKoc)
    0, .415,  34 (cm3 g'1).   From Eq. (8), this gives velocities
    VE of  2.0,  0.89, 0.019 (m d~1 ) ,  respectively.  The  travel
    time,  t,  to  move  a distance,  L, through  the aquifer is
    simply  t - L/VE, or t -  500,  1124, and  5.26  x  10* days  for
    chloride,  benezene,  and  n-octane to reach the well.

The average travel times do not represent the earliest arrival
times of the  chemical  pulse or  front,  which  could  be much
shorter than  the average  time.  Prediction of  the earliest
times  requires a quantitative understanding  of  the soil
geometry variations which  is  usually not possible  to obtain in
the field.

    Also  important in  characterizing dissolved  chemical
transport  is the degradation rate of the  compound  which  in  the
absence of detailed   information about specific  reaction
mechanisms  is  commonly  represented by the half-life, TVa defined
as  the  time  at which the  mass of the compound  drops to 50
percent of its initial level while decreasing exponentially
with  time. For example, if a compound has a  travel time to an
observation  well  equal to  twice its half-life,  the  mass  at
arrival  should be degraded  to V*  of the initial mass at the  time

                              102

-------
of  injection.  Table  3.3t  adapted chiefly from Jury, et  al.
(1984),  gives values  of Koc  and T j / 2  for various  organic
chemicals together with  a  calculation  of the travel time  to
reach 1000 m for  the  conditions given in  the  example above.   In
cases where the  documented  half-life  is considerably smaller
than the travel time, as for  example with methyl parathion,  it
is  unlikely that the compound will persist  long enough to  reach
the well.  However,  it  should be stressed that  the ground-water
conditions may differ considerably from the conditions  under
which the compound half-life  was estimated.

   The  half-life and organic carbon  partition coefficient
represent so-called chemical  benchmark  properties for organic
compounds.  These single indices roughly describe the tendency
to  degrade and adsorb in  soil systems.   They mask much of  the
complexity of these processes  and, for that  reason, should be
regarded as lumped parameters.   Nevertheless, the values  of
these coefficients do provide  valuable information about  the
possible behavior of  the  compound in the  environment.   For this
reason,  the benchmark properties are  useful tools to use  in
screening large numbers of  compounds and in placing them into
smaller  numbers  of groups which behave  similarly (Jury, et al.,
1984).    After  such  a  screening process,  experimental
observations of  specific chemical behavior may be used to make
assessments of the expected  behavior of  other  compounds in  the
same  behavior group for  which  no direct experimental evidence
is  available.

Transient Movement of Dissolved Chemicals

   From the  above  information, it is possible to make some
general  comments  about  a  spill of chemical which is  completely
dissolved in water.   The  mass velocity of the  chemical,
retarded by adsorption  compared to the velocity of  the water,
is  roughly given  by  Eq.  (8).  The extent of lateral  and
vertical spread in the  unsaturated zone, as in the  case  of an
oil spill, is very dependent  on specific soil conditions  and
cannot  be predicted in  detail  with any certainty.   General
shapes  such as   those shown in Figures 3-1-2 for  oil spills
represent  plausible  profile shapes  during  vertical
infiltration.

   However,  unlike  the largely  immiscible oil volume,  the
dissolved  chemical mass does not  immobilize in  soil,  and has no
stabilized profile as  long as water  is  flowing through  the
system.   However, it may  move very  slowly  downward  because of
                               103

-------
  TABLE 3.3.  Koc AND T1/2  VALUES FOR VARIOUS MISCIBLE  ORGANIC
 CHEMICALS, ALONG WITH AN ESTIMATE OF THE TRAVEL TIME REQUIRED
  TO MIGRATE L - 1000 m THROUGH GROUND-WATER USING Eq.  8  WITH
        Jw-1md~1, 4> - 0.5,  Pb - 1*5 gcm~3, foc . 0.005

CHEMICAL

ATRAZINE
BENZENE
BROMACIL
CARBON TETRACHLORIDE
CHLORIDE
DDT
DIELDRIN
EPTC
EDB
LINDANE
METHYL PARATHION
MONURON
NAPROPAMIDE
NAPTHALENE
NITROBENZENE
N-OCTANE
PARATHION .
PHENANITRENE
PHENOL
PHORATE
PROMETRYN
SIMAZINE
TCE
1.1.1 -TRICHLOROETHANE
TRIALLATE
TRIFLURATIN
VINYL CHLORIDE
* no value available
A Jury, et al . . 1984
B Josephson, 1983

K«r»
tcrn^g ')
160
83
72
110
0
2.4E5
12000
280
44
1300
5100
180
300
1300
71
6800
11000
23000
27
660
610
140
150
113
3600
7300
400




TI/P
(d)
71
*
350
«
«•
3800
868
30
*
266
15
166
70



1


82
60
75
*
*
100
132
»



TRAVEL
TIME
(yr)
4.6
3.1
2.8
3.6
1.4
4900
250
7.1
2.3
28
106
5.1
7.5
28
3.6
141
227
474
1.9
14.9
13.9
4.2
4.5
3.7
75
152
9.6




REFERENCE

A
A
A
A

A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
A
A
A



adsorption  reactions or may cease to move entirely  if  the water
input to the soil  is stopped for a prolonged period  of  time and
the  chemical  is  below  the  first few meters  where  it might
migrate upward  with water moving towards a dry soil  surface.

Oil Migration After Stabilization

    When the liquid  oil  phase is  stabilized in the  soil, the
immiscible  oil body  shares  a large  interface  with  the
surrounding liquid water.  Thus, any water percolating  through
an oil spill either in  the saturated or unsaturated soil-water
                               104

-------
zones will  pick up dissolved  components  from the oil-water
interface  and  will carry them  downstream  at  the end  of  the
spill.   Fried,  et al. (1979).  analyzing experimental  results
and  applying theoretical calculations, concluded  that water
percolating  through a body  of immobilized  spilled oil will
reach saturation levels with respect  to  the  dissolved
components  after a short period of  time or equivalently  after a
short travel  distance,  of  the order of several tens  of
centimeters.   Thus,  in a region where  water is flowing,  the oil
apill acts  as  a distributed  source of  dissolved organic
chemicals  as long as the immiscible material remains in place.
Once present as a dissolved  component of  water,  the organic
compound is transported by  convection and dispersion in the
manner described in the previous section.

    To obtain  rough  estimates  of the release of chemical from
the  stabilized spill  into  ground-water,  one may neglect
hydrodynamic  dispersion and  write the  mass flux,  J3,  of
dissolved chemical as the  product  of  the  water flux, Jw, and
the  saturated concentration of organic  C& (g m~3)

       Ja - JwCt                                         (9)

This equation together with knowledge  of ground-water flow
rates and oil component solubilities may  be  used to roughly
estimate the flux of dissolved  material from the residual spill
into the ground-water.

    EXAMPLE

       For  Illustration,  the  previous example is used  where a
    gasoline spill spread  into a thin 1 cm  thick film over the
    ground-water and covered a radius of  147  m.   When Figure
    3.7 adapted from  Somers  (1974)  is  used, the solubility of
    the  gasoline mixture  is roughly  estimated as Cs  - 10(mg
    l~M - 10 (g m"~3).  Assume that the  aquifer has a  ground-
    water  velocity of  V -  2  (m  d~1) and a water-filled porosity
    of 
-------
               800
               400
               300
               200
               100
                                         • — n-ALKANES
                                         — AROUATIC8
                                         — OLEFINS
                                         — • CYCLO-AUCANES
                  C1
                           4     6     8    10    12   14    16
                                       NUMBER OF C-ATOMS
                             PETROL
                                    KEROSENE
                                         .QASOIL/DIE3ELFUEL
                                         HEATINGOIL
Figure 3.1,   Solubility  of hydrocarbons  in water (Somers,
             1974).
                               106

-------
because of  dispersion.  Fried, et al. (1979),  in  order to
produce  a simple conceptual model  of the  release  of
chemical from  the spill, defined an equivalent dissolved
chemical layer  thickness, D, which  contained the same total
mass of chemical  as the actual profile but which had all of
the chemical  at the saturated solubility concentration (see
Figure  3.8).

    In  analyzing  this  problem by  using a two-dimensional
form of the  dissolved chemical transport equation (A.11 in
the Appendix), Fried,  et al. (1979), calculated  that the
equivalent  thickness, D, in ground-water below the oil lens
containing  a  saturated  concentration of the dissolved
chemical was on the order of 1 m for the  spill geometry
given  in the previous  example.   In the example discussed
above, the  gasoline spill formed a circular  pancake of
radius R in  contact with ground-water.  Thus, by using the
ideal model  of  Fried, et al. (1979),  we calculate that the
front  of dissolved chemical flowing in the ground-water as
it leaves  the spill region  will  be approximately 1 m deep
and  2  R  wide  (plus a  small  amount  of additional  lateral
width from  dispersion).  By using Eq.  (8),  we can estimate
the mass flux  of dissolved chemical  as Js  -  JwCe -  10 (g
m~2d~1).   The cross-sectional area  of flow is A - 2RD - 291
m2  since  R -  147  m.   Thus, the spill  discharges the
dissolved  chemical into the ground-water at a rate Q - JSA
- 2910 (g  d~1) or Q -  1073 (kg yr~1).  Since the total mass
of the  spill in ground-water (assuming  a density of 900 kg
m~3) was about 27.05 m3 x 900 -  25000  kg,  this discharge
rate would  trasnsfer about 1/23 of  the  spill to the ground-
water  in one year.  This would dissolve the spill  in  about
23 years  if  the  entire spill volume were equally soluble
and if  the  lateral extent of the spill  remained unchanged.

    Obviously, the extent of dissolution into ground-water
is much greater when  the  layer of  oil  at the ground-water
interface is  spread thinly  over  the  surface than if the
immobilized layer is  thicker and has  a  smaller contact area
with  ground-water.   In  the second case from an earlier
example,  the spill thickness  was increased to  10  cm which
decreased  the spill radius to 16.5  m.   If  all calculations
above  are  repeated, the thickness, D,  of the  dissolved
organic zone at the exit boundary calculated by the method
of  Fried,  et al . (1979), drops to  about  0.70  m,  and the
dissolved  chemical mass flux decreases  to 238 kg/yr.  At
this rate of dissolution,  it  would require over 100 years
to  remove  the residual spill If the  dissolution process and
area extent  remained  constant during the lifetime of the
oil event In the soil.
                           107

-------
    Chemical Concentration
 Q.
CO
 o
CD
 0.
 o
Q
                                                      EQUIVALENT
                                                      THICKNESS
                                                           D
                   ACTUAL DISTRIBUTION
Figure 3.8.
Comparison  of actual and idealired concentration
depth profiles below a waste  spill in ground water.
The  equivalent  thickness, D,  is defined so that
the rectangle has the  same area between  the axes
as the curve.   (i.e., the same mass of chemical).
ct  is the chemical solubility in water.
                               108

-------
Oil Dissolution Within  the Vadose Zone

   Water  percolating  downward  through the portion of the
tadose zone volume contaminated by the residual portion of the
oil  spill  will  also  dissolve  into solution, creating a new
source of dissolved oil material  to contaminate ground-water.
The  extent of mass flow from  this source will depend on the
spill shape, the  percolation  rate, and the extent of lateral
•igratlon  of the dissolved material.   This mass flow rate  is
too uncertain to be estimated  quantitatively unless the spill
shape is  known.  As  a  rough  estimate, the mass flux may  be
calculated  as  the product  of  water  percolation  rate,
solubility,  and the spill cross-sectional  area normal to  flow
in the  vadose zone.   In the case of  the previous  example,
assuming 1  (m yr~1) vertical drainage through the spill and A «
9 m2  area,  the  calculated mass flux  is  JwC3A -  .09   (kg
yr~M which is very small compared to the release rate  into
the ground-water.  This  is because the  interfacial area between
undissolved oil and water is  very much larger in the latter
case  and because the  drainage  flux is much smaller than the
ground-water flux.

Sinking of Heavy Insoluble Components into Ground-Water

   Organic  liquids which are immiscible In water  and  denser
than water could sink through  the saturated ground-water  zone
if not  bound in  a residual state (Anderson and Jones, 1984;
Mackay,  et al., 1985).  There has been some laboratory evidence
of this phenomenon  (Schwille,  1975),  and it  probably
contributes to the vertical transport of material within the
saturated  zone.  These "sinkers" as they are called could
eventually reach the  bottom of  the aquifer if they  are not
immobilized en route  (Mackay, et al., 1985).

MOVEMENT OF HYDROCARBON VAPOR THROUGH SOIL

   The  volatile components of  the  oil  material  will release
chemical to the vapor phase which will in turn migrate toward
the  soil surface by diffusion  (lateral and vertical) or  will
sink through the soil air if the partial pressure of the vapor
is sufficiently high  and If the vapor  is denser than air.  The
rate of migration will  be a function of the soil resistance  to
vapor flow, of the amount which is redlssolved into the  liquid
phase, and of the amount which  is adsorbed or degraded.  The
mathematical description of the transport process, described  in
detail in the Appendix, combines a mass balance (Eq. A.1)  with
the  flux equation for vapor  movement called Pick's Law  (Eq.
A.H).

   Equation (A.I) describing vapor flux assumes that  vapor  is
transported only  by  diffusion, and  this is a reasonable

                              109

-------
approximation  if  the air phase is  stagnant and if  the chemical
vapor is a dilute  component of  the  soil  atmosphere.   However,
some mass  flow of vapor could occur, particularly  if a large
quantity of chemical vapor which is  denser than air  is evolving
in the  unsaturated zone.  In  this  case,  the  overlying  dense
vapor could sink as well as diffuse  through  the  air  phase and
would collect at the ground-water  interface (Schwille,  1984).

    Although chemical  transport  occurs  in the vapor  phase when
mass flow  of dissolved  chemical  compounds  is negligible,
chemical  vapor molecules still  interact  with the liquid and
adsorbed phases by redissolving  when  the  molecules  come into
contact with water which is low in dissolved concentration.
The relationship which describes  the equilibrium partitioning
between the vapor and liquid  concentrations is  called Henry's
Law:

        Cv - KHCt                                         (10)

where KH  (cm3  solution/cm^ air) is the dimensionless form of
Henry's constant (Jury, et al . , 1983).   Since  Henry's Law has
been shown to  be  valid all the way to  saturation  for a number
of organic chemicals  (Spencer and  Cliath, 1970), it  is commonly
calculated as  the ratio of saturated vapor  density, Cy, to
water  solubility,  cj.   Values  of Cc,  c{, and KH
for a number of organic chemicals  are given in Table 3.4.

    The Henry's Law  constant is  also expressed occasionally as
a ratio of vapor pressure to dissolved concentration, as, for
example, in

        PV - *HC                                          (11)
where  P«  (Pa or  J  m**3) la vapor pressure, and kjj has  the units
of (Pa-nj3g~1). The conversion factor  between kjj  and KH in Eq.
(10), obtained by using the ideal gas  law is:



          -KHJT                                        (12)

where  R  - 8.3 (J  mole"1 •K"1 ) is the  universal gas  constant, T
is the absolute temperature, and M (g  mole'1) is  the molecular
weight of the compound.  An extensive compendium of  values of
KH for * variety of organic  chemicals are  given  in Mackay, et
al.  (1982).

     As shown in the Appendix,  the effective soil  vapor
diffusion coefficient for  an organic chemical  which is also
present  in the adsorbed and dissolved phases is  reduced, often
significantly, compared to  the  diffusion  coefficient of a gas

                               110

-------
which  is  insoluble.  Gas dissolution and subsequent adsorption
has  the  effect of greatly  slowing down the transport of the
chemical  in  the vapor phase  and also of exposing the gas
•olecules to degradation processes which may only be occurring
in solution.   However,  the  Henry's Law partition model Eq.  (10)
implies  that  the vapor  concentration is proportional to the
dissolved concentration  (and, as shown in the Appendix,  to the
total concentration).  Thus, to the extent that the equilibrium
relations are valid in  soil, the vapor phase concentration may
be used  to quantitatively  monitor a  volatile waste spill.
Furthermore,  even if the relationships are only approximately
valid because of rate-limited nonequillbriurn between phases,
the  vapor phase profile will  still give qualitative information
useful in detecting  a  spill  and mapping its spatial extent.

Diffusion Travel Times

    For  any spill  with volatile components, a vapor phase  will
evolve  above the  dissolved phase  as it migrates through
ground-water.  The  maximum  vapor concentration will be given  by
Eq.  (10) where Cj  is the  concentration  of dissolved organic
aaterial  at the ground-water  interface  with the vadose zone.
This  vapor  will  move upward through.the vadose zone  by
diffusion  (reduced by  dissolution and adsorption), by  possible
•icrobial degradation, and by  chemical transformations.

    A  qualitative measure  of diffusion for a  given chemical,
called the characteristic  diffusion  time, tp,  is the time
required for an organic chemical with  an  effective diffusion
coefficient,  Dg, to diffuse  through a  distance  L  (Jury,  et
al.,1983):

|        tD  - L2/DE                                       (13)

where Dg  is given  by Eq.  (A.17) of the  Appendix.   Table 3.5
summarizes  values of to  to  diffuse L -  1m,  calculated  by using
the  procedures in the  Appendix.

    The  values  in Table  3.5  are qualitative but  do allow
compounds  to be grouped into mobile and  relatively  Immobile
categories of vapor  diffusion  potential.  Thus, for example,
ethylene dichloride  has a  relatively high saturated vapor
density  (Table 3.1) but only  a  modest vapor mobility because  so
much of  its total mass partitions into  dissolved  and  adsorbed
phases.   Conversely,  n-octane,  with even  lower vapor
density.moves much faster because of its  low solubility.

Steady-State Diffusion Profiles

    For  those compounds with reasonably  short  diffusion times,
the vapor  concentration profile  with depth should reach a

                               111

-------
    TABLE 3.4.  SATURATED VAPOR DENSITY,  WATER SOLUBILITY AND HENRY'S
        CONSTANT Kg FOR VARIOUS VOLATILE  AND SEMIVOLATILE ORGANIC
                                   CHEMICALS
Chemical

Benzene
Biphenyl
Carbon Tetrachloride
Chlorobenzene
Chloroform
Ch 1 orome thane
DDT
Dieldrin
EPTC
EDB
Ethylene
Ethylene Dichloride
Lindane
Methyl Bromide
Napthalene
Nitrobenzene
N-Octane
Phenol
Trifluralin
TCE
Toluene
1,1, 1-Trich loroe thane
Vinyl Chloride
Saturated
Vap.
Density
(gm~>)
400
0.49
750
71
960
1.2E4
6.0E-6
1 .OE-4
0.22
120
4.7E4
320
l.OE-3
2.0E4
1.6
1.8
94
0.57
2.0E-3
440
150
1.4E3
8.7E3
Water
Solubility
(gm-J)
1.8E3
7.5
8.0E2
4.7E2
8.0E3
5.4E3
3.0E-3
1.5E-1
3.7E2
3.4E3
1.3E2
8.0E3
7.5E2
1.3E4
3.2E1
1.8E3
6.6E-1
8.2E4
0.3
1.0E3
5.2E2
9.5E2
9.0E1
Henry ' s
Constant
(KH aio*
2200
600
9400
1500
1200
22000
20
6.1
5.9
350
3600000
400
1.3
15000
500
1000
1400000
16
67
4400
3000
15000
970000
Reference*

A
A
A
A
A
B
A
A
A
A
C
C
A
A
A
A
A
A
A
C
B
C
A
*A - Jury, et al., (1984)
 B • Mackay, et al., (1982)
 C - Thomas (1982)
                                     112

-------
characteristic final value  under  certain circumstances  (e.g.,
stationary source of vapor,  time  independent  biological and
chemical reactions, reasonably  constant moisture status over
time).   A hypothetical but  plausible case  of  interest to
examine  is the steady-state  distribution of gas concentration
above a source of saturated vapor at the ground-water interface.
It  will  also be assumed that the  chemical  undergoes a first
order decay process  characterized by a  decay constant y.

    For  this  case,  as  shown in the Appendix,  the steady-state
gas concentration profile as a function of depth is given  by:


                  Slnh
                 *  Sinh (qL)
where
                                                          (15)

Dv  is  the  soil  gas diffusion coefficient, and L is the  depth to
ground-water.  Profiles of concentration  for  various values of
Q  - qL are given In Figure 3.9.  Those  curves with  large Q
represent compounds whose diffusion  time  through the  soil is
comparable to,  or larger than, the  half-life of the chemical.
Hence, the vapor concentrations  drop to low values in the soil.

Vapor Monitoring as a Detection  Method

    From  the  preceding  discussion,  several  chemical
characteristics may  be Identified which  indicate whether a
contaminant plume will be accompanied  by a  measurable vapor
concentration.   First, the chemical must  have  a non-negligible
vapor  pressure  and density as  part of the total concentration.
-Thus, compounds  with very low values of Henry's constant KH or
compounds  which adsorb strongly  (large Kg) will be unlikely to
have a large vapor  density in soil.  Second,  the compound must
be  sufficiently mobile in the vapor phase  to allow vapor to
migrate significantly  beyond the  spill boundaries.   The
diffusion  travel times,  given in  Table 3.5,  are useful in
determining whether this criteria will be met.  Ultimately, the
same conditions which limit vapor density (small KH, large Koc)
will cause  large diffusion travel times.
    Finally, the compound must be persistent  enough to travel
beyond  the  spill boundaries without degrading into a form which
is  not  detectable.  Depending on  the  relation between the
diffusion  travel  time and the compound half  life, the vapor
profile will drop  off gradually or sharply  between the spill
and the soil surface, as in Figure 3.9.
                               114

-------
N

£
**
a
o
•o


o
CO
                                                                         Soli «urfao<
                    2         .4          .6          .8


                     Vapor concentration C/Co
i.o
                                                                         Ground-water
          Figure 3.9.  Steady  state vapor concentration profiles between
                     groundwater and the soil surface,  for  a  compound
                     undergoing first order degradation.  Oimensionless

                     parameter Q"qL ia given by eq 15.

-------
    of experimental  evidence.  J. Envir.  Qual., Vol. 13, No.  1.
    1981.

12.  Karickhoff,  S.  W.  Semi-empirical  estimation of sorption
    of  hydrophobic pollutants on  natural  sediments.
    Chemisphere 10:833-816.   1981.

13.  Kenaga,  E.  E.   Predicted bioconcentration factors and soil
    sorption coefficients  of pesticides and other  chemicals.
    Ecotoxicol.  Environ.  Saf. 1:26-38.   1980.

14.  Lambert, S. M.   Omega, a  useful index of  soil  sorption
    equilibria.  J.  Ag.  Food Chem. 16:310-313.  1968.

15.  Mackay,  D.  and  W. Y. Shiu.   A critical review of Henry's
    law constants for chemicals of environmental  interest.   J.
    Phys.  Chem. Ref. Data  10:1175-1199.   1982.

16.  Mackay,  D.  M.,  P. V. Roberts and J. A. Cherry.   Transport
    of organic contaminants  in  groundwater.  Env.  Sci.  Tech.
    19:381-392.  1985.

17.  Pfannkuch,  H.  0.   Hydrocarbon spills  -  retention  in
    subsurface and propogation into shallow aquifers.   1983.

18.  Pratt,  P.  F.   and  26 coauthors.   Agriculture and
    groundwater quality  CAST task force  report 103.   CAST,
    Ames,  Iowa, 1985.

19.  Rao,  P.  S.  C. and J.  M. Davidson.  Estimation of  pesticide
    retention  and  transformation  parameters required  in
    nonpoint source pollution models.   In M.  R.  Overcash and
    J.  M.  Davidson  (ed.) Environmental impact of  nonpoint
    source  pollution.  Ann Arbor Science  Publishers,  Inc. Ann
    Arbor, Michigan, 1980, pp 23~67.

20.  Schwille,  F.   Petroleum contamination of  the sub soil - a
    hydrological  problem.  In  P.  Hepple (ed.) The  joint
    problems  of  the  oil  and  water industries.   Proc.
    Symposium,  The  Institute  of  Petroleum,  Brighton,
    January  18-20, 1967.  PP 23-51.

21.  Schwille,  F.   Groundwater pollution  by mineral  oil
    products.  Ground Water Pollution Symposium, 1971.   AISH
    Publ. No. 103,  1975.

22.  Schwille,  F.   Migration of  organic fluids  immiscible with
    water in the unsaturated zone.  From  B.  Yaron,  G.  Dagan
    and S.  Goldshimd  (eds.) Pollutants in Porous Media:  The
    Unsaturated Zone between Soil Surface and  Groundwater,
    Springer-Verlag, 1981.  pp 27-18.

                              117

-------
23.  Somers,  J.  A.  The  fate of  spilled oil  In the  soil.
     Hydrologleal Sciences  Bulletin.  19:4:501-521.  1974.

24.  Spencer,  W.  F. and M. M. Cliath.   Desorption of lindane
     from soil as related to vapor density.  Soil Scl.  Soc.  Am.
     Proc. 34:574-578.   1970.

25.  Thomas,  R.  G.   Volatilization from water.  In W. J. Lyman,
     et  al.,  (ed.)  Handbook of chemical property estimation
     methods-environmental  behavior of  organic compounds.
     McGraw-Hill  Book Company, New York, 1982.  pp. 15-1-15-34.
                              118

-------
                           APPENDIX

      MATHEMATICAL THEORY OF DISSOLVED ORGANIC CHEMICAL
                    TRANSPORT THROUGH SOIL

   When  a chemical which is present as a dissolved constituent
of soil solution is also  adsorbed to solid soil material  and
has a non-negligible vapor pressure, the total concentration C?
(yg cm~3)  of the  chemical  in  units of mass  per volume  of  soil
aay be written as

          CT - Pbca * ecl  * aCv
                                                        (A.1)
where

   Ca  (yg g~^)  is  the mass adsorbed per mass of soil
   G£  (yg cm~3)  is the mass  dissolved per volume of solution
   Cv  (yg cm~"3)  is the mass  in  vapor per volume of soil air
   pb  (g  cm"3)  is  soil dry bulk density
   9 (cm3 cm~3)  is soil  volumetric water content
   a (cm3 cm~3)  is soil  volumetric air content

Conservation of  Mass

   The equation which represents conservation of mass for the
chemical,  called a  continuity equation, may  be written  as  (for
one-dimensional  flow)
                            *  r  -  0
                                                         (A.2)
(Jury,  et al.,  1983)

where J3£ (ug  cm~2 d~1)  is the flux  of  dissolved chemical
      J3V (ug  cm"2 d"1)  is the flux  of  chemical vapor

and r  (yg  cm~3 d'1)  is  a general  reaction  term representing the
net rate of transformation of  chemical  to another form.

Flux Equations

    The  one-dimens i onal flux of  dissolved chemical  through
porous media is customarily written
                               119

-------
          Jsi " •
                                                        (A. 3)

where  Jw (cm d-1) is the  volume of solution  and DI  (cm2d-1) is
a combined diffusion-dispersion term representing the spreading
of  chemical  by molecular collisions within solution and by
moving around soil solid  obstacles.  In the field,  dispersion
is usually more important than diffusion.

    The  one-dimensional flux of chemical  vapor  through soil,
called Pick's Law,  is  usually written as
                                                       (A.U)

where Dy  (cm2d-1)  is  the soil gaseous diffusion coefficient.  A
commonly used model  for Dy is the Millington-Quirk model
                                                       (A. 5)

where D^ir  (cm2d~M is the  gaseous  diffusion coefficient of
the chemical in  free  air, and * is soil porosity.  Jury,  et al .
(1983).  concluded  that  the  value  Dalr - M300  (cm2d~1)
satisfactorily described  intermediate  molecular  weight
compounds.

Phase Relations

    It is common  to use simple equilibrium models to describe
relationships  among  C y , C £ ,  and Ca  in a  three  phase
soil-water-air system.   The  simplest model  to  describe
adsorption is the  linear model.

          Ca - KdCt
                                                       (A. 6)
where K
-------
    Because Kd is soil-specific  and because organic chemicals
predominantly adsorb to  soil organic matter,  a modified
distribution coefficient  per  unit organic  carbon  fraction is
also used to describe the  chemical  adsorption affinity

         KOC - Kd/fOC
                                                      (A.8)

where  KQC  (cm3g~1)  js  an  organic  carbon  distribution
coefficient  and foe  *s tne soil organic carbon  fraction.  In
cases where only  the  organic  matter content TOM is known, one
may convert approximately to  organic carbon fraction  foe  by
using the equation

         fQC - fOM/1-73
                                                      (A.9)

Values  of  KOC vary less  than Kd  between  soils  for  a  given
chemical  (Hamaker  and Thompson,  1972).   Thus, KQQ  *s  a
preferable benchmark  property  to use  to represent the
adsorption  potential of a  given  compound.

    The  equilibrium relationship between Cy and Cj_ is called
Henry's  Law

         CV - KHCj,
                                                     (A.10)

where KH ( dimensionless )  is called Henry's constant.   Since
this linear  relationship commonly  persists to saturation,  K^ is
usually calculated  as the ratio  of  saturated vapor  density and
water solubility.

Partition Coefficients

    It  is  useful  to express direct  relationships  between the
total concentration and the concentration in  each phase.   This
is accomplished  by  combining  the concentration  relation  (A.1)
with the equilibrium relations  (A.6)  and  (A.10).   Thus,  for the
dissolved phase

         CT " PbKdc£ * QCi * aKHci

            - (pbKd + 6 +  aKH)  C£  s  RjCfc
                                                     (A.11)

where

         Ei m PbKd + 9 +  aKH
                                                     (A.12)


                              121

-------
 la called the liquid partition coefficient  (Jury, et al., 1983).
 In practice, the third term  aKH may be neglected in most cases.
 In addition, for strongly  adsorbed  chemicals (large Kd), only
 the first term Pt>Kd ls nonnegllglble.   For  the vapor phase,
               PbKdcV/KH  * KHCV * aCV
 (PbKd/KH  *  e/KH
                                      RVCV
                                                      (A. 13)
 where
- PbKd/!CH
                               a
                                                      (A.1H)
 is  the vapor partition  coefficient.   For strongly adsorbing
 chemicals, only the  first term PbKd/KH  ia nonnegligible.

    The general  transport equations above may  be combined by
 plugging  the flux equation expressions  (A.3)-(A.M)  into  the
 continuity  equation (A.2) and by  expressing all concentrations
 in  terms  of the total concentration CT  when the  partition
 coefficient  definitions (A. 12)  and  (A.14)  are used.  This
 results in the equation (assuming uniform soil properties)
                 r  -
                                   z
where
          VE -
                                                      (A.15)
                                                      (A.16)
is the effective  chemical connective  velocity, and
                                                      (A.17)

is the  effective diffusion-dispersion  coefficient.  For
volatile organic  chemicals which have  a  high vapor density, the
second  term  in  e.g.  A.17 dominates  the  first if the soil air
content is  high and  if the  water  carrying the  dissolved
chemical is  not  moving rapidly through  the soil.  Thus, in  this
case,
             Sat
             RV
                                                      (A.18)
                               122

-------
Conversely, If the  chemical has  a  low vapor density or if the
air phase is negligible (i.e.. ground water flow), then  the
first term dominates the second term  and
       DE =
                                                      (A. 19)

Degradation Rates

   The combined processes  of biological  and  chemical
degradation of organic chemicals are extremely complex, and  can
depend on a variety  of factors  such  as  temperature, organic
aatter  content, water content, and microbial population density.
Thus,  the  specific form  of  the reaction  term r  is  often
difficult to identify in a given  situation.   For this reason,
simple idealized forms  are  often  used to  give approximate
estimates.  The most common form is the first order degradation
nodel
       r -
                                                      (A. 20)

where y (d~M is  a  first order  degradation  rate coefficient.
It is related to  the effective half life  T-| /2  (d)  of the
compound by the equation

      T1/2 - 0.693/w
                                                      (A. 21)

Steady State Profiles

    If  the  compound is  present  at saturation level  in ground
water and diffuses upward while  undergoing first order decay,
the profile will  eventually reach a  steady  state value  whose
shape is described by the steady state  form of Eq . (A. 15)  with
Vg -  0 and r - iiC-j-, or
                                                      (A.22)

with  C(L) - C0 and C(0) - 0.

   The solution to this equation may  be written as


      r, x   r  sinh(qZ)
      Hz; • <-o sinh(qL)
                                                      (A.23)
                               123

-------
where



       q  -
                                                         (A.2H)
                                124

-------
                         CHAPTER
                  MEASUREMENT METHODOLOGIES
     The  following sections discuss  sampling  and  analytical
methodologies for  monitoring  volatile organics'in the
subsurface.   The  sections  are: Sampling Methods, Sampling
Design, Quality Assurance, and Analytical Methods.

SAMPLING METHODS

     This  section  presents various sampling methodologies used
to monitor subsurface  contamination.  These methods include:
headspace measurements,  ground  probes,  flux chamber
measurements, and sampling  with sorbents (usually  passive
sampling).

     The  techniques identified  are capable of  providing a yes
or no answer to whether subsurface hydrocarbon  contamination  is
present.   However,  the techniques  do not all  provide  an
equivalent measurement.   The ground probe  and  headspace
measurement techniques  measure a soil gas concentration, the
flux chamber technique  measures an  emission  rate,  and the
passive sampling technique measures some function of an average
soil gas concentration.  The  technique(s) selected  for  a
particular application  will  be dependent on  the objectives  of
that study.

     All  the  techniques can be  divided into  two steps, sample
collection and analysis.   Analytical instrumentation for
hydrocarbon  analysis  is  commercially available  but  is
relatively complex  and expensive.   The sample  collection
equipment discussed is  not  generally commercially available,
but is  usually simple to construct and operate, however  some  of
the equipment described is protected by patent.

     Various  techniques  have  been  successfully used for
ground-water contamination investigations at a  variety of sites.
However,  the techniques discussed  below are not standard
methods and have not yet been  adequately evaluated.   Therefore,
best results will  be  obtained  when the techniques are used  by
experienced investigators  who  are familiar with the methods
used and the local  geology  and hydrology. All the techniques
are dependent on the movement  of  volatilized organic  species  up


                               125

-------
through the  overlying soil.  Any limitations of this  transport
will limit  the  utility of these  techniques.  The suitability of
each  technique for various  types  of site  conditions is
discussed and compared in this section.

Headspace Measurements

      This  section discusses the  determination, of hydrocarbon
concentrations by analyzing the headspace  gas from  samples
collected  in a  dry well or from  soil cores.

Headspace Measurements in Dry Wells—
      Sampling  the headspace in  existing  subsurface structures
is  a  simple technique  that  can yield valuable preliminary
information. The technique involves collecting grab samples or
using a portable hydrocarbon analyzer to measure the  headspace
concentration  in monitoring  wells,  storm sewers,  utility
vaults, or other subsurface structures.  The results obtained
provide information regarding the composition and extent of any
contaminant plume and assist in developing  an optimal sampling
strategy for subsequent investigative work.

      Recommended  use--It is recommended  that headspace
measurements be made  at existing subsurface  structures as the
first phase of any subsurface contamination  investigation.  The
technique is -quick,  simple, and economical.  Furthermore, it
can save  substantial amounts  of time  and  money by providing
input data  for  selection of an appropriate sampling strategy.

      Technique  applications--Headspace sampling is  typically
employed as part of any Remedial Invest!gation/Feas1bi1ity
Study (RIFS).  One example is given  below.

      A tanker truck spill  caused  5,500 gallons of jet fuel to
contaminate an  area of high ground-water.  A  preliminary  study
installed  ground-water monitoring  wells  along two perpendicular
lines.  A subsequent  study was  under  taken to develop  sampling
methods  and  to define the  contaminant   plume   (Radian
Corporation,  1984).  The first stage  of  sampling  involved
removing the well caps and collecting  headspace samples  in gas
syringes  for  on-site  gas chromatograph/flame  lonization
detector  (GC/FID) analysis.  Additional measurements  were made
by  using a  portable  total  hydrocarbon analyzer.  The results of
the headspace  analyses indicated  that  the  plume had  increased
in  area  since  the  initial study.   Therefore,  the gridded
sampling area  was expanded accordingly, prior  to an intensive.
follow-up  investigation.

      Llmltatlons--The limitations  associated with this
sampling technique include:


                               126

-------
   o   interferences  (e.g., methane in sewers);

   o   subsurface  structures or wells are not always present
       at investigation  sites  (or not optimally located);

   o   volatile  hydrocarbon species  can diffuse out  of
       unsealed subsurface  structures; and

   o   negative test  results are  inconclusive,  i.e.,  the
       absence  of  hydrocarbons in the headspace of  a
       subsurface  structure  does not guarantee that
       hydrocarbons are  not present in the surrounding soil.

Headspace Measurements of Soil  Cores—
     The headspace gas  or extracted solids of a soil  core  can
be analyzed to determine hydrocarbon concentrations.  To obtain
a sample, a technique  known  as grab sampling can  be  used.   An
"undisturbed" soil core is collected by using an auger or by
driving a tube into the  ground and is then sealed  in  a  sample
container.  Using  this  technique, liquid as well as  gaseous
hydrocarbon contamination can be  detected directly.   Two
approaches can be  taken.  First, the sample container can be
half filled with soil.  Hydrocarbons  can  then volatilize  into
the vacant headspace.  Care should be taken to ensure headspace
and soil without providing any headspace.  Soil  gas  is  then
extracted directly from  soil pores.

     Recommended use—This method of measuring hydrocarbons is
recommended when  the sampling crew has a modest  level of
technical expertise or  when sophisticated sampling  equipment is
either not available or  not  cost-effective.  The  method works
best when  sampling  sandy soils  containing little organic
matter.

i    The  technique of  grab-sampling  of soil cores  is  typically
both simple and  quick  to  perform.   Minimally,  the  method
requires only  one  person,  one hand  auger, and sample  storage
containers.  Analyses  can  be performed  off site at a later
date.

     Technique applications — Grab sampling of soil  cores can
be accomplished using a variety of  equipment  as  illustrated by
the recent review from the  U.S.  EPA  (198»»).  The  review is
summarized  in Table U. 1 .   A  number  of  researchers  have applied
this technique  to  detecting  hydrocarbon soil  gases  at  various
subsurface levels.   In  most  cases,  shallow soil  gasses  were
collected to assess  very deep sources of vapors.   Variety is
evident  in  sampling depth,  collection equipment,  storage,  and
analytical  methods.
                               127

-------
TAILI 44.  CtlTElIA 101  SELECTING BOIL SAMPLING EQUIPMENT
   ObUlM
     C«r«
Matt
  Cora TypM
   U
ttMy toil*
Matt
•oil NoUur*
ACCM* C*
•••Tiki
  HIM
  DacU|    NUtiv*
r«or loll    •••*I•
           flit
                                                                   Ufc«r
TIM «f fMfUr To Me
At •*•< A«|«r
1. Iet«¥-Ty>« A«|«t« 1
I. lirt«l AM|«M
•. roit-dolt Anger X
b. Dutch Anftff I
c. ••••!•( k*ct«l An|«r I
4. l«*4 Aw|«r« I
to •. Mu4 An|tc« I
). Tub«-TyM •••yl«ct
•. l«il rr*k««
(1) V«t Tlpi I
(2) Ofj T1M >
b. V*(U«ytc THkM I
c. Tkta-W«U«4 T»k« ••••Ur* I
4. r««t-«««pi«i i
1. fovir Au|«r
1. M«o4-MiU Betm lit* fo««r A«|«r I
2. Truck Houat*4 Auf*r
• . ierrw Typ« I
b. Drl»t liapUr I
). Tripod HouatH DrUt liapUr X
CUy l««4 I»t«r
I
I
1
I
I
I

I
I
X
1
X
X
X
X
X
X
*•* *(•*
1
I
X
X
X
X

X
X

X
X
X

X
X

tot Bry Iittr
X
s
X
X
X
X

X
X
X
X
X
X

X
X
X
Vtt la
X
X

X
X
X

X
X

X
X
X

X
X
X
ta I*
X
X

X
X
X


X
X
X
X
X

X
X
X
|«||« J/Hor*
X
X

X
X
X


X X
X
X
X
X

X
X
X

-------
     For example,  Horvitz (1951) reported using a  hand auger
or  drilling equipment to search for oil  and gas.   Samples were
collected  at 8  to 12  foot  depths and  brought to  the
surface.  The samples were  stored in glass jars or  cans,  and
were later analyzed by using a  vacuum analytical-combustion
technique.  In another study, Horvitz  (1951*) used a piston-type
coring  device to  collect shallow soil  samples during offshore
oil  and gas prospecting activities.   For  this study, samples
were stored  in plastic bags (with  a  reported  shelf-life of
several months), and then  extracted  samples were analyzed by
using gas  chromatography  (GC).   Horvitz noted that onshore
samples need to be  collected at a depth  greater than  6 feet to
ensure  a quality  sample whereas shallower depths were
sufficient for offshore sample collection.

     Devine  and Sears (1977) collected  and analyzed  over 1,000
cores in a search  for  oil and  gas  deposits in Australia.
Samples  were collected  by mechanically drilling  to  a 9 to
10-foot depth and then by Inserting a coring device.  Samples
were stored  in  heat-sealed  polyethylene bags.   Sample
preparation for analysis by GC  included  acid  leaching  and
cryogenically trapping the hydrocarbons.

     Smith and Ellis  (1963)  also collected  soil cores (0-40
foot depth) in cans soldered to  seal  in  the  sample.  For GC
analysis, the headspace gas was  removed  from the cans through a
hole in the top by using a glass syringe.   They  conducted
studies  concerning the  effect  of  refrigerated versus non-
refrigerated sample storage.  Their results  indicated that
hydrocarbon  concentrations decreased in samples  stored at
68-95°F  versus refrigerated  samples  stored  at 32°F  or
pasteurized samples (exposed to  185°F for  short periods).  They
also concluded that the organic  content  in soil samples  may
interfere with hydrocarbon detection  from underlying  oil  and
gas  deposits.  However, Horvitz (1972)  refuted  this conclusion
when he reported  that grass and  roots contribute negligible
amounts of saturated hydrocarbons  to  the soil  atmosphere.   In
addition, Horvitz  disagreed  with  the need for  sample
refrigeration by  stating that  minimal headspace  above  the
sample was the key  to a long shelf-life.

     In  other  oil  and  gas exploration efforts, Pogorski  and
Quirt (1981) collected soil samples at a 2  foot depth by using
hand or  power augers.   Instead of using  plastic bags  for
storage, they used specially designed aluminum cans.   The
samples  were sealed  in  the air-tight  cans and later were
analyzed by GC.  Similar methodologies  have  been used by
several oil and gas  exploration companies (Eklund, 1985).

     Limitations — The  primary  limitation  of this technique is
that it  is better suited for measuring adsorbed organics  rather

                              129

-------
than  free organics  in the interstitial pore spaces.  Hanisch
and McDevitt  (1984) reported that any  headspaee present in  the
sample container  will lead to  desorption of organics from  the
soil  particles.   Unless the  soil type, headspaee volume,
temperature, sample handling  techniques, and storage time  are
held  constant, relative concentration levels between soil
samples are not comparable.

       Another limitation to this technique is the possible loss
of volatile hydrocarbons when  the  sample  is  removed from  the
ground or transferred for analysis.   Sample exposure to  the
atmosphere has  been  successfully  avoided  by  capping the soil
core   tubes.    Be.dnas  and Russell (1967) capped tubes with
sealing wax and reported a shelf-life  of  at  least 12 months.
Their work involved detecting natural gas leaks by trenching.to
the desired sampling depth and by driving  tubes (26 in x  2  in)
into  the trench walls.  A carrier gas was used to flush soil
gas from  the  samples to  a GC analyzer.   Hanisch and McDevitt
(1984) reported a technique used  at several hazardous  waste
sites.  The core sampler used  (see Figure 4.1) consists of  a
brass core  sleeve which  is  pressed into the soil  to  a
sufficient depth  to fill the  sampler  but not so  deep as  to
compress the sample.  The method works  best for clays  and  silts
of medium moisture content.  After excess  soil is removed,  the
sleeve  is  sealed with  a Teflon-lined cap.* The samples  are
stored at room  temperature.  Headspaee  (i.e., pore space)  gas
collected by a  syringe through a port  are analyzed by GC.

       In  addition  to loss of volatile hydrocarbons, degradation
of organic compounds may also  occur because of  time delay
between  collection and analysis.  This  collection method is  not
appropriate for rocky soils nor  is  it well  suited for loose
sandy soils  that may  not be  adequately held  in  the  tube
sampler.  Sample retaining rings can be  used with some samplers
to retain coarse samples.

Driven Probes

       For the  driven ground-probe technique, a drive tip  is
attached  to a ground probe which is then forced into the  ground.
This  minimizes disturbance  of  the sampling environment.
Openings  in the tube near the leading  edge  allow soil gases  to
enter the tube.   Sample gas is extracted from a port at  the
upper  end of the tube using a gas-tight syringe.  Analysis  is
performed by using  GC.

      An  improvement  to the  ground-probe  technique  is  to
minimize  the internal volume  of  the sampler.  This means  a
smaller  sample volume is necessary  to purge the system, and,
consequently,  a  more  representative sample is obtained.
                               130

-------
                  ENO CAP
 -WING NUT         y-

 \     1/4*        /.       /-S
  \  VSWAQELOK/       A

  \  n\FITTIMQ  r*— i     //
— fl-HAr ---- F==^ — /r

  *  "             "
                           CREENS
                                                THREADED  ROD
                                BRASS CORE SLEEVE!
 LOCK
 WASHER
                            TEFLON RING


                    TEFLON CAP LINER
Figure 4.1.  Soil core canple sleeve (Hanisch  and McDevitt,
            1984).
                               131

-------
Applications of small-volume,  driven ground  probes  are
presented in a later  section.

Recommended Use—
      The ground  probe technique  has been successfully used at
a variety of sample sites.  For  ground-water  contamination
investigations, small-volume,  driveable ground  probes  are
preferred, although  larger-volume  probes have  been used with
good results-.  The  ground-probe technique is well-suited for
ground-water investigations except in the presence of  wet or
clayey soils or near surface rock  strata.  In addition, this
technique has been accurately used  to map plumes.  In general,
the ground-probe technique is relatively sensitive  and can be
used to measure subsurface  gas  concentrations  while avoiding
surface interferences.   Driven ground probes  also offer  the
user the ability to sample below impermeable soil and to  modify
the sampling depth to increase sensitivity.

Technique Applications—
      Applications reported in  the  literature for both
large-volume ground probes and  small-volume ground  probes  are
presented in this section.  The  internal volume of  the ground
probe significantly affects the measurement process and  the
utility  of  the resulting data.   The small Internal  volume
ground probes can be used  to attempt to measure  the "true"
soil-gas concentration.  The small volume permits the  air
inside the probe to be  purged and  a small  (e.g., 1  mL)  sample
to be collected without  substantially altering  the  equilibrium
of the soil-gas concentration.  Alternatively,  the use of  large
internal volume ground-probe typically involves sampling
several liters of soil  gas.  This  sampling may not permit  a
"representative" soil-gas  sample  to be collected  under most
conditions but allows for the soil  gas to be concentrated  prior
to-analysis  or multiple aliquots  to  be extracted.   The
large-volume ground  probes  are typically  used  for
investigations that  seek to  determine relative soil  gas
concentrations or that are concerned  with whether or  not
contamination affects a  given area.

Large-Volume Ground Probes—
      Ground probes have  been  widely  used.   For example,
Russell and Appleyard (1915) used a 2 foot  long probe (shown in
Figure 4.2) which was  hammered into the soil  to the desired
depth, and then the inner rod was  pushed down another 1/4  in.
Neglie and  Favretto (1962) used a  similar design  (see Figure
4.3) and  procedure except that the  outer tube was raised  0.7 to
1.0 foot after placement to provide a pathway  for  soil  gas to
enter the  sample.    They detected maximum  hydrocarbon
concentrations in the soil gas immediately after inserting  the
probe.  However, Neglie and Favretto  concluded  that grab
sampling of soil cores provided  better results  than  their

                              132

-------
Figure 4.2.
Ground-probe design  used by Russell and Appleyard
(1915).
                                 133

-------
                   •I "»
                   ILil

                   IL
IL
Figure 4.3.  Ground-probe design used by Neglia and Favretto
          (1962).
                          134

-------
technique.   Tackett (1968)  used a  probe with  a  slightly
different design  to  avoid plugging the sampling port during
insertion.  A slot was  cut in the side near the tip to  allow
soil  gas  to be pulled  into the  probe (see Figure  1.1).
DeCamargo (197D  modified Tackett's design so  that samples
could be collected in  glass ampules which could  then be sealed
and stored for later analysis.

      Other  researchers have used  driven probes  with perforated
ends  to  measure  landfill gases  (Thorburn,  et  al.,  1979;
Colenutt. et al.,  1980), carbon dioxide and oxygen (Lovell, et
al.,  1980) and to detect hydrocarbon spills  (Spittler,  et al. ,
1985).   For example,  Thorburn, et al.  (1979),  sampled landfill
gases by  using a probe in which a pointed rod was  placed  Inside
the  tube during  probe insertion  and  was  then  removed before
sampling  (see Figure 1.5).

      Tracer Research  Corporation  (TRC)  has used hollow,
perforated metal  probes for a number  of  site  Investigations
with  potentially  contaminated ground-water  (Lapalla, et al.,
1981; Marrin, et al., 1981).  For depths of less than 10  feet,
probes are driven to the desired sampling depths.   For  depths
greater than  10 feet, the  probe is  driven ahead of the bottom
of a  hollow  stem  auger that has been  advanced to just above the
desired depth.  Soil gas  Is pumped  from the sampling location
at a  rate of 0.5  to  0.8  gal/min  for  several  minutes.   Then a
syringe  is  used  to collect samples  for  GC analysis.   Others
(Lovell, et  al..  1980; Walther, et al.,  1983) have  reported use
of this method.

      Tracer  Research  Corporation  (Lapalla, et al.,  1981,
Marrin,  et al.,  1981) has  documented sampling results from over
12 sites with  varying  site  conditions such as ground-water
depth of 10  to  125 feet,  varying  clay and moisture Levels In
the  soil,  and  different  organic  contaminants present.   Tracer
Corporation  found  that this  technique detected organic
compounds in almost all  situations,  even above one site  with a
30 foot  caliche layer overlying the  ground-water  table.  The
technique  could  be  used to  map known  plumes accurately ;
however, It is  not suitable for wet,  clayey soils  or where an
uncontaminated  aquifer  overlies one that  is polluted.  Sampling
results from  the soil  gas  and  the  polluted ground-water
correlated  well even with repeat samples on successive days.
TRC  (Lapalla, et al.,  1981; Marrin, et al . , 1981) has reported
 that gasoline  vapors In  soil  act differently than chlorinated
 organic vapors.  Marrin  (1985) reports that the TRC  ground
 probes  can  be used to  map gasoline  plumes at sites where the
 water table is relatively shallow or where probes can be  driven
 below the oxidation zone In  soils.  Petroleum hydrocarbons are
 often absent from the  shallow soil  gas overlying  gasoline-
 polluted ground-water; this  is  believed to  be  due to


                               135

-------
                                 SEPTUM
                                     PIPE  UNION
Figure 4.4.  Ground-probe design used by TackkeCt (1968)
                               136

-------
                  SAMPLING OF QA8ES FROM LANDFILL
         DIAMETER 0.2 IN
U.X IN.
^m
6.6
^m
SAMP
TU
(ALUM
••
FT.
^
LING
BE
HUM,
•
1
L
8T
R
t,
1
El
01
J j
6.7
^^
EL
3
IN.
^»
FT.
••^
Figure 4.5.  Ground-probe design used  by Thorburn,  et al.(1979)
                                 137

-------
biodegradation  of  the  gasoline vapors in the near-surface  soil
layers.

      Several  researchers have used novel ground-probe
variations.  For example,  Jones and Drozd (1983) searched for
oil and gas deposits  by  augering a hole to a 13 foot depth,
Inserting  an inflatable rubber packer (probe) to isolate the
bottom of the hole, and  then pumping soil gas to a portable  GC
for analysis.  They also  sampled at shallower depths (1  to  2
feet) and detected more  high molecular weight compounds.   They
reported that  the deeper sampling depths provided more reliable
results (Schmidt,  1985).   Other  researchers have used similar
techniques.   For example,  Lovell, et  al .  (1980),  used the
equipment shown  in Figure  4.6.  Swallow's (Eklund,  1985)
technique was similar except  that  the void volume of the
sampler was much larger.   As  seen in Figure 4.7,  the sampler
was a plugged  corehole.  Spittler and Clifford (1985) used  a
method similar  to  Swallow's.  A hole was augered to a depth  of
12-18 in.  The  hole was  capped,  and a probe constructed of
plumbing fittings  was  inserted  into the hole.   Approximately
0.0035 ft  3/min  of  soil  gas  was removed for 4.6 minutes  until
the  soil-gas concentration  became constant.  Analysis was
performed  in the field  by  using a portable GC.

      Pogorski  and Quirt (1981) manually collected helium gas
samples by using an apparatus described as a low-dead-volume,
nonclogglng  steel probe.   The  apparatus is a type of  auger
which allows sample gases to be  pumped from the bottom.  Van
Bavel (1965)  has described  sampling soil gases by Inserting the
needle of  the  sampling  syringe to the desired depth.  Though
exceedingly simple, this technique has obvious limitations.

Small-Volume Ground Probes—
    .  As  mentioned previously, small-volume probes  are
believed to be advantageous  in obtaining a more representative
sample.  Use  of this type of probe has been reported by several
researchers.   Variations  in designs are  shown in Figures 4.8
(used by Swallow and Gachwend, 1983) and 4.9 (used  by Walther,
et al . , 1983).  The latter  was used  to obtain measurable
benzene concentrations  across a transect line that corresponded
to a plume of  known area.   LaBrecque, et  al.  (1984), modified
Walther's design  and  used  it  for the sampling of  a gasoline
spill at Death  Valley  National  Monument (see Figures  4.10,
4.11, and 4.12).   The  sampling  manifold shown  in  Figure  4.12
was shown to  give  carryover  between samples.   Sample  entry
holes were covered with  8  x  10-4 in. sintered stainless steel
disks to avoid blockage.  They concluded that the ground  probe
technique provided better results than geophysical methods  used
to define the dimensions of  the  plume but the  ground  probes
were  prone to false positive  readings.  A similar design was
used  by  Kerfoot,  et   al.  (1986),  to  investigate a  site

                              138

-------
      Equipment  for determination of mercury
      and  radon in soil  air  as  used  at
      Cachinal,  H.  Chile.   (1) 100 mm dis.  3-4
      m deep auger  drill  bole (2)  loose soil
      (3)  rubber packer (4) 25 mm dis. Dural
      prope (5) packer pressure line (6) PTFE  5
      mm dis.  line  (7)  dust filter (8) line to
      radon monitor  (9) radon monitor  (10) line
      to mercury  spectrometer  (11) mercury
      spectrometer (12) line to pump  (13) 1  1.
      pump (14) outlet.
Figure 4.6.  Ground-probe  design of Lovell, et al. (1983)
                          139

-------
                SEPTUi
         TUBIN
                                      VACUUM
                                       PUMP
                                 GROUND SURFACE
Figure 4.7.  Ground-probe design used by Swallow and
             Gschwend, 1983,
                           140

-------
                                   SYRINGE

                                   3  WAY VALVE
                                   TENAX QC TRAP
                                   FITTING
                                          FOUNDING  PLATE
                                  COUPLING

                                  AIR HOLES

                                  POINT
Figure 4.8.  Ground-probe design  used by Swallow and Gschwend (1983).
                                   141

-------
                            'HAMMERING  CAP
                            'IRON  PIPE
                     SAMPLING   PIPE
                            DRAEQER QA8
                         DETECTION TUBE
                            O-RINQ3
                         IRON  SAMPLER
                            AND TIP
                V
      • - Configuration for haaaering  aaapler into coil


      b - Placement of saapling pipe with Draeger tube into the sampler
Figure 4.9.   Ground-probe design used by Walther, et  al. (1983).
                                   142

-------
               PROBE TIP
                                              PROBE SHAFT
             CROSS-SECTION
                  A-A*
Figure 4.10.   Ground-probe design used by  LaBrecque,  et al.  (1984)
                                 143

-------
  INSERTION
      TOOL-
I EXTRACTION
TOOL
Figure 4.11.  Ground-probe driver  «nd extractor  used by
             LaBrecque, et al. 1984).
                               144

-------
         MSA SAMPLAIR PUMP
           ATTACHES HERE
MANIFOLD./
ASSEMBLY^
           3 mm TEFLON TUBE
           FROM SOV PROBE
            ATTACHES HERE
                                               MSA SAMPLAIR PUMP
Figure 4.12.
Sampling manifold and pump used by LaBrecque, et
al. (1984).
                                145

-------
contaminated  with  halogenated  organic compounds.  They  found
that the soil  gas concentration varied significantly (relative
standard deviation -  42  percent)  for adjacent locations  (less
than 7 feet  apart).  This  sampling variability was not constant
and  greatly  exceeded the  analytical variability.  A  later
survey at the  same  site showed  a much smaller variability of
12% rsd (Kerfoot and Mayer, 1986).

      Radian Corporation  (1984) has  designed and used  several
different small-volume, driveable ground probes.  Sampling was
conducted  at  the same site  in Death  Valley studied by
LaBrecque,  et al.   The sampling method  used by the  Radian
Corporation Involved  inserting ground probes (the design is
illustrated  in Figure 4.13) to a depth of 3 feet, raising the
outer tube  2  in.,  and allowing  the  probe to sit for 2 hours.
Samples were  collected  using syringes  and evacuated
stainless-steel canisters.  The canisters  were fitted with flow
regulators  to provide a  constant  sample flow of 3.5  x  10-4
ft/min over the 4-hr .sampling  period.  [The study showed that
the plume had  advanced  compared to earlier studies by the U.S.
Geological  Survey  and that  only  the  lighter molecular weight
compounds in the gasoline  were present at 1 ppbv-carbons  level
or greater  at the  near surface level.]  The ground probes used
in this  study were adequate  to define  the  general area of
contamination;  however,  an  insufficient  number were available
for  demarcating  the actual  plume dimensions  (Radian
Corporation, 1984).

      In another investigation. Crow, et al.  (1985),  used 32
smal1-vo1ume, driven  ground  probes to  determine  the
effectiveness of  soil-venting  techniques.  A pilot hole was
drilled  within  2 feet of the final  depth  by using a  4 in.
hollow-stem auger.  The  ground  probes (see  Figure  4.14)  were
inserted and driven to  16  to  22  feet below the ground  surface
and  then were sealed  in  place by using cement grout. For this
particular  site, Crow indicated  that sampling at  shallower
depths would  not have  provided acceptable  data.  Compressed air
was used to  clear any blocked sample entry holes.  The ground
probes sat in  place  for 24  hours before daily sampling was
begun.   Reproducible  results were obtained  from analysis of
repeat samples  from a single probe  and  samples from duplicate
probes.

      Llmitations--The  major  limitations in using soil gas
probes are  that  they are best  suited for shallow sampling;  they
are  not well suited  for rocky or  wet,  clayey soils; and
obtaining a  representative sample is difficult.  Other problems
include: the  method is labor  intensive, sample parts  may become
occluded during  probe  Insertion,  and ambient air can  in some
cases  migrate down the outside  of the probe shaft  and dilute
the sample.

                              146

-------
8YRINQE  SAMPLING PORT
  1/2 INCH  TEFLON  TUBE
                                         CABLE
                                   I
                                               TEFLON  PLUG
                                                  2 INCH  SPACE
                                                 DRIVE TIP
   Figure 4.13.
Ground-probe  design  used by Radian Corporation
(1984).
                                  147

-------
                SEAL
                             8WAQE
                             UNION
GAS TIGHT SYRINGE
             18".
                   zo
                   UIK
                   OO
                 O.J

                 S5
                 tt«
                 3
                 H
                 0

                 O
                                     SEPTUM
                                /PLASTIC LINER
                                    1/8* STAINLESS
                                    STEEL TUBING
                                     4' BORE  HOLE
                                   1* PIPE
                        o o
                        o o

                        o o
                     V
Figure 4.14.  Ground-probe design used by Crow, et al. (1985)
                             148

-------
Surface Flux Chambers

     The  surface flux  chamber technique  involves the  use of an
enclosure device to sample  gaseous emissions from  a  known
surface area.  Sweep air flows  through  the chamber.   The exit
gas is  analyzed  on-site or  collected  for  later analysis.
Knowledge of the flow  rate  of air  through the chamber and .of
the concentration of the exit gas enables  the emission rate to
be calculated.

Recommended Use—
     The  surface flux  chamber is particularly applicable to
measuring population exposures  since  gaseous emissions  are
being  measured at the surface level.   However,  surface
hydrocarbon levels are generally lower than subsurface levels;
therefore, the usefulness of the flux chamber method is limited
where the soil-gas  concentration is low.   Best results  are
obtained when sophisticated  sampling  techniques (e.g., stain-
less steel evacuated canisters)  and/or  sensitive detection
systems  (e.g., GC)   are  used.  The  technique minimizes
disturbance of the soil  or of any associated emission processes.
Sampling  is quick  (normally  requires  1/2  hr per sampling
point), requires simple  equipment, and is suited to most soil
types.   Radian Corporation  (Schmidt,  et al., 1983)  reported
88.5 percent to  124 percent  recoveries  when  a 36-component
organic standard was used and was introduced into the  sampling
system.  In addition,  a range of 2.3  x  10-11 to  1.1 x  10-1
Ib/ft2/min was determined  (Radian  Corporation, 1981;  Schmidt,
et al.,  1982).  The  Radian Corporation (Balfour, et al.,  1985)
also reported accuracies  of  better than  ±10 percent  and
analytical variabilities within  ±20 percent determined  by an
independent audit.

Technique  Applications—
     In  1983,  Eklund  and  Schmidt (1983) performed a  review on
the development of the flux  chamber  sampling technique.   This
review revealed  only three  groups  of researchers using this
method  for measuring hydrocarbon emission  rates.  However,  this
method  has long  been used to measure fluxes of non-hydrocarbon
gases.  Sekulic and  Delaney (1980) suggested the application of
this method  for  measuring  hydrocarbon emissions from a
wastewater treatment lagoon.   Their  device  consisted  of a
floating truck Inner tube with translucent  plastic covering one
end.  Sweep air  was used to  force a sample to  a  portable
organic  vapor analyzer  (OVA)  equipped  with  a FID.   Another
researcher, Zimmerman  (1977), used  a  2.1-foot diameter  flux
chamber with a collapsible top to measure  blogenic hydrocarbon
emissions.  Analysis took place  within  21 hours  using  three
separate GCs to  examine a broad  range of  hydrocarbon  species.
The  flux  chamber  used by  Schmidt (1983),  consisted  of a
stainless  stee1/acry1ic  chamber (see  Figure  1.15)  with

                              149

-------
                          CD I  TEMPERATURE
                                 READOUT
SAMPLE COLLECTION
       AND  ANALYSIS
                                                                                   ON/OFF
                                                                                   FLOW
                                                                                   CONTROL
                                                                                  GRAB
                                                                                  SAMPLE
                                                                                  PORT
                                                                               PLEXIGLASS
CARRIER QA8
          Figure 4.15.  Surface flux chamber  and  peripheral equipment (Eklund,  et al., 1984),

-------
Impeller,  ultra-high  purity sweep  air and rotameter for
measuring flow  into the chamber, and a sampling manifold for
monitoring and  collection of the specie(s)  of  Interest.
Portable  FID-  and photoionization  detector  (PID)  based
analyzers were  used  to continuously monitor  total hydrocarbon
concentrations in the chamber outlet gas stream.  Samples were
also collected  for subsequent GC analysis once  a steady-state
emission rate was obtained.  Air and soil/liquid temperatures
were measured by using a thermocouple.  The system pressure was
monitored by using  a magnahelic pressure gauge.  This  technique
has been applied  to  determine the area of contamination at two
Jet fuel  (JP-4) spill sites  (Radian Corporation, 1981;  Radian
Corporation-T, 1981).

      Limltations--The surface flux chamber  technique has
several  limitations  which should  be considered prior  to
selection.   The sweep  air dilutes  the gas  sample which
decreases the sensitivity of the method.  This  technique  is not
suited for sites  with caliche or other semi-impermeable soils
or when the soil sampled is  saturated with  water which  blocks
gas transport  pathways.   In addition, the method has  an
inherent effect  on the emission rate being measured.  These
effects have been  investigated (Koerner, et  al., 198U; Zohdy,
et al., 1974),  and modified chambers have  been  developed
[Mathis, et al. (1980)] to minimize these effects.  Finally,  it
should be noted  that gas concentrations at  the surface are
normally lower than at subsurface  locations.

Sorbent Samplers

      Sorbent  samplers can be used to collect  soil gases during
a given time period.  The sampling time  is  adjusted  to  provide
a  sufficient  quantity  of trapped  gas for  analysis.  This
technique provides an integrated sample  that  compensates for
any short-term fluctuations in soil gas concentration.

Recommended Use—
      Sorbent  samplers are well-suited  for  almost  any  site.
The sorbent sampler technique is best suited for cases when the
soil gas hydrocarbon concentration is expected to be very low.
The sampling duration can be varied to  ensure  that  sufficient
material is collected for analytical detection.  This technique
is useful for determining whether  contamination is present, but
other  techniques are  more appropriate for obtaining more
specific information.

Technique Applications—
      A wide  variety  of  sample gas extraction and accumulation
procedures  have  been reported.  In general,  they involve the
addition of a  sorbent sampler to sampling techniques such  as
ground probes, collection cans, or flux chambers.

                               151

-------
      Several researchers  have used passive samplers  In
conjunction  with ground  probes to detect  the presence  of
hydrocarbons.   For example,  Boys (1967) pumped sample gas from
a ground probe for 30 sec to 10 min and recorded the  volume.
He then  used his patented method of using a refrigerated  GC
column for analyzing the soil gases.   In  other work, the
exhaust  gas  from ground probes has been  trapped by  using
activated charcoal  (Colenutt,  et al., 1980), Tenax (Swallow,  et
al.,  1983).  and Draeger  tubes (Walther,  et al.,  1983).
Colenutt and  Davies (1980) cite other researchers using  silica
gel,  graphitized  carbon black, and porous polymers.

      Other researchers  have used collection cans or some sort
of enclosure  to obtain gas samples.   For example, Pearson,  et
al. (1965),  pumped gas  from an enclosure through a series  of
chilled impingers containing absorbing solution to measure the
radon-222  flux from soil.  During  their  investigation,  they
experienced  a large  variability in their results which  they
attributed to soil disturbance from the edge of the enclosure.
Thereafter,  caulk was used to  seal the  enclosure with the
sampling surface.   Kristlansson  and Malmqulst (1982)  also
measured radon by using detectors in Inverted cups placed  in
shallow  holes  and  refilled  with  soil.   Ryden,  et al.
(1978),measured  the nitrous  oxide flux from  soils  by
continuously pumping gas from an enclosure and by trapping the
sample  in  molecular  sieves.   Karimi  (1983)  used enclosure
devices to  sample at hazardous waste sites.  The procedure
Involved  pumping  gas   out of the  enclosure, trapping
hydrocarbons in  a column of activated charcoal, and analyzing
samples by GC and GC/MS.  Fluxes were measured for ten selected
organic  compounds and ranged from  4.3 x  10"11 to 1.2  x  10~9
Ib/ft2-sec.

      McCarthy  (1972)  collected mercury emissions by  using
enclosures on the ground surface (for 2-hr  periods) with  an
amalgamation on  a silver screen placed  inside the enclosure.
As reported by Kanemasu, et al.  (1974),  some investigators  have
used hydroxide solutions in inverted cans  to measure the carbon
dioxide flux.  These  investigators  reported that this  static
collection method  yields fluxes 20 percent lower than a dynamic
(flux chamber) method.

      Rouse  (1984)  used a slightly  different procedure  to
passively measure soil surface gases.  Glass vials filled  with
an absorbing solution were buried  6 inches deep in backfilled
holes and left in the  field for  1 month.  Rouse found  that the
depth  at which  the  vial was  placed  (a  few inches to  a few
yards)  did not affect  the  results.  In  addition, he concluded
that  this procedure produced similar  results to the  grab


                              152

-------
                              CUTAWAY  VIEW
 Ground  Surface
              INVERTED CAN
                                              CHARCOAL


                                              ABSORBENT/  x
                                                         ' 0 /
                                             FERROMAGNETIC
                                             WIRE
Figure 4.16.  Curie-point wire accumulator device (Voorhees,  1984).
                                  153

-------
                0.20 gal gasoline at 20 ft  depth
                              Buildup and decay
                                through aand
                                      Buildup and decay
                                        through wot clay
                                              Total Ion counti
                                               remain 10X
                                                 background
                                                  on  fifth day
Baekground
Monitored
for 10 daya
* 2OO T.I.C./day
               DAYS ELAPSED
   Figure 4.17,
Build-up and attenuation of volatile! from
gasoline through a  aand  column and through
undisturbed wet  clay  soil aa measured by
Curie-point wire  sampler (Bisque,  1984).
                                154

-------
A

CO
3
O
o
Distance In meters 1 2
/ / ''
/ / /
III
III/
^^ ^^^^m '
'/>'
' / /^
/ /
1
1
i
-c\\
~"\ \ .\
^\ \ \ \
\ \ \ 1
-~— ^-^— ™^— ^— .^™.
3
\ Ground
\surface
\\
\ \
\ \
                                      Source  at  10 ft depth
  Figure 4.18.   Hypothetical diffusion  pattern  (below) and
                measured surface  flux anomaly (above) as measured
                by  the Curie-point wire sampler (Bisque, 1984).
                                 155

-------
 sampling  soil  core  technique  and  to  an  ambient  air  sniffer
 survey.

      A  passive sampling technique using activated  charcoal was
 developed by Petrex  Corporation  and has a  patent pending.
 Reported advantages  of this device  include:   (1)  the samplers
 are simple and rugged, (2) sampling and  analysis are relatively
 inexpensive, (3) light hydrocarbons can  be  detected, and (4)
 the technique appears  to be   relatively unaffected by  weather
 and  site conditions.   Bisque (1984)  describes  the sample device
 as  a thin  ferro magnetic  (Curie-point) wire  coated  with
 activated charcoal.   The procedure  Involves  placement of the
 wire  in a glass tube which is buried 6-12  in.   below  the
 surface  and left  for several weeks.  When the sample  is
 retrieved, the wire  is placed in a vacuum  chamber,  is  heated,
 and  the desorbed hydrocarbons  are analyzed by Curie-point mass
 spectrometry.  Although this method is limited  by frozen  ground
 and  saturated soils,  meteorological  and hydrological conditions
 have a minimal effect.

      Applications  of the  technique  have been widely reported.
 Voorhees  (1984),  described  the use of Petrex tubes  to
 investigate ground-water which was  41-feet  below the surface
 and contaminated with tetrachloroethylene  (TCE) at the  Rocky
 Mountain Arsenal.   Twenty-five  Petrex samplers (one is shown in
 Figure 4.16) were placed along  two traverse lines.  The  plume
 boundaries  and agreed  well with  the results obtained  from
 monitoring wells.  Chloroform was the  only  major component not
 detected in the ground-water by the  trapping  device.  Voorhees
 (1984)  described another study conducted at  Rocky Mountain
 Arsenal  to  detect various hydrazines and  their oxidation
 products.  None of  these species could be positively identified
 because of  high  background  concentrations of naturally
 occurring petroleum  deposits in  the area.

      Bisque  (1984) presented  the results of  a quality control
 study.  Approximately  one quart  each  of  gasoline, diesel  fuel,
 and crude oil were introduced at 10-foot depth in soil media
 that varied from tight clay containing 10 percent free water  to
 dry  colluvial material.  In all  cases, trace emissions  were
 detectable at the surface within hours.  Figure  4.17 shows
 concentration versus  time  data  for each  soil  type.
 Contamination could  still be detected after 60  days.  The
observed diffusion pattern from the  point source contamination
 is shown in Figure  4.18.

      Limitations—The Petrex tube sampling technique and other
passive techniques  require a long sampling time and  disturb the
sampling site.   In  addition, high  background concentrations may
 interfere with obtaining accurate measurements.  The  efficiency
of this  collection method cannot be determined since, unlike


                               156

-------
the surface flux  chamber technique, the sample enters the tube
by passive diffusion,  and the volume of gas  sampled  is  not
measured.

      Sample  collection by pumping soil gas  from collection
cans or ground probes also has limitations.   The  pumping  may
disturb the equilibrium between the soil gas and the  gas sorbed
on soil particles.  It may cause dilution and/or contamination
of the sample  by ambient air.

      Kerfoot  and Mayer  (1986) give  the  following limitations
for passive charcoal soil-gas  samplers with analysis by thermal
description:    it is not quantitative, there may be thermal
decomposition  of tightly sorbed  and less volatile compounds,
and chemical decomposition of  the  samples is  promoted.

SAMPLING DESIGN  AND SAMPLING QUALITY ASSURANCE TECHNIQUES

      This section discusses the  approach necessary  to
make  optimal  use  of the available  resources and to  ensure
adequate data  quality when making  soil-gas measurements.

Sampling Strategy

      The sampling strategy  should be devised to obtain  all
necessary and  required Information with a minimal  expenditure
.of  time  and  resources.   Prior  to  developing  a  sampling
strategy,  any  available information pertaining to the following
items  should  be collected and evaluated:  type of contaminant
present; amount  of contaminant  present;  length of time
contaminant has been present; direction and  rate of flow of
ground-water;  depth to ground-water; geological soil properties
of the  site;  number, type, and location of existent subsurface
structures (e.g., wells and  sewers); existent sampling  and
analytical  results;  and  any  anecdotal evidence  of
contamination.

      The above information,  along  with  the objectives of the
test program,  should be used to tailor  a  sampling  strategy to
the  specific circumstances  encountered.   An example
contamination  problem  is discussed below  and  serves  to
illustrate the process of  developing  a  sampling strategy.
Examples of specific sampling  strategies may  be found in the
references to  the Individual sampling  methods discussed in
Section III.A.

      At  a hypothetical site,  an  underground storage tank
containing industrial/organic  wastes  is  suspected of leaking
based upon inventory control records and tank-tightness tests.
Little information is available regarding  the site,  and no
existing  observation wells are  present.   Furthermore, nearby

                               157

-------
wells  (approximately  1 mile  distant)  In  every direction
provide community drinking water,  and the press is aware of the
situation.   The  objective  is  to determine the  extent of any
contaminant  plume.

      The first step after background information collection
and evaluation  is to establish a grid system over  the site with
the suspected  source  at the center.  The number  of points that
should  be sampled  is  very site  and project  specific.   The
number will depend  on  the  resources available,  the project
objectives,  and  the  level of  risk that is  acceptable.   In
general, it  is much  more resource-intensive  to prove that an
area is not  contaminated  than to  roughly  delineate the extent
of a contaminant plume.  In our example, a grid of 10 rows of
10 sampling  points  each  with points and  rows  separated  by 50
fe-et  might  be used.  Once all the design  considerations
discussed  below  had  been met, measurements  made  at
approximately  a  dozen selected  sampling points  circumscribing
the point source would indicate  the  direction of  any
contaminant  plume.   A more refined secondary phase of sampling
could  then  be  planned.  A transect of equi-spaced  sampling
points in the direction of  the plume would  indicate  the
location of  the plume front.   Once this  has  been determined,
several  lines  of equi-spaced  sampling points perpendicular to
the transect line  will  indicate the lateral extent of  the
plume.

      The next consideration is  to select a sampling method.
The selection  will  depend upon  the contaminant species,  the
site  characteristics, the expertise available, and any time and
budget  constraints.   In the  example case,  a  sampling  and
analytical  system such as ground  probes and on-site analysis of
the gas samples might be  selected to provide rapid feedback to
site investigators.

      A third consideration is  what species to  monitor.  This
is, of course,  dependent  on the sampling method selected.  The
choice  (tracer gas)  should  be a compound that is detectable at
low concentrations, has low molecular weight so that it will be
present near  the  plume front, but not  so  light that  it is
rapidly lost to  the  atmosphere.  The tracer gas should be
relatively  inert and insoluble  in water so that  attenuation is
not a problem.  Finally,  the tracer compound  must be readily
attributable to  the  contaminant plume and not  to background
sources or  analytical  interferences.   In  our  example,  a
compound would be  selected  that was typically  present  in the
storage tanks.  Good candidates might be a  chlorinated solvent
such as TCE, benzene, or  total hydrocarbons.

      A  fourth consideration is the selection  of a sampling
horizon or depth.  Soil-gas measurements are typically made at

                              158

-------
depths from 0  to  10 feet unless existing wells are present.
Sampling at depths  below 10  feet  negates the advantages in
terms of  time  and  effort of  using  soil-gas measurements
relative to other  options.  Certain sampling methods such as
surface  flux chambers  or Curie-point wire samplers can be  used
at or  near the surface.   Other methods, such as ground probes,
can be used at  variable depths.  When using  ground probes, it
is recommended that several vertical profiles be performed at
sampling points known to be contaminated.   This information can
then be  used to  select a sampling depth that yields conclusive
results  for subsequent  sampling  points.  In our example,  ground
probes would be  driven to 6 to 10 feet until  contamination is
first encountered.  Once a  contaminated  point  is  found,
additional ground probes could be  driven nearby.   Samples
collected at one-foot  depth  intervals would then show what a
suitable  sampling  depth would be.   It is useful,  during
subsequent sampling,  to periodically perform  vertical profiles
rather than always to sample  at a fixed  depth.  The vertical
profiles  will  Indicate  whether the typical sampling depth
should be modified; deeper to  improve sensitivity, or shallower
to improve productivity.

      A  final   consideration  is the length of  time spent
sampling at any  given point.  This is dependent on the sampling
method selected and on the  sensitivity desired.  Methods  that
require  a probe/sampler to be  placed in the ground may  result
in  a  disturbance of the equilibrium among free  product,
adsorbed product,  and gas in  the soil-pore spaces.  The  time
required to allow this equilibrium to be reestablished prior to
initiating sampling can be  as  much  as one  day.   Also,  for
methods such as accumulator devices, the sampling duration can
be extended to  lower the detection limit  of  the method.  The
final choice  of sampling  duration  must be  based  upon
experience, preliminary results,  and site  and  project specific
factors.  In our  example, since  absolute concentrations are not
required, the  ground probes  could be driven to  the desired
depth,  allowed  to equilibrate  for as little as one hour, and
then the samples  could be collected.

Quality  Assurance

      This  section addresses sampling quality  control.
Analytical  quality control is discussed  in the next  chapter.
Quality control must be an  integral part of  any sampling  plan
and is necessary so  that the results obtained for  the  project
are meaningful, i.e.,  the data are of a  known quality.  The
exact  quality control checks required  for  any project  will be
dependent on the  sampling and analytical methods selected.  The
following  list  of  quality control considerations is applicable
to most  soil-gas measurement programs:


                               159

-------
    o   detailed  sampling  procedures and schedules  clearly
       written and consistently followed;

    o   samples labeled with  all pertinent  information;

    6   data  collected on appropriate data sheets and reviewed
       dally;

    o   sample  logs,  ehain-of-custody forms,  and other
       paperwork kept up-to-date and reviewed  daily.

    o   sampling blanks collected at least  daily;

    o   repeat measurements  at  a control point;

    o   background measurements made at least daily;

    o   a  minimum of  10  percent of  samples  collected  in
       duplicate; and

    o   a  minimum of  10  percent  of samples  analyzed  in
       dupli cate.

      The value  of  keeping  good  records and  of using  a
consistent technique  is obvious.   Sampling blanks  are  useful
for detecting contamination within the sampling system  that
would not be detectable  from analytical  blanks.   Background
measurements allow  comparison to measurements  in the
contamination zone to  ensure  observed contamination  is  not due
to problems with the sampling method or implementation.

      Duplicate  and repeat  sampling permit  statistical analyses
to determine the  variability associated  with  the sampling
procedure.  For sites  where the  temporal variability exceeds
the spatial variability, repeat sampling may be  replaced  with
side-by-side duplicate sampling.  Side-by-side  samples should
be expected to show greater variability than repeat  sampling
since no two discrete sampling locations are  perfectly
identical.  With the sampling methods that  require the  sampling
location to be disturbed, e.g., ground probes, the side-by-side
sampling  locations must be sufficiently separated so  that the
placement and sampling at  one probe  does not effect the soil
gas concentration  at  the adjacent probe.   This  distance  will
vary  depending  on  the soil  characteristics,  but  should  be
assumed to be at least 3 feet.

      When  employing  sampling methods that permit the reuse  of
sampling  components for multiple sampling points, care must  be
taken to avoid cross-contamination  of samples.  This can best
be prevented by  cleaning  sampling  components  before each

                              160

-------
sampling point and by performing sampling blank and background
measurements after any measurements  showing high  levels of
contamination.   Should the  blank/background data  show  a
significant contamination problem, the  sampling components need
to be  thoroughly cleaned and retested,  or replaced.
                              161

-------
                        REFERENCES
 1.  Balfour,  W.  D., R.  G.  Wetherold and  D.  L.  Lewis.
    Evaluation  of Air  Emissions  from  Hazardous  Waste
    Treatment, Storage,  and Disposal Facilities.  EPA-IERL
    68-02-3171.  Task Number 63, 1984.

 2.  Bednas, M.  E.  and D. S.  Russell.   Determination of  Natural
    Gas Leakage via Gas Chromatography of Drill Core Samples.
    Journal  of Gas  Chromatography,  Vol.  5, No.  11,  pp.
    592-594,  1967.

 3.  Bisque, R.  E.  Migration Rates  of  Volatiles  from Buried
    Hydrocarbon Sources  through  Soil Media.   Proceedings of
    Petroleum Hydrocarbons  and  Organic Chemicals in Ground
    Water, Houston, Texas, November 5-7,  pp. 267-271, 1984.

 4.  Boys, F.  L.   Method  of  Analyzing Soil Gases. U.S. Patent
    3,336.792, 1967.

 5.  Colenutt,  B.  A.  and  D.  N. Davies.   The  Sampling  and Gas
    Chromatographic Analysis of  Organic Vapours in Landfill
    Sites.  International Journal of Environmental Analytical
    Chemistry, Vol. 7, pp. 223-229, 1980.

 6.  Crow, W.  L.,  E.  P.  Anderson and E. Minugh.   Subsurface
    Venting of Hydrocarbon Vapors  From an Underground Aquifer.
    Draft Final  Report to American Petroleum  Institute, Radian
    Corp., May,  1985.

 7.  DeCamargo,  0.  A., F. Grohmann,  E.  Salati  and  E.  Matsui.
    A Tech-nique  for  Sampling the  Soil Atmosphere.  Soil
    Science,  Vol.  117, No. 3, PP.  173-174,  1974.

 8.  Devine,  S. B.   and  H. W.  Sears.  Soil Hydrocarbon
    Geochemistry,  A Potential  Petroleum Exploration  Tool in
    the  Cooper  Basin,  Australia.   Journal of Geochemical
    Exploration,  No. 8, pp. 397-414, 1977.

 9.  Eklund, B.   Detection of Hydrocarbons in  Groundwater by
    Analysis  of Shallow Soil  Gas/Vapor.   API Publication 4394,
    May, 1985.

10.  Eklund,  B.  M. and  C.  E.  Schmidt.  Review of Soil Gas
    Sampling  Techniques.   In  Soil Gas Sampling Techniques of


                              162

-------
     Chemicals for Exposure Assessment.  Radian  Corporation,
     Interim Report.  EPA-EMSL 68-02-3513. Work Assignment 32,
     1983.

11.  Hanisch,  R. C. and M. A. McDevltt.   Protocols for Sampling
     and Analysis of  Surface  Impoundments  and  Land
     Treatment/Disposal -Sites for  VOCs.   Technical Note.
     EPA-EMB 68-02-3850, Work Assignment  11, 1984.

12.  Horvitz,  L.   Near-Surface Hydrocarbons and Petroleum
     Accumulation  at Depth.   Mining Engineering, Vol. 6, pp.
     1205-1209, 195*.

13.  Horvitz,  L. -Vegetation and Geoehemical  Prospecting for
     Petroleum.   Bulletin of  the American Association of
     Petroleum Geologists.  Vol. 56, No.  5, pp. 925-940,  1972.

14.  Jones,  V. T.  and R.  J. Drozd.   Predictions of Oil or Gas
     Potential by Near-Surface  Geochemistry.   Bulletin of the
     American  Association of Petroleum Geologists, Vol.  67, No.
     6,  pp. 932-952, 1983.

15.  Kanemasu, E.  T.,  W. L.  Powers  and  J. W.  SiJ .  Field
     Chamber Measurements of  C02 Flux from Soil Surface.  Soil
     Science. Vol.  118, No. 4, pp. 233-237, 1974.

16.  Kariml,  A.  A.  Studies  of the Emission and Control of
     Volatile  Organics In Hazardous  Waste Landfills.   Ph.D.
     Engineering  Dissertation, University of  Southern
     California, February 1983.

17.  Kerfoot,  H. B.t J.  A. Kohout and  E. N. Amick.  Detection
     and Measurement of Ground-Water Contamination by Soil-Gas
     Analysis.  Presented at 3rd National Hazardous Wastes and
     Hazardous  Materials Conference, Atlanta, Georgia, March
     4-6, 1986.

18.  Kerfoot,  H.  B. and C.  L.  Mayer.  The Use of Industrial
     Hygiene Samplers for Soil-Gas Measurements.  Ground Water
     Monitoring Review, Fall 1986.

19.  Koerner,  R. M. and A. E. Lord, Jr.   Nondestructive  Testing
     Methods for Detecting  Buried  Wastes.  AIChE 1984 Summer
     National  Meeting, Philadelphia, Pennsylvania, August
     19-21,  1984.

20.  Kristiansson,  K.  and  L.  Malmquist.   Evidence  of
     Nondiffusive Transport  of  222  Rn in the Ground and a New
     Physical  Model for  the Transport.   Geophysics,  Vol. 47,
     No. 10, pp. 1444-1452,  1982.


                               163

-------
21.   LaBrecque,  D.  J.f  S. L. Pierett, A. T. Baker  and J. W.
     Hess.  Hydrocarbon  Plume Detection at Stove Pipe Wells,
     California.   EPA-EMSL. 68-03-3050, Draft Final Report,
     1984, 31 P.

22.   Lappala,  E.  G. and  G.  M. Thompson.  Detection of
     Ground-Water Contamination  by Shallow Soil-Gas Sampling in
     the  Vadose  Zone,  Theory and  Applications.   National
     Conference on Management of Uncontrolled Hazardous Waste
     Sites, Washington, D.C., November 1984.

23.   Lovell,  J.  S.,  M.  Hale  and  J.  S. Webb.   Vapor
     Geochemistry in Mineral Exploration.  Mining Magazine, pp.
     229-239, September,  1980.

24.   Marrin, D.  L.   Delineation of Gasoline Hydrocarbons in
     Ground  Water by  Soil  Gas  Analysis.   Proceedings of the
     1985  Haz.  Mat. West  Conference,  Long  Beach, California,
     December 3~5. 1985.

25.   Marrin, D.  L.  and  G.  M. Thompson.  Remote  Detection of
     Volatile Organic Contaminants in Ground Water Via Shallow
     Soil-Gas Sampling.   Proceedings of Petroleum Hydrocarbons
     and  Organic Chemicals in  Ground  Water,  Houston, Texas,
     November  5-7,  1984,  172-187  p.  National Water  Well
     Association, Dublin, Ohio.

26.   Matthias,  A.  D.,  A.  M. Blackmer and J. M. Bremner.   A
     Simple-  Chamber  Technique  for  Field  Measurement of
     Emissions  of Nitrous Oxide  from Soils.   Journal of
     Environmental Quality, Vol. 9, No. 2, pp.  251-256, 1980.

27.   McCarthy,  Jr., J.  H.   Mercury Vapor  and Other Volatile
     Components in the Air as Guides  to  Ore  Deposits. Journal
     of Geochemical  Exploration,  Vol. 1,  No. 2,  pp. 143-162,
     1972.

28.   Neglia,  S.  and L,. Favretto.  Study of an Analytical Method
     for  the  Execution of Surface Geochemical Prospecting for
     Petroleum  and  Natural  Gas.   Proceedings  of the
     International Meeting on Advances  in Organic Geochemistry,
     Milan, Italy, Vol. 15, pp.  285-295, 1962.

29.   Pearson, J. E.,  D.  H. Rimbey and G. E. Jones.   A Soil-Gas
     Emanation Measurement System  Used  for Radon-222. Journal
     of Applied Meteorology, Vol.  4,  pp. 349~356, 1965.

30.   Pogorski,  L. A. and  G. S. Quirt.  Helium Emanometry  in
     Exploring  for Hydrocarbons: Part  I.   In, Unconventional
     Methods in  Exploration for  Petroleum and Natural Gas II,


                               164

-------
     September  1979,  Dallas,  Texas, Southern Methodist
     University Press,  1981, pp. 124-135.

31.  Radian Corporation.   Soil Gas Sampling  Techniques of
     Chemicals for Exposure Assessment, Bonifay Spill  Site Data
     Volume.  EPA-EMSL  68-02-3513, Work Assignment 32,  198*.

32.  Radian Corporation.   Soil Gas Sampling  Techniques of
     Chemicals for Exposure Assessment, Stove  Pipe Wells Spill
     Site  Data Volume.  EPA-EMSL 68-02-3513, work assignment
     32, 1984.

33.  Radian Corporation-T.  Soil Gas Sampling  Techniques of
     Chemicals for Exposure Assessment, Tustln Spill  Site Data
     Volume.  EPA-EMSL  68-02-3513. Work Assignment 32.  1984.

34.  Rouse,  G. E.   Sulfur Gas Geochemlcal  Detection of
     Hydrothermal Systems, 1984.  DOE/ID/12063, 27 p.

35.  Russell,  E.  J.  and A.  Appleyard.   The Atmosphere of the
     Soil:  Hs  Composition and the Causes of Variation.
     Journal  of Agricultural Science, Vol. 7,  Part 1,  pp. 1-43,
     1915.

36.  Ryden,  J. C.,  L. J.  Lund and  D. D.  Focht.   Direct
     In-Field  Measurement  of Nitrous Oxide Flux from  Soils.
     Soil  Science  Society  of America Journal,  Vol. 42,  pp.
     731-737. 1978.

37.  Schmidt, C.  E.,  B.  M.  Eklund,  R. D. Cox and  J.  I.
     Steinmetz.  Quantltation  of Gaseous  Emission Rates from
     Soil  Surfaces.   In  Soil  Gas Sampling Techniques of
     Chemicals  for Exposure  Assessment.   Radian Corporation,
     Interim Report. EPA-EMSL 68-02-3513, Work  Assignment
     32, 1983.

38.  Schmidt,  C.  E.,  W. D.  Balfour, and R. D. Cox.   Sampling
     Techniques for Emissions  Measurement  at Hazardous Waste
     Sites.   In  Proceedings  of 3rd National Conference  and
     Exhibition on Management  of Uncontrolled  Waste  Sites,
     Washington, D.C.,  1982.

39.  Sekulic,  T. S. and B.  T. Delaney.   Assessing  Hazardous
     Waste Treatment Facility Fugitive Atmospheric Emissions.
     In,  Proceedings of 4th  Symposium  on Fugitive  Emission
     Measurement and Control, pp.  119-135, 1980.

 40.  Smith, G. H. and  M. M. Ellis.  Chromatographic Analysis of
     Gases from Soils  and  Vegetation, Related to Geochemical
     Prospecting for  Petroleum.  Bulletin of the  American


                              165

-------
    Association  of  Petroleum Geologist, Vol.  47, No. 11, pp.
    1897-1903,  1963.

41.  Spittler,  T.  M.,  W.  S.  Clifford, and L. G.  Fitch.  A New
    Method for  Detection of Organic Vapors in the  Vadose  Zone.
    Presented at National  Water Well  Association  Meeting,
    Baltimore,  Maryland, September, 1985.

42.  Swallow,  J.  A.  and P.  M.  Gschwend.   Volatilization of
    Organic Compounds  from Unconfined Aquifers.   Proceedings
    of the  3rd National  Symposium on Aquifer  Restoration and
    Ground Water  Monitoring, Columbus, Ohio,  May  25-27,  1983,
    PP. 327-333.

43.  Tackett,  J.  L.   Theory  and Application of  Gas
    Chromatography in Soil Aeration  Research.   Soil Science
    Society  of America  Journal,  Vol. 32, No.  3,  pp. 346-350,
    1968.

44.  Thorburn,  S.,  B. A. Colenutt,  and S.  G.  Douglas.  The
    Sampling and Gas  Chromatographic Analysis  of Gases from
    Landfill  Sites.  International Journal  of  Environmental
    Analytical  Chemistry, Vol.  6, pp. 245-254,  1979.

45.  U.S.  Environmental Protection Agency,  Environmental
    Monitoring Systems  Laboratory,  Office  of  Research and
    Development  and Office of  Solid  Waste and Emergency
    Response.  Unsaturated Zone Monitoring for Hazardous  Waste
    Land  Treatment Units. Draft Permit Guidance Manual for
    Public Comment,  September 1984.

46.  Van Bavel, C.  H.  M.   Composition of Soil Atmosphere.  In:
    Methods  of Soil Analysis Part 1, C.A.  Black Editor-in-
    Chief, American Society of  Agronomy, Inc.,  Madison,
    Wisconsin,  Chapter 22, pp. 315-318,  1965.

47.  Voorhees,  K. J.  Application  of a  New  Technique for the
    Detection  and Analysis of Small Quantities of Contaminants
    in the  Soil.   Petrex  Sales Literature  (Unpublished
    Manuscript, 1984).

48.  Walther, E.  G., D.  LaBrecque, D. D. Weber,  R.B.  Evans and
    J.  F.  vanEe.  Study of  Subsurface Contamination With
    Geophysical Monitoring Methods  at  Henderson, Nevada.
    Proceedings  of  National  Conference  on  Management of
    Uncontrolled Hazardous Waste Sites,  Washington,  D.C.,
    October 31-November 2, 1983, PP- 28-36.

49.  Zimmerman,  P.    Procedures  for Conducting Hydrocarbon
    Emission Inventories  of Biogenic Sources and Some Results
    of Recent Investigations.   Presented at the  1977 EPA

                              166

-------
    Emission  Inventory/Factor Workshop, Raleigh, North
    Carolina, September 13-15,  1977.

50.  Zohdy, A. A.  R., G. P. Eaton and D. R.  Mabey.  Application
    of Surface  Geophysics to  Ground-Water Investigations.
    Techniques  of Water  Resources  Investigations.  U.S.  Geo-
    logical Survey, BK. 2, Chap. D1, 1971,  p.  116.
                              167

-------
                           CHAPTER 5

                   ANALYTICAL METHODOLOGIES
SELECTING THE PROPER METHODOLOGY

     The method chosen to  analyze soil gas  is dependant on  the
pollutant being monitored,  the concentration of the pollutant,
the matrix accompanying  the  pollutant, and the information
expected to be obtained from the analytical results.  Expected
concentrations for  organic species in soil gas can vary from
the ppt (trillion)  volume  level  (well below most analytical
detection limits)   in background measurements to high percent
by volume levels in a measurement  made directly  over a  highly
volatile liquid lens such  as gasoline.  The  concentration level
actually measured will depend on the sampling method and  on  the
amount of  dilution  or concentration which occurs during  the
sample  collection.   Flux chamber  methods dilute the soil  gas
while  accumulator  device methods concentrate the soil  gas
components.  The  analytical sensitivity of the  method  chosen
for soil  gas analysis must be consistant with the sampling
method, the soil  type, pollutant  quantity and volatility,
ground-water plume depth,  and the data requirements.  Measuring
emission rates at the surface with acceptably low variance  or
mapping the fringe of a plume at  the surface where the plume is
of low  volatility, is at a  great depth, or where the soil has  a
high adsorptivity or low  permeability, would require  a very
sensitive analytical technique.  The same technique would need
to be  greatly  modified to analyze a sample over a gasoline
plume  in sandy soil.

      In some  cases,  the  pollutant  species which requires
monitoring  is not the major component in  the  soil  gas,  and  the
determination  of that species  is complicated  by the  sample
matrix. A  very sensitive technique may  not  be appropriate  if
it responds to  the components  in the  matrix as well  as  the
species of  interest.  In such cases,  selectivity is required
and can be obtained by Isolating the desired species  during
collection, separating the desired species from  the sample
matrices  during the  analysis  (i.e.,  chromatography) ,  or
detecting  only the compound of  interest (selective detection).
Specific exam-pies of such cases  are concentrating hydrocarbons
on a porous polymer  absorbent  while excluding  the highly
volatile hydrocarbons, permanent gases, and water; using  high
resolution capillary columns to separate  benzene from other

                               168

-------
hydrocarbons in  gasoline;  and  detecting  trace levels of toxic
halogenated compounds in soil gas by using an electron capture
detector  or a Hall Electrolytic Conductivity  Detector (HECD).

      In  addition to sensitivity and selectivity, several other
considerations  are required  to determine the  degree of
analytical  sophistication that is needed for  a  specific
measurement.  The questions that need to be asked are  listed
below.

    (1)   Is  detailed  speciation  required  or  will a  total
         organic value  provide the data  required?

      Sometimes a total organic value is  all that is needed to
monitor the movement of a  ground-water  plume or to determine
the  emission  rates  from surface or downhole  emission
experiments.  Detailed speciation is required if the migration
and flux  rates of individual components  are desired.  Often  a
few detailed speciation  analyses can be used as a profile, and
individual  results can be extrapolated from total organic
values.

    (2)   Is the  analytical technique to be used to determine
         the  relative  concentration or  will  absolute
         concentration  values be required?

      Real-time portable analyzers (e.g.,  FIDs  and 'PIDs) are
very cost effective and easy to use  for obtaining  relative
levels of organic compounds of similar  compositions.  This can
be extremely useful in screening the sampling points before
deciding  on more  resource  Intensive  remote analysis or  before
plotting  the relative concentration across  a  homogeneous  plume.
If the sample  components  in the plume vary with distance or
depth, the portable analyzer can fail to give correct relative
values as will be shown  later.

    (3)   If the  samples collected are to  be analyzed  in the
         lab, will the  sample require that  the lab be on site
         or can  it  be  in  a centralized laboratory remote from
         the sampling site?

      This question  is  dependent on two factors: the stability
of the components of  interest and the analytical sophistication
required to do the  analysis.   If the sample's shelf life is
less than 24 hours,  then  analysis  must be performed  in the
field.  If the analysis  is so  complicated that logistics or
cost prohibit its use in the field,  then the sample will have
to be sent  to the  lab.   When  a sample cannot be stored for
transport to the lab  or  the  analysis is too complicated to take
to the field,  another  method for analysis  or sampling  must be
used.

                              169

-------
Portable VOC Analyzers

     The  use  of portable VOC analyzers for fugitive emission
screening, source  identification,  and  industrial  hygiene
monitoring has  proven  to  be  very valuable, and they have  also
been used to analyze VOCs in soil gas.  Of the commercially
available portable analyzers,  several types are useful for
soil- gas  measurements.   These are the nondispersive  infrared
detectors, flame ionization  detector (FID) analyzers, and the
photoionization  detector (PID) analyzers.  The manufacturer's
descriptions of selected  portable analyzers are given in Table
5.1.  These portable analyzers  generally provide  real-time
measurements without  performing separations; however, some  have
chromatographic  capabilities as an option.  The advantages  of
portable analyzers include:

    o   The analyzers  are easy to  transport  in the field.

    o   The  operation  of the  analyzer requires minimum operator
       skill.

    o   The elimination  of the sample collection steps
       minimizes the   uncertainties and expense  of sample
       collection,  storage,  and transport.

    o   Data  are provided  immediately which  enable the
       investigators  to make  timely discussions  in the field.

     The  disadvantages  of the portable analyzer  include:

    o   The limited  sensitivity  because of  the  lack  of a
       concentration  step.

    o   The limited  selectivity  and  interference  problems
       because  of the lack of a separation step.

    o   The limited  accuracy because of the  inability  to
        calibrate  adequately for the mixtures found  in  soil
       vapors.

      The various  analyzers on  the  market (Anastas,  et  al.,
 1980)  have their unique advantages,  disadvantages,  and areas
 for use   as monitors  of  soil gas  for VOCs.  The  following
 section will  discuss  the types  of analyzers best  suited for
 soil gas  analysis.

 FID analyzers—
      The FID is the most widely  used detector  for the analysis
 of  VOCs by  gas  chromatography.   It  is also one  of the  most
 widely used  for portable analyzers.  The Century Organic Vapor

                               170

-------
TABLE 5,1.  DESCRIPTION OF SELECTED PORTABLE ANALYZERS

»*..
Teckaleu*
rreciiloe
leBiltlvlty
•**poa«* Tla*
•••I*
•••pit He*
rower fopplf
Service Lift
(coat IBVOU*
vM/ch(r(t)
Mel|kt
Appro* laece
Cott
CoBUct
iBforaitiea
Century fyete.
OVA-IOI
ri*a* loBli*tioB
11 OS
1 pp.* (actkHe)
2 •*«
1-10,000 pp.*,
1 -1000 .000 ppav
lof.tr itkaic
1 L/ai«
DC
• kr*
14 Ik*
li.200
(417) 541-17)0
Analytic*! UftcMM«t
••0 Developenal, IM. rkoto«M, IM.
ri ioi
rkotOloltMtloB
ill ot n
0.2 ppajv (k«Bi*B«)

-
0-2000 pp.
0.27} L/ata
DC
4 kr*
J Ik*
11.100
(J14) »l-MOt
l*ch*r*ek
nv iBiffor
C*t«ljpti« Co*BV*ttoa
-
2.0 pp. (heiBao)
-
0-100
0-1000
0-10.000 pp.
2 L/ain
DC
• kr*
S Ik*
11, >00
(412) MJ-221J
Hired
II
Inrroree
-
-
-
".-X
-
DC
4 kr*
21 l»*
II2.SOO
(417) Ml-

-------
Analyzer (OVA) la  an  example of an FID Instrument.  The FID
responds to organic compounds  with a sensitivity of <1 ppmv for
methane  but does not  respond to inorganic components  in  air.
The Century OVA can be  operated continuously with a  sample
probe  collecting a sample  at  approximately  1 to 2 L/min  or  as a
GC with an isothermal  (ambient  or ice bath)  column and with gas
sampling valves.   The total  range can be scaled to 3 ranges or
as a single scale from  1  to 10,000 ppmv and higher  by using a
dilution  probe.   The OVA  is. powered by a battery pack and
contains a hydrogen tank  which supplies the fuel for the  FID.
The FID oxidant is ambient  air.

     An advantage of the OVA is  its  response to a wide range
of organic compounds  with sub-ppmv sensitivity.   It is  also
easy to  use, has a wide  range, and has optional chromatographic
capabilities.  The greatest disadvantage of the  OVA  is its
variable  response to different organic  compounds.  Several
studies  have been  made to  determine response  factors for
organic  species  (Brown,  et  al., 1980; Dubose  and Harris, 1981;
Dubose , et al.,  1981;  Willey,  et al., 1976).  The response
factors  were found to vary  from 0.2 to more than  100.   Table
5.2 shows how  the varying response  factors affect the
concentration  reported  by the OVA.  The large range in values
reported in Table 5.2  also indicates the OVA response is  highly
variable.  The  equal  carbon response typical  of  the FID is not
found  for  the OVA  when used in the non-chromatographic  mode.
However,  hydrocarbons display less  variation  than organic
compounds  containing heteroatoms  such as  oxygen or  nitrogen  as
shown in Table  5.2.   Using  an OVA for gasoline detection would
result in  smaller errors due  to calibration than other portable
analyzers.  The  OVA  has been used extensively  for gasoline
detection.

      This problem of calibrating the OVA can be  minimized by
using the  GC option  and by calibrating for  each individual
compound.  However,  environmental samples contain  such a large
number of  components that adequate chromatographic separation
cannot  be  obtained with the  ambient  temperature  OVA  column.
Another disadvantage  of the OVA  is the high sample  flow
required  (1  to 2 L/min).  For  soil-gas measurements  such  as
soil-core  or  ground-probe measurements, removing  soil  gas  at
this rate  would be difficult without  disturbing  the soil/gas
equilibrium or drawing  in  air  from above  the soil.   The
sensitivity . of flux chamber measurements will  be  limited by the
OVA since  the diluting sweep gas  flow  rate  needs  to  be equal  to
or greater than the OVA sample flow rate.

      Despite  the problem  mentioned  above,  the OVA has  been
used  successfully as  a  screening  tool  for  ground-probe
measurements, to correlate other  measurement techniques
 (Glacum,   et  al., 1983), to determine  when  steady-state

                               172

-------
     TABLE 5.2.  INSTRUMENT RESPONSE TO SELECTED ORGANIC COMPOUNDS
          F«aily/Compound
  Actual Concentration (ppnv)
      Required to Cause a
10,000 ppmv Instrument Response*
             OTA
       Alkanes/Ethane
               n-Butane
               n-Bezane

       Cyclo-alkanes/
         Cyclohexane

       Alkenes/Etbylene
               1-Butene
               1-Hexene

       Aromatics/Benzene
                 Toluene
                 o-Xylene

       Alcohols/Methanol
                1-Propanol
                1-Butanol

       Ac ids/Formic Acid
             Acetic Acid
             Butyric Acid

       Halogenated Conpounds/
         Chloroethane
          1,1-Dichloroethane
          1,1,2,2-Tetrachloro-
           ethane
    6,500 (4,400 to 15,800)
    5,000 (4,600 to 5,500)
    4,100 (3,800 to 4,500)
    4,700 (3,900 to 5,800)

    7,100 (6.300 to 8,200)
    5,600 (5,100 to 6,200)
    4,900 (3.900 to 6,600)

    2,900 (2,800 to 3,100)
    3,900 (3,600 to 4,300)
    4,300 (2,800 to 8,500)

   43.900 (36.100 to 56.000
    9,300 (7,700 to 11,600)
   14,400 (8.900 to 23.400)

  142,000 (106,000 to 198.000)
   16,400 (11,100 to 26,500)
    8,000 (3,800 to 31,400)
   53,800 (18,700 to 264,000)
    7,800 (6,200 to 10,200)

   78,900 (50,100 to 138,000)
*Both instruments were calibrated to methane at 8,000 ppmv
  (Dubose.et alt, 1981).
                                     173

-------
conditions have  been reached in  flux chamber experiments
(Radian Corporation,  1984), and to adjust the sweep-air  flow
rate to achieve the desired concentration range in flux chamber
experiments.  When using the OVA for these purposes, one  should
remember  that changes in concentration readings could be due  to
a change in the organic composition instead of a change  in the
actual concentration.  Likewise, a stable concentration reading
may be due to a change in the total  organic concentration
complemented with  a  change in composition.   However, in cases
where  the composition of the ground-water plume is homogeneous,
the OVA  should provide  a convenient  screening tool  of the
relative total organic concentration.

PID analyzers—
      The  PID has  become popular as a  detector for  portable
analyzers because of its high sensitivity to certain compounds,
its ease  of operation, and because no fuel gas or  flame  is
required.

      The  PID  ionizes compounds whose ionization potentials are
close  to or lower than the energy of  the  lamp  used.   Different
energy lamps are  available for analyzing different compounds
allowing for some selectivity.   With  the  10.2  eV lamp,  alkanes
have  little  or  no  response while  alkenes,  aromatics,
organosulfur compounds,  and carbonyl compounds  have  high
responses.  Alcohols, halogenated alkanes,  and most inorganic
gases have no response.   With a  11.7 eV lamp,  all  of the
selectivity for organic compounds except methane is lost.  The
reported sensitivities for benzene are 0.1 ppmv in air  and have
a range up to 2000 ppmv for benzene.   Another feature of the
PID is that it is essentially a non-destructive detector which
allows samples  to be  collected after passing through the
detector.

      The three  major  portable analyzers available are the
Photovac  10A10 portable GC, the HNU  photoionizer, and the AID
Model 580  PID system.  The Photovac  is designed as a GC  but can
be operated in a continuous mode with the  addition of a pump.
Samples are injected  into the column and are separated by using
a  dry air  carrier  gas.  Packed columns up to six  meters  in
length and capillary columns can be used in the portable GC.
Sensitivities down  to  0.1 ppbv for hydrogen  sulfide are quoted
by the manufacturer.  Photovac also now markets a portable PID
Instrument  called "TIP" for gas analysis which is designed  to
be used as  a continuous analyzer.

      The HNU photoionizer is a  continuous analyzer with the
lamp  contained  in the sample probe.  The small probe  volume
allows the detector to respond in  as little as 3 seconds and
requires only  a  small sample size.  The  sample flow  rate  is
approximately  0.5  L/min.  Three lamps  (9.5  eV, 10.2 eV, 11.7

                              174

-------
eV) are available.   The  AID  580  is  similar to the HNU except
the lamp  is  enclosed in the  unit and is not in the probe.  The
sample flow rate  is  approximately 0.5 L/min, and the exit  is
plumbed for  easy collection  of samples for laboratory analysis.

      The advantages of the  PID analyzers are the  greater
sensitivity for  many compounds compared to the  FID, the
selectivity  for  certain  classes  of  compounds, the
nondestructive nature of the PID, and the absence of fuel gas.
The disadvantages are the  highly variable  response factors for
each compound, matrix effects such as quenching of the response
because of  oxygen (Freeman, 1980), and the  low response or lack
of a response for some compounds such as the fluorocarbons .
The Photovac used with a column  could minimize some of these
problems.

      The  usefulness of the PID analyzers  is  in screening soil
vapors containing non-methene  hydrocarbons in the presence  of
high  levels of methane  or when ambient levels of  methane
Interfere  with monitoring  low levels of nonmethane pollutants.
They  can  be used  to screen  for compounds by class  such  as
aromatics  and organosulfur or for individual components  such  as
vinyl chloride.  As with  the FID analyzers, the values obtained
for the PID should  not be considered  absolute.   When the
response   is used as  a relative  value, the organic composition
and matrix must be  homogeneous  over the area or  time being
compared.

Other  analyzers-
      Two  other  types of  portable analyzers have been used  to
detect trace levels  of VOCs  in  air. The long-path-length  IR
analyzers have sensitivities in the low ppmv.  The MIRAN-IA gas
analyzer has a variable path length cell (range 0.75 m to  20  m)
so that a range from less  than  1 ppmv to 10,000 ppmv  can  be
analyzed.  The wavelength  is  also  variable and  allows for
moni-toring almost  any organic  component.  A microprocessor
model  can  be used to monitor up to 11  components.  The analyzer
is concentration-dependent and  has a sample flow rate  of  30
1/min.

      The  advantage of the MIRAN-IA  portable analyzer is the
ability to monitor many of the reactive organic compounds such
as phosgene, ethylene oxide,  and  formaldehydes which are
difficult  to sample and to analyze with other techniques.  The
disadvantage of the  MIRAN-IA is the large  cell volume which  is
required  for ppmv measurements.  A large volume of soil  gas  is
needed to obtain  that  kind of  sensitivity.  Such a volume  is
seldom available except with diluted flux  chamber measurements.
Total  organic  values are  relative at best, while individual
compounds  can be selectively monitored and quantitated  at the
chosen wavelength.  The  MIRAN-IA portable model weighs  32

                              175

-------
pounds  which  is  substantially mtjpe  than the  PID  and FID
analyzers.  In  addition, the stability and the ruggedness  of
the optical analyzers would be  expected to  be less than the FID
or PID analyzers.

     Another  commonly  used portable analyzer is the Bacharach
TLV Sniffer.  It is a hot-wire detector which measures  vapors
which can be catalytically combusted.  It has a range of 2 ppmv
to 10,000 ppmv  and a nominal sample flow rate of 2 L/min.   Like
the FID,  the  TLV  Sniffer responds similarly to all  organic
compounds and  has  variable response factors (Brown,  et  al.,
1980; Dubose,  et  al. ,  1981).  As a result, it has  the  same
advantages, disadvantages, and  uses as the FID analyzer  listed
above.

Conclusions—
     The  portable analyzers  designed for  ambient-air analysis
have limited uses as analyzers  of soil gases. Their  advantages
are the elimination  of  sample collection and transportion, and
theimmediate availability of results.  The major disadvantages
are variable response  to different classes of compounds and
large sample volume  requirements.  Having a detection  limit
of 1 ppmv can  limit  use in some cases. The uses for portable
analyzers  include screening wells  and   ground  probes  to
determine if more  accurate and expensive  sampling  efforts are
needed  and to optimize  flux  chamber  conditions  before
collecting a sample for detailed analysis.

Remote  analysis—
      In almost all  cases,  to  accurately  determine the amount
or composition  of organic compounds in  soil gas,  a  sample has
to be collected  and  taken to  a laboratory  where conditions are
stable  enough to support the level of sophistication  required
to analyze the sample.   The laboratory could be a mobile field
lab on  the site or a modern analytical lab  on the other side  of
the country.   With either situation,  a representative sample
must be  collected, and its Integrity must be maintained until
it  can  be analyzed.   The sampling method  must be  compatable
with the analytical method.   If  the  analysis  is not  very
sensitive, a  large sample must be collected.  If  the sample
must be  sent  across the country, the container must  be Inert
and  rugged.   The following sections  will discuss first the
sample  collection and storage methods and  then the  analytical
methods which  are  used  or could  be  used for soil-gas
measurement.

Sample  collection—
      Sample collection methods  of  VOCs  in gases are divided
into two classes:
                               176

-------
    o  Adsorbent methods  where  the  gas is passed through a
       solid adsorbent which removes the VOCs  from  the
       inorganic gas matrix.

    o  Whole  air  methods where the entire sample is  placed in
       a container and  is transported  to  the lab;

      Adsorbent methods--The  adsorbent  method is attractive
because it concentrates  the  components  of  interest and removes
many of the components  known  to  add  to  the instability of the
sample and which  interfere with  the  sample analysis.   The
adsorbent  containers  are  generally  small  and can be easily
transported to and  from  the field.   The limitations of  the
adsorbent  methods  are irreversible adsorption, incomplete
adsorption  (breakthrough), and  artifact  formation.
Irreversible adsorption  occurs when adsorbed components cannot
be completely desorbed.  Ideally  for the analysis  of VOCs,
adsorbents  should be  thermally  desorbed  since  solvent
desorption increases  artifacts,  dilutes the sample components,
and  interferes with the  analysis.  Incomplete adsorption is
characterized as a breakthrough and results  in the loss of  the
more volatile sample components.   Artifact formation  can occur
during thermal desorption  or from  reaction with the adsorbent
material  and the  sample.  All three of these possible problems
must  be  fully investigated  during the  sampling method
validation, and a strict quality  control plan must be followed
to Insure the method  is  performing  within  acceptable limits

      The adsorbent  materials most often  used for  sampling VOCs
are activated charcoal  and porous polymers such as Tenax.
Other adsorbents  which  have been used  are  molecular sieves,
silica gel, and activated  alumina.   Charcoal  has been used
extensively, in industrial  hygiene application for monitoring
VOCs, and NIOSH has  published  a standard method  for  charcoal
(White,  et  al.,  1970; NIOSH, 197*).  Charcoal has a  high
adsorbent efficiency for all  organic compounds  but  requires
solvent desorption.  Desorption efficiencies can  vary with the
lot of the manufacturer  (Saalwaechter, et al.,  1977).  Carbon
disulfide  is used  as  the solvent  and interferes  with  the
determination of components more volatile  than n-butane.  Since
the solvent cannot  be  concentrated without volatile  loss,  the
sensitivity of the method  is limited.   The use  of charcoal as
an adsorbent for soil-gas collection  has been reported by
Colenutt  and Davies  (1980) and Karlmi  (1983).   A  patented
method using a Curie-point wire coated with activated charcoal
is used by Petrex Corporation  (Bisque, 1983).   Instead of
passing the sample  through an adsorbent  bed, Petrex allows the
sample to diffuse into the coated wire  over 3 to 15 days.   The
wire  is then analyzed  by Curie-point mass  spectrometry.  The
technique is  very  sensitive and does  not affect  the  gas/soil

                              177

-------
equilibrium.  Absolute quantitation determined  by  using this
technique  is difficult, and long sampling times are required to
obtain  high sensitivities.

     The  porous polymer adsorbent Tenax-GC  has  been used
extensively in ambient air measurements.  Several  good reports
have characterized  the breakthrough  volumes for  a large number
of compounds, the affect of moisture and  sample  flow,  of
desorption efficiency, and of  artifact formation (Brown and
Purnell, 1979; Krost, 1982).  Tenax  has been found  effective
for  most  organic  compounds  except  for hydrocarbons  and
halocarbons that  are more volatile  than n-hexane,  low  molecular
weight alcohols,  amines, and  aldehydes.   Compounds  can
effectively be thermally desorbed to obtain  sub ppbv  detection
limits.  Varying amounts of artifacts have  been reported for
Tenax;  some of  them are believed to be due  to  oxidation  of
Tenax and to improper cleaning of the adsorbent.  Because of
the artifact peaks and the inability to  trap the  less volatile
compounds, Tenax cannot be used to obtain total VOC values, but
it would be an ideal method to  monitor individual components
when it is shown that they are not affected by these problems.
Tenax-GC has been used to analyze soil gas  from ground probes.
Swallow and Gschwend (1983) monitored benzene,  toluene, and
trichloroethylene in ambient air and at  two  depths in the soil
above the ground water.  They  reported values  as low as 0.2
ng/L with  a precision of ±15 percent.

     Whole air  methods— Whole air methods of analysis of VOCs
in soil  gas have been used  extensively.   The method could be
divided into two subcategories.  One  involves  removing the soil
gas from the soil and transporting  it  to the lab,  and  the  other
involves transporting the soil  with the gas.  For  collection of
soil gas,  three  different containers can be used:

    o   plastic  bags made of Tedlar or Teflon,

    o   passivated stainless-steel  canisters and syringes, and

    o   glass syringes.

When evaluating which container to use, several  factors should
be considered:

    o   Sample hold time and stability over the hold time

    o   Sample  handling  and shipping and the  durability of the
        container.

    o   Sample  container cleaning procedures and memory
        effects.
                               178

-------
      The use of plastic bags for air  samples  Is  Inexpensive
and convenient.  However,  several studies have been made using
Tedlar and Teflon  bags  (Sella, et al.,  1976; Lonneman,  et al.,
1981) and have found that  contamination was significant If the
bags were exposed  to  light.  Other problems found with  plastic
bags are  permeation of  compounds Into and out of  the  bags
during storage  and sample  leakage encountered during handling
and  transportation.   Plastic  bags are not recommended  for
soil-gas  analysis  unless the  storage times are less than 4-8
hours and the concentrations are high.

      Stainless-steel canisters with a passivated interior
surface have  been used  for a wide variety of VOC measurements
Including soil-gas  monitoring.  The canisters have  been
described by  Harsch (1980),  and hydrocarbon, halocarbons,  and
carbonyl  compounds have been found to be stable for as  long as
three  weeks  (Oliver,  et al.,  1985;  Westberg,  et al.,  1982).
The  canister can  be  pressurized to hold more than 20  L of
sample and can be shipped  easily without any sample loss.   The
canisters  can  easily  be cleaned  since  they withstand
temperatures up  to 150°C and  can be evacuated to  very  low
pressures.  Because  no light  can enter the canisters,  the
possibility  of  photochemical  reactions is minimized.   The
passivation  process used to produce canisters has been  applied
to the production  of  stainless-steel  syringes (Scientific
Instrument Specialist,  Inc.,  Moscow, Idaho).   These syringes
would have the same advantages as the canisters plus the added
value  of  easy sample  collection.   A  soil-gas  sample could be
drawn slowly  when the  syringe  is used with the disruption of
the  soil-gas equilibrium being  minimized.  Slowly drawing a
small  sample when  a canister  is used  requires vacuum  flow
regulation.    Sample  dilution is required to obtain samples
smaller than  canister  volume  (generally greater than 0.5  L).

      There  are  two  methods  for  collecting a sample  in
canisters.   One  involves using  a pump  to fill the canister.
This technique  requires a  clean,  inert pump and enough sample
to purge the  canister  before filling it.  The other method is
to first  evacuate  the canister  and  then to bring the canister
to atmospheric pressure with sample.  The first technique would
not  be compatible  with  most  soil gas techniques except  flux
chamber methods.  The  second method has been found to work  for
both  ground  probe and  flux  chamber  methods  (Radian
Corporatlon-S,  1984; Crow, et al. , 1985;  Radian Corporation-T ,
 1984).  By  using a vacuum  flow regulator,  the rate of sampling
 can be  controlled to minimize both  soil/gas equilibrium
 disturbances and migration  of atmospheric gas into the sampler.
 Detection limits of  1 ppbv have been obtained by using  the
 canister  method.
                              179

-------
      Glass  containers such as syringes  (Radian  Corporation,
1981;  Thorburn, et al. .  1979;  Marrin.  et al., 1981;  Radian
Corporation-S, 1981; Crow,  et al., 1985;  Radian Corporation-!,
1981;  Wood, et al.,  1980;  Dowdell, et al.,  1972) and evacuated
flasks (Thorburn,  et  al.,  1979) have been  used to collect  soil-
gas samples.  Glass  has been  found  to  be   inert  and
noncontaminating  for most  organic compounds.   Most glass
syringes and flasks  require  a Teflon valve or  seal to be
gas-tight.   However,  the Teflon can be a source of
contamination.  Ground-glass Joints have  been used; however,
they are not completely gas-tight, and samples cannot be stored
for any length of  time.   Because of the fragile  nature of
glass, it cannot be  shipped very easily and is used mostly for
on-site  analyses.  Since glass can transmit  light,   it has  been
suggested  that samples  could be affected by photochemical
reactions and should be  kept  out of direct  sunlight.

      Another method for sampling soil  gases is  to pump the
soil gas directly into a sample loop for injection  into the GC
(Weeks, et al., 1982).   This  requires  a  large volume  of  soil
gas from a vacuum  pump.   This would disrupt the  soil/gas
equilibrium and can  pull  in atmospheric  gas around the ground
probe.

      The  collection  of soil cores is described  in  chapter 1
(Head-space Measurement) and  is an alternative to collection of
soil gas alone.  The soil cores  are  sealed and are sent  to a
laboratory where soil  gas can  be removed and analyzed.   The
advantages of this technique are that the soil itself can  also
be  analyzed by other physical  or  chemical methods,  the
technique requires relatively  little expertise or equipment,
the samples appear to have  a  long  shelf life,  and  a  more
accurate soil gas  measurement can be  obtained under laboratory
conditions than in  the field.  The disadvantage of  this method.
is  loss  of  volatiles during  the coring of the sample and
transport.  In some cases, obtaining  a representative sample
may be difficult.    The sample can also be  affected by
biological activity  in soil.

Sample Analysis—
      Analysis of  soil-gas samples is performed  in  either an
on-site  mobile laboratory  or a remote laboratory.   The
instrumentation  of the  mobile laboratory is  limited to
equipment which is easily  set up,  rugged, and  requires a
minimum amount of  power  and support  equipment.  Methods  which
require  subambient  temperature  programming,  cryogenic
concentration, and detectors  with vacuum systems  such  as  mass
spectrometers are generally excluded from  mobile laboratories.
Theoretically, any  instrument  discussed here   can  be  made
mobile,  but experience has shown that it is only cost effective


                              180

-------
 to  use  an instrument  specifically designed for  the  mobile
 laboratory.

      Mobile  laboratory  instruments--The simplest  mobile
 laboratory Instruments are the portable analyzers  which  have
 chromatographlc options.  This  includes the Photovac with a PIO
 and the  Century OVA with a FID.  These have simple injection
 systems,  either  a gas sampling valve.(OVA) or  a  syringe
 Injection  port  (Photovac).   The  Photovac has an  internal  column
 space which  can  hold a 1/8" column up to 6 meters long  or  a
 25 meter capillary  column.   There  is no temperature control  for
 the column.  The  OVA has  an external column which could be of
 any size and  is at  ambient  temperature.  These  instruments  are
 easy to transport  and use  and,  in the case  of the Photovac,  can
 be very sensitive with ppbv detection  limits.   The added
 advantage  of the  chromatographic  capabilities  is that
 Individual compounds may  be monitored and quantitated with
 accurate  response  factors.   Because of the complex nature of
 most environmental  samples  including soil gas, the uncontrolled
 temperature of the column in these portable analyzers  does  not
 allow the  high resolution separation needed to  identify  and to
 quantitate  for  most components accurately.  Reproducible
 retention  times are difficult to achieve in the field because
 of temperature fluctuations, and separation of components  with
 a wide range of volatilities is difficult without temperature
 programming.  These analyzers can be useful if  the  sample
 contains large amounts of easy to  separate  contaminants such as
 those found in a  chemical spill or  if only a  volatility range
 is needed  such as total C2-C4 compounds.  Gasoline samples  are
 often characterized by comparing the function of the total
 organic  compounds in volatility  range.

     A  step  above the  portable  GCs  are the field GCs.   These
 are small, sturdy GCs that  contain temperature-controlled ovens
and a variety  of  injectors  and detectors.  Some of the modules
which have been  used for  soil  gas are the Varian Mobel 6000
 (Marrin,  et al.,  1984),  Shimadzu  Model GC-Mini 2 (Radian
Corporation,  1984; Radian Corporation-S , 1984;  Radian
Corporation-T, 1984),  Carle AGC (Wood, et al., 1980), and  the
HHU GC 301 (Radian Corporation-S, 1984).   Table 5.3  contains
the descriptions  given  by the manufacturers of selected
portable GCs.  This list is not  exhaustive  and does not mean to
exclude other  GCs which could Just  as easily be used  in  the
field.   The components of the GCs such as Injection ports  and
detectors can  usually  be  selected.  In some cases, temperature
programming is available.  Because these GCs are smaller than
standard  models, the number  of  columns and  their lengths  can be
limited.  The  best results  are obtained  with an instrument
 having  a  heated  gas sampling valve  for injection  of the  gas
 samples.  The detectors most commonly found are the FID, PID,
 and the ECD.   The  HNU  GC  301 is unique in that it has both an

                               181

-------
                 TABLE 5.3.  DESCRIPTION OF  SELECTED PORTABLE GAS CHROMATOGRAPHS

Hod* I
Dttcctori
Free it ion
Stnt itivUy
U.«.
Topiritur*
ProfracBiof
Carrier Cat
Height
Co*t
C3
K)
Coo tact
Inforaatioo
SbiMdau HMU
CC-Nini 2 301
no riD/rio
llOI ilOl
O.I ppvv (bcnitnc) 0.01 pp«» (btnitnt)
O.I ppav-1,000 ppav 0.01-10,000 ppav
Y«i* Tti*
Hitro|«o Hitro|*n
31 Ibt 25 Ibt
13.900 $7.300
™ ™
(713) 170-2495 (617) 964-6690
rhotovtc
10)10
no
-
O.I ppb (btBt«n«)
1 ppbv-100 pp«»
Ho
Air
26 lb(
J7.000
InitruMnc can ba
uaad aa portabl*
THC anal]riir
(516) 351-5809
S.....
Sctntor
ICD/AID/riD*
-
0.01 pp> b*ni«n* (AID)
1 ordara of •a|nltu<>*
Ho
Arfoa
40 Iba
SI 3. 000
trie* iicludaa •icro-
procataor, racordar,
•nd ttaioing court*
(201) 9*5-369*

910
r»
-
AID



O.I ppa (b*ni«n«)
0.1-1.
Ho
Air
11 Ib*
$4.600
-
000 pp.





(21)) 268-1181
•Opt ion*1

-------
FID and  a  PID  which can be operated either separately  or In
series.   The  PID can be added  to other instruments  if  there is
room.  The ECD is extremely sensitive and is  selective for most
halogenated compounds with detection limits of 1.0 ppbv with no
concentration  of  the sample  (Marrin, et al.,  1984).   In one
study, both a cryogenic concentration step and an ECD were used
to analyze trichlorofluoromethane and dlchlorodifluoromethane
below the 100 pptv (parts per  trillion) level  (Weeks,  et al. ,
1982).

      The chromatography  columns chosen for  field  analysis
depend on the types  of  compounds and on their  volatility.  A
nonpolar methyl silicon liquid phase, such  as  SE-30, has been
used  extensively  for hydrocarbons and halocarbons.  Columns
packed in stainless  steel  have been used more  extensively for
field use than those packed  in glass and capillary  columns.
The packed  columns have  a  higher  capacity  than capillary
columns,  and  large volumes of  samples can be Injected  without
the need for cryogenic  concentration steps.  The lower carrier
gas-flow rates of capillary  columns would require  almost a
minute  to inject a 1  mL sample and would create very  broad
peaks  while a packed  column at a typical carrier flow rate of
30 mL/min would take 2  seconds to inject 1  mL.   Stainless steel
is chosen over glass  because it is easier to Install  and does
not break during transport or  use.

      The chromatographic  data can be acquired  in  the field on
stripchart recorders, on Integrators, or on a  portable computer.
Most of  the  portable GCs can be supplied with  small stripchart
recorders.  Small integrators  such as the HP 3390 can be  useful
to store calibration information and to integrate peak areas or
heights.  A  portable computer  with chromatographic  software
becomes  very useful  when the raw data needs  to  be  stored, when
more than one detector  is being used, and when the data from
different analyses or detectors needs to be compared.

      Off-site  laboratory instruments — When positive
identification  is  needed, when very low detection limits are
required, when difficult sample matrices are encountered, or
when  environmental  conditions  prohibit an on-site analysis,
soil gas  samples will have to  be sent to an off-site laboratory.
The off-site  lab may  have the  instrumentation discussed  above,
but generally  a higher  sophistication of instrumentation is
used.

      If  low  detection  limits are required,  the samples can be
collected on  a solid  adsorbent or in a stainless-steel  canister
and sent to  a lab  where  the  sample  can be  cryogenica11y
concentrated.  The  organic components of a  larger volume  of
soil  gas can be  trapped at  cryogenic temperatures  by  using
liquid oxygen, liquid argon, or  dry ice/acetone  baths.  To trap

                               183

-------
large volumes of  sample,  water needs to be removed from  the
sample.   Permapure  driers,  potassium carbonate  (K2C03)
(Colenutt, et al.,  1980),  and  magnesium dichlorate [MgCClOii^]
(Schmidt,  1983) have been used to remove water from air samples.
Recoveries  of halogenated hydrocarbons are good when these
drying methods are used,  but  for  polar compounds  such  as
alcohols and aldehydes,  variable recoveries have been found.
For injection of volumes  smaller than approximately 250 mL,  no
drying  method is  required.  The sample is trapped on glass
beads  or glass wool packed in 1/8 in. stainless-steel  tubing.
The frozen  samples are  thermally  desorbed  by  using heat
cartridges and boiling water and are injected onto a column  by
switching  a gas-sampling  valve.

      Adsorbent  samples are injected by placing the cartridges
into  a  block heater and  thermally  desorbing at  a  high
temperature into  a flow of an  inert gas such as helium.  Since
most adsorbents do  not  have an affinity  for water, no drying
method is required.  For  high-resolution analysis of adsorbent
samples, a second  concentration step is needed  to  keep  the
desorbed  effluent from  "broadening" before  it reaches  the
column.  A cryogenic trap or a  smaller secondary  adsorbent  is
generally  used to focus  the sample before analysis.

      The chromatographic separation  available  in  the
laboratory can be  superior to  that in the  field.  The use  of
subambient temperature  programming and capillary columns allows
most volatile compounds  to be separated.  The chromatographic
condition and columns  used   have been  described in  several
papers (Krost, et al., 1982; Westberg, et al.,  1982; Jeltes,  et
al. ,  1977; Cox, et al.,  1982; Westberg, et al.,  1984).  In most
analyses,  a non-polar methyl-silicon, fused-silica  capillary
column has been  used.   These  columns can  separate up to  300
different compounds.    The oven  temperature  is usually
programmed  from  subambient   temperatures to over 100°C  to
separate  the full range  of VOC found in environmental samples.

      The detection methods  used  for  analysis of soil-gas
samples  include:

    o   Flame  ionization detector  (FID)  for  the  full range of
        organic compounds;

    o   Photoio n i z ation  detector  (PID) for the  aromatic
        hydrocarbons and sulfur  species;

    o   Electron  capture detection  (ECD)  for  selective
        detection of halogenated hydrocarbons;

    o   Hall  Electrolytic  Conductivity detector (HECD)  for the
        specific detection  of  halogenated  species,  nitrogen

                               184

-------
        elements  containing   organics,  or sulfur containing
        species; and

    o   The flame  photometric detector  (FPD) for sulfur and
        phosphorus compounds.

Good  reviews have been written describing the merits and
limitations of these detectors  (Farwell, et  al., 1981; Hill, et
al.,  1982; Sevick,  1976).  In addition  to these standard GC
detectors, several other configurations  have  been  used
specifically for soil gas.  These  include the combinations of
the FID-PID-HECD described by Earp and Cox (1982)  and used for
soil-gas measurements by the Radian Corporation (1981, S-1981,
T-1981).   The combination of the FID  and PID  provides
qualitative information to  aid in the  identification of
hydrocarbons while the HECD can selectively  detect  and
quantitate halogenated  hydrocarbons.  The multiple detector
method combined with high resolution  chromatography  is a
powerful  tool providing  accurate quantitation for  all
hydrocarbon and halogenated species with a  high  degree of
confidence in the  compound  identification.  Unidentified
compounds can be  easily classified as an  alkane, alkene,
aromatic, or halogenated hydrocarbon from  the responses of the
detector.

      A  new technique which has been applied to the analysis of
hazardous  waste  Is GC  with  Fourier  transform  infrared
spectrometry (GC/FTIR)  (Shafer,  et  al., 1981).   The FTIR can
quickly, with good sensitivity,  scan the  infrared spectrum of
an eluting peak which can then  be  quantitated by using an FID
or other GC detector.  The Infrared  spectrum,  after some  data
manipulation, can  be  compared to  a spectral  library  to  make
identifications.  A  similar technique which has  been  used
extensively in environm.ental  analysis is  GC mass spectrometry
(GC/HS).  A mass spectrometer is used to obtain a mass spectrum
of the eluting GC peaks.  A mass spectrum can sometimes make
position identifications of unknown compounds,  and, in  the
single ion mode, it  is  extremely sensitive and  selective for
the compound of interest.  A good review of  the GC/HS technique
has been published (tenNoever,  et al., 1979). The  cost of these
two techniques, compared  to the possible benefits,  has been the
main factor  that has  limited  their  use  in  soil-gas
measurements.

      Concluslon--Hodern analytical laboratory  methods  have
been  developed to  the point where  VOCs in  soil gas can be
separated and quantitated in the sub-ppbv  concentration range,
and identification can be made with a high level of confidence.
The   researcher's job  is  to determine  what  level  of
sophistication is necessary.  The use of portable analyzers and
field GCs  to screen  the sample  can  often  provide  the answers

                              185

-------
needed to make this  decision.   Areas where better methods are
needed  include the collection and determination of oxygenated
compounds,  of sulfur  compounds, and of nitrogen containing
compounds.  Determinations of  these compounds are difficult
because of their  polar  nature compared to the non-polar  nature
of the hydrocarbons  and halogenated hydrocarbons.   Drastic
measures  are generally required to obtain a reasonably low
detection limit  for these compounds  such as the  field
collection of sulfur species that makes use of the deactivated
cryogenic  traps described by Farwell (1979).

Quality Assurance/Quality Control  (QA/QC)

      For  an  analytical procedure  to  have any value, a QA/QC
program must be designed  so that the quality of the data is
defined (i.e., confidence limits) and so that assurance  exists
that  the method is performing at that level.  Most analytical
method QA/QC plans  contain a  calibration step, a linearity
check,  a QC standard  analysis,  a blank analysis, a duplicate
analysis, and an audit  sample or  inter labor at ory  sample
analysis.  A typical calibration and quality control schedule
for  various  analytical systems is given in Table 5.H.  The
QA/QC plan should address the sampling method, the analytical
method,  and  the  data  reduction and  reporting steps.  The
acceptance criteria chosen  are limited by the  available
analytical and sampling techniques performance, but they  should
be set  by  the data requirements needed to make the necessary
decisions.  These  requirements should be determined for each
project during the initial Data Quality Objective  planning
phase of the project before any data is acquired.

Method  Calibration—
      Calibration methods  vary depending on the instrument used
and  the level of  confidence required.  The portable organic
analyzers have  single calibration capabilities which  limit
their use when accurate values are  needed.   Most  analyzer
methods  use  a  single  component  standard  at several
concentrations such as  methane  or hexane for FID and benzene
for  PID analyzers.  Since  none of the analyzer response factors
are universal for VOCs,  calibration procedures using a  single
component do not  provide  accurate values for the entire range
of VOC  compounds.  The values obtained will also  vary greatly
in keeping with  the compound  that is used to calibrate the
analyzer.

      Calibrating GCs  can be  more specific, and the  actual
method or  standard used depends on the detector.  For field GCs
with FIDs,  a single component standard such as propane or
hexane can be used to calibrate the instrument.   If the VOCs
can  be separated  into carbon number class, concentration  can be
calculated by assuming  an equal carbon response  for the FID.

                               186

-------
  TABLE 5.4.   SUMMARY OF SUGGESTED CALIBRATION AND QUALITY CONTROL REQUIREMENTS  FOR ANALYTICAL SYSTEMS
O)
Type of
InttruMent
far I (bit TUC
Analyier







Port cblt Cat
ChroM«tO|raph


Type ol
Detector
no I)
2)
J)
4)
no I)
2)
1)
4)
no i)
a)
i)
Type of
Cellbration/qc Tot
Multipoint calibration
(taro plu* thro up-
acale concentration*)
tero/epan callbratloa
Control ample
aaalyale
Drift chock
Multipoint calibration
(aaro plua tbraa up*
acala concentration*)
lare/apan calibration
Control aaMpla
analyeia
Drift check
Multipoint calibration
(aero plua three up-
icala conctntratlona)
taro/apan calibration
Control Maple
analyil*
Praouetty
At atart of
program
Daily
Daily, prior
to teatini
Dally, at
conclusion
of taatlni
At atart oC
proiram
Dally
Dally( prior
to teatini
Daily, at
cone Itia ion
of teatini
At atart of
proiram
Dally
Daily, prior
to taatini
Cae
Itaadard(o)
Methane or
other aliphatic
coop auaa1
UUP Air or •./
•ethane
Nathan*
Methane
•enaane or
other aromatic
compound
•anaan* or
other aromatic
compound
•enaene or
other eroBatic
compound
leniene or
other aromatic
compound
lenaene or
toluene
IMP air or »,/
•ethane
Nniene
Ace apt ante Criteria
Correlation coefficient
tf>.**}
letponae (actor agreement
vlthln IJOX of aea* U
lor nultlpolnt calibration
Heaanrae' concentration
ulthin ilOt of certified
concentration)
Drift SJOt of the input
value
Correletion coefficient
2,0.**}
loaponao factor aireraant
vlthl* ±101 of Man U
for Multipoint calibration
Meaaured concentration
vlthla ±101 of certified
concentration
Drift 1201 of the Input
value
Correlation coeffieiant
^0.**}
•eaponao factor a|reement
vitbin 120X of Man HP
for Multipoint calibration
Measured concentration
vithin llOS of certified
concentration
Corrective Action
Repeat Multipoint calibration
after checklni calibration dilu-
tion ayatem
1) Repeat laro/apan calibration
2) If atlll unacceptable, repeat
oultlpoUt calibration
1) Repeat lero/apan calibration
2) Repeat control aaapt* analy-
ala
1) rial day'* data aa eueit ion-
able
2) Repair or diacontinue uae of
analyaar
Repeet Multipoint calibration
after cheeklni calibration dilu-
tion ayate*
1) Repeat aero/apan calibration
2) If atlll unacceptable, repeat
nultlpolnt calibration
1) Repeat tero/apan calibration
2) Repeat control aaaple eaaly-
al*
1) Pla| day'a data aa fueatlon-
•ble
2) Repair or diacontinue uae of
enalyaar
Repeat Multipoint calibration
after checklni calibration dilu-
tion ayatem
1) Repeat aero/apan calibration
2) If attll unacceptable, repeat
Multipoint calibration
1) Repeat aero/apen calibration
2) Repeat control aanple analy-
aia
                                                                                                (Continued)

-------
                                                TABLE  5,4.(continued)
oo
00
Type of
Inatrument
fortabta Cea
Chromatograph
(Continued)










Type of Type of
Detector Cettbratlon/QC Teet
FID 4) Drift check
i) Retention time check*
4) Analytical blenka
7) templing ayatem blanka
() Duplicate lamplea
») Control point aample*
10) lick) round eamplea
riD I) Multipoint calibration
daro plu* thrae up-
•cale concantratlona)
2) Zaro/apan calibration
3) Control •••pie
•nalyai*
4) Drift check
Frequency
Dell*, at
conclusion
of letting
Daily
Daily
Daily (plu*
after very
high samplea)
101 of eam-
pllng polnta
(minimum)
After every
10 iimple*
or once per
day, which-
ever 1*
greater
One aample
per day
At atart of
program
Dally
Dally, prior
to teitlng
Daily, at
conclusion
of teatlng
Caa
Itanderd(a)
Renaene
Rente** or
toluene
UHF air or N{
temple gat
temple ga*
temple gaa
taaple gee
Renaene or
toluene
UHF air or »,/
•ethane
Rentene
Renaene
Acceptance Criteria
Drift iJOI of the input
velue
None
Heeeured concentration iJI
of the Inatrument *p*n
value
Meaaured concentration ill.
of the initrunent (pa*
value
Nonej provide* a meaeure
of total aampling vari-
ability
•one; provide* a meaaure
of temporal variability
Nonet provldee e meaaure
of background concentra-
tion
Correlation coefficient
Reaponae factor agreement
within ilOI of mean RF
for multipoint calibration
Neeaured concentretlon
within ilOX of certified
concentretlon
Drift i20X of the Input
value
Corrective Act lorn
1) Flag day 'a data a* queat lorn-
able
2) Repair or dlacontlnue use of
enelyier
None
Clean/replece eyitem component a
until acceptable blank can be
obtained
Clean/replace ayatem component*
until acceptable blank cen be
obteined
None
None
None
Repeat multipoint calibration
after checking calibration dilu-
tion ayatem
1) Repeat aoro/epam calibration
2) If ntllt unacceptable, repeat
multipoint calibration
1) Repeat aero/span calibration
2) Repeat control aample analy-
eia
1) Flag day'a data aa question-
able
2) Repnlr or discontinue HII of
enalyaer
                                                                                                          (Continued)

-------
                                              TABLE  5,4,(continued)
00
1C
type of
tnatruaiant
Portable Cea
ChroMCograpfc
(Continued)




Ofl-lltt Gaa
Chroaiatograph



Type ol
Detector
no 5)
t)
7)
1)
*)
10)
no i)
2)
})
0
Type el
Caltkratlon/QC Tt«t
Intention tin* checke
Analytical blenke
•••pi Ing ayetem kltnkt
Duplicate ttmfltt
Control point •••»!••
Background ttmfltt
Multipoint calibration
(uro plui three up-
tcilt concantrationa)
lin|l* point callbre-
tlan cktck
Retention tlM ehtck
Control aeaple anelyele
ttti»tncr .
Daily
Pally
Dally (pitta
attar vary
M|h aaaplaa)
101 «f aaai-
pllna polnta
Ulnlw.)
Altar ovary
10 aaaplaa
or one a par
dty. whlch-
avar la
iraatar
On a aanplo
par 4ay
l/Mntb
Dally, prior
to laapla
analyaaa
Dally, prior
to aaapla
analyaaa
Dally, prior
to aaaplo
analyaaa
6aa
•t«M«H(a)
•aaaana or
taUaaa
HUT air or »2
(••pi a gaa
laaplo gaa
laaplo gaa
laapla gaa
Prepana/kaiana
Fropana/kaiaia
Mdttlcoaponant
ttto4tt*
laapla gaa
Atcaptanca Crltarla
•ona
Haaavra4 eoncantratlon iJX
ol tka InttruMnt apan
valua
H«Mur«4 cone ant ration ill
of tko Inatrunant apan
*al«o
•ono| pro«!4aa a maaaura
ol total aampllng vari-
ability
•ona| provlfoa a Btatura
of traporal variability ' '
•onai provl«*aa • Maaura
ol background concintra-
tlon
Corratatlon coaff ielant
29. tM
•aaponaa factor agraaaont
•Itkln 1201 of wit racant
•varaga Ut for aultlpoint
calibrationo
•graaawnt vitb proaatab-
llahad ralatlva ratantlon
tUaa
1) Cor r act Hanllflcatlon
of (OS of coaiponontt
2) for 90S of cooiponanta,
•aaaurad concantrationa
within tJOl of actual
concantrationa
Corractlva Action
•oat
Claan/raplaea ayatta cevponanta
until accaptabla blank can bo
obtainad
Claan/rapUco ayataai covponanta
until accaptabta blank can ba
obtainad
Mono
•ona
•ono



Rapaat linearity ckack
Rapaat alngta point calibration
Adjuat CC eondltlena
IT ckack
•apaat control aaapla
•nd ropoat
analyala
                                                                                                         (Contlnuad)

-------
                                               TABLE 5.4.(continued)
vo
o
Type of
Inatruaent
Off-Site Caa
ChroBatograph
(Continued)







Type of Type of
Detector Callbration/qC Teat
FID )) Duplicate analyaee
i) Blank analyaia
FID I) Multipoint calibration
(tero plua three up-
acale concentrationa)
2) Single point calibra-
tion check
3) Retention time check
*) Control eanola analyaia
}) Duplicate analyaei _
6) Slank analyala
Frequency
Hlninuni of
lOt of •••-
plee (all
duplicate
caniater
aaaplea will
be analysed
In duplicate)
Deily, prior
to ample
analyaia
l/eionth
Daily, prior
to ample
analyiet
Daily, prior
to ample
analyaae
Daily, prior
to ample
analyaea
NinlMa] of
lOt of •••-
plea (all
duplicate
canlater
aaaplaa will
be analyaed
in duplicate)
Daily, prior
to aaaiple
anatyaia
Caa
Stendard(a)
tmple gee
On air or Bj
Fropane/hexano
Propeoe/heiane
Hultlcomponent
atandard
Imp la gee
Saaiple gaa
UNF nlr or »j
Acceptance Criteria
C* iJOI for ten aujor
eaBple coaponente
Total 120 ppbv-C
Correlation coefficient
20.995
leaponee factor agrenent
vlthin ±20t of Boat recent
average RFi for multipoint
calibratlona
Agreeaent vith preeatab-
liahed relative retention
tlena
1) Correct Identification
of 90t of coBponanta
2) For 90t of coaponente,
aeaaured concentretlone
vlthin 1301 of actual
concentrationa
Cf 1201 for ten aajor
aaaple cOBponente
Total 120 ppbv-C
Corrective Action
Repeet •••pie annlyaia
1) Clean eyeteei
2) Repeet blank analyaia
Repeat linearity check
Repeat eingle point calibration
Adjuat CC condition* and repeat
RT check
Repeet control eaaple analyala
Repeat ample anetyelo
1) Cleen eyatam
2) Repeat blank analyala
                                                                                                       (Continued)

-------
TABLE 5.4.(continued)
Type of
Instrument
Off-lite Caa
Chrra*tO|raph
(Continued)




Type of Type of
Detector C*libr*t lon/QC Te*t
MtCD I) Quontitativ* etendord
2) Retention time check
3) Control eemple enalyaia
4) Duplicate analyaaa
)) Hank anelyai*
Cee
•regency Itandard(a)
Daily, prior Hultlco»poe«nt
to (••pie ataidard
analyaie
Daily, prior Nvttlcoapomemt
to •••pi* etandard
analyeei
Deity, prior Sample gee
ta iimple
anelyee*
Minimum of temple |a*
IDS •••pi**
(ell dupli-
cate canltter
•aeiplea will
be analysed
in duptlcote)
Daily, prior ONP air or •,
to aaeiple
•nalyaea
Acceptance Criterie
lUaponie fector etreement
• Itbln ±J« of tnree-de*
rolling ewen RT* for «ll
component a
•one) «lll provide baale
for comperieon of rtD/PID
retulte to NECD reenlte
1) Correct identification
of all eomponemta
2) For MS of componento,
m***«red concentration)
vlthln 130X of actual
cone entrat lone
C* 120 for ten major
•••pie componenta
Total iJO ppbv-C
Corrective Action
Rapeet callbritlon
....
Rcpeet control aompl. onelyeie
Repeat (ample anal yd*
1) Clean eyeteei
2) Repeat blank aoalyda

-------
Otherwise, the concentration values are  reported in the  units
of volume ratios  of  carbon (i.e., ppmv-C).  For example, one
ppmv  of hexane would be reported as 6 ppmv-C.

      Only two primary  gas  standard  are available for GC
calibration, and they are the  NBS propane and benzene standards.
The propane standard  is  available in concentrations of  1  ppm,
3ppm, 10  ppm, 50  ppm,  100  ppm, and  500 ppm.   All  other
standards can be certified by using the  NBS propane standard.
These standards have been found to be very  stable even  at  the  1
ppm level.

      The PID detector  is much more difficult to calibrate
because the response factors vary more than those observed for
the FID for each compound.  To accurately  quantitate samples,  a
response  factor  for each component  of interest would be
required.  In most  cases, a  single compound such as benzene is
used to calibrate the response. If a known mixture  of organic
compounds is being  monitored, such as that found in a gasoline
spill, the mixture can be used to  calibrate the instrument and
to provide a number quantitating the  total amount of  that
mixture in a given sample.  This works well if the  composition
at  the  site  is  homogeneous  and  if  there are no  other
significant sources of the  compounds in the mixture.  Dr. Tom
Spittler of the U.S. EPA Region 1  calibrates a PID  for gasoline
by analyzing the  headspace above- known  amounts  of gasoline
dissolved in water  (Clark, et  al.,  1983).  Working  level
standards are prepared according  to  the procedures in EPA
Method  621  (US  EPA,  1982).    The  dilute (e.g.,  40  ppb)
gasoline-in-water standards are stored under liquid mercury in
serum vials.  When needed, air is  introduced  into the vial, and
a headspace sample is collected.  This can also be used as  a
qualitative  check  for  matching  retention times  and for
fingerprinting the sample with the  source.

      Once the  instrument  is calibrated, a quality control
standard  should be analyzed which  comes  close to approximating
the  expected  concentration  and matrix of the samples.   This
sample is a check  to see  if  the calibration will  accurately
provide a concentration value for the components  of  interest.
For the  FID, a mixture of  components  is analyzed  by using the
single component  response factor to see  if it can accurately
identify  and  quantitate the  components  within  a  set limit.
This  QC  standard analysis provides a  good indication  of the
day-to-day variability of the instrument.

      Duplicate analyses  and samples are required  to  determine
the  variability  of the sampling  and analytical  technique.
Nested  duplicate samples,  where  samples are  collected in
duplicate and  analyzed  in  duplicate,   provide  a means to
statistically  determine total variance  of the method and the

                               192

-------
amount  of variance  which results  from both the  analytical
method and the sampling  method.

      Blank analyses are  required  to determine  the level of
contamination which results from  the sampling and  analytical
methods.  Field blanks are generated by passing a  gas from a
clean source through the sampling  apparatus and collecting it
by the method being used.  This  sample  is sent to the lab and
is analyzed as if it were a real sample.   Contamination because
of the analytical  system is determined by injecting a volume of
clean air  or nitrogen into the  instrument.  Blanks  should be
run periodically and analytical  system  blanks run between the
analysis of high-level samples and  low-level samples.

      To determine  the  absolute  accuracy  and lab-to-lab
variability,  audit  sample  analyses and 1nter1 ab or a tory
comparison studies are  required.  Performance audit samples are
unknown  samples  provided by one lab  and  submitted  to another
lab to be analyzed simultaneously with  the  soil-gas samples.
Interlaboratory comparisons  consist of  collecting a large
sample,  dividing that large  sample into  smaller samples,
sending  them to  several labs  for  analysis, and comparing the
results.  Audit samples have not been developed specifically
for soil-gas measurements; however, the EPA has established an
extensive repository of organic  gaseous compounds  at  a  wide
range of concentrations to.be  used  as  audit materials  for
emissions analysis  (Jayanty, et  al.,  1983).  There  are no
published  results for  interlaboratory comparison  for soil-gas
analysis;  however,  many of the same  techniques  have  been
compared for the analysis of ambient air (Balfour, et  al.,
1984).  The results of this  study of five laboratories using
the same method  (GC/FID) with  differing analytical  procedures
showed a coefficient of variance  of  11 percent in the value of
total nonmethane hydrocarbons.

Conclusions—
      The  level  of QA/QC  effort required  depends on the data
accuracy and precision  requirements.  In any case,  the  QA/QC
program  should  establish  the  limits  of  both the sample
collection and  analysis methods  and should  ensure  that  they
continue to perform within these  limits.   Accurate calibration
methods for portable analyzers have not been developed,  and
calibration of a large  number of compounds  for PID, ECD, and MS
is difficult.  There  is  a need for  standard reference material
for  VOCs  in  soil  as well  as  accurate QC standards  and
interlaboratory  comparison studies.
                              193

-------
                        REFERENCES
1.   Anastas, M. Y. and  H.  J.  Belknap.  Summary of Available
    Portable VOC Detection Instruments,  PEDCo Environmental.
    EPA  Report No. 340/1-80-010,  March  1980.

2.   Balfour, W.  D.,  R.  G.  Wetherold  and  D.  L. Lewis.
    Evaluation  of  Air  Emissions  from  Hazardous Waste
    Treatment,  Storage, and Disposal Facilities.   EPA-IERL
    68-02-3171, Task Number 63,  198U.

3.   Bisque, R. E.  Migration  Rates of Volatiles from Buried
    Hydrocarbon  Sources through  Soil  Media.   Proceedings of
    Petroleum Hydrocarbons and  Organic Chemicals  in Ground
    Water,  Houston, Texas, November  5-7,  1981,  pp.  267-271.

1.   Bisque,  R.  E.   New' Geochemical  Technique  Used in the
    Denver  Basin.  Western Oil  Reporter,  pp. 23-26, June 1983.

5.   Brown,  G. E., D.  A. DuBose,  W.  R.  Phillips  and  G. E.
    Harris.  Response Factors of  VOC Analyzers Calibrated with
    Methane for Selected Organic Chemicals.  EPA 600/2-81-002,
    NTIS PB 81-13619M,  1980,  U.S.  Environmental Protection
    Agency,  Research Triangle Park,  NC.  491 pp.

6.   Brown,  R. H. and C. J. Purnell. Collection and Analysis of
    Trace Organic Vapor Pollutants in Ambient Atmospheres, The
    Performance of a Tenax-GC  Adsorbent Tube. J. Chromatogr.
    178, pp. 79-90, 1979.

7.   Clark,  A. E., M. Lataille  and  E.  L. Taylor.  The  Use of a
    Portable FID Gas Chromatograph for Rapid Screening of
    Samples for Purgeable Organic  Compounds in the Field and
    in the  Lab.  Presented at the Northeast Regional  Section
    Meeting  of  the Association  of Analytical Chemists,
    Bennington, VT, June 28-29,  1983-

8.   Colenutt, B. A.  and  D.  N.  Davies.  The Sampling and Gas
    Chromatographic Analysis of  Organic Vapours  in  Landfill
    Sites.   International Journal  of  Environmental Analytical
    Chemistry, Vol. 7, pp. 223-229,  1980.
                              194

-------
 9.   Cox,  R.  D. and R. F. Earp.  Determination of Trace Level"
     Organics  in  Ambient  Air by  H 1 g h - R e s o 1 u t i on Gas
     Chromatography with  Simultaneous Photoionization and Flame
     lonization Detection.  Anal.  Chem. 54, pp.  2265-2269, 1982.

10.   Crow, W.  L. ,  E.  P.  Anderson and E. Minugh.   Subsurface
     Venting of Hydrocarbon Vapors From an Underground Aquifer.
     Draft Final Report to American Petroleum  Institute, Radian
     Corp., February, 1985.

11.   Dowdell,  R.  J., K.  A.  Smith,  R.  Crees,  and S.  W.  F.
     Restall. Field Studies of Ethylene in the  Soil  Atmosphere
     -  Equipment and Preliminary Results.  Soil Biol. Biochem.
     1,  pp. 325-331, 1972.

12.   DuBose,  D. A., G. E. Brown and G. E.  Harris.   Response  of
     Portable VOC Analyzers  to  Chemical  Mixtures.
     EPA-600/2-81-110 ,  1981,  U.S. Environmental  Protection
     Agency, Research Triangle Park, NC. 54  pp.

13.   DuBose,  D. A. and G. E. Harris.  Response Factors of VOC
     Analyzers  at a Meter  Reading of 10,000 ppmv  for Selected
     Organic   Chemicals.   EPA-600/2-81 - 05 1 ,  1981,  U.S.
     Environmental Protection  Agency, Research Triangle  Park,
     NC. 29 pp.

11.   Earp, R.  F. and R. D. Cox.  Identification  and Quantitation
     of Organics   in  Ambient Air  Using  Multiple Gas
     Chromatographic Detection. In Identification and Analysis
     of Organic Pollutants in Air.  L.  Keith, ed.  Ann Arbor
     Science Publ. Inc., Ann Arbor, Michigan,  1982.

15.   Farwell,  S.  0., D.  R.  Gage  and R. A.  Kagel.   Current
     Status of Prominent Selective Gas Chromatographic
     Detector:  A Critical Assessment. J.  Chromatogr. Scl. 19,
     PP. 358-376, 1981.

16.   Farwell,  S.  0., S. J.  Cluck,  L. Bamesberger, T.  M.
     Schutte and D. F.  Adams.  Determination  of Sulfur-
     Containing Gases by  a Deactivated Cryogenic Enrichment and
     Capillary  Gas Chromatographic System. Anal. Chem.  51,  pp.
     609-615, 1979.

17.   Freeman,  A.  M.  The- Photoionization Detector  Theory,
     Performance and Application  as a Low-Level Monitor  of Oil
     Vapor. J.  Chromatogr.  1980,  pp. 263-273.  1980.

18.   Glacum,  R.,  M.  Noel,  R.  Evans  and  L.  McMillion.
     Correlation of Geophysical and Organic  Vapor Analyzer  Data
     Over a Conductive  Plume Containing Volatile Organics.
                              195

-------
     Proc.  of 3rd National  Symposium on Aquifer Restoration and
     Groundwater Monitoring,  Columbus, Ohio,  pp. 121-127,  1983.

19.   Harsch,  D.  E. Evaluation of  a Versatile Gas Sampling
     Container Design.  Atmos.  Environ. 11,  pp.  1105-1107,  1980.

20.   Hill,  H. H., Jr.  and  M.  A.  Balm.  Ambient  Pressure
     lonlzation Detector for Gas Chromatography Part 1: Flame
     and  Photolonization Detectors.  Trends in  Anal.  Chem.  1,
     pp.  206-210, 1982.

21.   Jayanty,  R.  K.  M.,  C. B. Parker,  C.  F. Decker, W.  F.
     Gutknecht, D. J.  vonLehmden   and  J.  E. Knoll.  Quality
     Assurance for Emissions  Analysis  Systems.  Environ.  Sci.
     Technol..  17, pp. 257A-263A, 1983.

22.   Jeltes,  R., E. Burghardt, T. R. Thijssee, and W. A.  M. den
     Tonkelaar.  Application of Capillary Gas Chromatography  in
     the  Analysis of  Hydrocarbons  in  the Environment.
     Chromatographia, 10,  pp.  130-137, 1977.

23.   Karimi,  A.  A.   Studies of the Emission and Control  of
     Volatile Organics in Hazardous  Waste  Landfills.   Ph.D.
     Engineering  Dissertation,  University of  Southern
     California, February  1983.

21.   Krost,  K. J., E.  D. Pellizzari,  S.  G. Walburn and S.  A.
     Hubbard.  Collection  and Analysis of  Hazardous  Organic
     Emissions. Anal.  Chem.  51, pp. 810-817, 1982.

25.   Lonneman,  W.  A., J.  J. Bufalini, R.  L. Kuntz and S.  A.
     Meeks.  Contamination from Fluorocarbon  Films.  Environ.
     Sci. Technol. 15,  pp.  99-103, 1981.

26.   Marrln,  D.  L.   Delineation of Gasoline Hydrocarbons  in
     Ground  Water by  Soil Gas Analysis.   Proceedings  of the
     1985 Haz. Mat. West Conference,  Long  Beach,  California,
     December  3~5, 1985.

27.   Marrin,  D.  L.  and G.  M. Thompson.  Remote Detection  of
     Volatile  Organic Contaminants in Ground Water Via Shallow
     Soil Gas Sampling.  Proceedings of Petroleum Hydrocarbons
     and  Organic Chemicals  in Ground  Water,  Houston,  Texas,
     November  5-7. pp.  172-187, 1981.

28.   NIOSH  Manual of Analytical Methods. National Institute for
     Occupational Safety and Health, U.S.  Government Printing
     Office, Washington, D.C.  1971.

29.   Oliver,  K.  D.,  J. D.  Pleil and W. A. McClenny.   Sample
     Integrity of Trace Level Volatile Organic  Compounds  in

                              196

-------
     Ambient Air Stored in Summa Polished Canister.  Atmospheric
     Environment. 20, pp 1403-1411.  1986.

30.  Radian  Corporation.  Soil Gas Sampling Techniques of
     Chemicals  for Exposure Assessment, Bonifay Spill Site Data
     Volume.  EPA-EMSL 68-02-3513, Work Assignment 32,  1984.

31.  Radian  Corporation-S.  Soil  Gas Sampling Techniques of
     Chemicals  for Exposure Assessment.  Stovepipe Wells Spill
     Site Data Volume.  EPA-EMSL 68-02-3513, Work Assignment
     32,  pp.  26.  1984.

32.  Radian  Corporation-T.  Soil  Gas Sampling Techniques of
     Chemicals  for Exposure Assessment, Tustin Spill Site Data
     Volume.  EPA-EMSL 68-02-3513. Work Assignment 32,  1984.

33.  Saalwaechter, A. T.,  C.  S.  McCammon, Jr., C. P.  Roper and
     K. S.  for the  Determination  of Air Concentrations of
     Organic  Vapors. Am.  Ind. Hyg. Assoc. J. 38, pp.  476,  1977.

34.  Schmidt,  C.  E.,  B.  M.  Eklund,  R. D.  Cox and  J. I.
     Steinmetz.  Quantitatlon of Gaseous  Emission  Rates from
     Soil Surfaces.  In, Soil  Gas Sampling Techniques of
     .Chemicals  for  Exposure Assessment.   Radian Corporation,
     Interim  Report. EPA-EMSL 68-02-3513, Work  Assignment
     32.  1983.

35.  Seila,  R. L., W. A.  Lonneman   and S. A.  Meeks. Evaluation
     of Polyvinyl Fluoride  as  a Container Material for  Air
     Pollution Samples.   J.  Environ. Sci. Health,  Part  A, 11,
     pp.  121-130, 1976.

36.  Sevick,  J. Detection in  Gas  Chromatography,  Elesevier
     Scientific Publishing Company,  New York. New York, 1976.

37.  Shafer,  K.  H., T.  L. Hayer, J.  V. Brasch and  R. J.
     Jakobsen. Analysis of  Hazardous Waste  by Fused  Silica
     Capillary Gas Chromatography/Fourier Transform Infrared
     Spectrometry and Gas  Chromatography/Mass Spectrometry.
     Anal. Chem. 56, pp.  237-240,  1984.

38.  Splttler,  T. M., personal communication,  1985.

39.  Spittler, T. M., W.  S.  Clifford  and L. G. Fitch.  A New
     Method  for Detection of Organic Vapors in  the Vadose Zone.
     Presented at National Water  Well  Association  Meeting,
     Baltimore, Maryland, September,  1985.

40.  Swallow,  J. A.  and P. M.  Gschwend.  Volatilization of
     Organic Compounds from Unconfined Aquifers.   Proceedings
     of the  3rd National  Symposium on Aquifer Restoration and

                               197

-------
     Ground Water Monitoring, Columbus,  Ohio,  May  25*27,  pp.
     327-333. 1983.

41.   tenNoever  de Braun,  M. C. Combined Gas Chromatography-Mass
     Spectrometry: A Powerful Tool in Analytical Chemistry. J.
     Chromtogr. 165, pp.  207-233. 1979.

U2.   Thorburn,  S., B.  A.  Colenutt  and  S. G. Douglas.   The
     Sampling and Gas  Chromatographic Analysis of Gases from
     Landfill Sites.  International Journal of Environmental
     Analytical Chemistry, Vol.  6, pp. 245-25U,  1979.

•13.   U.S.  Environmental  Protection Agency.  Methods for Organic
     Chemical Analysis of Municipal and Industrial Wastewater.
     EPA-600/U-82-057, July  1982.

M4.   Weeks,  E.  P., D.  E.  Earp   and G.  M. Thompson.  Use of
     Atmospheric  Fluorocarbons F-11 and F-12 to Determine  the
     Diffusion  Parameters of the  Unsaturated  Zone  in  the
     Southern High Plains of Texas.  Water  Resources  Research,
     18, pp.  1365-1378, 1982.

U5.   Westberg,  H. H.f  M.  W. Holdren   and H. H.  Hill,  Jr.
     Analytical Methodology for  the  Identification  and
     Quantification of  Vapor  Phase  Organic Pollutants. Final
     Report to  the Coordinating Research  Council,  CRC-APPRAC
     Project  No.  CAPA-11-71, 1982.

H6.   White, L. D., D. G. Taylor, P. A. Mauer  and R. E. Kupel.
     A Convenient Optimized  Method for  the  Analysis  of  Siluted
     Solvent Vapors in the  Industrial Atmosphere. Am. Ind. Hyg.
     Assoc. J. 31:225, 1970.

M7.   Willey,  M.  A.  and C. S. McCammon, Jr.  Evaluation of
     Direct Reading Hydrocarbon Meters  (Flame  lonization,
     Photoionization, and IR).  HEW  (NIOSH)  Pub.  #77-137.  1976.

U8.  Wood,  M.  B.  An  Application  of Gas Chromatography to
     Measure Concentrations of Ethane,  Propane,  and Ethylene
     Found in Interstitial  Soil Gases. J.  Chrooatogr. Sci.,  18,
     pp. 307-310, 1980.
                               198

-------
                          CHAPTER  6
                 STATISTICAL TREATMENT OF SOIL
                 ORGANIC VAPOR  MEASUREMENTS
INTRODUCTION
    Soil organic  vapor  (SOV)  measurement is usually  performed
in an exploration  phase of an investigation.   The statistician
has two  duties  at the start of such  an investigation.  They are
(i) to determine a method for taking  the measurements in such a
way as to  meet  requirements for data precision in a  realistic
and cost effective  manner, and  (ii)  to help  determine  the
locations  for the initial exploratory survey in accordance with
prior knowledge  about the site and the objectives and budget of
the  project.   The'determination  of  the  precision  of  the
measurements  and  the size of the  contributions of various
sources  of error  is  the starting point in the planning of all
good sample surveys.  The measurements must  represent the SOV
concentrations  in the soil at the  sites where the samples are
taken.   The measurements will be  worthless  if they  merely
represent the  errors generated in obtaining, handling,  and
analyzing the  samples.  One should not wait  till the  end of a
survey to  determine the precision of the measurements, for then
it may be too  late to save the study  from  bad data.  Obviously
it is impossible  to  completely eliminate errors.  However, by
use of proper  statistical techniques, it is possible to measure
the effects of  the various sources of error on the precision of
the measurements and, if necessary,  to find the most economical
means for  reducing  the effects  so as to attain a desired level
of precision.

     The proper methods  for  choice  of locations for the taking
of measurements  is unique to each site because  of the changes
in objectives,  budgets, and prior knowledge between sites.  The
decision concerning location of sample points  must be made in
concert  with  the  other principal  participants in the study and,
because of the exploratory nature of the SOV  investigations,
may  in  large  part be  subjective.  Some words of advice on
possible sampling patterns have been  given  in a preceding
chapter .


                              199

-------
     Techniques  for assessing  measurement  precision  and
contributions  of  various sources  of  error  to  total  error are
discussed in  this  chapter along  with some comments  on the use
of the final  survey data in  interpolation  and contouring
procedures.


COMPONENTS OF  VARIANCE ANALYSIS

     The  central statistical procedure involved in the process
of measuring and  improving precision  is  called components of
variance analysis.   This procedure is based on a model for the
measurements that Includes an equation of the form,

               Yr-vi * en- + e12 + ...+ elk

where YI is a measurement of  the concentration  of  an organic
vapor  in sample  1, y^  is  the expected value of  such  a
measurement (I.e.,  the  average of a hypothetical population of
repeated  measurements on samples taken at the same location and
using  the  same  technique),  and  e^  is the error In  the
measurement coming from source J (J-1	k).  The measurements
are said to be precise if the sum  of the e's on the right-hand-
side of  the equation is  near  zero  with  high probability.
Hence,  to  obtain precise  measurements,  one  wants  the
contributions  ejj from each of the  various  sources of error to
be small  with  high probability.

     To  develop  the  above  model  equation,  one must  be
sufficiently familiar  with the  sample acquisition,  handling,
and analysis process to  search out and  list all  the  major
sources of error.  Since the  size of errors vary from  sample to
sample and cause  variation in  measurements,  it  is  common to
refer to sources  of error as  sources of error variation  or
sources  of variation.   For example, analytical  errors  cause
variation between  analyses of  subsamples of  a  sample  of soil
gas.  If one  considers soil  organic vapor measurements taken by
drawing soil gas  into  a soil  probe, withdrawing  gas from  the
probe with a  syringe  and finally Inserting  the  gas from  the
syringe into a gas chromatograph for measurement of a chemical
concentration,  then the  variation  in gas  chromatograph
measurements of gas from several syringes taken from the same
probe  after  a single purge  and relaxation  of  vacuum  would
represent the  variation caused  by such things  as air  leakage
into   the  syringes  and  analytical  errors in  the gas
chromatography.  If several syringes of gas  are  withdrawn from
each of  several closely  spaced probes at a sampling location,
the variation  of the average  measurements for  the  probes  would
give information  concerning  combined variation  caused by short
range spatial  differences In  concentrations of the  SOV  being
measured, by  differences  in  insertion of  probes,  by differences

                               200

-------
in the  vacuum  achieved in  the purging  of the probes, and
perhaps  by leakage of surface  air into the probes.  It is wise
to determine the  sizes  of  the contributions of the various
sources  of error variation to  the total variability of the data
prior to the investigation  of  the  spatial  distribution  of the
SOV.   Knowledge of the contribution of the sources of variation
allows  the  investigator to determine how best to allocate
resources to obtain precise estimates of SOV  concentrations (or
perhaps to determine  that the study will not be able to  obtain
desired levels  of  precision  and  therefore should  not be
completed).   For  example,  if there  is  a large  irregular
variation between measurements  at the same sampling location
taken at substantially different times (say  a week apart), the
investigator will have to  determine whether this variation is
due  to  differences in  measurement procedures or is due to
changes  in environmental conditions such as  ambient temperature
and  atmospheric  pressure.   If  the cause  is  change in
measurement procedures, this  variation might be substantially
reduced by additional training and practice.  If the  cause of
the large variation over time  is due to environmental factors,
it may be necessary  to establish one or more control sites and
to measure at the control  sites whenever measurements are taken
at other sites.  This would  provide the information needed to
make  measurements taken at different times comparable.

     As  is implied by its  name, the components of variance
technique uses variance (the  second moment  about the mean) as
the  measure  of  the variation caused by errors.   If the
probability distribution of the errors is normal (Gaussian) as
is assumed in variance components analysis  (the error terms are
assumed to be independent random variables with  ejj  having
normal distribution  with mean  zero and unknown variance  Oj2),
the  shape  and  spread of the  distribution is  completely
determined by the  variance.   Under the assumption of  a  normal
distribution for  Yj, the precision  of the measurement is
measured by its  variance; the smaller the  variance, the  better
the precision.   However,  for most other types of probability
distributions,  variance  does not completely characterize the
spread of the distribution.   In addition,  for many nonnormal
distributions,  the  spread of the distribution changes  as the
mean, m, changes.  It is  impossible to estimate variances of
the  EIJ if the variances  are  not constant (i.e., if they  change
with the magnitudes of the measured concentrations).  Hence, if
variances change as the mean  changes, it is  vital that  the data
be transformed in such a way  as  to stabilize  variances relative
to the mean.  Typically,  if  one can stabilize the variance with
a  transformation,  that transformation also  makes  the
distribution more  symmetric  (see Section IF  of Hoaglin  et al. ,
1983), and thereby  a better approximation to the normal.   A
frequently used  variance-stabilizing transformation  for data


                              201

-------
   -5  -i
   -4  -
   -3  -
    -2 -
                       Normal Density Curve
                -1
   -6  -i
                        Skewed Denelty Curve
       -2       -1        0
Figure 6.1.   Density curves  for normal and skewed distribution.
                             202

-------
from  distributions that  are skewed  to the right  is the
logarithmic  transformation  of the measurement  Z  into Y with

                   Y-log(Z * m)

where m is  a nonnegative constant such as the minimum detection
limit, 1,  or, if all the  measurements are considerably larger
than  1,  the minimum measurement minus 1.  Another commonly
employed  transformation for data that are skewed to the  right
is the square-root transformation (i.e.,  X  *  /Z) which does not
reduce large values relative  to  the smaller observations  quite
as  much  as does the  logarithmic  transformation.   For
discussions of  methods for  determining when  and  how to
transform data see Hoaglin,  et al.  (1983),  and Scheffe (1959,
Section  10.7).

     The planning of an  experiment  to obtain data for the
estimation of variance components is very  much dependent on the
nature  and costs of the  various operations  involved in the
taking of  SOV measurements.   A  review and bibliography of the
literature  for the design and  analysis  of experiments  planned
for variance component estimation is given by Anderson  (1975).
A detailed discussion  of the  theory of variance components
estimation  is given  in Scheffe (1959, Chapters  7  and 8).
Computer programs are  available for components of variance
analysis  in  most mainframe  statistical software  packages (e.g.,
PROC VARCOMP in the SAS package  - see Ray. 1982, p. 223).

     As an  example  of  the  use of  the  results  of variance
components  analysis, suppose  one can take k syringes of gas
from  each  of m  probes  located on the  nodes  of a small grid
centered on a sampling  location.   If the  model for the
measurement  of the gas from syringe i taken  at probe h is
                Eh * ehi        n "
where  w  is  the  expected value of Yhi, cn is the error which
results  from  measuring  at  probe h,  eni  is the  error which
results  from  sampling probe  h  with syringe i, then the variance
of  the  mean of  the km measurements  gives  a  measured
concentration  for this sampling location that has variance

                V(Y)-V(en)/m  +  V(ehi)/(mk).

If  the variance components  analysis has  estimated V(cn)  to  be
10  and V(cni) to  be  20, then  the estimated precision of  Y is
given in terms of the variance  (10/m * 20/(mk)).  The numbers k
and m may  be  adjusted within the limits  of  costs and
feasibility to  reduce the  variance  to a desired  level and
thereby  to  Increase the  precision of the measured concentration
obtained from each sampling location.  For m-1 and k-M,  V(Y)

                              203

-------
is estimated to be  15  which is also obtained  for m-2 and k-1 .
If the cost of a  probe is about the same as the cost of the
analysis of a sample,  the second option would  be preferable to
obtain a desired variance of 15.  The second option  might be
essential if it is not feasible  to take k as large as i».

Variance Components Analysis Example

     In  this section,  data  from the Case Study 2.0, hereafter
called the Pittman Case Study, of the chapter  on case  studies
that  follows, will  be used to  illustrate some of  the points
that were made concerning variance components analysis  that was
made  above.   The  study was  run  to  test an SOV sampling
procedure by comparing its results with data obtained from
nearby  wells.  The SOV  data  consist  of gas chromatograph
measurements of chloroform concentrations (in ppbv) in gas
obtained in 250 yl  syringes from probes.  Although no special
experiment was performed to measure  variance components,  this
study does  provide data from which some variance components may
be estimated.

     The first step  in a  variance  components  analysis of this
data is  a search  for an appropriate transform of  the data to
move  it toward normal distribution  characteristics.  Table 6.1
is obtained from  the data in Table 6.2  of the  Pittman Case
Study by ordering the 23 sets  of replicate analyses  according
to the size  of their sample means.  In Table 6.1, the  standard
deviations of  log  transformations  and  of  square-root
transformations of  the sample  data are  given along  with the
standard deviations for the raw data provided  in Table 6.2. If
one looks at the  standard deviations of  the raw data for the
smallest 8  means  and  for the  largest 8  means, it  is obvious
that the sample standard deviations  are  increasing  along with
the  means.  If  one does the  same thing for  the standard
deviations  of the log-transformed  data,  one  finds that the
standard deviations  are getting smaller  as  the  sample mean
increases which  implies  that  the  transformation has  over
compensated for  the skewed distribution.  Finally,  when one
checks the  first and last  8 standard  deviations  for  the square-
root  transformed  data, one finds no evidence of  trend of the
sample standard  deviations with the sample means.   Hence, In
this  case,  the  square-root transformation seems  a reasonable
choice.

    The analytical  (between syringes)  error variance for the
square-root  transformed chloroform measurements of  the Pittman
Case  Study  may  now  be estimated with the  pooled  estimator of
the variance,

                    3p2.z(ni-i)sxi2/E(ni-1)

                               204

-------
where  sxi2 ^s  the sample  variance calculated from the n$
square-root transformed measurements  taken at location  1
(i-1 ,...,23) .  From the square-root transformations of the  data
In Table 6.2, one  finds  sp2-0.0836  and sp-0.29  (based on 43
degrees of  freedom).

    In the case study, measurements  were taken after each  of  a
series  of purges of the same  probe.  These measurements allow
estimation of  "between-purge"  variance.  Two syringes were
loaded  with gas  after  each  of four purges of the probe.  The
results  of the analyses of the gas in the syringes are given in
Table 6.2.  The usual model for the analysis of the transformed
measurements is

                     xij  - H  + Pi + Eij

where y  is the mean,  pj  is  a random variable representing the
deviation associated with  following purge i and is distributed
N(0,0p2),  c£j is  the  random variable representing effect of
syringe  (analytical  error)  J  taken  after purge  i  and is
distributed N(0,oa2),  and the random variables are independent
of  one another.    (There  is some doubt  concerning  the
independence and  identical  distribution of these measurements
in that  vacuums achieved declined from purge to succeeding
purge.   Perhaps it is  impossible to obtain a true  between purge
variance  estimate  because of this problem.)   The analysis of
variance of the transformed  purge data gave a mean  square for
analytical  error of 0.13 based on 4 degrees  of freedom.  This
estimate of oa2  is remarkably  close,  considering the small
number of  degrees  of  freedom, to the value  0.0836  obtained
earlier.   The between  purges mean square is 0.33,  based  on  3
degrees  of freedom.   Since  the  expected value of  this  mean
square under the  above model is  oa2+2op2, the estimate of the
between purges component of variance would be

                   3p2-(0.33-0.13)/2-0.10,

or, when  the more precise  estimate of  oa2 found earlier is used

                   3p2-(0.33-0.08)72-0.12.

The  problem with  vacuum and the small number of degrees of
freedom  available  for  between-purges-component estimation
should make one suspicious of the accuracy of either estimate,
•P*.

    The extreme variability  of the  estimator s2  of  a variance
o2 based on small  numbers of degrees  of freedom  is  seldom
adequately appreciated.  Table 6.3 indicates how  the length of
confidence  Intervals for o2  based on  estimates s2 varies with


                              205

-------
TABLE 6.1.  SAMPLE  STANDARD  DEVIATIONS FOR RAW AND TRANSFORMED
   CHLOROFORM MEASUREMENTS ORDERED  BY SIZE OF SAMPLE MEAN
(Z - Chloroform measurement, Y - log(Z), X - /Z)


























Rank
1
2
3
1
5
6
7
8
9
10
11
12
13
11
15.
16
17
18
19
20
21
22
23
Sample
Mean
5
6
10
10.5
12.3
25
27
27
27
28
30
32.1
15.6
55
72.9
112
115
161
171
266
326
376
511
Sample
Size
3
2
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3
3
2
3
3
3
3
sz SY

0 0
3 0.19
2 0.19
0.3 0.06
0.5 0.01
2 0.09
5 0.17
2 0.07
5 0.20
5 0.18
1 0.03
0.2 0.01
0.2 0.01
2 0.03
0.1 0.00
12 0.11
6 0.05
6 0.03
0 0
6 0.02
10 0.03
6 0.02
17 0.03
SX

0
0.59
0.31
0.09
0.06
0.22
0.16
0.20
0.53
0.15
0.09
0.02
0.02
0.12
0.01
0.55
0.28
0.22
0
0.17
0.29
0.15
0.37

























(Kerfoot. 1985)







(
TABLE 6.2.
CHLOROFORM



Purge


Kerfoot, 1985)
Syringe
1 10.7(3.
2 6.6(2.
3 10.5(3.
1 10.0(3.

CONCENTRATIONS (ppbv) MEASURED
SERIES
1
271 )*
569)
210)
162)

degrees of freedom. The table
normally distributed data whereas
data are nonnormal, and usua
estimator s2 is
even greater than
OF PURGES OF
Syringe 2
16.1(1.050)
8.5(2.915)
9.8(3.130)
6.7(2.588)

THE SAME PROBE






is based on an assumption
real (and even transforme
lly the variability of t
that indicate
d by the tabl
ON







of
d)
he
e .
206

-------
Also  note  the  rapidly diminishing rate  of decrease  in
confidence  interval lengths once  the number of degrees  of
freedom exceeds 20.   This makes  20 degrees of freedom a
reasonable goal in planning an  experiment to estimate a
variance.

        TABLE 6.3.  CONFIDENCE INTERVALS FOR o2 BASED
     NON s2  AS A  FUNCTION OF  DEGREES OF FREEDOM (D.F.)  AND
            ASSUMING A NORMAL DISTRIBUTION FOR DATA

              D.F. _ 95% CONFIDENCE  INTERVAL _

                 2      0.27s2S02S39.21s2
                 3      0.32s2*o2S13.89s2
               10      O.H9sS
               20      0.58s2So2S2.08s2
               30      0.6132:£02S1 .78g2

               50      0.7032*o2S1 .61S2
              100      0.77s2*o2S1 -35s2
    It is important  to  estimate the between-probe variation of
the measurements  to  determine  how  well  an  individual
measurement characterizes the concentration  of the organic
vapor  in the  immediate vicinity  of the sampling  point.  In the
Pittman study,  several  probes  were placed at locations only a
few feet apart near well No. 627 (see Figure  7.26 of the  case
study).  The results from the  three locations  (E23; E23, S3;
and E20, S3) are  employed to obtain an estimate of spatial
variation.  The  model employed is:
where  AI is a  random variable with a N(0,OL,2)  distribution
representing  the effect of location i, PIJ is a random variable
with a N(0,0p2)  distribution representing the random effect of
purge J  within  location i, and the EIJ  is  defined as  in  the
previous model.   The  random variables are  assumed  to  be
independent.  The mean square for error in the analysis  of
variance of  the square-root transformed data is  0.182 and is an
estimate of  (op2*aa2) based on 6 degrees of  freedom.  This
estimate agrees  surprisingly well with the previous estimates
for the sum of the  two variances.  The  mean square for between
locations  is  21.036  and  represents  an  estimate  of
(op2*oa2)*3oL2.  Hence the estimate of oL2 is (21 .036- . 1 82) /3 -
6.95.  This  estimate is based on only two  degrees of freedom
and is extremely unreliable.  In addition, the investigators in
the Pittman  study believe that the variation  observed  in the
data from the three probes was primarily due to the time  delays

                               207

-------
in getting the  samples analyzed  rather than being due  to
short-range spatial variation  in  chloroform concentrations.   If
the estimate had  been  a  more reliable estimate of short-range
spatial variation,  it would indicate that to  improve the
estimated concentration  at a sample point significantly,  one
must  average measurements  over the  results from several closely
spaced (but non-interfering)  probes at that location.  Based  on
the variance components model  given above,  the variance of  the
average of the measurements on s syringes from each of m purges
from  each  of n  closely spaced probes at  a  sample location  is
given by the formula,
            Var(X) - a^/n + Op^/mn + oa2/mns

When variance  between  probes is  the major source of variation,
as it is  in this case, the only way to substantially  reduce  the
variance of the  mean X  is to  increase the  number,  n,  of
probes.   When variance  between purges  is  the major source of
variation, one may increase m or n, and  the choice  could be
made on  the basis of costs.  Similarly,  when analytical error
is the major source, one could  choose on the basis of  cost to
increase  m, n,  or s, in  order to decrease the variance  of X.

    Once  again it  should be  pointed  out that  the  above
discussion of  the components of  variance analysis  of  the
Pittman  data  is  an illustration  of  procedure.  In  actual
practice, one should note that the measurements of  samples from
different purges  of the same probe are neither  independent nor
identically  distributed.  Therefore, one  should  use  only  the
sample from after  the  first purge  in measuring  concentrations.
This effectively  removes "between  purges" as an estimable
variance  component.  For each possible variance  component, it
is necessary  to consider whether the assumptions  of  variance
components analysis are reasonable.   In  addition,  as was  done
above, it is necessary to decide or  to investigate  further  the
causes  of large  variance components (i.e., was the large
between probe variance due to short-range spatial  variation or
differences  in  times  the  gas waited  in syringes  before
analysis) .


INTERPOLATION AND  CONCENTRATION CONTOURING

    One  of the principal reasons for  taking  SOV  measurements is
to estimate  the location of  a pollutant  plume.   Another  reason
may  be  to indicate a possible source  for a pollutant  plume.
Data analyses used  to further these objectives usually involve
interpolation between sample points  and the drawing of
concentration   contours  (isopleths).  Before discussing these
analyses, it  is important  to  point  out a few things  about the
data.  SOV measurements  are  measurements of concentrations of

                               208

-------
certain gaseous chemical  compounds  near  the surface.  They do
not necessarily represent the concentrations (or some monotone
transformation of  those concentrations)  of the compounds
directly below the  sample points  at the  level of the plume.
A positive SOV measurement  at  a  sampling location may be a
false positive (i.e., the plume does not extend  below  the
sampling point) in  that the positive measurement may be  due to
lateral  movement in the earth of the gas from the plume around
a lens  of  impermeable clay  or  rock, or  it may be caused by
errors  in  the sampling and  analysis.   A  "none-detected"
measurement may  be a  false  negative (i.e.,  plume is  below
sampling point) caused by  an  Impermeable  layer between  plume
and sampling device, by biodegradation of  the compound, by slow
transport rate, or by sampling  and  measurement  error.   Thus a
very irregular spatial pattern  of  SOV measurements may  be due
to one or more of  several causes.   Even a fairly  regular
spatial pattern of  SOV measurements may  not be Indicative of
the actual  location of the  pollutant  plume because of  lateral
drift of the vapor.  Under the very best  of circumstances, the
SOV measurements  represent  some monotone transformation
distorted by measurement  errors of  the concentrations  of the
compounds below the sampling points  at the  level  of the  plume.

    There  are many methods of interpolation available  to the
investigator such as linear, inverse squared distance, splines,
and kriglng.  Most such methods,  such as the first  three
mentioned above, are deterministic  (i.e., do  not  rely on a
probability model)  while  some,  usually denoted as kriging, do
depend on probability models.   Typically the  various common
interpolation procedures  give similar results concerning the
general pattern of  SOV  concentrations.   The advantage  of
kriglng (basically a regression procedure  that uses information
about  the spatial correlation of observations) is that  it also
provides an estimated standard  error for each Interpolation.
However, that estimated standard  error Is  highly dependent on
the probability model  (commonly  referred to  as the  spatial
structure model in geostatistics).  The probability model must
be estimated anew  for each  SOV study because of the unique
characteristics of different sites.   Good estimation of  a  model
requires more and better  data than  are usually obtained in an
SOV study.   For this reason,  it  seems better to use a simple
spline  or  inverse-square  interpolation  procedure, and  if  an
indication  of the amount of error that may be involved in the
interpolation is  desired, cross-validation techniques (see
Efron,  1982, Chapter 7) may be employed.

     Finally, care must be  taken  in the  use of computer
interpolation and  contouring  packages so as  not  to obtain
misleading  contours.  A typical situation in which misleading
contours occur is one in which an observation is  several orders
of magnitude larger  than  those  obtained  at neighboring  points

                              209

-------
(see Figure 6.2).   In this case there will usually be several
contour  lines  circling the point with  the  large observation;
the locations of  these contours reflect almost nothing other
than the idiosyncrasies of the contouring package.  It would be
better  to remove  these circling contours  and  merely flag the
large observation.
                13       10


                15       10


                12       11


               Figure 6.2.  Misleading contours
                               210

-------
                         REFERENCES
1.    Anderson,  R.  L. Designs and Estimators for  Variance
     Components.   In A  Survey of Statistical  Design  and
     Linear  Models, J. N.  Srivastava,  ed. North-Holland
     Publishing Co.,  New  York, NY, 1975.  p.  1-29.

2.    Efron,  B.   The Jackknife, the  Bootstrap and  Other
     Resampling Plans.  Society for Industrial and  Applied
     Mathematics, Philadelphia, PA,  1982.  92 pp.

3.    Hoaglin,  D.  C.,  F.  Hosteller  and J.  W. Tukey.
     Understanding Robust and Exploratory Data Analysis.   John
     Wiley & Sons, New York, NY, 1983.  *47 pp.

4.    Scheffe,  H. The Analysis of Variance.  John Wiley & Sons,
     New York, NY, 1959.  477 pp.
                              211

-------
                            CHAPTER 7

                          CASE STUDIES
INTRODUCTION

      Volatile  compounds  are  components  in  ground-water
contamination at many, if not most, Superfund  sites. Soil  vapor
concentration  serves as a surrogate for  actual measurements of
the concentrations of the compounds of interest in ground-water.
The  usual objective in measuring organic  vapors in soil is to
map the  lateral extent of soil  and ground-water  contamination
or both while  at the  same time minimizing the  number  of
conventional monitoring wells which must  be drilled.   Maps of
soil  vapor concentrations  can  be used  to site ground-water
monitoring  wells more efficiently.

      The basic  approach in  a  soil-gas  investigation  at  a
particular  site is simple in  concept.  The vertical profiles of
organic vapors present in the soil pore spaces are measured and
plotted  for several locations at the site.   Selection  of tracer
gases for  the  site  is  aided  when prior information  on
contaminant concentrations in ground-water is  available.   Based
on the  vertical  profiles, the  particular  organic soil  gases
present,  and  the sampling and analytical methodologies
available, one  or more tracer gases are selected.  A  sampling
depth  is also  selected,  based on the measured  vertical
profiles, which  is  expected to produce soil  gas  concentrations
well above the  minimum concentrations  detectable  with  the
analytical techniques at hand.  By using  this constant sampling
depth, soil gas samples are collected and measured across  the
site preferably  on  a regular  grid pattern.   These values are
then plotted on a  map and are contoured either by hand  or  with
a computer algorithm.   The  desired result  is a contour plot of
soil-gas concentrations at a constant depth across  the  site;
the investigator  hopes that this plot is  related in a more or
less linear way to contaminant  concentrations in ground-water
or in the  buried waste stratum of interest.

      Two case studies  are presented  as illustrations of this
basic approach.   While neither  case is a Superfund  site,  the
techniques used  are  very similar to investigations which might
be carried  out  as part of  a  Superfund Remedial  Investigation
Feasibility Study.   The first case study  illustrates the use of
soil-gas measurements  as a  means of delineating  a  plume  of
gasoline  from  a  leaking underground storage tank at  a  rural
                               212

-------
service  station.   In  the  second  case  study,  soil-gas
measurements are  used  to  examine the extent of ground-water
contamination originating  from surface impoundments  and
underground storage tanks  at a  large industrial plant.

HYDROCARBON PLUME DETECTION AT  STOVEPIPE WELLS, CALIFORNIA

Gasoline Plume History

      In May 1979 the odor  of  gasoline was detected  In an
unused well near  a Chevron service station adjacent to  the
Stovepipe Wells Hotel  in  Death Valley National Monument.  The
location of Stovepipe  Wells (La Brecque, et al., 1984)  is  shown
on Figure 7.1; Figure  7.2 shows the location of the  service
station and hotel  complex.  A sample collected from  the well
showed that a layer  of gasoline had accumulated on the water
table.   Service station  records indicated that between  October
7, 1978, and September 4, 1979, as much as 19,000 gallons of
unleaded gasoline were lost.   No records were available  prior
to October  7, 1978, but  there was evidence that the  buried
storage  tank was  already  leaking at  that time.   The  total
product lost was  probably considerably in excess of  20,000
gallons.  The National Park  Service requested that the  USGS
assess the spreading  and  the  hydrologic effects of the gasoline
leak on the ground-water system.

      The USGS  drilled several  wells  in  May, 1980; Figure 7.2
summarizes  the USGS results. In  1980  well K3  contained
sediments saturated  with  gasoline, but well J6 contained no
detectable gasoline.   By fall  1983t a strong gasoline odor  was
detected in well J6,  indicating that the gasoline had continued
to migrate eastward.   There was considerable  question as  to  the
exact  direction of the flow  of gasoline, its width,  and  its
areal extent.

Hydrogeologic Setting

      The Stovepipe  Wells study area is located on  north-
dipping alluvial-fan sediments.   Ground-water  is the only  local
source of water  available to the Stovepipe Wells Hotel  and
associated  facilities.   The  local  aquifer  consists  of
unconsolidated, gravelly,  sandy silt having  a transmissivity of
about 315 m^/d. Depth  to water ranges from about 25 feet  just
north  of the hotel to  about  145 feet at a supply well up the
alluvial west of the hotel.  Direction of ground-water  flow in
the  vicinity of the  hotel is to the east;  the  regional water
table appears to have  declined  slightly between 1977 and 1978.

      Ground-water quality in  the vicinity  of the hotel varied
from 5250 mg/1 to 8790 mg/1  total dissolved solids (TDS)  in
June 1980.   TDS consists  primarily of sodium,  chloride,  and
sulfate, with high concentrations  of boron and iron.
                               213

-------
           STUDY AREA
          STOVEPIPE WELLS
           DEATH VALLEY )
            ' NATIONAL   \
             MONUMENT   \
Los AngtU*


                  SCALE

               10         20
30
                  MILES


    Figure 7.1.   Location of study  area.
                        214

-------
N)
                                                                                                        m«tor»
                                                                                                              100

                                                                                                              d
                                                                                    	Predicted Plume Extent 19BO
                                                                                    • •••Predicted Plume Extent 1-1-84
                                                                                     Q  Uncon tarn (not od Well
                                                                                        Contaminated In 1984
                                                                                          Uncon torn Ina ted In 1980
                                                                                        Contaminated Prior to 1980
                                                                                        Contaminated Ground Water
      Figure 7.2.
Comparison of  Che  assumed  location of  gasoline
plume  from USGS wells  at  the beginning  of field
work (May,  1980)  with  the  plume extent  shown by
subsequent  Lockheed wells.

-------
Soil  Gas Measurements

      Concentrations  of  organic  vapors in soil at a depth of
approximately  5  feet were mapped with  a system which involved
driven steel  probes,  a special sampling  manifold,  gas sample
collection  with  a syringe, and field analysis via a  field gas
chromat ogr aph  with a photoioni zation detector.   The  probe,
shown in Figure 7.3,  was about 98 inches in length with  an
outside diameter  of 3/H - inch and an  inside diameter of 1/U -
inch; the  shaft was constructed of H130  carbon steel.   The
probe tip  was   a  cylinder of 316 stainless steel, 0.87 inches
long, with  a nose tapered at 30 degrees.  Six ports were  spaced
radially around the circumference of  the cylinder.  Frits -
porous stainless steel discs measuring 3-mm diameter  by  3-mm
thick - were  pressed into the ports to serve as screens.  Pore
size  was 20 microns, chosen to allow soil  vapor passage while
excluding  most soil particles.  A length of 3 mm  OD  Teflon
tubing was  connected to the ports in the probe tip;  the probe
was  completed by threading the 3 mm  Teflon tube through the
probe shaft.   [NOTE:  Teflon is no  longer recommended  for
soil-gas sampling  because some hydrocarbons tend to adsorb or
diffuse  into  the material.  The  LGAS  probe  used  in  the
following Pittman, Nevada, case study used  stainless  steel.]

      Modified  fence-post  drivers were used  both to  insert and
to remove  the probe from the ground.   The driver  was a  2.5
-inch OD   cylinder with a 1.5  -  inch ID.   The driver  and
extractor are  shown  in Figure 7.1.  Insertion of the probe was
difficult  at  Stovepipe Wells because much of the site was
covered by  cobble- and boulder-bearing sands and gravels.  To
reduce the wear and tear on the sampling probe, a  1.25 - inch
steel rod was  first  driven through most of  the sampling depth,
then  removed  so that the probe could be  driven to  the  desired
depth.  After  extraction, the  soil probe cone was removed for
cleaning.  A  clean  cone  was then  reinstalled  and used to sample
the next location.

      At a  chosen sampling location,  the  probe was  driven into
the  ground to the desired sampling  depth (about  5 feet  at
Stovepipe  Wells), the  fence  post driver  and steel drive cap
were  removed, and the free end of the 3-mm Teflon tube was
attached  to  the sample port of  the  gas sample collection
manifold as  shown in Figure  7.5.  The manifold was a  3~way
T-connector  which connected the sample port  to a Mininert valve
and  a shut-off valve which in turn was connected  to  a  Mine
Safety Appliances  (MSA)  Samplair manual pump.  To obtain  a
sample, the shut-off valve was opened and a 75-ml volume of gas
was  purged from the probe with the pump.  A  gas chromatograph
syringe was then used  to  extract a sample for analysis through
the  Mininert valve;  syringes  were   tagged and immediately
analyzed.
                               216

-------
       PROBE TIP
PROBE SHAFT
    CROSS-SECTION
          A-A*
Figure 7.3.  Diagram of Lockheed-QtSCO SOV probe tip and shaft.
                             217

-------
INSERTION
    TOOL-
I EXTRACTION
TOOL
        Figure 7.4.  SOV probe driver and extractor.
                            218

-------
          MSA SAMPLAIR
            ATTACHES HERE
MANIFOLD
ASSEMBLY
           3 Mm TEFLON TUBE
           PROM SOV PROBE
            ATTACHES HERE
                                                MSA SAMPLAIR PUMP
                Figure 7.5.  SOV sampling manifold.
                                    219

-------
      Analysis  was  performed  with a Photovac Model  10A10
Portable  Photoionization Gas Chromatograph  designed to analyze
volatile hydrocarbons  including  alkanes  above  ethane, and
cyclic compounds such as benzene, toluene,  xylene, and other
aromatics.   The detector of  this particular  GC was a vacuum
ultraviolet  photoionization system which  ionizes hydrocarbons
with  bonding  energies  of  10.2  eV or less.    Different
hydrocarbons  are detected  with  different efficiencies
(different  response per atom of  carbon)  in this  device; the
soil  vapor  data below are given  in terms of  the  equivalent
concentration of benzene.

      Background  concentrations  of   total non-methane
hydrocarbons and five individual  hydrocarbons in ambient air
were  measured at locations 2-km east and 2-km west of the  study
area;  the observed  values are listed  in Table 7.1.   Figure 7.6
shows  the  total non-methane  hydrocarbon  (NMHC) concentrations
in soil gas as ppmv-benzene equivalents  throughout the  site.
The hatched area outlines the  location  of  the gasoline  plume
determined  from drilling.  From  Table 7.1, background  NMHC
concentrations were on the order of 1.0  to 1.5 ppmv (benzene
equivalent). Assuming that a NMHC  soil-gas concentration of
twice  background defines  the limits of  the contaminant plume,
the 3.0-ppmv NMHC soil-gas  contour agrees  reasonably well with
the limits  of the ground-water  plume estimated from drilling.
However, the NMHC soil-gas  plume  appears to lie north of the
plume  outline  estimated  from  drilling,  and soil gas
concentrations in the vicinity of well SP6  are  lower  than  those
both upgradient and downgradient of well SP6.

      Figures  7.7  through 7.11  show  plots  of  the five organic
vapor components measured: ethane/propane, butane, pentane,
benzene,  and isooctane.   The  lightest  components,
ethane/propane, do not correlate well with the  gasoline plume.
Ambient  air background concentrations of  these components were
in the range of 0.6 to 1.1  ppmv  (benzene equivalents), and the
ethane/propane  values shown  on Figure 7.7  are in the range of
0.1 to 1.8  ppmv; these comparatively high  background  values may
explain the behavior of  these  compounds.  Pentane  and benzene
 (Figures 7.9  and  7.10)  were measured  at   concentrations  above
ambient  air background values  at only a  few locations and were
 also  poorly correlated with ground-water  concentrations.
Butane and  isooctane  (Figures  7.8  and 7.11,  respectively) were
measured in  concentrations  significantly above  background, and
 yielded smooth  contour plots which agreed  well with the  plume
 outlines estimated from  ground-water sampling.

 Vertical Cross-Sections

       A  vertical  cross-section of  relative  organic  vapor
 concentrations  was established by drilling  a  north-south line
                               220

-------
TABLE 7.1.  BACKGROUND CONCENTRATIONS AS PPMV BENZENE
                         2 km east         2 km west
                         of source         of source
   Total NMHC              1.43              1.04

 Ethane/Propane            1.43               .64

     Butane                 ND                .2

    Pentane                 ND

     Hexane                 ND

    Heptane                 ND                 -

   Iso-octane               ND                .2

    Benzene                 ND                ND

     Other                  ND                .2
                            221

-------
                                                                                              •.43
                                                                                               meter*
                                                                                                     100

                                                                                                    d
                                                                                  CONTOUR MAP OF
                                                                              TOTAL ORGANIC  VAPOR
                                                                                 Unit*  mg/l OS B«ni«n«
                                                                              •  .02  SOV Sompllng Point
                                                                                 OEnclo»«d Low Concentration
                                                                                  Zoo.
                                                                              ^Contaminated Ground Wotar
Figure 7.6.  Map of  total  organic vapor in ppm aa benzene from
             SOV aampling, August  1984,  Stovepipe  Wells,
             California.

-------
                                       .04
    • < .02
                • .08
   < .02
• '
                                             ••";;M-.-.v.-.-.v.-.-.-.-.-.v.v.-/v.':v
                                                                                                                100

                                                                                                               d
                                                                                                         meters
                                                                                            CONTOUR MAP OF
                                                                                         ETHANE/PROPANE
                                                                                          Unit*:  mg/l o« B«nz«n«
                                                                                          .02 SOV Sompllng Point
                                                                                          Enclosed Low Concentration
                                                                                            Zone
                                                                                          Contaminated Ground Wattr
      Figure 7.7.   Map of ethane/propane  from SOV sampling,  August
                    1984,  Stovepipe  Wells, California.

-------
                                                                           06
                                                                                              06 •
                                                                                                       100
                                                                                                      d
                                                                                                 maters
                                                                                    CONTOUR MAP OF
                                                                                       BUTANE
                                                                                  Unite  mg/l at Benzene
                                                                               •  .02 SOV Sampling Point
O                                                                                  Enclosed Low  Concentration
                                                                                    Zone
                                                                               Fivj Contaminated Ground Water
Figure 7.8.  Map of butane  from  SOV  sampling,  August  1984,
             Stovepipe  Wells, California.

-------
                                                    <.02
                                                                        • < .02
                                                                        •<-02
                                                                                       02
                                                                                             • <.02
                                                                                  CONTOUR MAP OF
                                                                                     PENTANE
                                                                                Unite  mg/l o» Bwmn*
                                                                              • .02 SOV Sampling Pofcit
                                                                                OEndOMd Low  Conc«ntrotlon
                                                                                  Zom
                                                                              3 ContomlnoUd Ground Wot«r
Figure 7.9.  Hap  of  pentane  from SOV  sampling,  Auguat  1984,
             Stovepipe Welli, California.

-------
I )
I 5
                                   < .01
    •< .01
               • < .01
                                                                                             •< .01
                                                                                 CONTOUR  MAP OF
                                                                                   BENZENE
                                                                                 Units:  mg/l as Bsnzsns
                                                                                  .02 SOV Sampling Point
                                                                                  Contaminated Ground Water
                                                                                                    100
      Figure  7.10.  Map of Beneene  from SOV sampling, August  1984,
                   Stovepipe Wells, California.

-------
                                                                                    .01
     • < .01
I I
• •
 I
                                        < .01
                                                                                                           < .01
100
d
                                                                                                           meter*
                                                                                              CONTOUR  MAP OF
                                                                                               ISOOCTANE
                                                                                            Unlttt  mg/l  01 Banz«nt
                                                                                         •  ,02 SOV Sampling Point
O                                                                                            Enclosed Low  Concentration
                                                                                              Zoo.
                                                                                         F;3 Contaminated  Ground Water
       Figure 7.11.  Map of  ieooctane  from SOV sampling,  August  1984,
                      Stovepipe Wells,  California.

-------
of five wells  across the ground-water  plume.  Four other wells
were drilled for other purposes.  Wells  were drilled by using a
hollow-stem auger to minimize disturbance of subsurface organic
vapor concentrations.  Samples were collected by driving a 2 -
inch ID split-spoon sampler, lined with four Shelby tubes  each
4.5 inches long.  Headspace  analysis was  used to evaluate the
samples collected.   This technique consisted of placing a  soil
sample in an airtight container,  then sampling the headspace
gases in the container after an equilibrium was reached between
sample and headspace.  The  concentrations  thus measured are
relative; the  values obtained are not  the actual organic  vapor
concentrations  in the  soil  but  should be  a linear function of
the  actual concentrations.  Samples were transferred from the
Shelby tubes directly into  one-pint mason  jars  fitted  with
Mininert valves.   Samples were allowed to stand for at least 3
hours at  21°C prior  to analysis by gas chromatograph.

      Figure 7.12  is a  qualitative summary of well-drill ing
results across the Stovepipe  Wells study site.  Figure  7.13
shows  vertical profiles  of  ethane/propane, butane, benzene,
pentane,  and isooctane observed in headspace samples  from  well
S P 5  which  is  considered to  be  a  background well.
Concentrations  of  these components remain relatively constant
at  all depths  both above  and  below the  water table.  Figures
7.11 through 7.18  show  vertical  cross-sections  of  headspace
concentrations of  ethane/propane, butane,  benzene, pentane, and
isooctane observed  in samples from wells 36J1,  SPU,  SP1,  SP3,
and SP2.   This five-well  transect lies perpendicular to the
axis of the  plume near its leading edge; at this location  no
free product  was observed standing on the water  table  while
only gasoline components  dissolved in ground-water  were
observed.   Along the cross-section, contamination  appears to
extend downward into the  water  table for several feet.  In the
vadose zone,  etnane/propane and  butane  occur in greatest
concentration just  above  the  water table,  as  expected.

Conclusions

      The survey  of organic vapors  in soils confirmed the
position of the contaminant plume  delineated  by  ground-water
sampling.   Butane, isooctane, and  total organic vapors  were
 correlated strongly with ground-water concentrations;   plume
limits determined  from  these soil-gas components  agreed well
with well-sampling results both in  plume  location and extent.
Ethane/propane had weaker correlations probably  because their
 ambient air background concentrations were of the same order of
magnitude as the soil-gas concentrations.
                                228

-------
                                                                        •
                                                                        t
                                                                        I
                                                                       !  a  36J1
                                                                               SP4
VO
                                                                                       U.S.O.S. Uncantamlnotod Wefl
                                                                                       U.S.O.S. Contaminated W«U
                                                                                       LEMSCO Wsll-Unoontamlnotad
                                                                                       LEMSCO VMI-VwIoM Zon«
                                                                                        Contaminated
                                                                                                      ContamlnaUd
                                                                                                  Ground Water
Contaminated
        Figure 7.12.  Summary  of  drilling  result*,  August  1984,
                      Stovepipe  Wells, California.

-------
                              WEU.SPS
                                                    gXHAMATIOII
             0.10
                          1.00
                                       10.O
                                                    too.
                                                               1,000.
Figure 7.13.
Levels  of volatile organics
function of depth, August 5,
Wells, California.
in veil  SPS as  a
198A,  Stovepipe
                                230

-------
                     ISOOCTANE n
                     ppw BENZENE
  TBS?1
           NORTH
                                                       SOUTH
                                          JP^ .. .r
Figure 7.14.  Cross-section of  isooctane levels (ppm)  across
             the  contaminant plume,  August 1984, Stovepipe
             Wells, California.
                                231

-------
          NORTH
                                                       SOUTH
Figure 7.15.   Cross-section  of benzene levels (ppm)  across  the
              contaminant plume, August 1984,  Stovepipe Wells,
              California.
                                 232

-------
                    PENTANEM
                    ppm BENZENE
 vumcM.
          NORTH
                                                        SOUTH
Figure 7.16.
Cross-section  of  pentane levels (ppm) across Che
contaminant plume, August 1984,  Stovepipe Wells,
California.
                                 233

-------
                                                          •P2 (MOUND
                                                             SURFACE
                       BUTANE ••
                      ppm BENZENE
            NORTH
                                                           SOUTH
Figure 7.17.  Cross-section  of  butane  levels (ppm) across  the
              contaminant plume, August 1984,  Stovepipe  Wells,
              California.
                                 234

-------
            ETHANE/PROPANE
                  BENZENE
          MJ1
           NORTH
                                                      SOUTH
Figure  7.18.  Cross-section  of ethane/propane  levels  (pom)
             across  the contaminant  plume, August  1984,
             Stovepipe Wells, California.
                                235

-------
STUDY OF GROUND-WATER  CONTAMINATION FROM INDUSTRIAL SOURCES AT
PITTMAN,  NEVADA

      The Pittman,  Nevada, site  is  a  portion of the east-west
right-of-way  of  a major municipal supply aqueduct  called  the
Pittman Lateral.  The  study  site (Walther,  et al., 1983)  (see
Figure  7.19) lies  in undeveloped desert about 11  miles
southeast of Las Vegas and downgradient of  a chemical refining
and processing complex.   The complex was originally  constructed
during World War II  to refine manganese ore; other activities
which have been performed there include titanium refining  and
the  production of  intermediate components of  pesticides.
Ground-water  contamination at  the site probably began shortly
after the construction of  the complex;  from the mid-1910's to
the late  1970's, unknown quantities of liquid and solid wastes
were routinely  disposed  of  in leach pits and in unlined  ponds
on  property belonging  to  the complex.  A major leak  was
detected  in  1976 in  an underground storage tank  on property
leased by one of the  companies operating in the complex.
Approximately  30,000 gallons of benzene are thought to have
been released in that  incident.

      Depth  to  ground-water v-aries  significantly over  the
length of the contaminant  plume, from 55 to 60 feet  at  the
southern regions of  the  plume to 10 to 12  feet in the  Pittman
area.  Figure 7.20 shows an idealized cross-section along  the
axis of  the  plume from the source to its discharge  in Las  Vegas
Wash.  A  hydraulic gradient of 0.012 has been reported for  the
area with a linear  ground-water flow velocity estimated in
excess of 1000  feet/year.   The surficial  geology has  been
characterized as unconsolidated sand and gravel alluvium  30 to
100 feet  thick.  Below these layers the  hydrologic bottom is
composed of  a comparatively  impermeable muds tone  interspersed
with thin layers of  sand and gravel.  Bands of low permeability
caliche  are found  throughout the region.  Paleo-channels
(buried  deposits of  sand and gravel)  and  surface drainage
channels have been suggested as conduits which may accelerate
contaminant movement.

      Unconfined ground-water occurs in the Pittman Lateral
area at depths of 6  to 12 feet in  calcified but unconsolidated
alluvium.  Since the  early  1970's, a series  of monitoring  wells
have been installed  by the company responsible for  the benzene
leak and by  the U.S.  Bureau  of  Reclamation.  Figure 7.21  shows
concentrations  of  total  dissolved solids in ground-water
measured  in  monitoring wells in the area downgradient of  the
industrial complex.   From  hydrologic studies by  the company
responsible  for the  benzene  spill, a plume  of organic solvents
and  pesticides  was  identified;  Figure  7.22  shows
isoconcentration lines of  benzene in ground-water.  The  areal
extent  of  the  solvent  plume  is about  0.6  square miles  and
underlies residential  and  commercial  portions  of Pittman.
                              236

-------
                   Lower
                   Tailing  *
                   Ponds
PITTMAN
LATERAL
TRANSECT
Upper
Tailing
Ponds
              Industrial
              Complex
                           Approximate  Scale
                                                  N
                      t
                                 Kilometers
        Figure 7.19.  General location map.
                         237

-------
               South
ISJ
U)
00
                                                                                      North
                       2000
6000
10000
14000
18000 It.
                      Figure 7.20.   Hydrogeologic cross-section with the locations  of
                                    sampling boreholes along the contaminant plume.

-------
                         Low«r
                         Telling
                         Ponds
           MAN/LATfRALT
                                          Upper
                                          Tailing
                                          Ponds
                                        Contours of
                                     Total  Ditaolvod
                                          Solid*
                                       (1000  mg/l)
                                            Approximate
                                               Seal*
                                              0     .6
Industrial
Complex
 V
                               Kilometers
Figure 7.21.  Ground-water quality based  on  total dissolved
            solids  (U.S.  Bureau of Reclamation,  1983).
                             239

-------
                                        /  TAILINGS
                                       '    PONDS
                  P1TTMAN
                  LATERA
                                  N  Approximate
                                        Scale
Figure 7.22.  Isocontour projection of benzene concentrations
             (ppm)  in  ground-water (from data 1982-February
             1983).
                              240

-------
Benzene concentrations  in ground-water have been reported to
range from  in  excess  of  500,000 mg/1 near the source  to  5-10
mg/1 in the vicinity of  the Pittman lateral.   These contaminant
plumes  move downgradient through the unconfined aquifer  from
the  industrial  complex through the study area to  discharge
ultimately  in  Las Vegas Wash, the major  surface and subsurface
drainage  path  for  the Las Vegas Valley.   Figure  7.23  is  a
cross-section  along the Pittman Lateral  (perpendicular to the
plume)  of  the  vadose zone, the unconfined  aquifer,  and the
underlying  clay  formation which forms  a barrier to downward
movement from the unconfined aquifer.  Figure 7.21 shows the
study area  and the locations of  monitoring wells, the drilling
logs for  these  locations provided the information  used to
prepare Figure 7.23.   The ground-water  contains a variety of
organic and  inorganic  contaminants.   However, the volatile
organic compounds of interest to this study are chloroform in
the  contaminant  plume on the eastern side of the site and
benzene and chlorobenzene in the plume on the western side.
Tables  7.2  and 7.3 list  the monitoring wells  which were sampled
together with  their  respective concentrations of chloroform,
benzene, and  chlorobenzene, as well as the  dates of sampling.
Figure  7.25 is  a  plot of  the ground-water concentrations  of
these  three  compounds as a function  of distance  along the
Pittman Lateral, perpendicular to the plume.   Figure 7.25 shows
that there  are  two  distinct plumes: one plume on the eastern
end of  the  study area exhibiting significant concentrations of
chloroform,  and  another  to the  west  exhibiting large
concentrations of benzene and chlorobenzene.

      As a  test  of the  soil-gas sampling  system  described in
the Stovepipe Wells  case  study (Figures 7.3-7.5),  soil-gas
sampling was  conducted  along the Pittman Lateral to  develop
profiles of volatile organic concentrations  (Kerfoot  and
Barrows, 1985).   Figures  7.26 and 7.27 show  the sampling  plan
for a series of  samples  collected at a  4-foot depth.   Figure
7.26 shows the  sampling plan used over the  chloroform plume
while  Figure 7.27  shows  the sampling plan used  over  the
benzene/chlorobenzene plume.

      For  chloroform  analysis, the  AID gas  chromatograph  with
electron-capture  detector (AID GC/ECD) was selected;  this
instrument  has  relatively high sensitivity to chloroform.   The
chromatograph  column was  a stainless steel tube, 1/8-inch in ID
and  6   feet  in  length packed  with  10  percent  DC-200  on
80/100-mesh Chromosorb  HP.  The  detector and injection  port
operated at 37°C;  the column, at J»3°C.  The carrier  gas was
approximately  20 cm3/min  of 5 percent methane in argon (P-5).
The chloroform  detection  limit for the AID  GC/ECD was 5 ppbv.
Sample sets from each probe location and  depth point  consisted
of two  or  three  Hamilton  Series  100 Gastight syringes.  These
were 250 ul capacity except for  the vertical profile study

                              241

-------
WMt
              Baruene/CMorobanzeM
East
                                                                       610
                                                     V WawrTabto
   Figure 7.23.  Hydrogeologic  cross section of  the  transect.
                                     242

-------
                                                                                » - o>
                      M
ho
                                 Hydro
                                 Conduit
                                  Corp.
                                           «* « «-
                                                        « «»
                                           LL.LiJ.il
                • Woll
                                                   French
                                                   Drain
QEC Industrie*


     U.S. Homea
                                                                     Sunael Road
                   o
                   oc

                   o
                   o
                                                     Marlayne Drive
                                                                   ---  PITTMAN TRANSECT
                                                                              N

                                                                              t
          Figure 7.24.  tocations of monitoring wells along  the Pittnan

                      Lateral  (LEMSCO, 1984).
        100

-------
(O
                 TABLE 7.2.  CONCENTRATIONS OF CHLOROFORM IN GROUND WATER SAMPLES COLLECTED
                           FROM WELLS ALONG THE PITTMAN LATERAL (nicrograui/liter)
WELL NUMBER
DATE 617
3/83 d
11/84 a
4/85 8
8/85 a
619 621 623
<10 28 500
a a a
8 a 570
a a 541
625
430
850
1000
732
627
181
a
175
a
629 631 633
11 <10 d
a a a
888
888
     d " not detected
     s • not sampled

-------
                             TABLE 7.3  CONCENTRATIONS OF BENZENE AND CHLOROBENZENE IN GROUND-WATER
                           SAMPLES COLLECTED FROM VEILS ALONG THE PITTMAN LATERAL («icrogra*a/lit0r)
Ol
WELL NUMBER
DATE 633

3/83 <10
11/84 a
4/85 a
8/85 a

3/83 <10
11/84 a
4/85 a
8/85 a
635 637 639 641 643 645 647 649
Benzene
340 3800 4700 3200 3400 3100 3500 1300
<10 a a a a a a a
<10 a a 5900 a a a a
888 2548 a 3820 a a
Chlorobenzene
520 4000 5100 3100 2400 4700 3600 2400
<10 a a sa a a a
<10 a 1300 7100 a • a a
as a 4521 a 5060 a a
651 653

37 <10
a •
a •
8 <10

800 21
a a
a a
a 60
       s = not sampled
       a = not analyzed (sample holding time exceeded)

-------
     900-1
o

>
A
a
<
K
III
O
z
o
o
o
U.
o
oc
o
600 —
     300-
         •21
                                          620
                  DISTANCE ALONG PITTMAN LATERAL

                          (lOO'e of !••!)
                             Soil- Gaa Concentration (ppbv)


                             Ground - Water Coneantratlon
                                                         631
Figure 7.25.  Ground-water concentrationa of chloroform,

             benzene and chlorobenzene.
                           246

-------
    631
     o
       N
629
•Q.
625
623
621
  f—200ft.—|

 O Well Location
 • LGAS Probe Location
Figure 7.26.   Probe locations  in Che area of  Che chloroform/
              tetrachloride.
653
O
O
649
0
o
645
O
         (^-200 n.
        O Well Location
   N     • LGAS Probe Location
Figure 7.27.  Probe locations  in Che area of  Che benzene/
              chlorobenzene contaminant plume.
                               247

-------
TABLE 7.4  OBSERVED CHLOROFORM CONCENTRATIONS OVER THE CHLOROFORM
              CONTAMINANT PLUME.  DEPTH STUDY DATA IN TABLE 2.2.
Date, No.
of Well,
Location
7-17-85:
623 S 20
623 N 20
623 W 20
623 E 20
7-18-85:
625 ENE 20
625 WSW 20
625 NNW 20
625 SSE 20

7-22-85 :
627 N 20
627 W 20
627 S 20
627 E 20. S 3
627 E 23, S 3
627 E 23,
627 E 20
7-23-85:
625 SSE 25
627 E 20, N 3
631 E 20
623 E 40
623 E 42

621 W 20
7-24-85:
629 W 20
629 E 20
629 N 20
629 S 20
Chloroform Concentration (ppbv)
Syringe
No. 1

24.9
12.7
119
8

378.2
272.3
330.9
-


73.0
32
45.4
54
166
119
—

492
171
5
32.2
32.7

11

12.4
33.1
22.3
30.9
Syringe
No. 2

26.2
11.8
108
4

381.1
264.7
313.6
-


72.7
23
45.8
54
161
119
—

516
171
5
31.9
24.5

10

8.5
23.8
25.3
31.2
Syringe
No. 3

28.9
12.3
117
bad sample

369.9
261.2
332.4
-


72.9
30
45.4
57
155
99
—

524
not analyzed
5
not analyzed
22.1

11

9.6
25.5
26.6
29.3
mean
[SD/RSD]

2712/72]
12.3[0.5/4Z]
11516/5Z]
6[3/50Z]

37616/22]
266(6/22]
326(10/32]
bad point


72.9(0.1/02]
28(5/182]
45.6(0.2/02]
55(2/42]
161(6/42]
112(12/112]
bad point

511(17/32]
171(0/02]
5(0/02]
32.1(0.2/12]
27(5/192]

10.5(0.3/32]

10(2/202]
27(5/192]
25(2/82]
30(1/32]
Comments






Sampling grid
rotated 22° to
avoid earlier
probe
locations



•
3 ft. x 3 ft.
square



to replace
bad points

two points
check spatial
gradient






                               248

-------
    6000-1
    6000—
    4000-
Ul
o
o
o
3000—
    2000-
o
    1000-
        610
                  620
630
640
                                                                          660
                         DISTANCE ALONG PITTMAN LATERAL
                                (100's of fact)
                                                         Chloroform (ug/1)
                                                         Benzan* (ug/1)
                                                         Chlorobanzana (ug/1)
       Figure 7.28.
                 Chloroform concentrations  at 4-foot depth as a
                 function of distance across plume.
                                       249

-------
where  1000 ul  syringes  were used  because of low soil-gas
concentrations.

      Table 7.4 lists the  concentrations of  chloroform found  in
soil gas over  the  eastern plume; Figure  7.28 compares the
averages of the chloroform  soil-gas concentrations measured  in
the  vicinity  of  each  monitoring well  with  available
ground-water  measurements  from Table  7.2.  Figure 7.28
indicates  good  qualitative  agreement between the shapes  of the
soil-gas and  ground-water profiles  despite the fact that the
ground-water samples were  taken at times up to several  months
earlier than  the  soil-gas  measurements.  Figure 7.29 is a
scatter plot  of soil-gas  chloroform  concentrations  versus
ground-water  chloroform  concentrations (from Tables 7.2 and
7.*, respectively).   The  correlation  coefficient  for this
scatter plot is 0.848.

      A vertical  profile of  soil-gas  concentrations  of
chloroform  and  carbon tetrachloride was obtained near Well 623
in the  eastern  plume;  Table 7.5 lists the data obtained from
successive  samples at 1-foot  Intervals to a depth of 6 feet at
a location where the  water  table depth was  12.5 feet.   Figure
7.12 is a  plot  of  the vertical profiles  of chloroform and
carbon tetrachloride at  this location.  Both gases exhibit
essentially linear  increases  in concentration with  depth;
straight  lines fit by least-squares fit both gases,  with a
correlation coefficient greater  than 0.99.

      Analysis  of  soil-gas samples  taken over the western
benzene/chlorobenzene plume was performed with two different
instruments:   (1)  a Photovac  GC with photoionization detector
(Photovac  GC/PID),  and (2)  an AID GC  with phot oi oni zat i o n
detector   (AID  GC/PID).  The Photovac GC/PID had a 5 percent
SE-30,  1/8-inch Teflon column and ultrapure air was used as a
carrier gas.   The  temperature-controlled AID GC/PID  used a
6-foot, 1-8-inch ID stainless steel column with 3 percent SE-30
on 80/100  mesh  Chromosorb HP.   The  injector port and detector
operated at 89°C and the  column at 82°C.  Ultrapure nitrogen
was  used  as  the  carrier  gas.  The Photovac GC/PID benzene
detection  limit was  1  ppbv,  and the AID GC/PID chlorobenzene
detection  limit was 3 ppbv.

      Concentrations of benzene  and dichlorobenzene in soil-gas
samples taken  over  the western plume were all less  than  10
ppbv and were  not  detected  at all in many  samples.  These low
concentrations  are  unexplained although  the investigators
speculated  that  biodegradation  might be responsible.

Conclusions

      These two case studies  illustrate the use of soil-gas
measurements as  a means of  delineating ground-water contaminant
                              250

-------
z
o
2
Ui
o
2
O
0
Ik
o
oc
o
.J
z
o
o
I
-I

o
CO
     400 -i
                                                          369.866 O
300  -
     200 —
100  —
                100    200   300   400    BOO   600   700   800


                 GROUND-WATER CHLOROFORM  CONCENTRATION
                                                                000
   Figure 7.29.
           Soil-gas  chloroform concentrations at 4-foot

           depth as a  function of ground-water chloroform

           concentrations (r " 0.848).
                                    251

-------
                             TABU 7.5  OtOROFORM AND CARBON THRACHLORIDE CONCENTRATIONS IN SOIL CAS
                                                       AS FUNCTIONS OF DEPTH
K)
(J1
Chloroform Concentration (ppbv)
Location
(Baaa at
No. 623*100


Baaa
3 rt north
6 rt north
Baaa
3 rt north
6 rt north
Baaa
3 rt north
6 rt north
Baaa
3 rt north
6 rt north
Baae
3 rt north
6 rt north
Carbon Tetrachloride


Depth
(rt)
1
1
1
2
2
2
3
3
3
4
4
4
5
5
5


Syringe
No.l
22.9
23.5
18.5
76,9
72.8
55.4
109.0
111.0
83.4
160.0
150.0
122.0
206.0
188.0
150.0
Heaaurad

Syringe Syringe Mean
No. 2 No. 3
23.0
22.3 21.2
18.3 20.1
75.3
69.2 68.3 68.4
60.2
110.0
110.0 105.2
111.0 102.0
146.0
148.0 144.7
142.0
206.0 202.0
177.0 187.4
183.0
Concentrationa



35.7
36.0
35.0
106.0
101.0
87.0
145.0
149.0
134.0
198.0
194.0
196.0
249.0
226.0
212.0
(ppbv)


Mean

35.6


98.0


142.7


196.0


229.0

          Base
256.0
216.0
236.0
                                                                                                   273.0
                                                                     273.0

-------
plumes.   In the first  study, vapors from  leaked gasoline
(dissolved  in  ground-water and possibly  lying on top of  the
water  table)  moved upward through approximately 40 feet of
alluvial  overburden  (largely  sands  and  gravels)  in
concentrations  large enough to be measurable at depths of about
5 feet.   By  sampling soil gas  over  the  entire  study area with
steel probes driven to 5 feet and by  analyzing the samples in
the field with  a transportable  gas  chromatograph,  the lateral
extent of the  plume could be mapped.  Not  every vapor component
appeared  to  be a good tracer  of ground-water contamination;
total non-methane hydrocarbons,  butane, and isooctane appeared
to  outline  a  plume which was  not  inconsistent  with  that
estimated from ground-water  sampling.   Contours drawn with
ethane/propane  and benzene  did not resemble the ground-water
plume.   Benzene concentrations near the surface were so low as
to  be nearly unraeasurable  despite substantial quantities in
ground-water;  this implies  that some  phenomenon depletes
benzene from the organic  vapors at some  point in the migration
path.  The choice of tracer  is therefore quite  important.

      In  the  second  study, soil-gas  concentrations of
chloroform defined a plume  of chloroform in the ground-water
downgradient of a large industrial plume.  Here  the depth to
water was about 20. feet in desert alluvium.  Ground-water
concentrations of benzene of 2-5 mg/1  caused no measurable
soil-gas  concentrations of benzene.  Again, benzene appears to
be  of  little  use  as  a  tracer.   Perhaps the most  important
result of the second study  is  the  linear  relationship between
depth  and  soil-gas  concentrations  of chloroform  and
carbon tetrachloride in Figure  7.30. This behavior  is exactly
what would  be  expected  if  simple  diffusion processes were the
primary mechanism governing  the vertical migration of the  gases.
Figure 7.30 lends hope that it  may  be  possible  to  derive  a
quantitative relationship  between organic concentrations in
ground-water and those  in soil  gas,  at  least in some locations.
                               253

-------
     uj    o

     O
            COMPOUND  CONCENTRATION


                        (ppbv)


             0       100      200      300

               i i i i i i  i I i i i i i i i I i i i i i i i I
     §
     CO     2 -
     O o   3 -

     we   A
     CD     4 -




     t     5-
     UJ

     0     6 -
^..
    \<'
                         CARBON TETRACHLOR10E
    CHLOROFORM
                 I
•  -v

\\

   X\
                 W«t«r  at  12-1/2  f««t
Figure 7*30* Chlorofor* and  carbon tetrachloride depth

          distribution. Coefficient of determination (r* «

          .99) (Kerfoot, in draft, 1986).
                         254

-------
                        REFERENCES


1.    Buecker,  D.   "Technical Summary  of  Soil and In Situ Gas
     Sampling Study, Basic Management, Inc.,  Henderson,
     Nevada."   Ecology  and Environment, Inc.,  EPA Draft
     Report, June, 198*4.

2.    Kerfoot,  H.  B. and L.  J. Barrows.  "Soil-Gas Measurement
     for  Detection of Subsurface  Organic Contamination."
     Lockheed Engineering and Management Services Company,
     Inc., EPA Draft Report,  1985.

3.    La Brecque, D.  J.,  Pierett,  S. L. and  A.  T.  Baker.
     "Hydrocarbon Plume  Detection  at  Stovepipe  Wells,
     California." Lockheed  Engineering and  Management
     Services,  Inc., and  J.  W.  Hess, Desert Research
     Institute, EPA Draft  Report,  198i».

H.    Walther,  E.  G.,  D. LaBrecque  and D.  D. Weber.  "Study of
     Subsurface Contamination with Geophysical  Monitoring
     Methods at  Henderson,  Nevada."  Lockheed Engineering and
     Management Services Company,  Inc., and R. B. Evans and J.
     J. van Ee,  EMSL-LV,  EPA,   Proceedings of the National
     Conference on Management of  Uncontrolled Hazardous Waste
     Sites,  Washington, D.C.,  1983.
                             255

-------
                            CHAPTER 8

                      SUMMARY AND  CONCLUSIONS
UTILIZATION OF SOIL-VAPOR MEASUREMENTS

      Soil-vapor  measurement Is a  useful  tool in subsurface
investigations.  Its most popular  use is  in  mapping the extent
of ground-water and  unsaturated zone contamination related to
surface spills, leaks  from storage tanks,  and leachates and
leakage from waste disposal sites.   The technique has potential
application in most Superfund site investigations because many
of the most frequently observed  contaminants at Superfund  sites
are volatile organics.   Volatile organics  are such a common
component of ground-water contamination  from Superfund and
interim status RCRA  facilities  that VOC  analysis  has  been
suggested as a detection indicator parameter for interim status
RCRA monitoring.

Transport  Processes

      Organic  liquids spilled or applied on the surface will
tend to migrate downward through  the  vadose zone and will leave
a cone of material  contaminated by residual amounts of the
organic.   Many properties of both the organic liquid  and the
subsurface will Influence the  downward migration of the organic
and its subsequent behavior, as discussed  in Chapters 2 and 3.
Once a low-density organic liquid has reached-the water table,
it will begin to create a "mound" and to  spread.  Behavior of
each organic contaminant will be governed by its individual
characteristics such as density,  water solubility, and tendency
to adsorb onto clays  and organic constituents of soils.  The
low-density fluids will float on  the water  table and continue
to spread; some fraction of the  various organic components will
dissolve  in ground-water and move  with  ground-water  flow as
governed  by Darcy's  Law.  Organic liquids which are immiscible
in water  and denser than water  could  sink  through the saturated
ground-water zone  if not bound  to  soil as  residuals.

      The volatile components of the contaminant will release
chemical  to the vapor  phase,  which will then either migrate
toward the soil surface by diffusion or sink through the soil
air if the partial pressure of  the vapor  is  high and the vapor
is denser than air.   The rate of migration will be a function
of the soil resistance to  vapor  flow, of the amount which is

                              256

-------
redissolved  Into the liquid phase, and of the amount which is
adsorbed or  degraded.  In most if  not  all  cases,  the rates of
vapor migration will be much faster than rates of ground-water
movement.  For  a situation where  the dissolved  components  have
migrated downgradient of the original spill location, initial
soil  vapor  concentrations of  the  organic  contaminant
immediately  above  the water table  can be predicted from Henry's
Law, as  discussed in  Chapters 1  and 3.  Chapter 1 discusses the
simple  situation where the organic vapor is "conservative" and
is not  degraded or absorbed, where the vadose (unsaturated)
zone  is homogeneous  in composition, and where the  rate of vapor
loss  f~rom the surface  is small  compared  to  the  aqueous
concentration.   An  example might be  a  chlorinated  organic
solvent  in a  homogeneous sand.  In  this case theory predicts  a
linear  decrease  from  the  initial Henry's  Law vapor
concentration just  above the  water  table to  near  zero
concentration  at the  surface.  As  discussed in Chapter 1, field
data exist which confirm this behavior in  certain situations.
Chapter 3 presents a solution of  the slightly more complicated
situation where  the  organic  vapor  undergoes  first-order
degradation  or adsorption.  These two idealized situations are
probably uncommon in nature since  the subsurface vertical
cross-section  is  rarely homogeneous.   In  general, the least
permeable layer of  the vertical  section  will  control the  rate
of  vertical  migration.   The  vertical profiles  of  gas
permeability,  moisture  content,  and soil  properties  will
usually be  hard  to  estimate  at  a  particular  site; thus,
determination  of  the relationship between the vapor  source
(residual amounts  in  soil and  concentrations  dissolved in
ground-water) and  the resulting vapor  concentration profiles
will generally  not  be  possible.

      The fact  that  a mathematical expression cannot usually be
written to  relate soil-gas  concentrations  to ground-water
concentrations of  volatile organics does  not obviate the
usefulness  of soil-gas surveys.   Detection of volatile organics
in  soil  vapors indicates that there is  clearly a source of the
organics somewhere in the subsurface,  and  increasing soil-gas
concentrations will usually indicate increasing amounts of a
substance.   Mapping of soil-gas  concentrations measured at
uniform depth, can be useful in defining the lateral extent of
subsurface  contamination.

Soil-Gas Surveys

      Figure 8.1  is  a flowchart of the planning and execution
of  a  soil-gas  survey.  Preliminary sampling  is  performed to
determine  vertical  profiles  of subsurface organic vapor
concentrations  at several  locations  within  the  site to be
surveyed.   A  tracer  gas must be  chosen;  selection of tracer
gases for  the  survey  is   aided  when  prior  information is

                              257

-------
I  Preliminary sampling:  vertical profiles |
      I  Selection of tracer gases |
                     *
        Selection of survey depth
   Sampling on a grid  at uniform depth
                    i
             Sample analysis
                    I
             | Data analysis
        Figure 8.1. Flowchart of coil-gas surveys.
                     258

-------
available about  the types  and concentrations of  volatile
organics found in  ground-water.  Halogenated  organics  are
generally  preferred as  tracers because they  tend to be
"conservative"  and  resistant to degradation; measurements of
total hydrocarbon  concentrations may also give good results.
Based on the preliminary vertical  profiles,  a sampling  depth is
chosen which appears likely to  provide.gas  concentrations large
enough  to be  readily quantified by available  analytical
techniques.   Soil-gas  samples are  then  collected  over  the
survey area on a predetermined grid at the uniform  sampling
depth.  The samples  are  analyzed on site or are  transported to
a laboratory for analysis.  Data analysis consists of  plotting
the  values  on  a map  of  the  survey  area  and  drawing
isoconcentration  lines  either by hand or with  a computer
algorithm.  The final  objective of the soil-gas  survey will be
the  siting of monitoring wells to  obtain representative
measurements of ground-water  concentrations  of the
contaminants  or  to obtain  core samples to determine  the
concentrations  of the contaminants in the soils.

Sampling Devices

      Figure 8.2  is a schematic of  the  sampling and analysis
process.   Soil  vapor samples are obtained  with any  of several
methods.  The  available methods include headspace measurements,
ground probes,  flux chamber  measurements, and passive sampling.
While any of these techniques  may be  used to detect subsurface
hydrocarbon  contamination, they  are not equivalent.   The  ground
probe and headspace  measurement  techniques  measure a soil-gas
concentration,  the  flux chamber  technique measures  an  emission
rate, and the  passive sampling  technique measures some function
of an average  soil-gas concentration.  The technique  chosen
will  depend on  study objectives, on  the  magnitude of  soil
hydrocarbon vapor  concentrations, and  on  the sensitivity of
available analytical instrumentation.   The flux  chamber is
generally  appropriate for applications  where a measure  of human
exposure  is to be  determined and where soil-gas  concentrations
are relatively large  because the introduction  of  sweep  air
effectively dilutes the organic vapor concentrations.   The
passive sampling  techniques  serve  to  integrate soil-gas
concentrations over some period of time  and could be  of  use
where detection  of very low  concentrations  is  necessary.
Ground  probes are probably  the most  widely used technique
because of speed and cost.  Headspace measurements  can be  used
in preliminary surveys to monitor concentrations in monitoring
wells, storm sewers,  utility  vaults,  or  other  subsurface
structures, and  as a first step in  planning more extensive
investigations.   The technique can also  be used  to measure
headspace concentrations of  organic vapors  in  containers
holding soil core  samples; this approach  has  been used to


                               259

-------
  SAMPLING DEVICE
SAMPLE  COLLECTION
          i
  SAMPLE  ANALYSIS
Figure 8.2. Flowchart of coil-gas aeasurenenta.
           260

-------
provide  vertical  profiles  of relative  organic vapor
concentrations.

Sample Collection

      Accurate  determination  of  the  concentration and
composition  of  organic compounds  in soil  gas usually requires
the collection of  samples which are  subsequently analyzed in a
laboratory where  conditions are  sufficiently stable to permit
GC  analysis.   Analysis  may take  place  in  a mobile  field
laboratory or  in  a  sophisticated  modern  analytical laboratory
at  considerable distance from  the  site.   In either case, a
representative sample must be  collected,  and its  integrity
must  be maintained until it can be analyzed.  The sampling
technique must be compatible  with  the  analytical  method.
Relatively insensitive analytical methods  require large samples.
Long  sample  transport  distances  require  rugged sample
containers.   Sample  collection methods for VOCs in gases  can be
divided into two  classes:

      o   adsorbent methods  where the gas is passed through a
          solid adsorbent which removes  the VOCs  from the
          inorganic gas matrix; and

      o  .whole-air  methods in  which  the entire sample is
          placed  in  a  container and   transported to the
          laboratory.

      Adsorbent materials most  often  used for  VOCs are
activated charcoal and  porous  polymers  such as Tenax.   Other
adsorbents  which have been  used are molecular sieves,  silica
gel,  and activated  alumina.  The adsorbent method is  attractive
because it concentrates  the  gas components of  interest and
removes many of the components known to add to the  instability
of  the  sample and to interfere with the sample analysis.  The
adsorbents are small  and can  be easily transported  to  and from
the field.   The  limitation  of the adsorbent  methods are
possible irreversible adsorption, incomplete adsorption,  and
artifact formation.   Irreversible adsorption  occurs when
adsorbed components cannot  be completely desorbed.   Incomplete
adsorption  is also  called breakthrough  and results in the loss
of  the  more volatile  sample components.  Artifact formation  can
occur during  thermal desorption  or from reaction  of the  sample
with  the adsorbent  material.   All  three of these possible
problems must  be  fully  investigated  during sample method
validation.

      Collection  of whole-air samples is probably  the most
widely  used technique in soil-gas  investigations.   In  selecting
 an  appropriate  container, the  investigator should  consider  the
 length  of time which the samples must  be  held in  the  container,

                               261

-------
the need for ruggedness If the samples are to be shipped long
distances, and  the  ability to clean  the container between
samples.  Three  different classes of containers  have been used
in such  investigations:

      o   plastic bags made of  Tedlar or Teflon;

      o   passivated stainless-steel cannisters and syringes;
         and

      o   glass syringes.

      Tedlar  and Teflon bags  are convenient  and inexpensive.
However, photochemical  reactions  can degrade samples  if  the
bags are exposed to light;  sample leakage during transport  can
occur; and gas species can permeate Into the  bags  and  out  of
the bags during  transport and storage.  Consequently, plastic
bags are not recommended for soil-gas surveys unless storage
times  are short, and concentrations are high.

      Passivated stainless-steel  cannisters are  sturdy  and
impervious to light, and they can be cleaned easily.  Samples
can be held for periods up to  several days in  properly prepared
cannisters.  Stainless-steel syringes are on the market  which
should have the  same  advantages as the cannisters.  However,
such  devices are  expensive  and represent  a  considerable
Investment.

      Glass  containers such as syringes have  also been used to
collect  gas samples.  However,  glass  containers  usually require
Teflon valves and seals which  can be a source  of  contamination.
Because  glass transmits light, they  should  be kept out  of
direct sunlight to avoid photochemical reactions in the samples.
Because  of their fragility, such vessels are most appropriate
for on-slte use.  Glass syringes are the most popular devices
for collection  of soil-gas samples when on-site analysis  is
planned.

Sample Analysis

      Analytical methods typically used in soil-gas surveys
include  portable VOC analyzers, field gas chromatographs,  and
laboratory-based GCs.   The method  chosen to analyze soil-gas
samples depends on the  pollutant being monitored, the
concentrations  of  respective pollutant species,  and the
Information to  be  obtained  from   the  analytical results.
Expected concentrations for organic species  in soil gas  can
range from the  pptv level (below most analytical  detection
limits) in  background measurements to several per cent  by
volume in measurements made directly over a liquid lens of  a
highly  volatile  organic fluid   such as  gasoline.  The

                              262

-------
concentration  level  actually  measured will depend on the
sampling method  used and on the degree  of  dilution (as in flux
chambers) or concentration (as in charcoal or Tenax adsorption
devices).  More sensitive methods may  be  needed  at  plume
fringes  than  at  locations  near the source.  Some of the
considerations  involved in selecting the appropriate analytical
technique(s) are

      o    the need for  detailed chemical speciation;

      o    the need for  relative values  or  absolute
          concentrations;
      o    the need for  on-site analyses;

      o    the methods of sample collection

      Portable  VOC  analyzers  can  be useful at sites where total
hydrocarbon  concentrations  are   greater than 1 ppmv.   Such
concentrations are not uncommon at leaking gasoline storage
tanks and gasoline spill sites.  Portable VOC analyzers  can  be
particularly  convenient  when used in conjunction with driven
ground  probes to  sample  directly  from the  probes.   The
advantages  of  portable VOC analyzers are the elimination  of
sample collection  and transport; their major disadvantages  are
calibration  problems, the  need  for large sample volumes,  and
lack of  sensitivity.

      Gas chromatography  is  used  where concentrations  less than
1  ppmv must  be  measured or where chemical  speciation  is needed.
Detection methods used  for GC   analysis of soil-gas samples
include:

      o   flame ionization  detector   (FID)  for  volatile
          hydrocarbons;

      o   ph o t o ionization detector  (PID)  for  aromatic
          hydrocarbons  and sulfur  species;

      o   electron capture  detector (ECD)  for  halogenated
          hydrocarbons;

      o   Hall electroconductivity  detector  (HECD)  for
          detection of  halogenated species, nitrogen-containing
          organics,  or  sulfur-containing  species;

      o   flame photometric detector  (FPD) for  sulfur  and
          phosphorous compounds.

      Field  gas chromatographs  are used where samples  can be
 collected and  analyzed on site.   The simplest field  instruments
 are portable  VOC analyzers with  chromatographic options  such as

                               263

-------
the Photovac equipped with  a  PID and the Century OVA  equipped
with a FID.  The useful detection limits for such Instruments
are  usually  on  the order  of 1 ppmv.   A step above  the
"portable" GCs are the "field"  GCs:  small, sturdy GCs  with
temperature-controlled ovens  and a variety of injectors and
detectors.  Best results  are  obtained with instruments  with
heated-gas sampling valves  for  injection  of the gas  samples.
The detectors most commonly found are the FID, PID, and  the BCD.
The  ECD  is  quite  sensitive, with  detection limits  for
halogenated compounds  in the  ppbv range.

      Soil-gas samples  will  need to be  sent to an  off-site
laboratory when positive  chemical  species  identification is
required, when very low detection limits  are required,  when
difficult  sample  matrices  are encountered, or when
environmental  conditions  prohibit on-site  analysis.
Combinations of detector types in multi-detector systems are
useful for specific types of  analyses. Mass spectrographs can
be used where positive identification of  chemical species is
otherwise difficult.  Modern  analytical laboratory methods are
readily available  with the  capability  to separate  and
ouantitate VOCs in soil gas at  concentrations less than 1  ppbv
and to identify  gas  species with  high  confidence.   The
investigator's task is to  choose a  level  of sophistication
which  accomplishes the survey objectives at a reasonable  cost.

Quality Assurance and  Quality Control

      The  desired product of most soil-gas surveys is a  contour
map of soil-gas  concentrations, in  either two  or three
dimensions.   The investigator usually  intends to  use the
contoured  soil-gas measurements as a surrogate for other  more
expensive measurements such as  ground-water or  soil residual
concentrations.  The important  factors are the  location  and
shape  of  the contours and their relative magnitudes, not their
absolute concentrations.  However, individual measurements  must
be comparable to one  another in order  to draw the  contours.  To
insure comparability,  a soil-gas  survey, like any other
environmental  measurements program,  needs  a  quality
assurance/quality control program.   EPA  policy requires  a
written QA/QC plan for any environmental measurements program.
The QA/QC plan should  address the  sampling method,   the
analytical method, and the data reduction and reporting steps.
Discussion of  analytical QA/QC  should  include  method
calibration and  primary and field  standards.  Duplicate
analyses and samples  are necessary to determine the variability
of the sampling and analytical  techniques.  Blank analyses are
required to determine  the level of contamination  attributable
to  the sampling  method and  to the  analytical method.   To
determine the absolute accuracy and  lab-to-lab  variability,


                              264

-------
audit sample  analysis and 1 nterlaboratory  comparison studies
are required.

Data Analysis

      Another  useful step in assuring that measurements are
comparable  is  a  components of variance  analysis,  a statistical
procedure  which  develops a model  to assess  the error  in
reported concentrations  attributable to  each step in the
sampling  and analysis  process.

      Data analysis  for  soil-gas  surveys usually involves
interpolation  between  sample  points and the  drawing  of
contours.   Because soil-gas measurements  represent
concentrations  of certain gas species near the surface,  they do
not necessarily represent the concentrations of the compounds
in an underlying area  of soil contamination  or of ground-water
plume (or  a consistent or "monotone"  transformation of  those
concentrations).   Under  the best  of conditions, soil-gas
measurements  represent a monotone transformation distorted by
measurement errors  of the  concentrations  of the compounds in
soil or  ground-water below the point of measurement.

      Many methods of interpolation  between data points are
available to the Investigator,  such as  linear, inverse  squared
distance, splines,  and kriging.   Most  such methods are
deterministic and  do not rely on a probability model  while
kriging  does need such a model.   The various  common
interpolation  methods  typically give similar  results concerning
the general pattern of soil-gas  contours  (which is the  final
objective  of  a survey).  Kriging has  the  advantage that it
provides an estimated standard  error  for  each  interpolation.
However, the  estimated standard  error in  kriging  is highly
dependent  on  the  probability model  which must be developed
again for each new survey  because each site is. unique.   Good
estimation of a model requires  more  and  better data than are
usually  available from a soil-gas  survey.  For this reason, use
of  a simple spline or inverse-square interpolation procedure is
probably more  efficient.   If an estimate  of the interpolation
error is  desired,  procedures  other than kriging can be
employed.

      Care  must be used in employing computer interpolation and
contouring  packages.  Such  packages are  ill-equipped  to deal
with situations where adjacent  data  points differ by several
orders of magnitude.   Such situations are  not uncommon  in soil
gas  data but  do not  necessarily  represent corresponding changes
in  the  variables  of  ultimate concern  (soil or ground-water
concentrations).   Because such large isolated values are  likely
to  represent  anomalies as  enhanced gas-flow  paths, it would be


                              265

-------
better  to  simply  flag these  values rather  than to incorporate
them into the  interpolation  process.

Case Studies

      The two case studies  presented here Illustrate the use of
soil-gas measurements as a  means of delineating ground-water
contaminant  plumes.   In the  first study, vapors from  leaked
gasoline (dissolved in ground-water and possibly lying  on  top
of the  water  table)  moved  upward through approximately 40 feet
of alluvial  overburden (largely sands  and  gravels)  in
concentrations large enough to be measured at  depths of about 5
feet.   By  sampling  soil gas  over the entire study area with
steel probes  and  by  analyzing  the samples  in the field with a
transportable gas chromatograph, the lateral extent of  the
plume could be mapped.   Not  every vapor component appeared to
be a  good tracer of ground-water  contamination.  The choice of
tracer  is therefore quite important.

      In the  second study, soil-gas concentrations  of
chloroform defined  a plume  of  chloroform  in ground-water
which was  downgradlent  of  a  large Industrial plume.  The most
important  result  of  the second  case  study  is probably  the
observed linear  relationship  between depth and  soil-gas
concentrations of chloroform and carbon tetrachloride which
would  be expected  if  simple  diffusion processes  were  the
primary  mechanism governing  the vertical migration of the gases.
This case  study lends  hope that  it may be possible to derive a
quantitative  relationship  between organic concentrations  in
ground-water and those in  soil gas, at least for  some  gas
species  and at some locations.

      Both case  studies Involved  the  use of  other  methods to
support  the data that were obtained from the soil-gas  methods.
Geophysics and hydrogeologlc  information from  existing wells in
the vicinity of the soil-gas surveys were  useful in the  design
of the  survey and the  Interpretation of the results.  Like  any
step  in a ground-water study  soil-gas  measurements  by
themselves are a  useful but  not exclusive  means for defining
the extent  of  ground-water contamination from organics.
                              266

-------