EPA/540/2-89/044
      SUPERFUND TREATABILITY
            CLEARINGHOUSE
               Document Reference:
Galson Research Corp. "Bengart and Memel (Bench-Scale), Gulfport (Bench and Pilot-
scale), Montana Pole (Bench-scale), and Western Processing (Bench-scale) Treatability
                Studies." 10 pp. July 1987.
              EPA LIBRARY NUMBER:

           Super-fund Treatability Clearinghouse - FCLC

-------
               SUPERFUMD TREATABILITY CLEARINGHOUSE ABSTRACT


Treatment Process:      Physical/Chemical - Dechlorination

Media:                  Sludge/Generic

Document Reference:     Galson Research Corp.  "Bengart and Memel
                        (Bench-Scale), Gulfport (Bench and Pilot-scale),
                        Montana Pole (Bench-scale), and Western Processing
                        (Bench-scale) Treatability Studies."  10 pp.
                        July 1987.

Document Type:          Contractor/Vendor Treatability Study

Contact:                Timothy Geraets
                        Galson Research Corp.
                        6601 Kirkville Road
                        E. Syracuse, NY  13057
                        315-463-5160

Site Name:              NCBC Gulfport, MS (Non-NPL)

Location of Test:       Galson Technical Services, Syracuse, NY

BACKGROUND;  This document presents summary data on the results of various
treatability studies (bench and pilot scale), conducted at three different
sites where soils were contaminated with dioxins or PCBs.  The synopsis is
meant to show rough performance levels under a variety of different
conditions.
    The sites discussed are the Naval Construction Battalion Center (NCBC)
site Gulfport, MS; Bengart & Memel site, Buffalo, NY; and the Montana Pole
site, Butte, MT.  No detailed site descriptions were provided.  There was
no discussion of laboratory analysis procedures, QA/QC plan, or the amount
of soils used in bench scale tests.
OPERATIONAL INFORMATION;  The APEG process for dechlorinating hydrocarbons
was utilized and the amount of reagents/time and temperature were varied.
Two different reagent loading rates were used.  Tests were conducted in
slurry form and in-situ at two of the sites (NCBC and Bengart & Memel).
Unit cost estimates for soil treatment are not provided.  Costs for each
bench-scale test run are estimated at $1,000 for PCBs and $2,000 for
dioxin.  Dioxin tests are more costly due to the complicated analytical
procedures.  The scope of work for the Montana Pole site treatability study
was to see if waste oil containing 100,000 ppb doxin and 2-3% penta
chlorophenol (PCB) could be treated with Galson Terraclene-Cl APEG
treatment.  The scope of work at the NCBS site was to determine the
kinetics of processing dioxin contaminated soil using 30 kg batches in a
modified 55-gallon drum reactor unit.  The scope of work for the Bengart &
Memel treatability study was to determine if PCB contaminated soils could
be treated.
3/89-33                                              Document Number:  FCLC

   NOTE;  Quality assurance of data nay not be appropriate for all uses.

-------
 PERFORMANCE;  The results of the tests on the NCBC site and Bengart & Memel
 soils are shown on Table 1.
     The results of laboratory tests at Montana Pole indicate the reduction
 had occurred,  reducing the dioxin levels from 100,000 ppb to less than 1
 ppb after operating the unit for 1 hour at 150°C.   The results of the NCBC
 study showed that the soil from Gulport, MS could  be decontaminated by
 mixing the soil with APEG reagent and heating to 120°C for 7 hours.  The
 results of the Bengart & Memel study indicates the PCB soil could be
 reduced to less than 50 ppm by adding reagent to the soil,  mixing and
 heating the soil/reagent mass to 120 C for 12-24 hours.   However, no
 significant correlation appears to exist between performance as measured by
 the amount of  contaminant remaining and reagents used,  reagent ratios,
 time,  temperature,  or reagent loading for all the  treatability studies.
 Contaminant destruction appears to take place in-situ or in soil slurry
 form.

 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/PCB        1336-36-3         Total PCBs
                               30746-58-8        1,2,3,4-Tetrachlorodibenzo-
                                                 p-dioxin
                               TOT-DF            Total dioxins  and furans
Note;  This  is a  partial  listing of data.  Refer  to  the document  for more
       information.
3/89-33                                              Document Number:  FCLC
   NOTE:  Quality assurance of data may not be appropriate for all uses.

-------
              C CU
                     0,0,0,0.0.0,0.0,0.0,0,0,0.0.0,0,0.0,0,0,
                     0,0,0.0.0.0.0.0.0.0.0.0,0.0.0.0.0.0.0.0.
                    VWW  -rHCMCOO  'V •  •  -V
                               rHWVOOO   COCMCM   COCN
                               V       r-(
                                                         0,0,0,0.0,0,0.
                                                         0.0.0.0,0.0.0.
                                                        -* \OvOCO -TOO
              C
              CO ,>>>>>,
                                                                        yyyrtaart
                                                                        o o o-o-o-o-o
                                                                               •o

                                                                                o.

                                                                                o
                                                                                N

                                                                                cu co
                                                                               •H >»
                                                                               T3 C
                                                                                O CU
                                                                                U.C
                                                                                o o.
                                                                               1-H-iH
                                                                               .C.O
                                                                                o
                    •4-CM CM CM CM CMin CM CM
                                   •

                                 O
              oo   QppQOOOpmmoooooooooo

              OC    CMrHrHrH
 bO
•H

"8
 o
                                                                        CM CM CM
                                                                        ooooooo
                      oooooo
                                     3OOOOOOOOOOOO
                                     SOCMCMCMCMCMCMCMCMmcMir)
                                                                            OCMCMOO
                                                                        t—ii—*i—t     i—ii—i
                                                                           IH CU
                                                                           *-> t->
                                                                           cu rt
                                                                           *-> G
                                                                            I -H

                                                                            • o
                                                                           oo-c
                                                                            • o
                                                                           CM >>
                                                                            .rH
                                                                           rH O
                                                                             o,
                                                                            II
                                                                              II
                                                                           Q
                                                                           a CQ
                                                                           00
                                                                           HCL,  O
                                                                                 o
 CO CO
 CO CO
 rt rt
 e e

 *-> I—I
 C-H
 CU O
 bo co

 CU
 IH
 X

o
o
                                                                                           bo
                                                                                           C

                                                                                          -o
                                                                                           rt
                                                                                           o
                                                                                                             o
                                                                                                             tu
                                                                                                             (V
                                                                                                             J3
                                                                                                             e
                                                                                                             B

                                                                                                             O
                                                                                                             O
                                                                                                             o
                                                                                                                                 V)
                                                                                                                                 10
                                                                                                                                 rt
I
                                                                                      O CU
                                                                                        •O W
                                                                                           IH
                                                                                         x *H u-i HH <*-m-i VM u-mn HH vw

     33333333333333333333
                        UC-3OUC-3UUUOC-3C-D

                        O rH CM CO •* lO VO r*. 00 ON O
                                                    3CaCrH3C3CrHrH

                                                     I  I-l  I  •• ••
                                                    1-H rH CM rH rH CM CM


                                                    CM CM CM CM CM CM CM

                                                    ONONrHCMCMrHrH


                                                    ONONrHCMCMrHrH



                                                           3333

                                                    IH U lH-iH*>H-iHTH
                                                    U U IHCOCOCOCO
                                                    3331  I  I  I
                                                    rHrHr-H GGGG
                                                    CO CO CO M M M rH
                                                        CQ CQ 03 CQ PQ BO 03
                                                        OOOOOOO
                                                         OOOOOOO


                                                         (tf 03 flj (T) (Tj (Tj  i j=
                                                                                   J3   rH
                                                                                    •M U -T3
                                                                                    CU rtrH C
                                                                                     1 1— i   rt
                                                                                    >>3 CU
                                                                                    X U C IH
                                                                                    O CU CU CU
                                                                                                                                 a
                                                                                                                                 o
                                                                                                                                 IH
                                                                                                                                 o,
                                                                                                                                 a.
                                                                                                                 Ji

                                                                                                                  O


                                                                                                                  i
                                                                                                                  rt
                                                                                                                 4-1
                                                                                                                  rt
                                                                                     > O O,'"-'
                                                                                    CU 5 O CU
                                                                                     I   "H
                                                                                    >. 'JSr-t
                                                                                    X bo*-1 >>
                                                                                    O > O .C
                                                                                   j= rt U4->
                                                                                    JJ   -O CU
                                                                                    CU  -
                                                                               CU CU ErH
                                                                               •o-o   o
                                                                               •iHT-t  II U IH O
                                                                               x x   >»*^ o
                                                                               O OrHrH CU >»
                                                                               UH i-l O bO*-1'— I
                                                                               r-l-O'-'      bO
                                                                               3 >ftH (U II
                                                                               MJ=JD C   CU
                                                                                    in CU CU C
                                                                               rH S rtrH C <1»
                                                                                    O >»rtr-H
                                                                           •u COrH'i-' OJ3
                                                                           CU CO >»CU«H 4J
                                                                           E rtJ= >rH CU

                                                                           -o o cu o co u
                                                                             Q.E O.  4->
                                                                           II        II
                                                                              II II II    II U
                                                                           O        Z    CU
                                                                           wo: HO -3 33"
                                                                                         rt
                                                                        CM CM CM CM CM CM CM
                                                                                               H  M  H n  n  H
                  CO
                  CO
                   I
                  ON
                  00
                  \
                  ro

-------
    ATEG TREATMENT OF DIOXIN- AND FL'RAH-COKTAKINATED OIL AT AN INACTIVE
                   WOOD TREATING SITE IN BUTTE, MONTANA*

                            Paul E. des Rosiers''

Background

     On February 20, 1986, technical 6tsff of Che Environmental Protection
Agency (EPA), Office of P.eeearch and Development's (OHIO Hazardous  Waste
Engineering Research Laboratory in Cincinnati, Ohio, met with the author,
who represented the ORD Office of Environmental Engineering and Technology
(OEET), to discuss the provision of technical assistance to EPA Region VIII.
Waste at the Montana Pole site is generated as the oily phase of groundwater
pumped from 21-foot deep wells; after separation by decantation, approximately
3-percent pentachlorophenol (PCP) in a diesel-like oil is obtained  at the  rate
of 30-50 gpd.  This site represents an inactive wood treating facility located
on a 20-acre, sloping abandoned mining site, where contamination by dioxins
(CDDs) and furans (CDFs) has reached an adjacent creek, including groundwater
and surface soil.  The PCP-oil waste contained CDD/CDF homologs ranging from
147 ppb of tetra- to 83,923 ppb of octa-congeners.  Because of the  presence
of these potentially highly toxic CDDs and CDFs, the raw oil could  not be
transported off-site for incineration.  The estimated cost of on-site mobile
incineration was $3.1 million.

     ORD agreed to assist Region VIII and proposed that a chemical  process
based on alkali polyethylene glycol (APEG) be employed to decontaminate
the oil.  Oil samples were obtained from the site and transported to the
Brehm Laboratory at Wright State University in Dayton, Ohio, for analysis.
Subsequently, parametric studies were conducted to determine treatment
conditions to effect optimum decontamination.  Tables I and II show that
the waste oil was effectively decontaminated by the APEG reagent in the
laboratory at conditions as mild as 70°C after 15 minutes.

     On May 1, 1986, the Region VIII on-scene coordinator agreed to allow
treatment of the oil by the AGEG process, with ORD technical assistance.
Arrangements were made to lease a mobile treatment unit from the Niagara^
Mohawk Power Corporation and contracts were signed in June 1986 with both
the IT Corporation and Galson Research Corporation to perform the requisite
work.  Decontamination of the oil was completed on July 31, 1986.

The APEG Process

     Potassium hydroxide Is reacted with polyethylene glycol (molecular
weight approximately - 400) to form an alkoxide (see Equation 1).  The
alkoxide in turn reacts initially with one of the chlorine atoms on the
aryl ring to produce an ether and potassium chloride salt (see Equation 2).
   Chairman, Dioxin Disposal Advisory Group
   Environmental Protection Agency (RD-681), Washington, DC 20460

   Paper presented at the Panel on New and Emerging (Waste Treatment and
   Disposal) Technology, Annual Meeting of the American Wood Preservers
   Institute, Washington, DC, October 28, 1986.

-------
                ROH + KOH
ROK -«- HOH
     ROK-
                              C!
                                                 >-^ —Nw^  >
                                               oi  To
              Cl

              CR
KC!
     The cobile field equipment employed to icplemeni  the  previous  chenical
process comprises a 2,700-gallon batch reactor mounted on  a  45-foot trailer
equipped with a boiler/cooling system and a laboratory/control  room area.
Heating of the raw oily waste/APEG reagent mixture was achieved  by  the
recirculation of the oil and reagent through a pump, a high  shear mixer,
and a tube-and-shell heat exchanger.  The heat transfer fluid on the shell
side of the heat exchanger was heated using a boiler or cooled  through  a
series of fin-type air coolers.  A schematic is shown  below:
      27DO GALLON
       REACTOR
                       HEAT
                     EXCHANGER
          UAIN REGULATION
               PUMP
                     HBH SHEAR
                       MIXER
                HTf RECIRCULATION PUMP
Results

     The PCP-oil was processed in five-batches,  each batch consisting of
1,400 to 2,000 gallons of waste oil and 600 gallons of APEG reagent.  The
mixture was heated to 150°C and allowed to react for 90 minutes before
cooling.  The reactions conditions were not optimal, however, and were
excessive based on results of the Brehm Laboratory parametric studies
(see_Table 1), but were considered a "safety measure" in light of the
infancy of the process.  Tne treated oil was then pumped from the reactor
into a holding tank from which composite samples were removed and sent to
the IT Corporation's analytical laboratory in Knoxville, TN.  .Table III
summarizes both the batch and reagent sludge analytical findings and the
destruction efficiencies for CDD/CDF homologs.  The data indicate that
all CDDs and CDFs were destroyed to concentrations below detection limits,
which were, on the average, less than 1 ppb.

-------
     The ATrlG process hes tnerefore successfully destroyed CDDs and CDFs
in waste PCT-oil.  Tne processing cost in the Butte,  MT,  decontamination
demonstration was less than 10-percent of the projected,  estimated on-site
incineration costs—$212 thousand.  Tne CDD/CDF-free  diesel fuel that was
obtained will be sold to a local dea.er,  further reducing fie process cost.
The purified PCP prodjced, estinvatc-d to he about 1600 pounds, will be
stored on-site until its ultimate disposition can be  determined.

     Additional parametric investigations &re planned on  the Montana Pole
contaminated emulsions, groundwater, and  soils.   The  results of these
studies will be reported or. next year.  The equipment has been transported
to Kent, WA, where it processed approximately 8,000 gallons of solvent
wastes (a mixture of solvents, oil, and water generated from pesticide
manufacturing) containing up to 14 ppb 2,3,7.S-TCDD.

     Based on its recent sucesses on PCBs, PCP-oil, and solvent waste—all
contaminated with CDDs and CDFs, the Agency's Dioxin Disposal Advisory Group
(DDAG) considers this an innovative and alternative technology, which is
also cost effective.

     Wastes defined under the Dioxin-Listing Rule (FR 50,  1978-2006) as
F021/F027 are considered as acutely hazardous unless  treated to <  1 ppb
for each of the tetra-, penta-, and hexa-CDD and -CDF honologs according
to the Land Disposal Restriction Rules, which become effective on
November 8, 1986 (FR 5J_, 1602-1766) and "delisted" accordingly relative
to RCRA Appendix VIII chemicals.  Table IV presents recent data concerning
2,3,7,8-TCDD contaminant levels for certain chlorophenols, including PCP,
and related organic compounds.  Given the "sensitivity" of CDD/CDF
contamination of manufactured chemical products and their uses, this list
is included here for informational purposes.

-------
     Table  1.   Parametric  Studies  of  KPEG  Treatment  of PCP-Oil,  Butte,  Montana3
I   Sa=Dle:    Raw  Oil  i
                           KFEG Treated  Oil .  nf/r  \TT-
  CDD/CDF
                                           70°C
ppb    '15 gin 130 cln  45 min 160 cin 190 cin 1120 tin 115 cm  I 30 cir
2,3,7,8-
TCDD /'
-TCDDs
PCDDs
HxCDDs
HpCDDs
OCDD
2,3,7,8-
TCDF 9
TCDFs
PCDFs
HxCDFs
HpCDFs
OCDF
28.2
422
822
2982
20671
83923
23.1
147
504
3918
5404
6230
KD
'(0.762)
ND
(0.456)
ND
( J.C1)
KD
(1.83)
11.2
6.50
12.1
33.3
ND
(0.427)
4.91
5.84
ND
(3.61)
KD
(0.738)
ND
(0.373)
KD
(0.922)
ND
(1.83)
2.14
4.01
1.28
16.3
ND
(0.468)
2.98
2.63
ND
(3.63)
KD
(0.644)
KD
(0.308)
ND
(0.635)
KD
(1.38)
5.82
KD
(3.90)'
ND
(0.210)
1.59
ND
(0.421)
2.82
-.61
ND
(4.86)
KD
(0.322)
KD
(0.305)
KD
(0.396)
ND
(0.487)
3.53
4.24
ND
(0.191)
ND
(0.135)
ND
(0.174)
2.65
2.79
ND
(2.99)
KD
(0.348)
KD
(0.221)
ND
(0.548)
ND
(0.893)
4.02
5.11
ND
(0.234)
ND
(0.233)
ND
(0.272)
2.42
3.72
ND
(3.40)
KD
(0.493)
KD
(0.292)
KD
(0.376)
KD
(0.516)
5.50
4.51
KD
(0.391)
ND
(0.212)
ND
(0.153)
1.63
2.85
ND
(3.69)
KD
(C.85E)
KD
(1.25)
KD
(2.09)
KD
(3.36)
2.25
4. 14
5.45
109
ND
(0.5°91
2.73
ND
(0.913)
ND
(3.15)
KD
(0.54:
KD
(0.27^
KD
(0.4i:
KD
(2.43;
1.43
2.56
ND
(0.2E:
ND
(0.35:
ND
(0.26?
ND
(0.757
ND
(1.06'
ND
(1.63:
   One should not make the  erroneous presumption  that  these data indicate
   contamination of the ?CP by  2,3,7,6-TCDD or 2,3,7,6-TCDF; rather, the
   most likely source is PCBs,  which are present  and contain chlorobenzenes,
   Furthermore, the site has had its share of pit fires further aggravating
   the situation.

a  Analyses performed by the Brehm Laboratory, Wright  State University,
   Dayton, Ohio.

-------
        Table 2.  Destruction Efficiency (DE) (X)a.b
   Sarr.- le:  Raw Oil
K.FEG Treated Oil
                                                 )00°C
                           Elapsed TIE.C • 30 ninutes
CDL»s/CDFfi ppn ppb DE ppb DE
2.3,7,8-
T-PD /•
TCDDs
PCDDs
HxCDDs
HpCDDs
OCDD
2,3,7,8-
TCDF /'
TCDFs
PCDFs
HxCDFs
KpCDFs
OCDF.
28.2
422
822
2982
20671
83923
23. 1
147
504
3918
5404
6230
ND
(C.73E)
ND
(0.373)
ND
(0.922)
ND
(1.83)
2.14
4.01
1.28
16.3
ND
(0.468)
2.98
2.63
ND
(3.63)
>97.7828
>99.91&0
>99.9319
>99.9524
99.9846
99.9942
93.9583'
87.3189
>99.9355
99.9284
99.9144
>99.9370
ND
(0.544)
ND
(0.274)
ND
(0.411)
ND
(2.43)
3.43
2.56
ND
(0.282)
ND
(0.351)
ND
(0.288)
ND
(0.757)
ND
(1.06)
ND
(1.63)
>96.2414
>99.8881
>99.9814
>99.9031
99.9832
99.9936
>98.4042
>99.5015
>99.8888
>99.9671
>99.9624
>99.9285
a  These values are not corrected for small concentrations of
   HpCDDs, OCDD, and HpCDFs detected in lab blanks, so actual
   DEs for these isomers are actually higher than indicated.

b  The notation > indicates that the isomer(s) were not detected
   and the lower limit cited for the DE is based on the analytical
   detection limit.

#  Source of the 2,3,7,8-TCDD/TCDF is PCBs, which contain
   chlorobenzenes,  and have experienced thermal stress, namely,
   pit fires.

-------
                                    -tr-
                3.   Results  of  APEG Chemical
                         Montana  Pole  Siteb,
Detoxification
Butte,  Montana
of PCP-0:
CDD/CDF
Homolog
TCDL»s

PCDDs
- HxCDDs
HpCDDs
OCDD
TCDFs

PCDFs
HxCDFs
HpCDFs
OCDF
Raw PCP-Oil
ng/g

422

822
2,982
20,671
83,923
147

504
3,918
5,404
6,230
KPEG-Treated Oil @ 150°C, 90 ain
ng/g
Batch No . :
12345
(0.64) (0.68) (0.82) (0.75) (0.80)

(0.97) (0.49) (0.78) (0.52) (0.74)
(0.92) (0.63) (0.43) (0.70) (0.90)
(0.41) (0.52) (0.32) (0.39) (0.47)
(1.3) (0.87) (1.1) (0.54) (1.1)
(0.77) (0.59) (0.83) (0.62) (1.5)

(1.3) (0.58) (0.66) (0.67) (1.1)
(0.84) (0.53) (2.2) (0.67) (2.1)
(0.49) (1.0) (1.6) • (0.56) (0.83)
(0.53) (1.3) (1.1) (0.61) (0.83)

Reagent
Sludge
(0.12)

(1.0)
(0.44
(0.18)
(1.0)
(0.14)

(0.51)
(0.49)
(0.28)
(0.22)


(0.73)

(0.70)
(0.76)
(0.4S)
(0.98)
(0.86)

(0.81)
(1.27)
(1.09)
(0.87)
DEs
ja

>99.98
I
>99.91
>99.97
>99.99
>99. 99
>99.41

>99.84
>99.97
>99.98
>99.99
(   )  - detection limit.  Analyses  performed  by  IT Corporation, Knoxville,  TN;
        QA/QC performed by  EPA EMSL-Las  Vegas, NV.

    a  - detection limit (DL)  averaged  for five batches;
        destruction  efficiency (DE) based on  average  DL  of homolog.
        (The notation  > indicates  that the homologs were not detected  and
        the lover limit cited for  the  DE is based  on  the average  of  the
        analytical detection  limit.)

    b  = oil processed  July  25-31,  1986;  in addition,  wipe samples of equipment
        after decontamination procedures completed showed: _np_  TCDDs/TCDFs,  PCDFs,
        HxCDDs/HxCDFs, or OCDF detected  at or above DL - 3.8 ng/m2;  only PCDD =
        36 ng/m2 and OCDD » 217 ng/m2  were found above the DL.  These  positive
        values are well below any  concentration  considered to  be  of  concern
        according to EPA ORD  ECAO-Cincinnati  and EMSL-Las Vegas laboratories.

-------
     Table A.  Levels of 2,3,7,8-TCDD in Commercial Chlorophenols
                        and Related Compounds**
Product
trichlorobenzene
Witophen Na
Preventol PN
Dowicide G
Witophen P (sample 1)
Witophen P (sample 2)
PCP sample
2,4-D
hexachlorophene
2,3,4,5-TCP
chloranil
2,3,7,8-TCDD
ng/g (ppb)
95
0.42
0.56
0.21
n.d .
n.d.*
n.d.*
6.8
0.3
0.3
n.dJ
Source
Dynamic Nobel
Dynamit Nobel
Bayer AC
Fluka
Dynamit Nobel
Dynamit Nobel
Rttkne-Poulenc
EGA-Chemie
Riedel-de Haen
Aldrich
Hoechst AG
*  Limit of detection for all three'analyses was approximately 0.03 ng/g.

*  Detection limit - 0.05 ng/g.

a  Na-pentachlorophenate.

b  H. Hagenmaier (1986).  Determination of 2,3,7,£-tetrachlorodibenzo-
   p-dioxin in commercial chlorophenols and related compounds.  For
   publication in:  Fresenius Z. Anal. Chemie.

   H. Hagenmaier and H. Brunner (1986).  Isomer specific analysis of
   PCDD and PCDF in PCP and PCP-Na samples in the sub-ppb range.
   Paper presented at DIOXIN 86, the 6th International Symposium on
   Chlorinated Dioxins and Related Compounds, Fukuoka, Japan,
   September 16-19.

-------
Bibliography
 1.  Memorandum of July 18, 1985, from John G.  Welles,  RA,  Region VIII,
     to Jack McGrav, Acting AA, OSWER, subject:   Immediate  Removal
     Request for the NPL Silver Bow, KT—Montana Pole Site,  Butte,
     Montana.

 2.  Memorandum of August 9, 1985, from Paul E.  dee Rosiers,  Chairman,
     DDAG, to Steve Heare and Colleen Carruthers, ERD,  subject:   DDAG
     Review of Immediate Removal Request for the NPL Silver bow,
     Montana,  Pentachlorophenol-Contaminated Site.

 3.  Letter of January 22,  1986, from David M.  DeKarini, HERL-RTP, NC,
     to Charles J. Rogers, HWERL-Ci, OK, subject:  Bioassays of  KPEG and
     KPEG-Treated Dioxin Samples.

 4.  OEET Briefing Document of February 24, 1986, subject:   Summary of
     February 19, 1986, OSW Wood Preserving Wastes Workgroup Meeting on
     Montana Pole Treating NPL Site.

 5.  List of Wood Preserving Plants, 1985, compiled by The American
     Wood Preservers Association.

 6.  H. Hagenmaier, H. Brunner, R. Haag, and K. Kraft (1986).  Selective
     Determination of 2,3,7,8-Tetrachlorodibenzo-p-dioxin in the Presence
     of a Large Excess of  Other Polychlorinated Dibenrodioxins and Poly-
     chlorinated Dibenzofurans.  Fresenius Z. Anal. Chen.,  323,  24-28.

 7.  R. Peterson, E. Milicic, and C. Novosad (1986).  Comparison of
     Laboratory and Field  Test Data in the Chemical Decontamination of
     Dioxin Contaminated Soils Using the Galson PKS Process.  Paper
     Presented at the 191st American Chemical Society Meeting, New York,
     NY, April 13-18.

 8.  Memorandum of March 7, 1986, from Alfred Kernel, HWERL-Ci,  OH, to
     Paul E. des Rosiers,  ORD-OEET, subject:  Use  of KPEG to Decontam-
     inate and Remove PCP  and Chlorodioxins from Oil.

 9.  OEET Briefing Document of May  5,  1986, subject:  Summary of
     Meetings with Region  VIII and  ORD Regarding Montana Pole Site and
     APEG Destruction Efficiency Data/Time-Temperature  Relationships for
     PCP-Oil.

10.  Brehm Laboratory, Wright State University Proposal for Analytical
     Support for Destruction of Toxic Chlorinated  Hydrocarbons, Includ-
     ing PCDDs, PCDFs, PCBs, PCBZs, etc. in Waste  Oils, in Carbon Used
     for Wastewater Treatment, and  in Soils, May  12, 1986.

-------
11.   IT Corporation/Galson Research Corporation  Proposal  for  Chemical
     Decontamination of  Dioxin Contaminated  Oil  in  Butte,  Montana,
     May 1986.

12.   Memorandum of  May 16, 1986, from Charles  J.  Rogers,  HVERL-Ci,
     OK, to Martin  Byrne,  OSC, Region VIII,  subject:   Treatment  of
     CDDs/CDFs  Contaminated Oil at Butte,  MT.

13.   Memorandum of  May 16, 1986, from Paul E.  des Rosiers,  Chairman,
     DDAG,  to Katherine  E. Daly, OSC, Region I,  subject:   Use of
     KPEG to Detoxify CDD/CDF-Contaminated Soils  at  the Tibbetts
     Road Site, Barrington, New Hampshire.

14.   M. J.  Byrne and D.  K. Wittenhagen (1986).  PCP  Contamination:
     A Removal  Action Case History—Montana  Pole Site, Butte, MT.
     TAT Conference, Edison, NJ, May, pp.  333-356.

15.   Draft  Memorandum of May 30, 1986, from Tim  Fields, Director,
     ERD, to WMDDs, Regions I-X and ESDDs, Regions  I,  VI,  and VII,
     subject:  Proposed  Alternative Treatment/Disposal Policy for
     Removal and Expedited Response Actions.

16.   Federal Register, 51, 1602-1766, January 14, 1986.

17.   Personal Communication with Martin Byrne, OSC,  Region VIII,
     June 20, 1986.

18.   K, R.  Rao, ed. (1978).  Pentachlorophenol:  Chemistry,
     Pharmacology,  and Environmental Toxicology. Plenum  Press,
     New York,  NY,  A02 pp.

19.   T. 0.  Tiernan, M. L.  Taylor,'G. F. Van Ness, D. Wagel, J. H.
     Garrett, J. G. Solch, C. Rogers and P. E. des  Rosiers (1986).
     Destruction of High Concentrations of PCDD/PCDF in PCP and
     Other  Chemical Wastes by Treatment with KPEG Reagent.  Paper
     presented  at DIOXIN 86, the 6th International  Symposium on -
     Chlorinated Dioxins and Related Compounds,  Fukuoka,  Japan,
     September  16-19.

-------
              DECONTAMINATION OF A SMALL PCB SOIL SITE
                    BY THE GALSON APEG PROCESS

                    C.F. Novosad, E. Milicic. R.L Peterson

                           Galson Research Corp.
                             6601 Kirkville Rd.
                          E. Syracuse, N.Y. 13057


           Presented before the Division of Environmental Chemistry
                        American Chemical Society
                New Orleans, August 30 - September 4,1987


The EPA Project Officer for this project was Charles Rogers,  Office of Research
and Development, U.S.  EPA, 26 W. St. Claire St., Cincinnati, Ohio 45220.
ABSTRACT

      In 1985, a total of fifty-two drums of PCB contaminated soil were
chemically decontaminated using an early version of the patented APEG
process developed by Galson Research Corporation.0) The fifty-five gallon
drums containing the contaminated soil were also used as reaction vessels.
Processing involved adding liquid reagents to saturate the soil.  The drums
were then heated to at least 100°C for 2-3 days without mixing.  The PCB
concentration in all of the drums was reduced to below 50 ppm, the required
"clean" level for this site.  The treated soil was neutralized by mixing with dilute
acid. The excess liquid was then decanted from the soil, with the treated soil
approved for return to the original excavation.

      Small sites could use this process as a cost effective alternative to
incineration or landfill. A more sophisticated multi-ton slurry reactor system is
currently being designed for larger sites.
INTRODUCTION

     Galson Research has patented (1) a chemical process which detoxifies
hazardous aromatic halides in contaminated soil.  The APEG (alkaline-
polyethylene glycolate) process has been proven successful for a variety of
soils in both laboratory and field studies. The procedure used at the Bengart
& Memel site was a small scale version of the process, which utilized the soil
storage drums as reaction vessels.

-------
      The APEG process entails the addition of liquid reagents to contaminated
soil and heating of the mixture until the PCBs in the soil decompose to lower
toxicity, water soluble materials. The reactions involved are shown in Figure 1.
ROM
KOH
                          DMSO
            Polyethylene Potassium  Dimethyl
            glylo.      Hydroxide  su,foxide
  «~,,    DMSO
+ ROK  	*»
Chlorinated
                      gycolate
                     ROK
                                                   HOH
                                                                      (1)
                                                            +  KCL   (2)
                     Polyethyleneglycol
                     biphenyl ethe?
                                                             Potassium
                                                             chloride
                         Figure 1 .  Reactions
      The reagent components include: a sulfoxide, e.g. sulfolane or dimethyl
sulfoxide (DMSO); a glycol or capped glycol, e.g. polyethylene glycol 400
(PEG), triethylene glycol methyl ether and highers (TMH), and/or methyl carbitol
(MEE); solid or aqueous potassium hydroxide (KOH); and water. The glycol is
reacted with KOH in the presence of DMSO to form an alkoxide.  The alkoxide
reacts with one of the chlorine atoms on the biphenyl ring to produce an ether
and potassium chloride salt. The sulfoxide acts as a cosolvent/catalyst  and
increases the overall rate of reaction.
SITE BACKGROUND

      Bengart & Memel, Inc. a wholesaler of non-ferrous scrap metals, was
originally founded in 1950 in Buffalo, New York. From the early 1950's through
1978 Bengart & Memel received and dismantled PCB (polychlorinated
biphenyl) transformers and capacitors, inadvertently releasing PCB into the soil.
In the mid-1970's, soil samples from the property were found to contain greater
than 50 ppm of PCBs. The New York State Department of Environmental
Conservation issued a Consent Order for remediation which required that the
PCB concentration in the soil be reduced to below 50 parts per million. (2, 3)

    Soil sampling and analysis indicated that seven sites on the property had
 excessive PCB concentrations.  These sites ranged in depth from 6 to 24
 inches and were defined to be 10 feet in diameter about the point where the
 soil core sample was taken.  As part of the remedial program, this soil was
 excavated and placed in a total of 166 fifty-five gallon steel drums.

-------
 PROJECT ORGANIZATION

    The processing of PCB contaminated soil at Bengart-Memel proceeded in
 3 stages;  analysis of each drum for PCB concentration, laboratory simulation
 of PCB destruction and neutralization, and actual treatment operations at the
 site.
 ANALYSIS OF EACH DRUM FOR PCB CONCENTRATION

    Galson Research Corporation sampled each of 188 drums and analyzed
 the soil for PCB concentrations using the McGraw-Edison PCB Field Test Kit.
 The kit was used to avoid using the more expensive gas chromatograph (GC)
 analyses for every drum. Twenty two samples which were near 50 ppm by the
 kit were analyzed by GC to establish a relationship between the analytical
 methods.  A conservative correlation derived from these results indicated that
 only those samples between 25 and 60 ppm as determined by kit analysis
 needed to be re-analyzed by GC.  A total of 39 samples fell into this range.
 These samples were analyzed by an EPA contract laboratory, Carolina
 Chemists  & Consultants (CCC).  Of the 61 samples analyzed by both kit and
 GC (GRC and CCC combined), only 14 drums were incorrectly classified as
 needing treatment by kit analysis  (false positives).  This amounts to 77%
 precision for kit results in the 23 to 60 ppm range.

    Both GRC and CCC ran three sets of  duplicates; CCC had an average
 deviation from the mean of 13.5%, GRC had an average deviation of 12.9%.
 Using a +/-14% variation applied to the cutoff point of 50 ppm, all drums 43
 ppm and higher were deemed to  require remediation; this conservative
 approach  insured an accurate classification of drums. This cutoff point is even
 more conservative considering that the samples analyzed consisted of soil
 devoid of the heavy, non-porous debris (e. g., gravel, rocks and metal) which
 made up a significant portion (ca. 20%) of the drum contents. The table below
 contains a summary of the drum classification.

       Table 1 - Analytical Methods by which Drums were Segregated

                           KIT     GRCGC    CCCGC    TOTAL
    Requires Treatment      67        17         10         94
    No Treatment Needed    38         5          29         72
      Ninety-four of these drums were determined to be above the 50 parts per
million PCB concentration which classified them as hazardous waste requiring
remediation.

-------
LABORATORY SIMULATION OF PROCESS

      Soil from the Bengart & Memel site was passed through a screen with
0.25" openings to remove rocks and thoroughly mixed before being used for
laboratory treatment simulations.  Five laboratory experiments were conducted.

      1.  Two Phase Reagent Tests
      2.  Reagent Formulation Experiment
      3.  Single Phase Reagent Tests
      4.  Effect of Sample Size on Reaction Rate
      5.  Effect of Temperature on Reaction Rate

      Each set of tests will be discussed separately.

Two Phase Reagent Tests

      Earlier laboratory work done by GRC (4) with soil from this site showed
that treatment at 70°C with 2:2:2:1 DMSO:MEE:KOH:water at 20% loading (10 g
reagent for 50 g soil) reduced average PCB concentrations from 77 ppm to 16
ppm (standard deviation = s = 5.0, number of samples = n = 3).  Treatment with
2:2:2:1 SFLN:MEE:KOH:water at the same temperature and reagent loading
reduced the PCB concentration from 77 ppm to 14 ppm (s=0.6, n=3). When 120
Ib batches of soil were treated at 70°C using the same reagent formulation, the
results were more variable, with little or no PCB reduction in  some cases.

     Because processing at 70°C was ineffective in the pilot test, additional
laboratory reactions were conducted at higher temperature (150°C), using a
more potent reagent formulation, (1:12:2:1 PEG:TMH:DMSO:KOH water), at
100% loading (50 g reagent for 50 g soil). This approach was quite successful.
After 3 days of reaction, the average PCB concentration in the soil was <0.1
ppm.  The reaction was repeated with a reaction time of 24 hours.  Again, the
PCB concentration was <0.1 ppm.  This general approach was used for the full
scale cleanup.

Reagent Formulation Experiment

      It was believed that much of the failure to scale up the unmixed soil
process was due to the fact that the solid KOH  used for the reagent was not
evenly distributed through the soil.  Previous work done  by GRC (4) had shown
that reagent formulated with saturated aqueous KOH (2:2:2:1
PEG:DMSO:KOH:water) instead of solid KOH  (1:1:1 PEG:DMSO:KOH) forms a
2 phase liquid mixture and that reagent containing 50% water (2:2:2:6
PEG:DMSO:KOH:water) is a single liquid phase. Use of a single phase reagent
promotes even distribution of reagent throughout the soil. However, since
water slows down the reaction, it would have to be removed  by heating and
evaporation. The less water added to the reagent, the less time and energy
needed to remove the water. In addition, 1:1 PEG.TMH mixture produces a
more effective reagent than PEG alone. Thus, a reagent of the formulation
1:1:2:2:X  PEG:TMH:DMSO:KOH:water was proposed for this project. The value
of X was to be the minimum amount of water that would produce a single phase

-------
reagent.  Experimentation showed that a minimum 44.6% water was necessary
to produce a single phase reagent.   A close approximation, with a slight safety
factor was a formulation of 1:1:2:2:5 PEG:TMH:DMSO:KOH (dry):water, which is
45.5% water by weight. Since the KOH used in the field was available as a
45% aqueous solution, the formulation used at the site became 2:2:4:9-5
PEG:TMH:DMSO:45% KOH:water.

Single Phase Reagent Tests

      Samples of screened soil weighing 10 g were treated with 4 g of the
single phase reagent and heated to 150°C for 24 and 48 hour intervals.  The
samples were cooled and analyzed along with samples of untreated soil. On
the average, the process reduced the PCB concentration in the soil from  116
ppm(s=9.2, n=3) starting to 6.3 ppm (s=1.4, n=3) in 24 hours and to 3.3 ppm
(s=1.2, n=3) in 48 hours.

Effects of Sample Size on Reaction  Rate

      A treatment simulation was done to investigate the effect of reactor size
on reaction effectiveness and to produce a single batch of soil for checking
spike recovery from treated samples. For this simulation, a 237.2 g sample of
screened soil was treated with 83.3  g of single phase reagent. Three portions
of the soil were extracted and analyzed, one portion was spiked with an
additional  2.2 ppm of Aroclor 1260 (added as a hexane solution). The larger
scale reaction reduced the PCB concentration to 8.8 (s*1.1) ppm in 48 hours.
Spike recovery from the treated soil was 153%. There is not enough data to
state that the difference between the larger and smaller reactions is statistically
significant.

Effect of Temperature on Reaction Rate

      A laboratory treatment simulation was done to simulate the lower
temperature and slow heating predicted for field operations.  The lowest  third of
the soil volume was immersed in an oil bath and the bath was heated slowly, at
20°C/hour, to 115°C.  The lower temperature and slow heating reduced the
reaction efficiency but still reduced the PCB concentration to <50 ppm within 48
hours.

Summary of Laboratory Results

      Table 2 is a summary of all of the laboratory testing done for this project.

-------
                      Table 2.  Laboratory Results
Description/Reagent

Two phase reagent tests
1:1:2:2:1PEG:TMH:DMSO:KOH:water
untreated
1:1:2:2:1 PEG:TMH:DMSO:KOH:water
1:1:2 PEG:TMH:DMSO (control)
untreated

Single phase reagent tests
1:1:2:2:5 PEG:TMH:DMSO:KOH:water
1:1:2:2:5 PEG:TMH:DMSO:KOH:water
1:1:2:2:5 PEG:TMH:DMSO:KOH:water
1:1:2:2:5 PEG:TMH:DMSO:KOH:water
1:1:2:2:5 PEG:TMH:DMSO:KOH:water
untreated
    %L = % Loading = g reagent/100 g soil
    n = number of samples
    s = standard deviation

Neutralization Testing
  %L  T  days n  ppm
100
na
100
100
na
150
na
150
150
na
3
na
1
1
na
2
2
2
1
1
<0.1
112
<0.1
27
83
0.07
6.4
<0.05
na
na
(small)
(small)
(large)
(small)
(small)

40
40
40
45
45
na
150
150
150
115
115
na
1
2
2
1
2
na
3
3
3
1
1
3
6.3
3.3
8.8
97
30
116
1.4
1.2
1.1
na
na
9.2
T = Bath Temperature in °C
ppm m Average mg PCB/kg soil
na = not applicable
      In order to discharge the treated soil, it was necessary to reduce the soil
pH from over 14 to between 5 and 9. Neutralization was simulated under
laboratory conditions.  Two 10 g samples from the 150°C treatment simulation
were neutralized by addition of 4 mL of 3 M HCI. Another pair of samples were
treated with 4 mL deionized water for comparison. The samples were allowed
to stand until foaming subsided in the acid treated samples. The samples were
then mixed with a sonicator for 5 minutes and shaken vigorously.  After
standing (capped) for about 60 hours the pH of the free liquid in the vials was
checked using Hydrion pH paper.

      Neutralization of the soil by addition of hydrochloric acid reduced  the pH
of the treated soil from greater than 12 to 6, but the procedure was somewhat
hazardous.  The reaction between treated soil and acid produced a great deal
of bubbling and foam.  For that reason, the field neutralization was done with
more dilute acid.

-------
 OPERATIONS/PROCESSING PROCEDURES

     The bulk of the decontamination processing took place in September and
 October of 1986. Fifty-one drums of soil were staged near the work area for
 decontamination.  Due to the consent order deadline, the remaining drums
 were not processed and were shipped to a landfill.  The only exception was
 drum #20 which could not be disposed in a landfill due to the >500 ppm
 concentration.  A confirmatory GC analysis of the soil gave a PCB level of 1300
 ppm.  Drum 20 was eventually processed in February 1987, during
 neutralization of the other drums.

     Soil decontamination at Bengart & Memel involved 6 basic steps;

       1. Set up of the vapor system & enclosure.
       2. Add premixed liquid reagent to soil drums.
       3. Attach heaters, insulation, and vent lines to the drums.
       4. Allow drum contents to heat and react.
       5. Cool and obtain sample for analysis.
       6. Neutralize the caustic soil.

     The soil filled the drums to within two to six inches from the top to allow for
 the addition of the liquid reagents.  An aluminum tube was driven through the
 dry soil to the bottom of the drum to allow air to escape during liquid addition,
 otherwise an air pocket would form preventing the downward movement of the
 liquid.  The single phase reagent mixture (2:2:4:9:5 PEG:TMH:DMSO:45%
 KOH:water) was added to the soil drums by weight, typically between 150 - 170
 pounds.  The KOH  reacted with aluminum, thus when the reagent reached the
 bottom of the drum, foam exiting the tube would indicate that the air pocket had
 been expelled. The tube was then removed and a bimetal thermometer was
 inserted into the 3/4" bung on the drum cover.

     The drums were heated from the bottom to insure that all of the soil was
 reacted.  Dents and deformities in the drums necessitated the use of heat
 transfer compound  in order to provide adequate contact between the heaters
 and the metal.  It was essential that liquid be contacting the metal on the interior
 so that the heat could be carried away. The soil temperature was  monitored 9
 to 12 inches below the lid with the heaters were at the drum's base.  Soil
 temperatures near the heater were higher than those measured near the top.
 Typically a drum would be heated to a maximum temperature readout between
 105 and 110°C for several hours.

      The drums were vented to a central condenser system with  a capacity of
 16 drums (See Figure 2). During heating, vapors would exit from the drum
through an insulated, flexible line connecting the 2" drum bung to the main
 header. Part of the header was ice jacketed so that much of the water vapor
was condensed and drained directly into a 55 gallon holding drum. The
 remaining vapor was drawn through an air cooled radiator (used as a
condenser) which also drained liquid to the 55 gallon holding drum.  An ice
jacketed scrubber containing a dilute sodium hypochlorite solution (for odor
control) trapped most of the remaining condensables. From the scrubber, the

-------
vapor passed through an air trap and a 55 gallon drum which was filled with a
mixture of activated carbon and molecular sieve.. A vacuum pump provided
negative pressure for the vapor control system.  •
                                                n
                                              condenser
                  PVC pipe ductwork
  drum
 heaters
to vacuum
pump
                                                                     i
                                                                   sorbent
                reaction drums
                       Figure 2.  -  Central Condenser System
      Treated soil samples were collected using a 30 inch auger attached to a
hand drill. Four wells were made in each drum, one in the center, one close to
the edge and two spaced between the center and edge in different parts of the
drum.  Soil was collected by digging as deep as possible with the auger and
bringing soil  up from the bottom of the hole. In digging out a sample well, it was
necessary to work around the large rocks in the drum - true core samples were
not obtainable.  Debris (rocks, metal, wood, etc.) larger than 1 inch in  diameter
was not included in the sample.  The soil from the cores was collected in a
plastic bucket and mixed thoroughly with a trowel.  Two jars were filled for each
sample, one for GRC and one for the EPA.  The soil remaining in the bucket was
returned to the drum and the auger, bucket, and trowel were cleaned  before the
next drum was sampled to avoid cross contamination.

    Once the treated soil had been decontaminated, the soil had to be brought
from a caustic state to a neutral pH. Mixing was necessary to avoid 'hot spots'
of concentrated acid or caustic. To provide controlled mixing, it was decided to
add the alkaline soil slowly to a drum partially full of dilute acid.  When the liquid
                                 8

-------
became neutral or slightly basic, more acid was added to compensate. Thus
when the acid and KOH reacted, there was sufficient mass to absorb the heat
and prevent splattering. This procedure required-frequent monitoring for pH
control.
PROCESS RESULTS

PCB Analysis

      The analytical results for PCB from this project can be divided into two
groups.  Group 1 consists of samples that were analyzed by the EPA soon after
processing and found to contain less than 50 ppm of PCB.  The drums from
which the Group 1 samples were taken were neutralized without any further
sampling or treatment. The concentrations in the untreated soil and EPA results
for the treated soil are summarized in Table 3.

          Table 3. Sample Group 1, Analyzed only by EPA
                      ppm PCB in
           Drum     untreated soil
             5            119
             6            119
             8            62
             10           128
             35           910
             36           91
             46           123
             53           64
             60           61
             62           65
             71           69
             81           74
             82           67
             87           62
             91           91
             92           138
             93           91
             102          69
             109          87
             119          64
             137          62
  ppm PCB in
treated soil 10/86
      43
      46
      45
      46
      33
      46
      47
      15
      44
      30
      28
      38
      43
      39
      37
      49
      44
      18
      42
       9
      23
             average     125
      36

-------
      Group 2 was analyzed by GRC in February of 1987. This group includes
fresh samples from the drums that were over 50 ppm when the EPA analyzed
them in October 1986,  samples from drums that .were never analyzed by the
EPA and one sample that was already less than 50 ppm when the EPA
analyzed it.  All of these drums were sampled in February 1987, after the
reagent had been in  contact with the soil for 5 months.  The low concentration
sample analyzed by  GRC to check the potential reaction progress in drums with
PCB concentrations  below 50 ppm.  Results of GRC's analyses  are presented in
Table 4.

               Table 4. Sample Group 2, Analyzed by GRC
ppm PCB in
Drum untreated soil
1
4
9
11
20A
20 B
37
43
44
45
54A
54B
63
64
kiA&b- '
83
85
86
101
111
112
113
114
120
121
122
131
135
Misc A
MiscB
MiscC
Misc D
MiscE
average
138
84
64
128
1300
1300
64
102
72
195
78
78
167
64
119
78
78
84
74
64
110
67
102
75
62
106
78
106
NA
NA
NA
NA
NA
179.
                           ppm PCB in treated
                             soil by EPA 10/86
                                   71
                                   60

                                   90
                                   74

                                   51
                                   59
                                   56
                                   74

                                   51

                                   78
                                   57
                                   26

                                   65
                                   62
ppm PCB in treated
 soil by GRC 2/87
       19
       20
        8
       30
       <5
       94
       12
        7
       31
       27
       10-
       49-''
       16
       13
       12'
       18
       16
       22
       11
       22
       33
       17
       27
       <5
       24
       22
       17
       28
       33
       17
        9
        6
       15
       22
*This drum not approved for non-PCB disposal.     NA = Not Analyzed
                              10

-------
      The reduced concentrations reported in February (compared with
October results) indicate that the reaction continued during the holding period.

    The only soil not to be reduced to below 50 ppm was from a drum of
particularly high concentration (drum #20 @ 1300 ppm) that the site owner was
not able to landfill.  Drum 20 and drum 54 were processed in February 1987
during the neutralization procedures. The soil from these drums were split and
transferred into two drums (20A & 20B, 54A & 54B) to insure good reagent
permeation and better heat transfer.  The PCB level was reduced by 93% in
drum 20B, to between 73 ppm and 104 ppm. Drums 20A, 54A & 54B were all
reduced to less than 50 ppm.

Soil Neutralization

      The drums of soil were neutralized between February 11 and 26, 1987.
They were allowed to rest until March 27 and the pH was re-checked.
Additional acid was added to those drums having a pH above 9. The pH of
those drums was checked the following day.  Results are presented in Table 5.
                               11

-------
      Table 5.  Results of Soil Neutralization at Bengart & Memel Site

                         pH of Saturated SoH

DRUM     27-Mar-87  28-Mar-87        DRUM   27-Mar-87  28-Mar-87
   18                         101       7
   3         7                         109      13          6
   4         8                         111      12          8
   6        11         6               113       7
   87                         114       7
   9         7                         119      10          6
  10        11         6               120       9          6
  11        10         5               121       6
  36         7                         122       7
  37         9         6               131      10          6
  43         6                         135      13          6
  46         8                         137       8
  53         7                          20A     7
  60         6                          20B     6
  63         8                        OF-1      11          6
  64         7                        OF-2       8
  71         8                        OF-3       7
  72         8                           I      10          6
  81        10         5                 II       8
  82        11         7                 III      10          6
  83         8                           IV      11          7
  85         9                          VII      10          8
  91        12         5           MISCA       7
  92         9         9           MISCB       7
  93         8                     MISC D       8
                                   MISC E       7

Drums designated OF were overflow from other drums during neutralization.
Drums that lost their Arabic numerals during handling were re-labeled with
     Roman numerals.
                             12

-------
CONCLUSIONS


      The early form of APEG processing used at the Bengart-Memel site was
successful in reducing PCB levels for 51 of 52 drums to below the 50 ppm
control limit set for the site. For the 51 successful drums, the average PCB
levels were reduced 75%, from  108 ppm to 27 ppm.  PCB levels for the sole
remaining drum were reduced by 93%, from 1300 ppm to 78 ppm.

      A more advanced form of the GRC APEG process is under development.
This new process features mechanical mixing of reagent and soil, along with
reagent recovery. However, the early form of the process described here
should prove cost effective for small sites, especially where the soil can be
allowed to  react for long periods of time.

ACKNOWLEDGEMENTS

      Portions of this work were supported by EPA contract # 68-13-3219.

      The assistance of Alfred Kernel of EPA Cincinnati in providing PCB
analysis is gratefully acknowledged.

REFERENCES

1.  U.S Patents, 4,574,013 and 4,447,541.
2.  Leese, K., Research Triangle Institute, personal communication.
3.  State of New York, Department of Environmental Conservation, Order on
   Consent, Index #917T010682.
4.  Galson  Research Corporation,  Progress Report,  EPA Contract #68-13-3219,
   September 1985.
                               13

-------
COMPARISON OF LABORATORY AND FIELD TEST DATA IN THE
CHEMICAL DECONTAMINATION OF DIOXIN CONTAMINATED SOILS

R. Peterson, E. Milicic, C. Novosad, Galson Research Corporation,
East Syracuse, NY,
C. Rogers, United States Environmental Protection Agency, Cincinnati, OH

      Gaison Research Corporation  has developed a series of
      patented (1) processes for chemical decontamination of soils
      contaminated with halogenated aromatics, including
      polychlorinated dibenzo-p-dioxins (PCDDs), chlorinated
      benzenes, polychlorinated biphenyls and similar materials.
      These processes allow reduction of PCDD levels to < 1 part
      per billion (ppb) in as little as two hours at moderate
      temperatures  and pressures.

introduction

Chemical decontamination  is an alternative to thermal processing or
landfilling of  soils contaminated with polychlorinated dibenzo-p- dioxins
(PCDDs) or other aromatic halides such as chlorobenzenes or
polychorinated biphenyls (PCBs). Chemical decontamination, like
incineration,  involves changes to the chemical structure of the dioxin
molecule.  While chlorinated dioxins are thermally stable, they readily
dechlorinate  to water soluble  compounds under relatively mild conditions of
temperature  and pressure.  For example, chlorinated dioxins in oil are
readily reduced to  the ppt level within 15 minutes at 80 degrees C. by
reacting them to a compound which is no longer oil soluble.  In soils
processing, the dioxin is dechlorinated to a water soluble form which is then
leached from the soil using countercurrent extraction  with water.
Dechlorination also affects the toxicity of the dioxin, with dioxins containing
fewer than three chlorine atoms generally showing low toxicity (2).

Process Chemistry

The proposed mechanism for these reactions is shown below  using 2,3,7,8
tetrachlorodibenzo-p-dioxin as an example;

          ROH +  KOH    DM3O     ROK + HOH                (1)


                                                        -i- KCI    (2)
                                                      -ROH      (3)

-------
An alkali metal hydroxide, usually potassium hydroxide (KOH) is reacted
with an alcohol or glycol such as polyethylene glycol 400 ( PEG 400) to form
an alkoxide.  The alkoxide reacts with one of the chlorine atoms on the
chlorinated dioxin to produce an ether and the alkali metal salt. This
dechlormation may proceed to complete dechlorination, although
replacement of a single chlorine is sufficient to make the reaction products
water soluble. The ether formed by the  dechlorination may  degrade to a
phenol form or may remain as the ether, depending on the reaction
conditions. The processing is carried out using dimethyl sulfoxide (DMSO)
as a solvent.  The DMSO catalyzes the reaciion by increasing the base
strength of the alkoxide. In addition, the DMSO aids in the extraction of the
PCDD from the soil.

 Toxicitv Considerations

Chemical decontamination of soil is a two stage process;

      1. Dechlorinate PCDD to lower toxicity/ water soluble form
      2. Wash excess reagents and  PCDD products from soil

A major concern in this type of processing involves the toxicity of any
reagents and/or reaction products which may inadvertently be left in the
decontaminated soil after treatment.  Some toxicity data on reagents used in
the process  are shown in Table I, along with comparison values for sodium
chloride and 2,3,7,8 TCDD.

Table I - Toxicity of Reagents and Comparison Materials

Material	LD50. Oral-rat (3)
 polyethylene glycol 400                           27,500 mg/kg
 dimethyl sulfoxide                                17,500 mg/kg
 sodium chloride (comparison value)                3,000 mg/kg
2,3,7,8 TCDD     (comparison value)                0.022 mg/kg

The reagents used in this process are some five times less toxic than table
salt, and roughly six orders of magnitude less toxic than 2,3,7,8 TCDD, the
dioxin isomer of major concern.  Polyethlyene glycol 400 is an FDA
approved material for use in foods and cosmetics.  Dimethyl sulfoxide is a
naturally occurring material in  foods such a potatoes, milk and coffee at the
part per million level.   Expected residual levels  of these materials in soil are
not expected to be a serious concern.

      Toxicity testing of the  reacted aromatic halides is currently underway
with EPA sponsorship.  Structural assessment of the theoretical toxicity of
the reaction products is favorable, ie the known reaction products would not
be expected to show significant toxiciiy.  Results of the Ames test for
mutagenicity are negative, ie the reaction products do not demonstrate
carcinogenic potential.  Bioaccumulation tests also produced negative
results, which is not surprising given the water solubility of these reaction
products.  Acute toxicity tests are currently (April, 1986) in progress.

-------
Process Description

The decontamination of soil proceeds in a series of six process steps.

      1.  Combine equal masses of soil and reagent to form a slurry
      2.  Mix and heat soil/reagent slurry to 100-180 °C
         Allow to react 1-5 hours
         Decant excess reagent
         Wash soil 2-3 times with water
         Discharge decontaminated soil
This process is shown in diagram form below;
                                       MAKEUPWATER •
CONTAMINATED
       SOIL
Figure 1 - Soil Decontamination Process

All of the process steps can be conducted in a single agitated reactor. The
number of wash steps required will depend on the effectiveness of each
wash step and on the degree of reagent recovery required.

 Results of Tests to Date

Three series of tests using dioxin contaminated soil  have been conducted to
date: laboratory tests at high and low rates of agitation and field tests at low
agitation only.  Each set is discussed separately.

Laboratory Testing - High agitation

      Initial laboratory tests used 250 g. soil samples soiked to a nominal
concentration  of 2000 parts per billion (ppb) of 1,2,3,4 TCDD.  The 1,2,3,4
isomer was used in  place of the 2,3,7,8 TCDD isomer to simplify
experimental and safety procedures.  These tests used an electrically
heated 1000 ml three neck flask equipped with a reflux condenser and high
torque agitator to provide a high degree of mixing.

-------
      In the initial series of experiments a Teflon paddle was used with the
agitator.  Analysis of'tne treated soil samples-revealed the presence of an
unknown halogenated contaminant which was later determined to be
partially cecomposed Teflon.  The combination of reagent and erosion from
the soil had broken down the Teflon used in the agitator. This interference
required  some additional sample cleanup.  Changing to a glass paddle
solved the problem for laboratory testing.

      Treated samples were analyzed by three dif^rent labs using either
gas chromatography/mass spectroscopy (GC/MS) cr GC/MS/MS  methods.
Analysis by gas chromatography alone was unsuccessful, partly due to the
Teflon interferences previously noted. The  results 01 this initial  testing are
summarized below, with all samples having a nominal initial concentration
of 2200 parts per billion;

         Table II - Results of LaDoratory Testing  with High Agitation
                     Initial Concentration  2200 ppb

Reaction                 Reaction         Final TCDD
Temperature. °C	Time. Hours	Concentration, ppb	

      260                      4                <1
      150                      2                <1
      100                      2                <0.2
      70                      0.5               15
      70                      2                <1
      50                      2                29
      25                      2                36

These tests indicated that while reaction rates for soils were lower than
those obtained in oil tests, the overall reaction times were reasonable for
large  scale application.

 Laboratory Testing - Low Agitation

After design of the field test equipment, it became apparent that the degree
of agitation obtained in the initial laboratory tests was not going to be
achieved in the field.  Therefore, tests were conducted at a low rate of
agitation to provide a prediction of the probable results of field testing.  This
testing used a flask equipped with a condenser and inserted into  a heated
oil bath.  Agitation was proviced by manually swirling the flask and contents
at periodic intervals. Soil for this test was the same soil to be used for field
testing, and contained 2,3,7,8 TCDD.  Analyses were made using
GC/MS/MS techniques. The results of testing are shown in Table III.

-------
Table III - Results of Laboratory Testing at Low Agitation, 125 °C
            Reaction time, nours    TCDD level. or
                  0
                  1.5
                  4.25
                  7.0
             175
             15.1
             2.06
             0.3
These data indicate tha* the reaction time for samples with low rates of
mixing are on the order of 2-3 times longe; than those for samples with high
rates of mixing, but still well below 8 hours.

Field Test Results

Field testing for this test series consisted of a series of runs using 30 kg. soil
samples taken from a dioxm site in Mississippi. Herbicide Orange had been
stored at this site, with some spillage, causing  soil contamination.

      Test equipment for this series of tests is shown in Figure 2 below;

     DRUM PUMP                    orernunio UAI UP           VENT
                                 RESERVOIR VALVE
   REAGENT DR WASH
   WATER DRUM
REACTION DRUM
ON DRUM ROCKER
CARBON
FILTER DRUM
 Figure 2 - Apparatus for Field Soils Processing

 A 55 gallon drum was modified by the addition of a steel plate halfway down
the drum. The plate was pierced by a valve and by a tap to allow pumping of
liquids out of the reservoir.   The steel plate was sealed with a silicone
sealant around the perimeter of the plate. Contaminated soil and reagents
were added to the drum  by weight.  The drum was then covered and a vent
line attached between the drum  and a carbon filter.  A heat tape was
wrapped around the drum and the entire drum was insulated with fiberglass.
The insulated drum was  placed  on a drum rocker and rocked from side to
side to mix the soil and reagent  during heating and reaction. The results of
the testing are shown in Table IV.

-------
                    Table IV - Results of Field 1 t-c:ing
Reaction time, hours
TCDD level. DDb
Initial Final
i.O
2.5
6.5
154 37.3
356 10.7
equipment failure
The equipment failure occurred at the seal around the perimeter of the steel
plate nolding the soil/reagent slurry out of the reservoir.  In the 6.5 hour njn
this seal '.ailed, allowing the reagent to separate from the soil and stop the
reaction. As noted in the laboratory testing, the reageni is very corrosive to
polymers, including Teflon.  Seal material selection will  be studied in depth
prior to scaleup of this process.

Discussion of Laboratory and Field Data

The data from the three series of tests is summarized in Figure 3 below;
         10000
          1000
 TCDD,
 ppo
Field data - Low mixing
                                                  Lab data - low mixing
                 Lab data - high mixing
              0123

                               TIME, MRS

Figure 3 - Results of Soils Processing

Three points can be noted from these data;

       1. For each reaction, an initial rate of reaction is followed by a
      second lower rate of reaction, vs. the single line reaction plot
      expected for a single order reaction.
       2. The laboratory da*a for the high mixing case indicate a
      higher rate of reaction than for the lower rate of mixing.  This
      difference is primarily in the initial reaction rate, while the
      secondary reaction rate is closer to that for the low mixing case.
       3. The rates of reaction for field tests and for laboratory tests at
      low agitation rates are very similar.

-------
      The bimodal reaction rate is characteristic for soils treated by this
process but not for oil treatment where a single line reaction plot is
observed. This dual  reaction rate  may be due.to the heterogeneity of soils.
Organics on soils may be adsorbed on the surface of the soil particles,  in the
micropores of the soil or even wrapped up in the helica' humic structures
present in some soils. A bimodal reaction rate would be consistent with a
process where extraction of the dioxin from the soil into the reagent is the
rate limiting step. Extraction of dioxin from the micropores would be
expected to be much slower than from the surface  of the soil particles.  This
is consistent with overall rate data showing that the rate of reaction is much
higher for liquids than for soils, indicating an extraction limited procesL.

      The micropore/soil surface  phenomeion may also explain why the
high mr.ing case shows a much greater difference in initial and final reaction
rates thun those for the low mixing case. Despite the fact that the soil for all
tests came from the same  site, the soils for the high mixing case were spiked
with dioxin on the same day as Uie soil was processed.  By contrast, the low
mixing case soils had weathered for more than five years. It may be that
weathering may change the micropcre/soil surface distribution of the
adsorbed dioxin, possibly by differential volatilization oi the dioxin from the
surface or by successive displacement of the dioxin by other materials.

      The low mixing data for laboratory and field data show a very  high
degree of correlation. If placed on a normalized graph, these data fall on a
single  line, as shown below;
        1000000i



FRACTION

 TCDCT   10000°
1000000



          10000
               1
                                  2        3
                                  TME.HOLRS
 Figure 4 - Comparison of Laboratory and Field Data at Low Agitation

This indicates that the 100:1 scaleup of the process was successful in
demonstrating that the procedure is not strongly dependent on sample size.

-------
Conclusions

       The results of this study can be summarizes as follow;

       1.  Chemical decontamination of dioxin contamina'.ed soils
      can reduce dioxin levels to < 1 ppb under laboratory
      conditions using either high or low rates of agitation.
       2.  Increasing rates of agitation yield increasing rates of
      reaction, although other factors may also be involved.
       3.  Field test data at low rates of agitation are very
      comparable to laboratory data at low rates of agitation.

Acknowledgments

This work has been sponsored by the United States Air Force and by the
United States Environmental Protection Agency under EPA contract 68-03-
321.

 References.

1. Peterson, U. S. Patent 4,574,013, March 4, 1986.
2. Esposito et. al., EPA-600/2-80-197, p. 187
3. Niosh Registry of Toxic Effects of Chemical Substances, 1981-2

-------
Gal son Research
  Corporation
   S(OIKtt**Ro«d
  EtstSyncuse, NY 1J057
 (31S)4SJ-S1SO(v*t*>NYS)
(aOO)PCBS-)23(outj(deNYS)


                                                   Decembers, 1988
David Hrebenach
COM Federal Programs Corp.
13135 Lee Jackson Memorial Highway
Suite 200
Fairfax VA 22033

Dear Sir:

      Tim Geraets informed me that you need updated information about our
APEG process. Enclosed are papers describing final treatment of dioxin
contaminated waste streams from Montana Pole and Western Processing Sites,
a paper describing a primitive drum-scale remediation of PCB soil at the
Bengart & Memel Site, and a paper comparing lab and pilot data for dioxin
contaminated soil from Gulfport Mississippi.

      We are currently writing the report on the PCB soil pilot study we did this
summer at the Wide Beach site and designing a full scale soil treatment unit.
We expect to have the treatment unit built by next summer.

      If you have any questions or need further information, please feel free to
call.
                                                   Very truly yours,
                                                   Edwina Milicic
                                                   Senior Chemist

-------