EPA/540/2-89/024
                                        PROTECTION
BALLAS, TE
                                              A,
      SUPERFUNDTREATABILITY
            CLEARINGHOUSE
               Document Reference:
International Technology Corp., AFESC, EG&G Idaho, Inc. 'Technology Demonstration
of a Thermal Desorption/UV Photolysis Process for Decontaminating Soils Containing
     Herbicide Orange." Prepared for EG&G Idaho. 14 pp. Technical report.
              EPA LIBRARY NUMBER:

            Superfund Treatability Clearinghouse -EWGE
          HOE DO HT IMO'/E Ffi

-------
                SUPERFUND TREATABILITT  CLEARINGHOUSE ABSTRACT


Treatment  Process:       Physical/Chemical  - UV  Photolysis

Media:                   Soil/Generic

Document Reference:      International  Technology Corp., AFESC, EG&G  Idaho,
                         Inc.   "Technology  Demonstration of a Thermal
                         Desorption/UV  Photolysis Process for Decontamin-
                         ating  Soils Containing  Herbicide Orange."
                         Prepared  for EG&G  Idaho.  14 pp.  Technical  report.

Document Type:           Contractor/Vendor  Treatability Study

Contact:                 Major  Terry Stoddart
                         U.S. DOD/AFESC
                         BLDG 1117
                         Tyndall Air Force  Base, FL 32403
                         904-283-2949

Site Name:               NCBC Gulfport, MS; Johnston Island; and Guam
                         (Non-NPL)

Location of Test:        Gulfport, MS and Guam

BACKGROUND;  This treatability study report presents the results of
laboratory and  field  tests on  the effectiveness of a new decontamination
process for soils containing 2,4-D/2,4,5-T and  traces of dioxin.  The
process employs three operations, thermal  desorption, condensation and
absorption of contaminants into a solvent  and photo decomposition.
Bench-scale tests were conducted  to establish the relationships between
time and temperature and  treatment efficiency.  A pilot-scale (100 Ibs/hr)
system evaluation was conducted at two sites to evaluate system performance
and develop scale-up information.
OPERATIONAL INFORMATION;  The  intent of the laboratory and pilot-scale
tests was  to reduce the  combined dibenzo dioxin and furan constituents,
which originate from Herbicide Orange  (HO), to  less than 1 ng/g.  This
level represents the anticipated soil  cleanup criteria.  The soils used had
similar concentrations of HO contaminants, but were different types  of
soil.  In  the laboratory  the contaminated  soil  is passed through thermal
desorber and the off gases from the soils, including the contaminants, are
passed through a scrubber that uses a hydrocarbon solvent.   Contaminants
dissolve in the solvent and the solvents are passed through a flow reactor
which subjects the contaminant to UV radiation  to decompose the contaminant
molecules.  Testing was conducted on soil samples from three HO contami-
nated sites; Johnson Island, Eglin AFB and NCBC in Biloxi,  MS.   The  soils
tested had 2,3,7,8-TCDD concentrations greater  than 100 ng/g of soil and
2,4,-D/2,4,5-T levels greater  than 1000 ng/g soil.  Tests were run at three
different  temperatures and two different power  levels using high intensity
UV quartz mercury vapor lamps.
    Pilot tests were conducted at the NCBC site using a rotary indirect
calciner as the desorber, an off gas transfer and scrubber system and a
3/89-43                                              Document Number:  EWGE

   NOTE:  Quality assurance of data «ay not be appropriate for all uses.

-------
photo chemical  reactor  to  irradiate  the contaminants contained  in  the
scrubber solution.  A 1200-watt high  intensity mercury vapor lamp  was used
to irradiate  the contaminated scrubber solution.  No QA/QC plan was con-
tained  in  the document.  No discussion of analytical techniques utilized  to
detect  HO  and associated compounds is contained in  the paper.  A detailed
list of soil properties (particle size distribution, surface area, organic
matter, etc.) from  the  three different sites is contained in the document.
PERFORMANCE;  Laboratory studies revealed that thermal desorption/UV
photolysis destroyed all compounds to below their analytical detection
limit (which was generally less than 0.1 ng/g).  The concentration of
2,3,7,8-TCDD was reduced from 200 ng/g to less than 1 ng/g.  Insoluble
brown tars (presumably  phenolic tars) were deposited on the surfaces of the
reactor vessel  and  lamp well.  Reaction kinetics quantum yields' and rate
constants  were  determined.  Pilot tests also produced soil containing less
than 1  ng/g of  2,3,7,8-TCDD.  Table 1 shows the results of the tests.

CONTAMINANTS!

Analytical data is  provided in the treatability study report.  The
breakdown  of the contaminants by treatability group is:

Treatability Group             CAS Number        Contaminants

W02-Dioxins/Furans/PCBs        1746-01-6         2,3,7,8-Tetrachlorodibenzo-
                                                  p-dioxin (TCDD)
                                  TABLE 1

          EFFECT OF TREATMENT CONDITIONS ON RESIDUAL 2,3,7,8-TCDD
                 DURING NCBC PILOT THERMAL DESORPTION TESTS

           Soil Feed    Residence         Soil           2,3,7,8-TCDDb
             Rate         Time3       Temperature           (ng/g)
Test No.    (kg/hr)       (min)           (°C)       Initial     Residual

   1          13.6         40             560          260         ND

   2          13.6         40             560          272         ND

   3          25           19             560          236         ND

   4          44           10.5           560          266         ND

   5          20           24             460          233         0.5
Notes: a)  Soil residence time in heated zone.
       b)  Detection level for 2,3,7,8-TCDD was generally less than 0.1
           ng/g with a range of 0.018 to 0.51 ng/g.
       c)  This is a partial listing of data.  Refer to the document for
           more information.
3/89-43                                              Document Number:  EWGE

   NOTE:  Quality assurance of data Bay not be appropriate for all uses.

-------
   Technology Demonstration of a Thermal Desorption/UV Photolysis
    Process  for  Decontaminating  Soils  Containing  Herbicide Orange

         R.  Helsel,  E.  Alperin,  T.  Geisler,  A.  Groen,  R.  Fox
     International Technology  Corporation, 312  Directors  Drive,
                     Knoxville,  Tennessee  37923

                          Major  T.  Stoddart
          U.S. Air Force, Engineering and Services Center,
                Tyndall  Air  Force  Base,  Florida 32403

                            H.  Williams
            EG&G Idaho, Inc.,  Waste Technology Programs,
                      Idaho  Falls,  Idaho 83415
         Laboratory and  field testing  determined the  effec-
         tiveness of a  new decontamination process for  soils
         containing 2,4-0/2,4,5-T and  traces  of dioxin.   The
         process employs  three primary  operations -  thermal
         desorption to  volatilize  the contaminants,   conden-
         sation  and  absorption   of  the  contaminants  in  a
         solvent,   and  photochemical  decomposition   of  the
         contaminants.    Bench-scale  experiments  established
         the  relationship between desorption conditions  (time
         and  temperature) and  treatment  efficiency.    Labora-
         tory   tests  using  a  batch  photochemical   reactor
         defined  the   kinetics  of  2,3,7,8-TCDD  disappear-
         ance.   A pilot-scale system was assembled to  process
         up  to  100 pounds  per  hour  of  soil.    Tests were
         conducted  at  two  sites  to evaluate  treatment per-
         formance  and  develop scale-up information.  Soil  was
         successfully   decontaminated   to  less   than   1  ng/g
         2,3,7,8-TCDD  at temperatures above 460°C.
As part of a major program being conducted by the U.S. Air Force  to
restore  to  normal use  several  Department  of  Defense  sites where
soils  have  been contaminated  with low  levels  of Herbicide Orange
(HO),  International Technology Corporation  (IT), .under subcontract
to EG&G  Idaho,  has been  conducting  a project  involving  laboratory
bench-scale and  field pilot-scale tests  to  demonstrate  a new soil
treatment process  -  thermal  desorption/UV photolysis (TD/UV).  The
intent  of  the  demonstration  was  to  reduce the  combined  tetra-,
penta-, and  hexa-chlorinated  dibenzodioxin  (CDD) and  furan (CDF)
congeners,  which originated  from the  HO,  to  less than 1 ng/g, which
represented  the anticipated  soil  clean-up  criteria.    Treatment
should also effectively  remove the  primary  HO constituents, 2,4-D

-------
 and 2,^,5-T.   Two sites were  included in the  field  demonstration
 project for the TD/UV  process,  each having substantially different
 types   of  soil  but  reasonably  similar,  concentrations  of the  HO
 constituents.   Testing at  the Naval Construction Battallion Center
 (NCBC)  at  Gulfport,  Mississiopi  was  conducted by  IT  during  May
 '985;  testing  at Johnston Island (JI) in'cne Pacific Ocean occurred
 in  July 1986.   Based  on the results of these field pilot demonstra-
 tions,  an engineering  and  cost  evaluation is  being  performed  for
 applying  TD/UV technology using  large, mobile systems  for these  two
 sites   or  other sites  having similar  contaminated soil  problems.
 This  paper describes the technology,  highlights the results of  the
 initial laboratory test phase,  and  summarizes  the  field demonstra-
 tion  results.

 Process Description

 The  thermal   desorption/UV  photolysis  process  developed by   IT
 accomplishes substantial  volume  reduction and toxicity reduction  by
 concentrating  the  hazardous constituents  contained  in  the soil  into
 a  small volume which  is  easier  to  treat  than  large  quantities  of
 soil.   The process  incorporates  three steps:

      Desorption -  heating  the soil  to volatilize the  organic
                    contaminants
      Scrubbing  -  collecting the  volatilized organics in a
                    suitable solvent
      Photolysis -  converting the  contaminants  to relatively
                    non-hazardous residues  through photochemical
                    reactions.

      A  schematic  block-flow diagram is  presented  as  Figure  1.
 Contaminated  soil  is  passed continuously through  an  indirectly
 heated  desorber which  can be  one  of many  types  of  conventional
 equipment applicable  for  thermal processing of  solids.   The treat-
 ment  performance of  the  desorber  is  controlled by  the  residence
 time and  temperature of  the soil.   Treatment  requirements (i.e.,
 operating conditions) are determined by the volatility of  the soil
 contaminants and the  required contaminant  removal efficiency (final
 versus  initial concentration).
      The  off-gas  leaving  the  desorber   contains  organic  vapors,
 water   vapor   originating   as initial soil  moisture,   and small
 quantities of  air  which  enter with  the  soil.   Scrubbing  using  a
 high boiling  hydrocarbon solvent  is used  to  treat the  off-gas  to
 remove  the organic  contaminants  and  water  vapor  by  cooling,  conden-
 sation, and  absorption.  Particulates (e.g.,  fine soil)  which may
 be  entrained  by  the  off-gas are  also collected by the  scrubbing
 solvent.   Scrubbed off-gas is passed through a  conventional emis-
 sion control system,  such as carbon  adsorption, to ensure  that  no
organic  contaminants  or  solvent  vapors  are released.    Scrubber
solvent  is  recirculated  to  the  scrubber  after  being  processed
 through a system  of  phase  separation,   filtration,  and  cooling.
Condensed water, which  is immiscible  with  the solvent,  is separated
and either directly treated using conventional  techniques,  such  as
 filtration  and carbon  adsorption,  or discharged  to  an  existing

-------
 wastewater treatment facility.  Filtered solids are recycled to the
 desorber   or   packaged  as  process  waste  for  off-site  disposal,
 depending  on  the relative quantity and composition.
       A  small  portion  of  the  recirculated  solvent  stream  is
 diverted  to a UV  photolysis  system to treat  (detoxify)  and remove
 the organic contaminants,  with  the treated solvent  purge recycled
 to the  scrubbing system.    The  equilibrium  concentration of  the
 contaminants   in  the  scrubber  solvent  is  maintained as high  as
 practical  to  minimize the purge  stream and afford higher  photolysis
 reaction   rates,  thereby  decreasing  the  size  of  the  photolysis
 treatment  system.   The concentration limitation is dependent on  the
 solubility properties and  partial pressure of  the contaminants  in
 the solvent,   and  the resultant  effect  on scrubber efficiency  and
 emission  potential.   The  photolysis system  contains a  specially
 designed  flow  reactor which  subjects  the  contaminant-laden solvent
 to UV radiation  to  induce  molecular  decomposition.   High intensity
 mercury  vapor lamps  produce'  a  band  of  wavelengths,  some  of  which
 match the  absorption energy of the specific organic molecules  being
 treated.   Cooling is provided to  the  reactor  to remove the thermal
 output  of the lamp.  The  photolyzed solvent  is treated  by  using
 selected  conventional  physical   or  thermal,  separation  processes,
 such  as  distillation,  to remove  the  reaction  product  residue.
 Alternatively, a purge of  the photolyzed solvent can be discarded
 as  waste  to  control  the   levels  of  reaction  products  in  the
 recirculated solvent system.
      Other  configurations of  treatment  processes  using  thermal
 desorption  as  the  primary  separation technique can  be  applied  to
 organically  contaminated   soils.     Alternative  physical/chemical
 processes  can  be  used  to  treat  the  desorber  off-gas  and  the
 contaminants.    To  achieve complete  contaminant  destruction,  the
 off-gas can  be treated by  using conventional  fume incineration  or
 other  thermal  treatment  technology.   The  choice of  the type  of
 desorber and off-gas treatment system depends  on  the  concentration
 and properties of the chemical  contaminants,  soil  characteristics,
 quantity of contaminated  material,  site  characteristics,  availabil-
 ity of off-site disposal, and  regulatory and related requirements.

 Laboratory Testing and Results - Thermal Desorption

Thermal  desorption   is  a  physical  separation  process,  although
 chemical  transformation  of  the  organic  contaminants  may  occur
 depending  on  the thermal stability  and  the operating  temperatures
 required to  achieve  adequate  decontamination  efficiency.   Thermal
desorption has been used only  in a  limited  number of cases  (1-4)
 for   treating  contaminated  soil,  and  these  applications  have
 involved relatively  volatile  organic  compounds, such as  solvents.
Because  of  the  extremely  low  volatility  of  CCD  and  CDFs,  the
development' of basic treatability  data  was  essential  to  confirm
that  1 ng/g levels  in soil  could be achieved and  that the required
desorption  conditions  were   practical,   considering  the  design
features and operating  rates  of  equipment available for  performing
such treatment.
      Desorption   treatability testing- was  conducted  on  samples  of
contaminated soil  from  three  HO  contaminated sites - NCBC,  JI,  and

-------
 Eglin Air  Force  Base.    The  goals  of  the  test  effort were  to
 evaluate the effect of time and temperature on 2.3,7,8-TCDD removal
 efficiency  and  to  establish  the  importance of  soil  type.    The
 samples  were  selected  by the  Air  Force based  on results of site
 surveys  to  yield high contamination  levels  in order  to investigate
 a broad range of  treatability.  This  testing was an  extension  of
 earlier  testing performed  for the  EPA on  two  dioxin-contaminated
 soil  samples  from  Missouri  to  support EPA's  mobile  incinerator
 trial burn  in  1985  (5).
       After each soil sample was blended,  air dried,  and screened
 (2  mm sieve  opening) triplicate aliquots  were  taken  and analyzed
 for  2,3,7,8-TCDD,   CDD  and  CDF congeners,   and  2,4-D  and 2,4,5-T.
 The   three  prepared  soils  had  2,3,7,8-TCDD levels  greater  than
 100 ng/g and 2,4-D/2,4,5-T levels of about  1000 ug/g.   The JI soil
 had  significant  concentrations of hepta and  octa  CDD  compared with
 the other two  samples.   In addition,  selected physical and chemical
 properties  presented in Table  I, were  measured  (6).    The EPA test
 program  (5) had  indicated  that  soil properties  had  only a  minor
 influence on removal efficiencies for 2,3,7,8-TCDD.

   Table I.  Physical-Chemical Analysis of  Prepared Soil Samples
            Used for Laboratory Thermal Desorption Tests
Parameter
pH
Conductivity (millimhos/cm)
Organic matter (percent)
Cation exchange capacity
milliequivalents/100g)
Oil and grease content
( grams/ 100g)
Surface area (nr/g)
Particle size distribution
(percent)
Medium sand
Fine sand
Silt
Clay (<5 microns)
Moisture
JI
8.4
5.0
4.2
0.73

0.19

6.7


41
37
19
3
2.3
Eglin
3.8
0.15
1.2
0.77

0.41

2.5


41
52
5
2
0.79
NCBC
8.6
0.21
2.3
2.4

0.34

12.3


26
59
12
3
1.1
      A  series   of  10   individual   tests  was  performed   using
temperatures between  430 and 560°C and  treatment  times of 8  to  30
minutes.  Table  II  presents the test results, which are  comparable
to the earlier results for  Missouri soils.  The  objective of  1  ng/g
2,3,7,8-TCDD  residual in  soil  was  achieved  for all  three  soils
subjected to the highest  temperature.  There  was some difference  in
treatability observed between  the  three  soils at the lower tempera-
tures.   Also, longer treatment times  were  required  for the  NCBC
soil because of the higher  initial 2,3,7,8-TCDD  level (500 ng'g vs.
100 ng.'g).   One  set  of  treated tesr  samples which  contained  less
than  1  ng'g  2,3,7,8-TCDD was  also analyzed  for  the  other CDD and
CDF congeners  and  2,4-0/2,4,5-T.   These results,  shown in  Table
III,  indicate  greater than  99.999 percent removal  of  the  initial

-------
 2.4-D/2,4,5-T and the effective  removal  of  higher  chlorinated CDDs
 and  CDFs.

 Table II.  Effect of Treatment Conditions on Residual 2,3,7,8-TCDD
         in Soil During Laboratory Thermal Desorption Tests
 Nominal  Test
 Temperature
Soil
Time at Test
Temperature
2,3,7,8-TCDD
Concentration
   (ng/g)
(8C)
430


U81





558


Identification
JI
Eglin
NCBC
JI
JI
Eglin
Eglin
NCBC
NCBC
JI
Eglin
NCBC
(min)
20
20
30
15
30
15
30
15
30
8
8
15
Initial
106
101
494
106
106
101
101
494
494
106
101
494
Final
38.5
4.4
26. 6a
4.5
1.6
1.1a
0.45
10.1
4.6
0.56a
0.71
0.76a
 Average of duplicate  tests or duplicate analyses.
  Table III.   Residual 2,4-D, 2,4,5-T, and CDD/CDF in Soil Samples
       Treated at 558°C  in  Laboratory  Thermal  Desorption Tests
Concentration (ng/g)
Compound
2,4-0
2,4,5-T
TCDF
OCDFb
JI
ND*
16
0.6
0.3 .
Eglin
ND
0.8
0.4
ND
NCBC
ND
3
ND
ND
^ND = Not detected.
bNo other COD and CDF congeners were detected.

Laboratory Testing and Results - Photolysis

Photolysis has  had limited application  for  treatment of hazardous
waste or  detoxification  of chemically contaminated materials.  The
susceptibility of  chlorinated  aromatics, including herbicides such
as  2,4-D  and 2,4,5-T,  to UV-induced decomposition  is well  estab-
lished  (7.8).    Photodecomposition  of  such   compounds  leads   to
successive  dechlorination  followed by  condensation  reactions   to
form phenolic polymers (7,8).  Other research-has demonstrated that
CDD  and  CDF decompose   in  the  presence  of  UV  light  (6\ 9 JO).
Development of a  photochemical process  for Destroying 2,3,7,8-TCDD
in  a  waste tar  indicated similar  dechlorination  anc condensation
reactions and products (8). The nigh-molecular  weight  end products.
which are similar  in structure to humic  acids,  would be expected  to
have low  toxicity and mobility.    Therefore,  essentially complete

-------
 conversion of toxic  constituents  could produce  a  potentially non-
 hazardous (according  to RCRA),  easily disposable residue.
       Laboratory photolysis  experiments were  designed to  confirm
 that 2,3,7,8-TCDD contained  in  the  selected  scrubber  solvent could
 be reduced  to  1 ng/g  and to determine  the  reaction rates  of the
 primary HO constituents and 2,3,7,8-TCDD in that solvent matrix.   A
 previous  photolysis  process  for  2,3,7,8-TCDD  used  hexane  as  a
 solvent (8).  The solvent selected for use in the TD/UV process was
 different"-  a  high  boiling (kerosene-like)  mixture of  isoparaf-
 fins.    This hydrocarbon solvent was  selected  because of its very
 low  vapor pressure and  water solubility,  nontoxic  and nonflammable
 characteristics,  relatively low cost,  chemical  stability, and good
 solvent properties  for  HO  constituents.   A second  major difference
 from earlier IT  photolysis studies was the presence in the scrubber
 solution of  significant concentrations of other chlorinated  organic
 reactants (2,4-D and 2,4,5-T)  which were also  subject  to photoly-
 sis.    In fact,   the  typical concentration ratio between  2,4-D  or
 2,4,5-T and  2,3,7,8-TCDD in the soil samples  used in the desorption
 treatability testing  was 2000:1.
       The three  steps of the laboratory program included generation
 of scrubber  solution, bench-scale  batch photolysis  reactions,  and a
 pilot  system trial.   In order  to generate a representative  sample
 of scrubber  solution for  photolysis tests, a small desorption and
 scrubbing system was  assembled. . A  portion of  the  prepared  samples
 of both NCBC and JI soil used for the  thermal  desorption  tests was
 used  to generate scrubber solution.   Contaminated soil (-100 g) was
 placed  in a standard  tube  furnace  apparatus  which  was heated  to
 about  500°C  for 15  to 30  minutes.   A  nitrogen  purge swept  the
 vapors   into  the  scrubbing  system,  which  consisted  of  several
 solvent-filled   impingers.    Analysis of the  prepared  scrubber
 solutions indicated  thermochemical  conversion  of the 2,4-D  and
 2,4,5-T in the contaminated soil to  the corresponding  chlorophenols
 at  molar  equivalents.     In  addition  to  using  prepared  scrubber
 solutions, solvent  spiked with  2,4-D and  2,4,5-T, the  corresponding
 chlorophenols,  or  2,3,7,8-TCDD  was  used  for  baseline photolysis
 tests.
      Most  photolysis  experiments  were  conducted  in  a 0.5  liter
 capacity   standard   quartz   photochemical   reactor   using   either
 recirculation or  bottom agitation  for heat  and  mass transfer.   Both
 100- and  450-watt high  pressure  quartz mercury  vapor  lamps (Canrad-
 Hanovia,  Inc., Catalog  Nos. 608A and 679A) were  used, depending  on
 the   initial  reactant   concentration  in   the   particular   solvent
 solution  being  tested.    The  wavelengths  of  interest  based  on
 spectrophotometric  absorbance measurements of  2,3t7,8-TCDD,  2,4-D
 and  2,4,5-T  were in  the 280 to 320 nm region.  Isopropyl  alcohol
 (-0.05  g/g solvent  solution)  was used as a  proton donor  to minimize
 formation of polymeric  reaction by-products which  tend  to foul  the
 light  transmission  surfaces (8).  The  bench-scale  photolysis  tests
 gave the  following results:

 '.  All  compounds disappeared  to below  .the  analytical  detection
    limits.
2.  The  concentration  of  2,3,7,8-TCDD  was  reduced   to  less  than
    1 ng/g from initial  concentrations as high as 200  ng/g.

-------
 3.   For  a  given  reactor  configuration  and  lamp  wattage,  the
     reaction rates  of 2,3,7,8-TCDD and  2.4,5-tricnloropnenol were
     proportional  to  the  concentration,  indicating  pseudo-first
     order kinetics in agreement with previous work (8).
 4.   Absorbence  of UV  energy by  the  solvent,  which increased during
     irradiation,  resulted  in   low  quantum  yields  and  low  rate
     constants.
 5.   Insoluble brown  reaction  products  (presumably  phenolic  tars)
     were deposited on  the surfaces  of the reactor  vessel  and lamp
     well.   This  expected  phenomenon plus the  high  solvent absorb-
     ence demanded a careful reactor  selection and photolysis system
     design.

 Trials   using a  pilot  reactor  system described  in the  following
 section were  performed   in  the  laboratory   to  establish  reactor
 efficiencies  and  operating characteristics  prior to transport  to
 the  field.   A  synthetic scrubber solution was  prepared  containing
 2,4,5-trichlorophenol at a concentration  (-2,000 ug/g)  projected  to
 be   representative  of  the  planned   field  tests.    Kinetics  were
 determined to be first-order with a  rate  constant  of 0.07 sec"1.

 On-site Pilot Testing and  Results

 Based   on  the   information  developed  from   the  laboratory  test
 program,  a  pilot-scale TD/UV  system was  designed  and  assembled.
 Three   skids  were  used  to  mount   the   desorber,   scrubber, and
 photolysis  systems;   the   largest  skid   was  1.5  meters   by 4.3
 meters.     A   conventional  pilot-scale,  rotary,   indirect-fired
 calciner was used  as the  desorber.   The  calciner  consisted of  a
 3.3  meter  long   by  16  cm  internal  diameter  rotating tube  through
 which  the soil  was  transferred,  and a gas-fired  furnace  which
 surrounded the  middle 2.0 meters  of the  tube length.  The  initial
 and  final tube  sections  were  used  for  soil  feeding and  cooling.
 The  flow  rate  and residence  time of soils  traveling  through the
 desorber  were  controlled  by  varying  the  tube  inclination "and
 rotational  speed.     Temperature of   the  soil   was measured   at
 different  locations  by a thermowell  probe   extending  inside the
 tube.   Soil   was fed to the  desorber from a  small hopper  using  a
 variable speed screw  conveyor.   Soil leaving  the  tube was collected
 in a sealed metal can.
      The  off-gas  transfer  and  scrubbing  system  was  designed  to
 enable  recirculation  of scrubbed off-gas  through  the desorber. The
 entire  off-gas   treatment  and   recirculation  system, including the
 desorber and  scrubber,  was operated  at a  slightly  negative  pressure
 to prevent potential  fugitive emissions.   A small  amount of air
 entered  the   system  with  the  soil  feed  or  through seal  leakage.
 Nitrogen was  added to  the  recirculated gas  stream to maintain the
 oxygen  concentration  below the  level  necessary to  support  combus-
 tion.   This was  an extra safety  feature since  the  vapor pressure  of
 tne solvent at normal  scrubber  operating  conditions  is very  low.   A
 portion  of the  scrubbed off-gas was vented  from  the recircuianion
 system  to maintain   proper  pressure  ;n  tne  system.   This  purge
stream  was passed through a small HEPA  filter and  caroon  adsoroer
before  being  discharged  to  the atmosphere.   The  soivent  system

-------
 consisted  of  a scrubber,  receiving and  separation tank,  storage
 tank,  recirculation  pump,  filters  for  removing  suspended  solids,
 and  solvent  cooler.
       The  photolysis  system was  independent of  the desorber  and
 scrubber  systems;  its design  capacity  was lower  than necessary  to
 match  the  desorber's soil-processing rate.   A portion (about UO kg)
 of  contaminant-laden  solvent  was  taken from  the scrubber  system
 after  completion of one  or  more  desorption  tests and transferred  to
 the  photolysis  system.     This   system  consisted  of  an  agitated
 storage tank,  solvent recirculation pump,  and photochemical  reactor
 with   associated  cooling,  DC  power supply,  and  controls.   The
 selected  type  reactor  was a  standard quartz  falling-film  unit,
 approximately  10  cm  in  diameter and 50 cm  long  (Ace Glass,  Inc.,
 Part No.  7898).   A 1200  watt high  intensity mercury  vapor  lamp was
 inserted  through  a  central  quartz tube  within  the  reactor  to
 irradiate  the  solvent as  it  flowed by gravity  down  the  circum-
 ference of  the reactor body.  The  solvent was  recirculated  through
 the  reactor  for  many  cycles   to  achieve  sufficient  irradiation
 (e.g., reaction) time.
       Five  desorption tests were  carried  out  at NCBC  at  various
 treatment  conditions.   A total   of 800 kg of soil  was  processed;
 soil was  prepared  by drying and  crushing  to  less than  1/2  inch  to
 allow  proper  flow  in the desorber  feed mechanism, and blending for
 uniformity.   Each  test lasted  5  to  10 hours,  including  the  heat-up
 and  cool-down  cycle.  Samples of  feed  soil  and treated soil  were
 taken  during steady-state  operation,  and  samples  of the  scrubber
 solvent and  vent  carbon  were taken at  the  conclusion of each  run.
 Samples  were  analyzed  for 2,4-D, 2,U,5-T,  other  HO   indigenous
 compounds,  priority  pollutant  organics  and  metals,   and  tetra-hexa
 congeners of CDD and  CDF.   In addition,  2,3.7,8-TCDD concentrations
 of treated soil and  photolyzed solvent samples  were  determined  on a
 quick-response  basis  to  enable  adjustment  of the operating  condi-
 tions  in subsequent  tests.   Fresh solvent  and carbon were used for
 each test, and  the entire desorber  and scrubber  network  was  cleaned
 out  between  tests.   This  cleaning enabled  thorough inspection  of
 the  condition  of   the  equipment  and  provided  several different
 compositions  of  contaminated  solvent   to  use  in  the   photolysis
 tests.
      Table  IV  shows  the effect  of different soil temperatures and
 residence  times on   residual  2,3,7,8-TCDD for NCBC  pilot  tests.
Table  V  presents  the analytical  results  for  2,4-D, 2,4,5-T, and
 total  CDD  and CDF.   Analytical  detection  levels for 2,3,7,8-TCDD
and  the  various congeners  were   generally  less than  0.1  ng/g but
varied from sample to sample, ranging from 0.018 ng/g to 0.51 ng/g.

-------
 Table IV.  Effect of Treatment Conditions on Residual 2,3,7,8-TCDD
             During NCBC Pilot Thermal Desorption Tests


Test No.
1
2
3
4
5
Soil Feed
Rate
(kg/hr)
13.6
13.6
25
44
20
Residence
Timea
(min)
40
40
19
10.5
24
Soil
Temperature
CO
560
560
560
560
460
2,3

,7,8-TCDD
(ng/g)
Initial Residual
260
272
236
266
233
ND
ND
ND
ND
0.5
 3Soil  residence  time  in  heated  zone.
          Table  V.   Residual  2,4-D,  2,4,5-T,  and  CDD/CDF in
                 NCBC Pilot Thermal Desorption Test
Concentration (ng/g)
Compound
2,4-D
2,4,5-T
TCDD
PCDD
HCDD
TCDF
PCDF
HCDF
CDD and
Test 1
180
500
NDa
ND
ND
ND
ND
ND
ND
Test 2
150
270
0.23
ND
ND
ND
0.14
ND
0.37
Test 3
20
60
0.11
ND
ND
ND
ND
ND
0.11
Test 4
.
_
0.61
ND
ND
0.13
0.54
ND
1.28
Test 5
170
1240
0.75
ND
ND
0.95
1.0
ND
2.70
  CDFC
aNDa not detected.
bTotal of quantified values  for detected cogeners.

      All test  conditions  produced  soil containing  less  than  1  ng/g
2,3,7,8-TCDD.   The total quantified tetra-hexa congeners were  less
than the treatment goal  of 1 ng/g for the first three tests, which
were  performed at the  lower feed  rates.   Test  4,  made  at  the
highest  feed rate, nearly  met this value,  whereas the much lower
soil temperature  used  for  the final test resulted  in almost 3  ng/g
combined residual  CDD  and  CDF.  A longer residence time could  have
improved this  performance.   Residual  2,4-D  and 2,4,5-T concentra-
tions were   less  than  1  ug/g for  all but  the final  test.    This
reduction represents greater than 99.97 percent removal efficiency
for these primary  HO constituents.
      Because of  the very  low  moisture content  of the prepared  soil
feed,  an  insufficient  volume  of aqueous  condensate  was collected
from the tests  to perform analysis or  treatability tests.   A  vent
gas sample  was  taken,  but no  valid analytical  results were gener-
ated because of delays in  sample processing.   However, analysis  of
the carbon   used   in the emission  control  adsorbers  enabled  some
evaluation   of   scrubber  performance  and  process  emission  poten-
tial.   Or.iy  tne front  (upstream)  portion  of  carbon  from one of  the

-------
 tests showed detectable  levels of any  CDD  or CDF.   Mo  HO consti-
 tuents were detected  in  the downstream portion of  carbon.   Calcu-
 lated scrubber removal efficiencies  exceeded  99-9  percent  for CDD,
 CDF,  2,4-D, and 2,4,5-T.  Vent gas volume was about 0.05 nr/minute
 for all tests.
       Results  of  the  photolysis tests  are  presented in Table  VI.
 The  total  solvent volume  (-10.5  1)  was  recirculated through  the
 reactor at 0.75 1/min  for  6.5 hr, resulting  in  28 cycles  with  an
 irradiation time  of  about  1.5 sec/cycle.   The photolysis  system
 operating  time  was  selected  based  on  the  laboratory  trials  to
 achieve less  than  1  ng/g 2,3,7,8-TCDD; the actual  residual  level  of
 0.36  ng/g  represented greater  than  99 percent  conversion.    The
 reaction  conversion  of the other CDD  and CDF  congeners varied from
 85  to  99  percent.  Photolysis  reduced  the  concentrations  of 2,4-
 dichlorophenol   (2,4-DP)   and  2,4,5-trichlorophenol   (2,4,5-DP),
 (corresponding  to  the 2,4-D  and 2,4,5-T  present  in the  initial
 soil) by  85 and 97 percent  respectively.   Figure 2  shows  the rate
 of  disappearance  of  2,3,7,8-TCDD,   2,4-DP,   and   2,4,5-TP.     As
 demonstrated  during  the laboratory  tests,  the  reaction  kinetics
 were  pseudo-first order over  the given  range of  concentrations.
 The  reaction  rate constants  were similar  for  the  three  species
 (0.11 sec"1,  0.04 sec  , and 0.08 sec" , respectively); the  rate
 constant  for  2,4,5-TP was  comparable  to  that  determined in  the
 laboratory  trials  of  the  pilot  system.

  Table VI.  Initial  vs  Final  Concentration  of Selected Compounds
        in Scrubber Solution  from NCBC Pilot  Photolysis Tests
Compound
2 , 4-Dichlorophenol
2 ,4 ,5-Trichlorolorophenol
2,3,7,8-TCDD
Total TCDD
Total PCDD
Total HCDD
Total TCDF
Total PCDF
Total HCDF
Concentration
Initial
490,000
977,000
43.3
46.3
15.7
0.84
31.0
3-7
1.7
(ng/g)
Final
82,000
31,000
0.36
0.92
2.3
0.037
3.8
1.1
0.0031
      Three desorption  tests and one  photolysis  test were  conducted
at JI  to  compare the effects of different soil  characteristics  and
investigate higher  processing  rates.   The coral-like soil used  for
the  tests contained  lower  levels  of  HO contamination  than NCBC
(about 50 ng/g  versus  250 ng/g).   As much as 95 kg/hr of soil  was
successfully decontaminated  to  less  than 1 ng/g 2.3.7,8-TCDD  using
desorption  temperatures  of  550°C.    Treated  soil  from  all  three
desorption tests  had  nondetectable  residual   tetra-hexa CDD  and  'JDF
cogeners,   2,4-D  and   2,4,5-T,  and  corresponding  chiorophenols.
Analysis of  carbon  removed  from  the  desorber-scrubber system vent
showed  no detectable  concentration   of  CDD   or  CDF.    Gas  samples
taken  downstream  of   the  carbon  adsorber   showed  nondetectable
concentrations  of  CDD  and  CDF,  2,4-D  and  2,4,5-T,   and  chloro-

-------
phenols.     Photolysis   test   results  were  comparable   with   NCBC
tests.   Initial  concentrations of HO contaminants were much higher
in the  scrubber solvent  due to  processing of considerably more  soil
and  use of  less  solvent.   The concentration  of 2,3,7,8-TCDD was
reduced  from 780 ng/g  to  less than  0.7  ng/g during  12 hours of
system  operation (representing  about 80 sec reaction or irradiation
time).   Total chlorophenols were reduced from  430 ug/g to less  than
6 ug/g,  and  tetra-hexa  CDD  and  CDF cogeners  were  effectively
treated.    Reaction  rate constants  for  specific  compounds   were
essentially  the same  between  the  NCBC  and  JI  photolysis  tests.  At
JI  as  at  NCBC,  brown   residues  were  deposited  on   the  reactor
surfaces, and  solvent discoloration was obvious, but  there was no
evidence of rate retardation.

Conclusions

The  effectiveness   of  thermal  desorption  to decontaminate   soil
containing HO and of  UV  photolysis to destroy  HO  toxic constituents
has  been demonstrated   in  bench-  and  pilot-scale  tests.     Some
additional technical  information  is needed for a complete evalua-
tion of  the process and  to  provide  the basis  for design  of a full-
scale system for on-site  remedial action.  This project illustrates
the  requirements  for  developing  and  implementing  new  process
technology  for  solving  contaminated-soil  environmental  problems.
Only through such demonstration efforts can more  cost-effective and
environmentally  sound    remedial   action   alternatives   be    made
available.

Literature Cited

 1.    Noland, J. W.;  NcDevitt, N. P.; Koltuniak,  D. L. Proc. of the
      National  Conference   on  Hazardous  Wastes   and  Hazardous
      Materials. Atlanta. GA.  March 4-6.  1986.  pp. 229-232.
 2.    Hazaga,  D;  Fields,  S;  Clemmons,  G.   P.  The  5th National
      Conference on  Management  of Uncontrolled Hazardous  Waste
      Sites.  Washington.  DC. November 7-9.  1984. pp 404-406.
 3.    Webster,   David  M.    J.   Air  Pollution   Control  Association.
      1986,  36,  PP  1156-1164.
 4.    Hoogendoorn,   D.   Proc.   of  the  5th  National Conference on
      Management of Uncontrolled Hazardous Waste Sites. Washington,
      DC, November  7-9,  1984,  pp 569-575.
 5.    Helsel, R.; Alperin, E.;  Groen, A.; and Catalaho, D. "Laboratory
      Investigation  of Thermal  Treatment of Soil Contaminated With
      2,3,7,8-TCDD,"  draft  report  to  U.S. EPA,  Cincinnati,  OH on
      Work Order BAD001,  D.U.D-109,  IT Corporation,  Knoxville, TN,
      Dec.  1984.
 6.  ..Arthur,  M. F.;  Zwick, T.  C.  "Physical-Chemical Characteriza-
      tion of Soils," Battelle Columbus Laboratories, Columbus, OH,
      1984.
 7.    "Report on 2,4,5-T,  A Report on  the  Panel on  Herbicides of
      the President's Science Advisory  Committee," Executive Office
      of the President,  Office of  Sciences  and  Technology,  March
      1971.

-------
  8.    Exner,  J.  H.; Johnson, J.  D.;  Ivins,  0. D.; Wass,  M.N.;  and
       Miller,  R. A.  "Detoxication of  Hazardous  Waste,"  Ann  Arbor
       Science  Publishers,  Ann Arbor,  MI,  1982,  p  269.
  9.    Exner,  J.  H.;  Alperin,  E.  S.;  Groen,  A.; Morren,  C.  E.;
       Kalcevic,  V.; Cudahy,  J.  J.;  and  Pitts, D. M.  "Chlorinated
       Dioxins  and Dibenzofurans  in the Total  Environment,"  Keith,
       L.   H.;   Rappe,  C.;   Choudhary,   G.;   Eds.,   Butterworth
       Publishers, Stoneham,  MA,  1985, p 47.
 10.    Exner,  J.  H.,  Alperin,   E.  S.;  Groen,  A;  Morren,  C.   E.
       Hazardous  Waste. J.,  1984,  pp 217-223.
WPR: thermal

-------
            CD
            0)
                           •   w
                              o
I	I
                             JD

                             O
 CO

 s
p
'to
                                           CD


                                          "5
                                          CO
                                       a

                                    CO CO

                                    O CD
                                    O-Q

                                    §1
                                    <0
 Q.
 0)
 O

 o
o
 (0
 0)
 o
 o
 o
                              CO
                              '55
 L	I
                                     !
                                     CO
                                     ®
                                     O
                          T
                                         
-------
   1000
a
^
o

3
i
c
o
*•
CB
 o
a

"5
 c
 o
£.

 o
O
     OOH
    10H
i.cH
    0.1
                       2,4,5-TP
                                       MOO
          1200 Watt Lamp
               5% IPA
       0     10     20     30     40

           Irradiation  Time (Seconds)
                                       r-1000
                                       Ho
     o
     o
     o
     I-

1.0   co
                                            CM
                                       0.1
                                                  AG6NCY


-------