-------
          FIGURE 4   PLOT OF OXYGEN UPTAKE RATE
                              ( Ou VERSUS  TIME )
              10     20     30     40     50   .60    70
                      UM-0.1. KS-1. YC-0.4. XO-5. SO-100
FIGURE 5  Ptot showing (AQu/At) versus Out curve containing a
             ifnoAP Oa/*tu*vn tAjHh a *tts**>ui.»  4* WA UMA^ L _._&*	A.'	
     6.0
     4.5
     4.0
     as
   £3.0
   g2£
   02.0
     15
     to
     0£
     0.0
             linear section with a slopes mto be used In estimating
             the values of K Jn cases whore K_values are nealioible
             In comparison t
-------
    500
en
             10      20      30      40



                 ELAPSED TIME (Days)




FIGURE 7    OXYGEN UPTAKE DATA ON ANILINE



  500





  400
«   300
z
a
u


U)
         — — — PenlMqrthrilol A .

         —— Ptntxrirthfitol B .

              PentMrythrilol It Aniline
    200
     100 •
               "TO      20      30      40



                   ELAPSED TIME <0ayo>
 FIGURE 8   BIOLOGICAL OXYGEN UPTAKE CURVE

 r'°U              (Run 1 Sample No. 10-12)
     100
 a
 a
 m
               Paitaajthritol A


          — — — PenlMnrthritol 0
                    ELAPSED TIME  (days)


FIGURE 9  BIODEGRATION (% BOD REMOVAL) CURVE

                     ( PENTAERYTHRITOL )
                       291
-------
             1. Inoculum Control

             2. p-Xjtcn.

             3. p-Xylm BOD

             4. p-X;l
-------
    TABLE 1.   SUMMARY OF RESPIROMETRIC  BIODEGRADATION DATA FOR SELECTED BENZENES
                  PERCENT BIODEGRADATION (BASED ON % BOD REMOVAL)
Time
(days) Aniline
0.0 0.0
1.0 2.58
2.0 3.43
3.0 3.94
4.0 14.95
5.0 102.0
6.0 102.0
7.0 102.0
8.0 102.0
9.0 102.0
10.0 102.0
11.0 102.0
12.0 102.0
13.0 102.0
TABLE Z.
COMPOUNDS
Aniline
(Experiment 1)
(Experiment 2)
Benzene
Ethyl benzene
Toluene
p-Xylene
m-Xylene
tert-Butyl benzene
sec-Butyl benzene
Cumene
Butyl benzene
1-Phenyl hexane
Benzene
0.0
2.11
2.11
4.15
4.97
71.5
74.22
81.85
93.37
93.89
95.45
95.45
96.13
97.46
SUMMARY OF
ThOD
for 100
310
310
308
317
313
317
317
322
322
320
322
326
To! uene
0.0
2.78
8.81
85.11
89.04
94.34
97.92
100.0
101.18
103.48
103.48
103.48
103.48
103.48
BIO-KINETIC DATA
mg (days)
4.00
4.00
4.50
4.00
2.00
3.90
2.00
4.40
3.50
2.40
3.30
4.00
Ethyl benzene
0.0
2.27
2.27
2.99
3.78
4.29
40.25
73.91
73.78
83.94
. 88.86
88.86
89.65
91.00
FOR SELECTED














BENZENE
t; t2
(days) (days)
4.70
4.65
4.87
4.21
2.20
4.22
2.35
5.12
4.00
2.79
3.92
4.55
4.83
4.79
5.00
4.83
2.42
4.83
2.50
5.70
5.70
3.00
4.56
5.15
m-Xylene
0.0
1.35
2.14
69.68
69.68
76.68
82.58
84.67
87.03
87.79
90.12
90.15
90.97
91.80
COMPOUNDS
/I
(day-1)
2.78
3.80
8.57
8.33
8.75
9.94
6.60
1.21
0.78
2.31
2.42
1.85
p-xylene
0.0
1.58
1.58
1.58
9.4
70.38
82.68
95.58
99.49
101.76
104.7
105.78
106.7
108.6

*'
(day-1)
3.29
8.60
25.30
44.80
14.93
4.50
29.60
1.69
0.73
2.24
2.42
1.90
It  - specific growth rate constant for exponential  growth  phase of BOD curve.
p' » specific growth rate constant for declining  growth  phase of BOD curve.
                                        293
-------
  TABLE 3.  SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED BENZENE COMPOUNDS
COMPOUNDS
Aniline
Benzene
Ethyl benzene
Tolouene
p-Xylene
•-Xylene
tert-Butyl benzene
sec-Butyl benzene
Cunene
Butyl benzene
1-Phenyl hexane
Lag Time
-------
  TABLE  5.   SUMMARY  OF  MONOD KINETIC  PARAMETER DATA FOR SELECTED  PHENOLIC COMPOUNDS
COMPOUNDS
Aniline
Phenol e
Resorcinol
p-Cresol
o-Cresol
m-Cresol
Catechol
2,4-dimethyl phenol
Lag Time
(t0)
days
4.00
1.00
1.50
1.00
1.20
1.44
0.85
2.00
Y9
mg biomass
mg substrate
0.38
0.58
0.48
0.33
0.41
0.46
0.49
0.39
Mm
(day-1)
6.15
9.82
12.22
6.11
4.10
7.97
12.80
5.62
mgs/l
6.10
9.43
6.31
27.78
16.41
17.62
43.87
14.07
   H - maximum specific growth rate.
   K, - half  saturation constant; concentration of substrate at
   Yg - growth  yield, mg biomass formed/mg substrate coriiumed.
    TABLE 6.  SUMMARY OF BIO-KINETIC DATA FOR SELECTED PHTHALATE ESTER COMPOUNDS
COMPOUNDS
                       ThOD          t0         t,         t2          it           v'
                     for 100 mg     (days)     (days)     (days)     (day-1)     (day-1)
Aniline
Dimethyl phthalate
Diethyl phthalate
Dipropyl phthalate
Butyl benzyl phthalate
310
168
195
211
226
4.00
3.46
2.00
2.40
2.00
4.70
3.98
2.97
2.87
2.28
4.83
4.25
3.30
3.40
2.80
2.78
2.76
2.16
2.04
4.12
3.29
4.71
2.92
2.00
2.33
       specific growth rate constant for exponential  growth phase of BOD curve.
       specific growth rate constant for declining growth phase of BOD curve.
                                           295
-------
TABLE 7.  SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED PHTHALATE ESTER COMPOUNDS
COMPOUNDS Lag Time
(t0)
days
Aniline
Dimethyl phthalate
01 ethyl phthalate
Dlpropyl phthalate
Butyl benzyl phthalate
H * maximum specific
4.00 .
3.46
2.00
2.40
2.00
Y j^
ma biomass (day-1)
mg substrate
0.38 6.15
0.43 7.07
0.46 3.00
0.48 5.78
0.61 7.80
«g'/i
6.10
41.68
11.67
15.81
36.25
growth rate.
K, « half saturation constant;
Y, - growth yield, ng
biomass
concentration of substrate at 0j/2.
formed/ing substrate consumed.


        TABLE 8.  SUMMARY OF BIO-KINETIC DATA FOR SELECTED KETONE COMPOUNDS

COMPOUNDS              ThOD           t0          t,         t,          n           n'
                     for 100 ng     (days)      (days)      (days)      (day-1)     (day-1)
Aniline
Acetone
2-Butanone
4-Methyl -2-pentanone
Isophorone
310
221
244
272
278
4.00
3.70
2.00
1.85
22.30
4.70
3.99
2.20
2.24
23.70
4.83
4.18
2.35
2.35
25.40
2.78
2.45
2.41
2.31
0.73
3.29
3.98
4.98
4.80
0.38
  p  « specific growth rate constant for exponential growth phase of BOD curve.
  It' • specific growth rate constant for declining growth phase of BOD curve.
   TABLE 9.  SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED KETONE COMPOUNDS
COMPOUNDS
Aniline
Acetone
2-Butanone
4-Hethyl -2-pentanone
Isophorone
Lag Tine
(to)
days
4.00
3.70
2.00
1.85
22.30
ma biomass
ng substrate
0.38
0.36
0.39
0.45
0.43
(day-1)
6.15
4.86
5.11
6.40
1.57
*s
«>9/l
6.10
9.76
10.79
24.70
27.42
  /i - naxlnua specific growth  rate.
  K, - half saturation constant; concentration of substrate at
  Y, - growth yield, *g bloaass formed/eg substrate consulted.
                                           296
-------
       PREDICTION AND MODELING OF BIODEGRADATiON KINETICS OF
                     HAZARDOUS WASTE CONSTITUENTS

                          Rakesh Govind and Sanjay Desai
                      Dept. of Chemical & Nuclear Engineering
                             University of Cincinnati
                              Cincinnati, OH 45221
                                      and
                                 Henry H. Tabak
                    RREL, U S Environmental Protection Agency
                              Cincinnati, OH 45268

                                  ABSTRACT

     Biodegradation is the  most important mechanism in  controlling the concentration
of chemicals in an  aquatic  system  because it  can mineralize  toxic pollutants to
innocuous forms.  So the  fate  of organic  chemicals  in  an aquatic environment,  is
dependant on  their  susceptibility  to  biodegradation.  Because  of  a large  number of
chemicals, it will  be expensive and labor intensive  to gather this information in a
reasonable amount of time. Hence, there is  need for a  prediction method to obtain
these data.

     Experiments  using  electrolytic respirometer,  were conducted to collect oxygen
consumption data  of some RCRA compounds and  first  order kinetic rate constants
were  obtained.   Using  these  data  along with those from literature, a  structure-
activity relationship was developed.
                                       297
-------
                                  INTRODUCTION

      It has been  estimated that 50,000 organic chemicals are commercially produced in
 the United States  and a  large number  of new  organic chemicals  are added  to the
 production each year  (1).  The presence of many of these chemicals in the environment
 is  a  serious public health  --•-•---   -••  •               •• •     .......
 disposal techniques.
ilth problem. Their presence could be  attributed to inadequate
      Since  many of these 'hazardous  chemicals can be  detected  in wastewater,  their
fate in wastewater treatment system  is  of great  interest.  Of the many factors that
affect the  fate  of these  compounds,  microbial  degradation  is  probably  the  most
important  (2).  The   high  diversity  of  species   and  the  metabolic  efficiency  of
microorganisms suggest  that  they play a  major role  in  the ultimate degradation  of
these   chemicals   (3).   Biodegradation   can   eliminate  hazardous   compounds   by
biotransforming  them  into innocuous  forms,  degrading  them by  mineralization  to
carbon dioxide  and water. So the information regarding the extent and  the rate  of
biodegradation  of  these  chemicals  is  very  important   for  regulation   of   their
manufacture and  use.  Due to a  large number of  these  chemicals, gathering of this
information in a reasonable amount of time will be  both expensive and  labor intensive.
Thus, the only practical way out is to  develop correlations  and predictive techniques
to assess biodegradability (4). Lack of an adequate database on  biodegradation kinetics
prevents the development of such techniques.

                           EXPERIMENTAL TECHNIQUES

     There are many techniques  for collection of biodegradation  data and  these  are
reviewed  in great detail by Howard,  et  al. (5)  and Grady  (6). Experimental techniques
for biodegradation  data collection fall  into tnree  broad categories :  continuous,  fed-
batch and batch reactor systems.

     Continuous culture reactors require an acclimated biomass.  They also require long
transitional  time intervals for  reaching quasi steady state condition (7). So  it is time
consuming, tedious and expensive. However, the analysis of data  obtained is simpler
because  the equations for continuous  reactors, operating at  steady  state,  reduce  to
algebraic  equations that are easily solved.  This technique is  used  more for evaluating
parameters  for the design of  treatment systems rather  than  for biodegradation kinetics
(8).   Because of the microbial competition, each continuous  reactor will have  a unique
microbial  community associated with it  (9); hence the evaluated  kinetic  parameters are
system specific and are not intrinsic.

     Fed—batch  reactors have also been used  to  estimate  the kinetic parameters  of
biodegradation. In these reactor  configurations, a quantity of biomass  is added into a
reactor and a substrate stream is  continuously added in negligibly  small amounts with
respect to  the  reactor  volume.  Because  of small amounts  of both  substrate  and
microorganisms, a  pseudo steady state is achieved. So this technique reduces  the time
required  and also alleviates the  problems associated with  changes in the composition
of the microbial community. This configuration requires an acclimated biomass because
the microbes must be capable of responding instantaneously  to the input of substrate.
However,  this technique cannot be used to determine the  Monod biokinetic parameters,
but it is  an excellent  procedure to determine the parameters for  treatment  plant
design or operation.

     The use of batch cultures  in biotechnology and  biological  wastewater treatment
                                          298
-------
represents  a less  expensive  and much  faster way  to model  biokinetics  in  fermenters
and  in  activated  sludge  tanks  (7). The batch  method  commonly  used for a  large
number of  compounds (10)  is  one in  which the substrate of  interest, at  different
concentrations, are inoculated with the small amount of biomass.  Then the  increase in
biomass  concentration  in  each  reactor is measured as  a  function  of  time.  Another
technique focuses  on the substrate removal rather  than the microbial growth. This is
commonly  used in engineering  studies. The batch reactors are inoculated  with  large
quantities of biomass and the  substrate  removal is  measured  as a function  of  time.
Both  of  these  techniques  have  been widely  used using general measures as  five day
BOD  and  COD.  Batch  reactor can  be used  with either acclimated or unacclimated
biomass. It requires that samples be taken at discrete time intervals during  the course
of biodegradation.  So if unacclimated biomass is  used, the number of samples required
may  be  large  depending  on  the acclimation  time. Tabak, et  al. (11)  have  collected
degradability and   acclimation  data on 96  compounds  by  static  culture  screening
procedure and culture enrichment process. Larson and Perry (12) and  Paris,  et al. (13)
have  done  biodegradation studies with  unacclimated biomass in  the batch reactors and
have evaluated the kinetic parameters.

      In the above  procedure, the number of data points  collected are less  because of
manual sampling. This  can be avoided by monitoring oxygen consumption as  an indirect
measure  of  biodegradation  using  an  electrolytic  respirometer.  The automatic   data
collection and  recording allows  sufficient accumulation of data,  so  the reliability  of
the kinetic parameters evaluated is maximized.

      In the respirometer methods of BOD measurement,  wastewater samples  are kept
in contact  with the gas phase source of oxygen. Oxygen uptake by the microogranisms
over  a period of time  is measured by  the  changes in  volume  or pressure  of the  gas
phase. An  alkali is included in the apparatus to absorb  carbon dioxide produced during
biodegradation.  Samples are  usually incubated at  constant temperature  and  are  kept
away from  light.  The latest development  in  respirometric techniques  has been  the
advent of  an electrolytic respirometer.  It supplies  oxygen, produced  from  electrolysis
of water, to the  air space above the sample in a completely  sealed reaction vessel.
The  production of oxygen is triggered  due to the  pressure changes in the reaction
vessel. Studies by Larson and Perry (12), Young  and  Baumann  (14), Tabak,  et al. (15)
and  Dosanjh and  Wase (16) have  shown that the electrolytic  respirometer eliminates
most of  the technical difficulties associated  with other  methods for determining  BOD.
It is  particularly  useful for  the rate studies because  it  provides both,  a  continuous
record of oxygen uptake and it  maintains an unchanging atmosphere over the sample
regardless of the length of the test.

                             EXPERIMENTAL SET-UP

      The electrolytic respirometer Sapromat B—12, consists of a temperature controlled
waterbath,  which  contains the  measuring units, a  recorder for digital  indication and
direct plotting  of the oxygen uptake curves  and  a  cooling unit. The waterbath has  12
reaction  flasks, each connected to the  recorder. Each  unit  as shown in  figure  1
consists  of a  reaction vessel  C,  with a  carbon dioxide absorber (sodalime) 3,  mounted
in  a   stopper,  an oxygen generator  B, and a pressure indicator  A;  The vessels
interconnected  by  hoses,  form  a  sealed measuring system so  that the  barometric
pressure  fluctuations do not  adversely  affect the result. The magnetic stirrer 1,  in the
sample 2,  to  be  analyzed,  provides  vigorous  agitation, thus ensuring an  effective
exchange of  gases.  The  activity  of  the  microorganisms in  the sample  creates  a
reduction in the  pressure  which  is  recorded by  the  pressure  indicator.   It controls
both,  the  electrolytic oxygen generation  and  the  indication  and the plotting of the
                                         299
-------
measured values.

     The consumption of oxygen by microorganisms  creates a reduction in  pressure in
the  reaction  vessel. As a  result, the  level of 0.5%  sulphuric  acid  in  the pressure
indicator  rises and  comes  in  contact with the platinum electrode. This  completes the
circuit and triggers  the  generation  of oxygen  by  electrolytic cell. The oxygen  gas is
provided  to the  reaction  vessel, alleviating the  negative  pressure.  So, the level of
electrolyte in the pressure indicator  drops down,  breaking  contact  with the electrode.
This switches  off the electrolytic cell. The amount of oxygen supplied to the sample is
recorded directly in  milligrams per litre by  the recorder. The recorder is  connected to
an IBM AT which records data from the measuring units every 15 minutes.
                         Figure 1. Diagram of a Measuring Unit

                      A.  Pressure indicator
                      B.  Oxygen generator
                      C.  Reaction vessel
                      1.  Magnetic stirrer
                      2.  Sample (250ml)
                      3.  CC>2 absorber
                      4.  Pressure indicator
                      5.  Electrolyte
                      6.  Electrodes
                      7.  Recorder
                            BIODEGRADATION KINETICS

     A  considerable  amount  of  information  concerning  biodegradation   kinetics  is
available  in  the  published  literature.  Early  literature  shows  widely  differing  kinetic
rates  in different  studies. The evaluation and prediction of  the  extent and  rate  of
biooxidation  is affected by methodological and  experimental  factors.  Regardless of the
different  assumptions  involved in  the  measurement  of  biodegradation rates,  it  is
                                          300
-------
 generally considered associated with microbial cell growth  and so most of the models
 for it are the same as those used to model growth and substrate removal.

      Although  many models  have  been proposed for microbial  growth, the  Monod
 relation  is the most popular kinetic expression (17). Monod model, in combination with
 the  linear law for substrate removal can provide an adequate description of microbial
 growth  behavior. It states that the  cell growth is  first order with  respect  to the
 biomass concentration  (X) and mixed  order with respect to the substrate concentration
 (S)

                            dX/dt = (S u™ X) / (Ks + S)  .  .  .  .  .  (1)

 Cell growth is related to the substrate removal by the linear law

                                dX/dt = - Yg (dS/dt)	(2)
      The  kinetic  parameters of interest  are  maximum  specific growth  rate |xm, half
 saturation constant Ks, (it is the  concentration of substrate when u,=0.5jxm),  and the
 yield coefficient Yg. The Monod equation has two limiting cases.  When the substrate
 concentration  is  much  greater than  the  saturation  constant  the term  (S/KS+S)
 approaches 1.0 and cell  growth and  substrate removal  are zero order with respect  to
 substrate concentration.   When the substrate concentration is much smaller than the
 saturation constant,  the term  (S/KS+S)  approaches  (S/KS)  and  cell growth and
 substrate removal  are first order  with respect to the  substrate concentration.  Many
 researchers have  used  either of  the above  two  approaches. The characteristics  of
 these kinetic expressions have been discussed by Sim kins and Alexander (18).

      The  electrolytic  respirometer  has been  mostly used to  measure the  extent  of
 biodegradation as a percentage of  the theoretical oxygen demand over a certain period
 of  time.  Several  researchers  have tried  to  extract kinetic  parameters from  oxygen
 uptake  data.  Larson  and  Perry  (12) used  empirical kinetic  expressions which  were
 system  specific.  Dojlido  (19)  divided  the  oxygen  uptake curve into  seven  different
 phases  and  then  proposed  an empirical model  for each  phase  and evaluated the
 biodegradability and toxicity of a test compound by measuring empirical rate  constants
 and time  intervals  associated with  each phase. Tabak, et al. (15) divided the substrate
 removal  region  of  oxygen  uptake  curve  into two  regions, separated by  an  inflection
 point. The  first  period  was  called  exponential phase where it was  assumed that the
 substrate  was  not  limiting and cell growth was occuring,  while the second period was
 called declining phase  and here it  was assumed  that the substrate  was limiting. So
 work done to obtain kinetic parameters from oxygen  uptake curve has been empirical.

                      STRUCTURE-ACTIVITY RELATIONSHIP

      The  structure—activity relationships (SAR) have  been widely used in pharmocology
 and medicinal chemistry. The different  methods  and  models  used are free energy
 models, Free-Wilson mathematical model, discriminant analysis, cluster analysis, pattern
 recognition, topological methods, and quantum mechanical  methods.

     The  free energy  model of Hansch, et al. (20)  is widely  used. They incorporated
octanol—water coefficient,  log P, in the linear free energy relationship as a measure of
lipophilicity. This provided a  general SAR model for biological  activity. The success of
this model has led many  researchers  to include additional physicochemical parameters
and properties,  structural,  topological  and molecular indices.  Using similar  principles
other researchers have  proposed models to include more  complex relationships between
                                         301
-------
the bioactivity and the  chemical  structure or the properties. Martin  (21) has dicussed
these models.

     Free and Wilson proposed a mathematical model to assess  the  additive effects of
substituents  and quantitatively estimate their magnitude  (22). According to their model
the structure of  a  compound is composed of different groups or  a  core which is
substituted in various positions, resulting in a  series of linear equations of the form
where  BA is  biological activity, X: is the j**1 group, aj is  the  contribution of the jtn
group  and (3  is the overall average activity.  These linear equations are solved by least
square method for a;  and (3. This method requires a large number of compounds for a
meaningful  analysis and  it  will breakdown  if there  are interactions between different
groups. Fujita and Ban (23) suggested that BA should  be expressed as log(l/C), where
C  is the  concentration of the compound  that produces a constant biological response.
This modified Free— Wilson model is in common use in medicinal chemistry.

     Discriminant analysis is used where  only semiquantitative or qualitative  data have
to  be  evaluated. In  this method,  a linear combination  of parameters  called  linear
discriminant function  is  formed,  which  classifies the  observations.  Martin  (21)  has
discussed the background of this method  and  has given examples.  Principal component
method is  used  as a  preliminary step in multiple regression  analysis  of  the  Hansch
type   (24).   Factor   analysis   is  used   to  gain  insight  into  the  structure   of
multidimensional data set and involves manipulation of the eigenvectors of  variance—
covariance  matrix of the dataset  (25).  Cluster analysis  is  used  to group similar
substituents when various combination of parameters are considered (26).
     Pattern  recognition   is   used  to  examine   structural   features  and   chemical
properties  for the  patterns  associated  with different biological activities  (27). In this
method, a set  of descriptors  is  generated  for  each  compound  and  then  suitable
algorithm  is applied to find some combinations  and weight  of the descriptors which
give a perfect classification  for a set of compounds. This classification is then applied
to another set of compounds of known classification and performance is judged by the
percentage of correct prediction.

     Various methods  have been proposed to relate topology of the molecule  with its
biological activity.  Verloop  (28)  has proposed  a method to treat directionality of  steric
effects. A  computer representation of  a  compound  is created  and then  measured by
tangential  planes which results  in five STERIMOL parameters. Kier and  Hall  (29)  have
used   molecular  connectivity   index,  a  number  calculated   from  graph  theoretical
principles for SAR correlation. Blankley (30) has reviewed all  these and other methods
used in SAR.

     In the field of biodegradation there are several studies which have attempted to
correlate  some  physical,  chemical or  structural  property  of  a  chemical   with its
biodegradation.  Based  on the type  and the location of the substituent groups, Geating
(31)  developed  an algorithm  to  predict  biodegradation. Qualitative relationships for
different compounds have  been investigated  by others,  but  quantification is  required
for regulatory purposes.

     In the literature,  first  order rate of biodegradation or five day BOD of  chemicals
have  been  correlated  with  physical  or  chemical  properties.   Paris, et  al. (32,33,34)
established a correlation between  second order  biodegradation rate constant  and the
                                          302
-------
van der Waal's  radius of substituent group, for substituted anilines and for a series of
para-substituted phenols. Wolfe, et al. (35)  correlated second order alkaline hydrolysis
rate  constant and  biodegradation rate constants for selected  pesticides  and phthalate
esters.  Several  workers  have  observed  a  correlation  between  biodegradability  and
liphophilicity,  specifically  octanol/water partition coefficients (log  P).  Paris, et  al. (36)
found a good correlation  between biodegradation rate  constant and log P, for  a series
of esters  of 2,4—dichlorophenoxy acetic  acid. Banerjee, et al.  (37) obtained a similar
relationship for  chlorophenols.  Vaishnav, et al.  (38)  correlated  biodegradation  of  17
alcohols  and  11  ketones with  octanol—water  coefficients  using 5—day  BOD data.
Pitter(39)  has found  a  dependence of biodegradation rate on electronic  factors,  like
Hammett  substituent constant, for a  series  of  anilines  and phenols.  Dearden  and
Nicholson  (40) have  correlated  5—day  BOD  with  modulus difference  of  atomic charge
across a selected  bond in a molecule for amines,  phenols, aldehydes, carboxylic acids,
halogenated   hydrocarbons   and  amino  acids.  A  direct  correlation  between  the
biodegradability  rate  constant and  the  molecular  structure  of the  chemical has  been
used  by Govind  (41) to relate  the first order  biodegradation  rate constant with the
first   order molecular  connectivity  index  and  by  Boethling  (42)   to  correlate  the
biodegradation rate  constants  with  the  molecular  connectivities  for  dialkyl  esters,
carbamates, dialkyl ethers, dialkyl phthalate esters and aliphatic acids.

                       GROUP CONTRIBUTION APPROACH

     Using a  group  contribution  approach,  a very   large number  of chemicals  of
interest can  be  constituted  from  perhaps  a  few  hundered  functional  groups. The
prediction  of  the  property is based on  the  structure  of the compound. According to
this  method the molecules of a  compound are structurally decomposed into functional
groups  or their  fragments,  each having unique contribution  towards  the compound
property.

     The  biodegradability  rate  constant  k, is expressed as  a  series  function  of
contribution a;, of each group of the compound. The  first order approximation of this
series function representing biodegradation rate constant can be expressed as
                         Ln (k) = 2
(3)
where  Nj is  the number of groups of type j in  the  compound, oj is  the  contribution
of group of type j and L is the total number of groups in the compound.

CALCULATION OF GROUP CONTRIBUTION PARAMETER, a

     Using equation 3  for each compound,  a  linear equation in ot's is  constructed.  For
a  given  dataset this generates a  series of linear equations which are solved for  ex's
using regression analysis.

     If there are n a's  and m  compounds (n
-------
                             m
       S = 2 { Ln(ki)
          1=1
                                               j }
(4)
The estimates of a's should  be such that it produces the least value of S. These
estimates are obtained by differentiating equation 4 with respect to a  and setting it
equal to zero
                       m
= -22{[Ln(ki)-2Njjaj]Nki} =
                                                                   (5)
This will generate a series  of n  linear  equations which  are solved for a's. If N is the
matrix of coefficients  of a's  and Y is the vector  of Ln(k)  values then the solution
vector a is given by

                                a=(N'N)~1 (N'Y)

where ' denotes the transpose of matrix.

                              MODEL VERIFICATION

     The  study of Urano and Kato (43) was  used to determine  the  contribution of
several different groups.  This  study  was selected because  it involved testing  of a large
number of chemicals using a consistent method. The requirement of consistent method
is  necessary  to ensure  that  the  difference  in  rate  of  biodegradation of different
compounds  is due  to the difference  in  their chemical  structure and not  the  test
conditions.  The  range  and average of Lnfk) values, for compounds used for this model,
are given in  Table  1.  The average values  were used for  calculation of  the group
contribution parameters given in Table 2.

     The biodegradation experiments were carried  out for cresols, phenol, 2,4— dimethyl
phenol,  2— butanone, acetone, 1— phenyl hexane and butyl  benzene using an electrolytic
respirometer Sapromat B— 12 (Voith— Morden, Milwaukee, Wl). The chemicals  were from
Aldrich  chemical company with 99+%  purity. These compounds were used to validate
the model, but  were not used in the calculation of the  group contribution  parameters.
The experimental conditions were : temperature 25°C, biomass concentration  30 mg/L
and  compound concentration lOOmg/L. The biomass was  obtained from Little  Miami
wastewater  treatment  plant in  Cincinnati, which receives 95% domestic waste.  The
biomass  was   aerated  for  24  hours  and  then   used  for   the  experiment.   The
biodegradation  rate   constants   for  these  compounds  were  determined   using   the
following empirical first order rate  equation given by Urano and Kato (44).
                             dBOD/dt = k BOD.
Integrating this equation
                        log (BOD) = kt + constant
                                                 (6)
                                                 (7)
is  obtained. The above equation was used  for only the rising part of the BOD  versus
time curve, and initial data and data for endogenous phase were neglected. Note that
the selection of the  rising  part of the BOD  curve  was  arbitrary and was  based on
visual  inspection of  the BOD  curve.  The biodegradation  rate  constants  were also
predicted using the group contribution parameters of  Table 1. The comparison of these
values  is given in Table 3  and  Figure  2. The % error in BOD values were calculated
                                         304
-------

                                                   
-------
by
                      % Error = (1/n) 2 {(BODe - BODp)/BODe}*100
where subscripts e and p  denote  experimental and  predicted values  respectively and n
is  the total number of data points used.  Figure 3  shows the best and  worst case of
prediction  obtained.  The  experimental  conditions  in  both  the  studies  were  similar
except the nature and the source of biomass.

                            RESULTS AND DISCUSSION

     The average error in  the prediction of BOD values varies from  13% to 85%. It is
important to emphasize that there are several reasons for the rather large discrepancy
between the predicted and experimental values. These reasons are as follows :

1.    The data used for calculating the group contribution parameters were  developed
by  Urano  and  Kato (refer Table 1).  The  experimental  BOD curves, obtained from
electrolytic  respirometry,  were   approximated  by  an  exponential   relation   (refer
equations 6  and  7}.  The  range  of curve  to  which this equation was applied was
selected arbitrarily. This resulted in some  error between the experimental data and the
calculated  value  of  the  kinetic constant,  k. Furthermore, a  range  of k values were
calculated  by Urano and  Kato  due  to  differences  in  the replicate samples.  In  our
calculation  of the group  contribution parameters,  an  average  value  for the  k  values
was used.  This introduced another error in  the input data  for the group contribution
estimates.

2.    The group contribution method described here is the first order approximation. It
assumes that the contributions of the groups  are  independent of each  other. So  the
rates predicted are within an order of magnitude.

3.    Biomass used in both the studies were different. Urano and Kato had  acclimated
their biomass to the compound  before using it in  the  experiment, while our biomass
was from  domestic wastewater  treatment  plant  and  was not  acclimated  to  the
compounds used in the study.

     Inspite  of  these errors, the  prediction is  within  an order of  magnitude. With
availability  of more  data  and using a  detailed  kinetic model rather  than a  first order
exponential equation, the prediction error may  be reduced considerably.

ACKNOWLEDGMENT  :This  research  was  supported by  co-operative agreement  CR
812939—01 from U. S. Environmental Protection Agency.

                                   REFERENCES

1.    Blackburn, J.W., Troxler, W.L. Prediction of the fates  of organic chemicals in a
     biological treatment process — An overview. Environ. Prog. 3: 163,1984.
2.    Howard, P.H.,  Saxena,  J., Sikka, H. Determining the fate of chemicals. Environ.
     Set. Tech. 12: 398,1978.
3.    Alexander,  M. Biodegradation of chemicals of environmental concern. Science 211:
     132,1980.
4.    Strier, M.P. Pollutant  treatability  :  A molecular  engineering  approach. Environ.
     Sci. Tech. 14: 28,1980.
5.    Howard, P.H.,  Banerjee, S., Rosenberg, A.  Review and evaluation of available
     techniques for  determining  persistance and  routes  of degradation of  chemical
     substances in environment. 560/5—81-011, U.S.  Environmental Protection Agency,
                                         306
-------
 o
 Q
 O
 m

 o
 o
 m
 s^»
 0>
o
Q
O
m

Q
o
m
•s^X
o»

3
0.9




0.8 -




0.7 -




0.6 -




0.5 -




0.4 -




0.3 -




0.2 -




0.1 -




  0
         0.05
      0.8
      0.7 -
      0.6 -
      0.5 -
0.4 -
      0.3 -
      0.2 -
      0.1 -
                 2,4-Dimethylphenol
                  0.15
0.25
0.35
                                                                    0.45
                    1-Phenylhexane
                            0.2
                                        —1—

                                         0.4
                                                                  0.6
                                     Time, days
         d   Experimental Data                      +   Predicted Data



               Figure 3.  Best and worst case  of prediction
                                    307
-------
     1981.
6.   Grady,  C.P.L.,  Jr. Biodegradation  : Its measurement  and microbiological basis.
     Biotech. Bioengg. 27: 660, 1985.
7.   Braha,  A., Hafner, F.  Use of lab batch reactors  to  model kinetics.  Wat Res. 21:
     73,1987.
8.   Chudoba,   J.,  Cech,   J.S.,  Farkac, J., Grau,   P.  Contro  of  activated  sludge
     filamentous bulking ;  Experimental  verification of a  kinetic selection  theory. Wat.
     Res, 19: 191, 1985.
9.   Bull, AT., Slater,  J.H. In: AT. Bull, J.H.  Slater feds.), Microbial Interaction and
     Communities, Vol. 1., Academic Press, New York, NY, 1982.
10.  D'Adamo,  P.C., Rozich, A.F., Gaudy, A.F., Jr. Analysis of growth data  with
     inhibitory carbon sources. Biotech Bioengg. 26: 397,1984.
11.  Tabak,  H.H., Quave,  S.A., Mashi,  C.I., Barth, E.F. Biodegradability studies  with
     organic  priority pollutant compounds. JWPCF 53: 1503,1983.
12.  Larson,   R.J.,  Perry,  R.L.   Use  of  electrolytic  respirometer   to  measure
     biodegradation in natural waters. Wat. Res. 15: 697,1981.
13.  Paris, D.F.,  Rogers, J.E. Kinetic  concepts for measuring  microbial rate constants
     : Effects of nutrients on rate constants. Appl. Environ. Microbiol. 51: 221,1986.
14.  Young,  J.C.,  Baumann, E.R. The electrolytic respirometer —  II —  Use in water
     pollution control plant laboratories. Wat. Res. 10: 1141,1976.
15.  Tabak,  H.H.,  Lewis,  R.F.,  Oshima,  A. Electrolytic respirometry  biodegradation
     studies,   OECD   Ring  test   of  respiration   method   for   determination   of
     biodegradability. Draft  Manuscript,  MERL,  U.S. Environmental  Protection Agency,
     Cincinnati, OH, 1984.
16.  Dosanjh, M.K.,  Wase, D.A.J. Oxygen uptake  studies on various sludges adapted to
     a waste containing chloro—,  nitro—, and amino—substituted xenobiotics. Wat.  Res.
     21: 205,1987.
17.  Howard, P.H.,  Banerjee,  S. Interpreting results from  biodegradability  tests  of
     chemicals in water and soil. Environ.  Tox. Chem. 3: 551, 1984.
18.  Simkins, S., Alexander, M.  Models for mineralization  kinetics with  variables  of
     substrate  concentration and population density. Appl.  Environ.  Microbiol.  47:
     1299,1984.
19.  Dojlido, T.R. Investigations of biodegradability and  toxicity of organic compounds.
     EPA-600/2-79-163, U.S. Environmental Protection Agency, Cincinnati, OH, 1979.
20.  Hansch, C. On the structure of medicinal chemistry. J. Med. Chem. 19:1, 1976.
21.  Martin,  Y.  C.  Quantitative  Drug Design  :  A  Critical  Introduction,  Dekker,
     NewYork, 1978.
22.  Free, S. M., Wilson,  J. W. A  mathematical  contribution to  structure-activity
     studies. J. Med. Chem. 7:395,1964.
23.  Fujita,  T.,  Ban, T. Structure-activity studies of phenylamines as  substrates  of
     biosynthetic enzymes of sympathetic transmitters. J. Med. Chem. 14:148, 1971.
24.  Mayer,  P. P. Drug Design, Ed.,  E. J.  Ariens, Medicinal Chemistry, Vol.  IX,
     p.187, Academic Press, NY, 1980.
25.  Weiner, M.  L., Weiner, P.  H. A  study of  structure-activity  realtionships  of  a
     series of diphenylaminopropanols by factor analysis. J. Med. Chem. 16:655, 1973.
26.  Hansch, C., Leo, A. J. Substituent  Constants for Correlation  Analysis in
     Chemistry and Biology, Wiley (Interscience), NY, 1979.
27.  Dunn,  W.  J.,  Ill,   Wold, S.  Structure—activity analysed by  pattern  recognition :
     The asymmetric case. J. Med. Chem. 23:595,1980.
28.  Verloop, A.,  Hoogenstranten, W.,  Tipker, J.  Drug  Design, Ed., E. J. Ariens, Vol.
     VII, p.165, Academic Press, NY,  1977.
29.  Kier, L. B.,  Hall,  L.  H. Molecular Connectivity in Chemistry and Drug Research,
     Academic Press, NY, 1976.
30.  Blankley,  J.  C.  Quantitative  Structure-Activity  Realtionships  of  Drugs,   Ed.
                                          308
-------
     Topliss, J. G., p.l, Academic Press, NY, 1983.
31.  Geating,  J. Literature study of the biodegradability of chemicals in  water.  EPA—
     600/2-81-175, U.S. Environmental Protection Agency, 1981.
32.  Paris,  D.  F., Wolfe,  N.  L.  Realtionships between  properties of a series of anilines
     and their transformation  by bacteria. Appl. Environ. Microbiol. 53:911,1987.
33.  Paris,  D.  F.,  Wolfe, N. L., Steen,  W.  C.  Structure-activity relationships  in
     microbial transformation  of bacteria. Appl. Environ. Microbiol. 53:971,1987.
34.  Paris,  D.  F.,  Wolfe, N.  L.,  Steen,  W. C.,  Baugham, G. L.   Effect of  phenol
     molecular  structure  on  bacterial transformation rate constants in pond and river
     samples. Appl.  Environ. Microbiol. 45:1153,1983.
35.  Wolfe, N.L., Paris,  D.F., Steen,  W.C., Baugham, G.L. Correlation  of microbial
     degradation rates with chemical structure. Environ. Sci. Tech.  14: 1143, 1980.
36.  Paris,  D.  F.,  Wolfe, N. L., Steen, W. C.  Microbial  transformation  of ester  of
     chlorinated carboxylic acids. Appl. Environ. Microbiol. 47:7,1984.
37.  Banerjee,  S.,  Howard, P. H., Rosenburg, A. M.,  Dombrowski,  A.  E., Sikka,  H.,
     Tullis, D.  L.  Development of a general kinetic model for  biodegradation and its
     application  to  chlorophenols and related compounds.  Environ. Sci.  Tech. 18:416,
     1984.
38.  Vaishnav,   D.   D.,   Boethling,   R.   S.,   Babeu,  L.   Quantitative   structure—
     biodegradability  relationships  for  alcohols,   ketones   and  alicyclic   componds.
     Chemosohere. 16:695, 1987.
39.  Pitter, P.  Correlation between the structure of aromatic compounds and the rate
     of their biological degradation. Collection Czecoslovak Chemical  Comm.  49:2891,
     1984.
40.  Dearden,  J.C.,  Nicholson, R.M.  The prediction of biodegradabilities by the  use of
     quantitative  structure—activity  relationships  : Correlation  of  biological  oxygen
     demand with atomic charge difference. Pestici. Sci. 17: 305, 1986.
41.  Govind, R., Treatability of toxics in wastewater systems. Hazardous Substances 2:
     16,1987.
42.  Boethling,  R.S.  Application of molecular  topology  to quantitative  structure—
     biodegradability relationships. Environ. Tox. Chem.  5: 797, 1986.
43.  Urano,  K.,  Kato,  Z.  Evaluation  of  biodegradation  ranks  of priority organic
     compounds. J>  Hazardous Mtl. 13: 147, 1986.
44.  Urano, K., Kato, Z.  A method to classify biodegradabilities of organic  compounds.
     J. Hazardous Mtl. 13:135, 1986.
                                            309
-------
TABLE 1. DATA USED TO CALCULATE GROUP CONTRIBUTION PARAMETERS

Compound
Ethyl alcohol
Butyl alcohol
Ethylene glycol
Acetic acid
Propionic acid
n— Butyric acid
n— Valeric acid
Adipic acid
Methyl ethyl ketone
Hexamethylenediamine
n— Hexylamine
Mono ethanol amine
Acetamide
Benzene
Benzyl alcohol
Toluene
Acetophenone
Hydroxybenzoic acid
Aminobenzoic acid
Aminophenol
Ln(k)
-2.9004 ~
-3.0791 "
-3.3524 "
-2.6037 "
-2.6736 "
-2.7031 ~
-2.6310 -
-2.8134 "
-3.4738 ~
-4.3428 "
-2.8647 "
-3.3242 "
-2.9957 ~
-2.8647 "
-2.7806 ~
-2.6037 "
-3.1701 "
-2.3538 "
-2.6592 "
-3.2442 "

-3.1465
-3.3242
-3.6496
-2.7181
-2.9759
-3.0576
-2.6593
-3.1235
-3.6889
-4.5099
-3.0576
-3.3814
-3.0576
-2.9759
-3.1701
-2.8647
-3.5404
-2.9188
-2.9374
-3.2968
Average Ln(k)
-3.0159
-3.1942
-3.4900
-2.6593
-2.8134
-2.8674
-2.6451
-2.9565
-3.5755
-4.4228
-2.9565
-3.3524
-3.0262
-2.9188
-2.9565
-2.7257
-3.3382
-2.6363
-2.7984
-3.2702
                             310
-------
        TABLE 2. GROUP CONTRIBUTION PARAMETERS
  Group
Methyl
Methylene
Ketone
Amine
Acid
Hydroxy
Aromatic CH
Aromatic Carbon
CH3
CH2
CO
NH2
COOH
OH
ACH
AC
-1.3667
-0.0438
-0.5073
-1.4654
-1.3133
-1.7088
-0.5016
1.0659
TABLE 3. COMPARISON OF ACTUAL AND PREDICTED Ln(k) VALUES

Compound
o— Cresol
m— Cresol
p— Cresol
Phenol
2,4— Dimethylphenol
2— Butanone
Acetone
Butylbenzene
1— Phenylhexane
Actual
Ln(k)
-2.6878
-2.3694
-2.4647
-3.0006
-2.8460
-3.1326
-3.1161
-3.1285
-3.3971
Predicted
Ln(k)
-2.9501
-2.9501
-2.9501
-3.1509
-2.7493
-3.2845
-3.2407
-2.9402
-3.0278
Error %
41.71
27.94
57.10
42.28
13.13
50.57
13.86
28.78
85.31
                         311
-------
           PRELIMINARY RESULTS ON THE ANAEROBIC/AEROBIC
      BIOCHEMICAL REACTOR FOR THE MINERALIZATION OF ORGANIC
                 CONTAMINANTS BOUND ON SOIL FINES
               Robert C. Ahlert, PhD, PE, Dist. Prof.
               David S. Kosson, PhD, Asst. Prof.
               William V. Black, Graduate Student

               Chemical & Biochemical Engineering
                       Rutgers University
               P.O. Box 909, Piscataway, NJ 08855
                         ABSTRACT

The goal of the overall research program, a part of which is
discussed in this paper, is to demonstrate a sequence of
aerobic/anaerobic microbial process steps for degradation of
contaminated soil fines and slurries of soil fines.  Toward this end,
it must be possible i) to assay individual organic species and total
contaminant organic carbon in soils of varying properties, ii) to
separate whole soils into fractions according to particle size, and
iii) to assay [as in i)] reactor slurries containing suspended soil
particles, microbial culture and dissolved, dispersed and sorbed
organic contaminants and metabolites.  These techniques are required
to define the nature of the contamination, devise operating
conditions to facilitate microbial contact, and assure complete
mineralization of target contaminations and "clean" residuals.

The first major section of this paper describes the development of
analytical methodology for whole soil and soil fractions; in
parallel, techniques for mixing/homogenizing, fractionation and
extraction used in sample preparation are discussed.  It has been
possible to separate soil fines and some "bulk" organic matter.  A
large part of the total organic chemical contamination is due to
sorption and physical pore interactions with the fine particle
fractions [clay minerals and humic substances] of whole soil.

A second section describes microbial degradation experiments.
Systems and procedures for microbial reactions were designed and
implemented to accomodate the properties and behavior of target
substrates.  Both shake flask and fermentation reactions are being
carried out on slurries of soil fines.  Low molecular weight
polynuclear and chlorinated aromatic hydrocarbons are readily
biodegraded.
                                  312
-------
INTRODUCTION

Distillation bottoms and sludges from benzene-toluene-xylenes [BTX]
production were impounded for several decades.  The production process
consisted of the catalytic cracking of naphtha, in the presence of fuming
sulfuric acid, and distillation.  Therefore, lagoon contents include
naphtha-related compounds, distillate residues and compounds resulting from
reactions of these species and sulfuric acid.  Possible contaminants
include, but are not limited to, simple aromatic species, polynuclear
aromatic hydrocarbons [PAHs], phthalates, as well as sulfonated derivatives
of these compounds.  Many of these species have slight solubility in water
and/or an affinity for some of the constituents of soil and have migrated
into and through the soil immediately adjacent to the lagoon.

During this study, soil samples were obtained from the containment area
surrounding the lagoon. The impoundment has been designated a CERCLA-NPL
site; it exceeds ten acres in extent and contains an estimated 100,000 cubic
yards of residues.  The contents of the lagoon have separated into several
distinct layers that include, in bottom-to-top sequence, a solid mixture of
organic and inorganic substances, a tar-like layer, a layer of viscous
organic matter, and a floating aqueous layer.

ANALYTICAL APPROACH

Initially, soil samples are mixed and homogenized.  The resulting material
is air-dried and sieved through a 3-cm brass screen to remove debris, rocks
and gravel; this procedure also breaks up macro-agglomerates.  A second
sieving, with a 5-mm screen, improves homogeneity, enhances mixing and helps
toward analytical reproducability.  Direct solvent extractions of
homogenized, contaminated soil utilize methanol, cyclohexane, or methylene
chloride.  Methanol has relatively high polarity, cyclohexane is a model
cyclic compound, and methylene chloride is a moderately polar, volatile
compound with broad solvent capabilities.

Gas chromatography [GC] is used to identify and quantify compounds in soil
extracts.  USEPA Test Methods 602 and 610 are employed in these analyses.
Method 602 is used to assay aromatic species in GC column effluent with a
photoionization detector [PID] in series with an electrolytic conductivity
detector [ELCD].  Method 610 is employed to detect PAHs and phthalates,
utilizing a flame ionization detector [FID].  Standard solutions are assayed
in sequence with solvent solutions to match retention times for compound
identifications.  Peak areas are used to construct standard curves,and
provide a basis for determination of contaminant concentrations.  Compound
concentrations are calculated from both PID and FID output to check
analytical consistancy.

Experimental Methods

Soil contaminant levels were initially estimated to fall between 2 and 5 %
on a dry weight basis.  Direct solvent extractions are carried on varying
masses of soil with the goal of limiting contaminant concentrations to
about 100 mg/L in extract solutions.  This target concentration was adopted
                                    313
-------
to  avoid  overloading  the  detectors.  Soil masses varying from 0.15 to 0.38 g
are extracted with  a  fixed  volume of solvent.

Duplicate amber  serum bottles, each with label, septum and aluminum cap, are
weighed with a Mettler PE3600 balance.  Bottles are 100 ml in volume  Soil
is  added  until target weights are attained; approximately 50 ml of methanol,
cyclohexane or methylene  chloride is added to both bottles.  Bottles are
sealed with the  septa [Teflon-coated neoprene] and reweighed.  Experimental
errors include the  small  discrepancies in obtaining target soil masses and
measuring solvent volumes.   In general, these are accounted for in
concentration calculations.  A second form of the experiment is carried out
to  facilitate compound identification.  It is the same in all respects,
except that 20 g of soil  are added; since quantification is not desired,
duplicates are not  performed.

Extraction vessels  are shaken for approximately one hour.  This time was
found to  be adequate  in earlier studies; however, it assumes that only
readily reversible, high-rate sorption processes are involved.  Higher
energy binding processes  and sorbate trapped by capillary forces would not
participate in such short-term partitioning.  Extract solutions are filtered
through 0.2-um MSI  Cameo  II 25-mm disposable syringe filters, into duplicate
5-mL vials.  Samples  are  stored at 4°C, to minimize volatilization losses,
and enclosed to exclude light and avoid photolytic chemical reactions.

Method 602 utilizes a 1.8-m long by 2-mm ID stainless steel GC column packed
with 100/200 mesh Supelcoport, coated with 5 % SP-1200 and 1.75 % Bentonite-
34.  Oven temperature is  held at 50°C for 2 min; a 6°C ramp takes the oven
to  90° for a final  period of 23 min.  The PID has detection limits of 1 to
10  pg for unsaturated carbon bonds found in aromatic compounds.  The ELCD
has detection limits  of 0.1 to 1 pg when used in the detection of
chlorinated compounds or  to verify PID results.  Sample size is 2 uL.
Method 610 is applicable  to PAHs and phthalates.  It utilizes a 1.8-m long
by  2-mni ID glass GC column packed with 100/200 mesh Chromosorb W-AW-DCMS,
coated with 3 % OV-17.  Oven temperature is held at 100°C for 4 min; a 8°C
ramp takes the oven to a  final temperature of 280°C.  The FID has detection
limits of 10 to  100 pg.   Sample size is 5 uL.

Chemical  Oxygen Demand [COD] is determined for some contaminated soil
samples.  This procedure  is identical to that described in Standard
Methods".  The COD  has some value for comparison with carbon in identified
species,  to estimate  extraction and identification efficiencies.

Analytical Results

Extracts  generated  in  Experiment 12888 were distinctly different in color.
After filtration, the  cyclohexane extract was translucent orange, methylene
chloride  gave an opaque brown liquid, and the methanol solution was clear
and tan.  This appeared to be evidence for susbtantial variation in
extraction efficiency.  In addition, the high polarity of methanol leads to
destruction of soil aggregates; thus, methanol is capable of extracting
contaminants held in  micro- and macro-pores by capillarity and interfacial
tension.  Solvents can be compared on the basis of the mass of naphthalene
                                     314
-------
extracted.  No naphthalene was extracted by cyclohexane; methanol and
methylene chloride extracted 1,089 and 1,403 mg/kg dry soil, respectively.
Thus, for naphthalene, solvent power varied considerably.

The PID sees cyclohexane and impurities in methylene chloride.  The ELCD
detects chlorinated compounds and is overloaded by methylene chloride
solutions.  The consequence is chromatogram baseline fluctuations and large
peak area integration inaccuracies.  Methanol was the only solvent suited to
Method 602.  Methylene chloride was used in conjunction with Method 610.
Eight major organic compounds were identified and quantified in methanol
solution.  In order of decreasing concentration [mg/kg], they are:
naphthalene - 1,090; 1,2-dichlorobenzene - 360; toluene - 150; xylene
isomers - 145; benzene - 113; and, ethyl benzene - 28.  The eight species
account for 64 % of the total peak area of chromatographic responses.

Benzene, toluene and the xylenes are primary products of naphtha
distillation.  GC residence times for standard solutions and extract
solutions, with PID detection, varied less than 0.004 sec for this group.
Naphthalene is also a major component of naphtha; residence times differed
by 0.004 sec.  Ethylbenzene [EB] is formed by catalytic reaction of benzene
with ethylene, an olefin found in industrial naphtha. In BTX production,
sulfuric acid is the catalyst.  EB residence times varied by 0.008 sec.

The appearance of 1,2-dichlorobenzene was signaled by the PID and verified
by the ELCD, a halogen-specific detector.  Chlorinated compounds are not
normally found in naphtha nor are they produced by sulfuric acid catalysis.
The presence of this compound may indicate disposal to the impoundment from
another manufacturing source or a spill clean-up activity.  Two substantial,
unidentifiable peaks were encountered with Method 610; neither was observed
with Method 602.  These peaks correspond to compounds that are believed to
be sulfonated aromatic hydrocarbons; operating temperatures for Method 602
preclude elution of such higher boiling species.  Naphthalene sulfonic acids
represent compounds of higher molecular weight and boiling point, requiring
increased GC oven temperatures  [Method 610] and extended residence times.
Standard  solutions
[see Table  la]  and
solvent  [see Table
Method 610  for  the
Similarly,  Figures
purgable  aromatic
is  the naphthalene
 included seven purgable aromatic compounds in methanol
 fifteen PAHs in a 50:50 methanol:methylene chloride mixed
 lb].  Figures 1 and 2 are gas chromatograms obtained by
 PAH standard solution and methanol  extract, respectively.
 3 and 4 are GC outputs obtained by Method 602 for the
standard and the methanol extract, respectively.  Figure  5
 case run with a modified version of Method 602.
 COD  has  been  measured  for  several  soil  samples.  This  assay  is  used  to
 determine  the total  oxygen required  to  fully  oxidize all  reduced  species;  it
 does not distinguish between  contaminant  organic carbon  and  hydrogen, soil
 organic  matter,  and  metals in reduced or  partially  oxidized  states.  The COD
 of sieved,  air-dried,  unextracted  soil  corresponds  to  about  50  mg C/kg  soil.
 In comparison,  the total organic carbon [TOC]  associated with the eight
 quantified contaminant species cited above  is  approximately  1,6 mg C/kg soil
 or 3.2 % of the COD  carbon equivalent.  This  is not a  reflection  of
 extraction efficiency. However, CODs run  on soil samples taken  at points
                                    315
-------
remote to the disposal lagoon average close to 50 mg C/kg soil.  Thus,
extraction efficiency with methanol is probably relatively good.
                              Table la

              Purgable Aromatic Standard Mixture [602-M]
                            (in Methanol)
              Compound

              Benzene
              Toluene
              Ethyl benzene
              Chlorobenzene
              1,2-Di chlorobenzene
              1,3-Di chlorobenzene
              1,4-Di chlorobenzene
Concentration (mg/L)

      2000
      2000
      2000
      2000
      2000
      2000
      2000
                              Table Ib

      Polynuclear Aromatic Hydrocarbon Standard Mixture  [610-M]
                fin 50:50 Methanol:Methv1ene Chloride)
              Compound

              Acenaphthene
              Fluoranthene
              Naphthalene
              Benzo(a)anthracene
              Benzo(a)pyrene
              Benzo(b)f1uoranthene
              Benzo(k)f1uoranthene
              Chrysene (93 %)
              Acenaphthylene
              Anthracene
              Benzo(g,h,i)pyrene
              Fluorene
              Phenanthrene
              Dibenzo(a,h)anthracene
              Indeno(l,2,3-cd)pyrene
              Pyrene
Concentration (mq/L)

      1000
       200
      1000
       100
       100
       200
       100
       100
      2000
       100
       200
       200
       100
       200
       100
       100
                                    316
-------
                         Method 610
figure i
            6.2S
                                                                   13  7.62
                                     ftuorene.
                        12.72. Phenanthrene and
          !
                                16.86
                      16.58
20-
             18.48




             19.4S
                            . 17 8en20 (a)anthracene and
            22.42
                                        ffTamb)fiuoranth&ne4i\d Senze
-------
 5--
10-
15-
20-
                                      Method 610

                                             2
               ,  16. 69
               )  17.03
               P  17,53
                  J&.20
                    .^*
                  >Z1*ft»

                      •8«2^
                   23. 3Z
                                          318
-------
   seoo-
 O
 -P
 U
   «<">-
1
   2JOO-
    JOO
   9200
                   Chromatograph  for Standard
                 Solution of Purgable Aromatics

                          Method 602

                           Figure 3
                                                    Ch lorobenzene
                                                    1 , 4-dichlorobenzene
                                                    1 ,3 -diehlorobenzene
                                                    1 ,2-dichlorobenzene
   6ZOO-
 S-i
 O

 O
 JJ

Pi

Q
M
PH
   zeno-
    zon
                                Benzene
                                Toluene
                                Ethyl benzene
                                Chlorobenzene
                                              -1,4-dichlorobenzene
                                              -1,3-dichlorobenzene
                "T"
                 T
                        'V _______
                           18
                                                       1 , 2-dichlorobenzene
                                    'T'
T"
"I"
 3£
T"
                                             319
-------
C37S-
                                    Hethanol  Extract

                                      Method 602

                                       Figure  4
 tn
 O
 4J
 O
 \.
                          ~r
                           10
                                          .-'~\--.
                                                      Naplithalene —-'"
                                                      1 , 2-dichlorobenzene
                                                          25

                                                                           I
                                                                           55
                                            320
-------
  3SBII -i
o rsn-
4J
o
0)

0)
o  .,
                      Naphthalene by Modified

                            Method 602


                             Figure 5
                                                             Naphthalene-
                                                              	j
   B'iWI-l
 U
'QJ
 4J
 (1)
                           .f


                           1C
"1	"
15
                                             	"1	
                                                  20


                                             Standard
r
T"
 4.0
                                                           Naphthalene
       :i
                                                                A.
                 5
                            14
                                      ._,	r—=_n.
                                       IS          20    •   •   ZS
                                         Methanol Extract
                                                                       T
                                                                       30
T-
                                              321
-------
 MICROBIAL  APPROACH

 The  solubilities  of PAHs  in  aqueous  media  decreases with  increasing
 molecular  weight;  for  four rings  or  more,  saturation  concentrations  are very
 low  and difficult  to measure.   Within  this context, rates of  aerobic
 metabolism of  PAHs are insignificant.   In  addition, this class of aromatic
 compounds  has  a high affinity  for several  constituents of natural soils,
 i.e., soil  organic matter [humic  substances] and clay minerals.  PAHs are
 lipophilic,  i.e.,  prefer  sorptive association and/or  dissolution in  organic
 rather than  hydrophilic phases.   In  addition, many multi-ring molecules can
 migrate into stable positions  inside clay  mineral structures, either
 diffusing  between  laminae in mica-like  structures or  entering the crystal
 lattice directly.   The latter  process  is a form of clathration.  These
 preferred  soil components are,  in general,  the smaller particle fractions of
 the  soil system.   Thus, it is  possible  to  introduce PAHs into a microbial
 systems in slurry  form, with substrate(s)  bound to particulate matter.

 A slurry form  of bioreactor has a number of possible  thermodynamic phases
 present, including:  aqueous medium with dissolved substrate and nutrients,
 substrate  bound to soil particles, suspended single and clustered cells,
 substrate  emulsions  and colloids,  and cells attached  to the exterior and
 macro-pore surfaces  of dispersed  particles.  Sampling and analytical
 methodology  are difficult throughout the design and implementation of
 biodegradation experiments.

 An earlier Section dealt  with  the problems  of accurate chemical contaminant
 identification and quantification for soils prior to  chemical or biochemical
 reaction.  Often tightly-bound  PAHs must be extracted from a soil mass, with
 uncertain  efficiency or recovery.  Method  verification is critical.  Extract
 solution is  assayed  by GC or High Performance Liquid  Chromatography  [HPLC].
 Multiple phases must be sampled and extracted, during and after biochemical
 transformations.   The  possible  appearance  of intermediate or final
 metabolites, i.e.,  incomplete mineralization, adds to the complexity of mass
 balances for substrate species.   COD is a  useful tool for quasi-continuous
 monitoring of  the  progress of  slurry-type  bioreactor  systems.  In well-
 aerated aqueous systems,  metals are not a  serious factor. However, variable
 cell mass and  natural  soil organic matter  severely limit this approach.

 Before slurry  reaction can be undertaken, whole soil must be reduced to
 several particle size  fractions.   Contaminant PAHs favor finer particles,
thus, fractionation  is a  useful way of concentrating these compounds prior
to degradation experiments.   A reproducible method for soil  classification
has been developed and has been demonstrated to lead to the desired
concentration of target substrates, as follows.   Whole soil  is separated
 into three phases: a tar-like organic phase, a coarse sandy phase and an
aqueous suspension of  soil fines.  The latter is suited to slurry reactor
experiments.  Experiments consist of small-scale studies in shake flasks,
performed in matrix  formats,  and  fermentation studies in reaction vessels  of
larger volume.
                                    322
-------
Background

The rates of microbial assimilation of PAHs have been demonstrated to be
functions of solubility, molecular weight, number of six-member rings,  ,
degree and type of substitution, as well as environmental conditions, such
as temperature, pH and oxygen concentration.  The solubility of
unsubstituted PAHs,in water, drops sharply as the number of rings increases.
It rapidly diminishes to levels that are too low to support significant
biological activity; see Table 2 for data.  Compounds of six or more rings
have vanishing solubility in water.

The number and type of substituents on or in a PAH molecule have a marked
influence on solubility.  The solubilities of phenols, nitrogen
heterocyclics, polynuclear polyols, sulfonates and other mono- and poly-
substituted PAHs are often significantly higher than the basic hydrocarbons.
Therefore, substituted compounds are more likely to be observed as solutes
in contaminated groundwater.  Also, surfactants increase PAH solubility.
However, these compounds complex or "react" with the high molecular weight
polynuclear species to create a composite hydrophilic exterior.  The result
is either a stable emulsion, colloidal suspension or micro-dispersion; it
cannot properly be classed as dissolution in the thermodynamic sense of a
homogeneous liquid phase.  Sodium  laurylsulfate increases the solubility of
2- to 7-ring PAHs by  2 orders-of-magnitude or more.

Biodegradation of 2-  to 3-ring  PAHs by pure microbial cultures has been
demonstrated;  naphthalene, phenanthrene and anthracene have been shown to be
assimilated quantitatively.  Higher molecular weight compounds, i.e.,
benzo(a)anthracene  and  benzo(a)pyrene, can  be degraded to simpler
intermediates  in the  presence of  supplementary carbon sources or
cometabolites,  i.e.,  biphenyl and  succinate.

Bacteria concentrate, grow,  and form  bioslimes  in  aqueous boundary layers at
liquid/liquid  and  liquid/solid  interfaces.  Organic  cosolvents can transport
PAHs  to  such  interfaces and  increase  the  rate of biodegradation.  There  is
no information to  support  microbial metabolism  of  solid  PAHs;  similarly,
there is little  data  on the  bioreaction of PAHs  sorbed on nonreacting
 surfaces.
       Compound
                    Table 2. Solubilities of PAHs in Hater
Mol. Wt.
Naphthalene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)pyrene
Dibenzanthracene
128
154
166
178
178
202
202
252
278
Solubility (ug/L)

     31,700
      3,200
      2,000
      1,300
         73
        260
        140
          4
# of Rings

    2
    3
    3
    3
    3
    4
    4
    5
    5
                                     323
-------
 Experimental

 i) Apparatus

 Aerobic biodegradation  experiments are carried out in 60-mL Ehrlenmeyer
 flasks, on a laboratory shake device, or in 3-L  [working volume]
 fermentation vessels.   The shaker studies are arranged in matrix format with
 the following common composition: 30 ml soil slurry, 20 ml inoculum and 10
 ml nutrient medium  [see Table 3].  The composition of the inoculum is varied
 to provide a basis  for  the evaluation of substrate volatility losses,
 reactor surface wetting and sorption onto biomass as mechanisms of substrate
 disappearance.  The 20  ml "inoculum" is live culture; in sterile controls,
 the 20 ml of seed is replaced by autoclaved culture or deionized water.
 Controls are intended to illustrate the extent to which volatilization and
 inorganic surface sorption influence sub-strate fate.  Given that biomass
 rendered unviable by autoclaving retains substantial sorption capacity,
 autoclaved culture  is designed to investigate this loss pathway.  Also, this
 control provides a  zero-time or baseline measurement for carbon or oxygen'
 demand.  The pH of  the  composite solution is adjusted to 7.15 by addition of
 a mixture of solid  potassium monobasic and dibasic phosphates.

 Fermentation studies are carried out with a working volume of 3 L prepared
 in the same ratio of 3:2:1 for soil  slurry:inoculum:nutrient medium as in
 the shake flask studies.  The reactor is sparged with air at the rate of 5
 to 6 L/min; it is stirred at 300 rpm.  The pH is maintained at 7.15 by
 periodic additions  of 0.25M sodium hydroxide regulated by a pH controller.
 Samples are taken at 24-hr intervals, to monitor the course of reaction.
Separate studies utilizing deionized water are used to evaluate
volatilization losses.
                     Table 3 - Medium Composition

                     Constituent    Concentration [ma/Li
                     (NH4)2S04

                     MgS04.7H20

                     FeCl3.6H20

                     MnS04.H20

                     CaClo
1,500

  100

  0.5

   10

  7.5
                                    324
-------
  ii) Soil Slurry

Soil slurry is prepared by a sequence of homogenization, extraction, and
fractionation steps.  This procedure is designed to create a suspension of
soil fines that displays minimum variation from batch-to-batch.   Whole soil
samples are homogenized by passing the air-dried material through a 5-mm
screen, quartering the resulting solids cone through the apex, segregating
the quarters, and sieving each quarter to form a new cone. Soil is sieved
three times.  A prescreening with a 3-cm sieve removes rocks and
miscellaneous debris and, also, serves to break-up larger clumps of packed
soil.

Homogenized soil [84 g on a dry basis] is extracted with 350 ml of water at
pH 7.  Extraction separates the contaminated soil into three phases: a tar-
like [smell, sticky, viscous, etc.] organic phase corresponding to 0.65 -
0,75 % of the initial dry mass; a mixture of larger, heavier particles
[sand]; and, an aqueous supension [slurry] of fine soil particles.  The
aqueous suspension  is separated from the settleable solids by screening
through a 10-micron sieve; filtered solids are washed with 650 ml of water
to dilute the filtrate slurry to 1 liter and the final concentration.  The
fractionation procedure has been found to retain approximately 65 % of the
initial dry mass of soil on the sieve.  The final slurry of fine particles
is stable and does  not show any evidence of settling under experimental
conditions.

iii) Inoculum

The  aerobic  inoculum is obtained as waste sludge from the Somerset/Raritan
Valley Sewage Authority.

iv)  Nutrient Medium

The  nutrient medium is prepared from a conventional recipe for aerobic
cultures  and does not employ a primary or supplementary carbon source; see
Table 3.  The soil  slurry supports microbial activity without the use of a
co-substrate; none  is added.

v) Analytical

The  progress of substrate conversion is monitored by COD and GC analyses.
Samples are  prepared for the latter by extraction with methylene chloride at
8 ml of solvent for 20 ml of slurry.  The remainder of the assay is carried
out  in accordance with USEPA Method 610.  A 1.8-m long by 2-mm ID glass
column is packed with a  stationary phase consisting of  100/200 mesh
Chromosorb  W-AW-DCMS coated with 3 % OV-17.  Oven temperature  is held at
100°C  for 4 min; a  8°C/min ramp  increases the temperature to  280°C.  An FID
is  used to  determine residence times and peak areas.  Gas pressures are 70,
40  and 65 psi for nitrogen,  hydrogen and dry air, respectively.  A  5-uL
sample is injected  into  the  GC.
                                     325
-------
   Results

   Table 4  summarizes  moisture contents  and  COD  analyses  for  whole  soil
   and  the  fractions generated by repeated sieving,  fractionation and
   extraction.   COD determinations are referred  to  1  kg of  air-dried whole
   soil. The COD balances  vary about 15 % for a typical  set;  this  is  a
   consequence  of an error  of at least ± 5 % in  this  measurement.   Slurry,
   diluted  with wash water,  has a COD of 7.1 g 02/L.
                 Table  4  -  Moisture  Content  and  COD

        Sample  Type            Composition  F%1      COD  Fa Oo/kol
        Whole  Soil
           Water
           Large Particles
           Soil  Fines
        Tar-like Residue
        SIurry
16
53
31
 89.6
  0
trace

 19.4
 84.5
When an active microbial  inoculum  is combined with a slurry of soil fines, a
lag period of approximately 6 hr is observed.  Acid  production in shake
flasks and fermentations, and carbon dioxide generation by fermentations,
are not observed until after the lag phase.  There is no loss of COD, as
might accompany volatilization or  sorption.  It is assumed that COD
attributable to biomass remains unchanged during the experiment, i.e.,
growth is negligible.  The reduction of COD in the flasks inoculated with
live cultures is indirect evidence for substrate mineralization, as opposed
to physical uptake  (sorption) by the biomass.  The results of a shake flask
matrix study are summarized in Table 5.  Flasks contain a working volume of
60 ml; total reaction time is 80 hr.  The slurry has an initial soil fines
concentration of 30 g/L and a COD of 7.0 g 02/L; inoculum COD is 4.75 g
02/L.

Table 6 sumarizes results of a larger scale fermentation study.  Reactor
working volume is 3 L and reaction time is 68 hr.   The slurry has an initial
soil fines concentration of 45 g/L and a COD of 7.5  g 02/L;  as in the shake
flask illustration, inoculum COD is 4.75  g 02/L.   Initial  reactor COD was
calculated to be 15.9 g 02; an experimental determination gave 18.9 g 02.
The difference is probably measurement error, due to the several  phases
present,  i.e., cells', contaminated fines, suspended tarry material  and
dissolved PAHs.   A volatilization loss study was carried out  with the
fermentor.  The COD of air sparged slurry did not change in 4 days; it
remained at 3.1  g 02/L.
                                    326
-------
                 Table 5 - Typical  Shake Flask Matrix Study
   Initial  COD [mg 02]
   Final  COD   [mg 02]

   COD Change [%]
   COD Change [%]
   {corrected for  inoculum)

   Final  Dissolved TOC
           [mg C]
           [mg C/L]
   COD Equivalent  [mg 023
      Live
    Inoculum

      310
      204

      -34
      -50
        8
      131
       18
   Autoclaved
    Inoculum

      310
      318

      nil
      nil
        8
      131
       18
      No
   Inoculum

      210
      240

      +14
      +14
        8
      133
       18
                   Table 6 - Fermentation Study
   COD [mg 02]
21 hrs

 13200
   COD Change [%]          30
   COD Change [%]          39
   (corrected for inoculum)

   Final Dissolved TOC
           [mg C]         318
           [mg C/L]       106

   COD Equivalent [mg 02] 848
46 hrs

 10200

    46
    61
                 315
                 105

                 840
68 hrs

 6900

   63
   84
                288
                 96

                768
  Figures 6 and 7 are gas chromatographs for the fermentation liquor
  after 21 hrs and at the end of  the experiment.  The characteristics of
the soil slurry without medium or culture added are described in Figure 8.
The contents of the fermentor and the original soil slurry were extracted
with methylene chloride; the volume ratio was 5:2 for aqueous
suspensionrsolvent.  Figures 6 and 7 show definite declines in the number
and size of peaks, especially those corresponding to low molecular weight
PAHs.
                                    327
-------
Figure 6:   21  Hours
                                                                (nt.n
                                                                til?.91
                                                                i •••M. 31
                                                               !«I*.*f   MM
                                                                             7.211
                                                                             t.flZ
                                                                             j.irr
Figure  7:   68 Hours
                                                     II.Jt
                                                     »•.**
                                                     '*•"
               *n« •-.

                1.972

               "tt'.t J
                •'.I »
               I?.9 9
                J.r 1
                :.i i
                •I.5-.I
                i.in
                3..-JI
 ^Figure  8:   Original  Slurry
                                                    If.11

                                                     1.1?
                                                     ]!'«•
 a»ii.ti  v*    /.jjj
ICIH.^f  vv    11.59:

ll-f.-|l«  vv    Il'll9

 n*i,:i  w    i!*,
-------
CONCLUSIONS

1) Bench-scale shake flask studies are performed with slurries of soil
   fines and mixed microbial seed.  COD, corrected for the presence of
   the inoculum growing at a trivial rate, is reduced by 50 % in
   approximately 80 hours.  Similarly, TOC is measured on settled [filtered]
   aqueous phase and remains low throughout.  The latter assays are a
   reflection of limited hydrocarbon solubilities.

2) Larger-scale fermentations are carried out in 3-liter, stirred, air-
   sparged reactors.  Inoculum and nutrient medium are mixed with slurry.
   Biodegradation is monitored by assays on samples of aqueous dispersions
   and measurement of carbon dioxide generation rates.  COD reductions
   exceed 84 % in 68 hours.

3) Whole soil is separated into size fractions to characterize contaminant
   distribution by soil constituent type and particle size.  It is possible
   to separate whole soil into larger, settleable particles [primarily sand
   and silt], slurried fines, a clarifiable aqueous phase and a bulk organic
   phase.  Recovery of initial whole soil COD in the slurry and a tar-like
   organic phase is nearly quantitative.

4) Analytical techniques have been developed and demonstrated with whole
   soil, soil fractions, slurries of fines and filtered liquids.  These
   techniques are essential to the identification of contaminant species and
   quantification of individual and total contaminant concentrations.
   Assays are necessary to define initial and intermediate conditions and to
   demonstrate that contaminant destruction [mineralization] by microbial
   reaction  is effective and approaches completeness.
                                   329
-------
                  FATE AND EFFECTS OF RGRA AND CERCLA TOXICS
            IN ANAEROBIC DIGESTION OF PRIMARY AND SECONDARY SLUDGE

                 by: Richard A. Dobbs,  2Rakesh Govind,  2Peter A. Flaherty,
                     3Thomas  L.  Crawford,  2Kaniz Siddiqui, ^arry M. Austern

                      xRisk Reduction Engineering Laboratory, United States
                      Environmental Protection Agency, Cincinnati, OH  45268

                     Department of Chemical and Nuclear Engineering,
                      University of Cincinnati, Cincinnati, OH  45221

                     3Department of Civil and Environmental Engineering,
                      University of Cincinnati, Cincinnati, OH  45221
                                     ABSTRACT
    Municipal wastewaters have been shown to contain toxic organics,  many of
which are anthropogenic.  Sorption onto solids is one of the primary means of
removal of these chemicals.  This study investigates the steady-state fate and
effects of organic priority pollutants sorbed on primary and secondary sludge
during subsequent treatment in anaerobic digesters.  The investigation
consisted of two separate studies of pollutants:  volatile organics chosen from
the RCRA list, and semi-volatile organics from the CERCLA list.  Simulating
typical digestion processes, three bench-scale, semi-continuous, complete-mix
units with a total volume of 40 liters each were operated with a solids
retention time of thirty days at 35°C.  Primary and secondary sludge were
combined in equal weight ratios prior to feeding to the digesters.
Conventional operating parameters were monitored for the one control and two
organic-spiked digesters to assess differences in performance.
    The fates of seven of the volatile compounds which showed consistent
behavior, including several chlorinated aliphatics and ethylbenzene,  were
determined by purge and trap gas chromatographic analyses.   Steady-state fates
of twelve of the semi-volatile organics, which included dichlorobenzenes,
phthalates, p-cresol and lindane, were determined by gas chromatographic/mass
spectroscopic analyses.
                                     330
-------
                                 INTRODUCTION
    Municipal wastewaters have been found to contain a number of toxic
organics (1).  In response, effluent regulations have also restricted the
particular pollutant discharge levels, thus increasing the interest in the
fate of specific toxic organic compounds in wastewater treatment processes.
One area which presents potential problems is the handling and disposal of
sludge.  Sorption of toxics onto solids is one of the fundamental removal
mechanisms of pollutants from wastewater.  Previous investigations have shown
that organic compounds tend to accumulate in sludge at concentrations several
orders of magnitude higher than the influent concentration (2,3).  One study
reported some organic component concentrations in sludge greater than 10,000
times that found in the aqueous phase (4).  A dilute toxics-laden wastewater
could potentially generate hazardous primary and/or secondary sludge, creating
more difficult and costly final disposal options such as landfilling or
incineration as a hazardous waste.
    Anaerobic digestion is often utilized in the stabilization and reduction
of wastewater treatment sludge.  A survey of ninety-eight municipal wastewater
treatment plants in the United States found that seventy-three incorporated
anaerobic digestion as a means of sludge stabilization and volume reduction
(5).  Anaerobic treatment systems offer several advantages over aerobic
treatment in terms of lower energy requirements, higher process loading, and
the potential energy recovery in the form of methane gas production.
Anaerobic processes may also control the discharge of volatile compounds to
the atmosphere via biotransformation or capture of the volatilized material
with the off gas.
    Research on the fate of toxic organic material in aerobic systems has been
extensive,  yet relatively little work has been conducted to determine the
fates of organic toxics in typical anaerobic processes.   Investigations
generally have been conducted with bench or pilot scale processes,  serum
bottles with digester sludge, and serum bottle studies with specific or
enriched anaerobic cultures.  Most studies have been conducted by spiking the
compound or compounds directly into the influent process stream or serum
bottle.  This investigation, however, attempts to better typify actual
processes by determining the fate and effects of compounds already sorbed onto
the digester feed sludge.  The goal of this study is to determine the steady-
state fates of several volatile and semi-volatile organic priority pollutants
in the anaerobic digestion of primary and secondary aerobic sludge, and to
determine the effects of these toxic sludge on the digestion process.

                              EXPERIMENTAL METHOD
AEROBIC PILOT PLANT

    The United States Environmental Protection Agency (USEPA) has evaluated
the removal and fate of selected toxic organic compounds during primary-
activated sludge treatment of municipal wastewater (6).   Volatile and semi-
volatile organics were investigated under separate studies with the same
aerobic treatment system. Two parallel pilot plants were studied: one
continuously spiked to allow for biomass acclimation, and the other
                                      331
-------
Intermittently spiked to simulate unacclimated operation.  The treatment
systems chosen as a representative municipal wastewater treatment plant design
consisted of primary clarification followed by conventional plug-flow
activated sludge treatment.  Each system was operated with a sludge retention
time of 4.0 ± 0.3 days over a period of six months.  Screened and degritted
raw wastewater from the Cincinnati Mill Creek Treatment Plant was used as feed
for the aerobic systems.  Mill Creek is a combined residential/industrial
treatment facility.
    Volatile and semi-volatile toxics studied were selected from the RCRA and
CERCLA lists of organic priority pollutants, respectively.  A miscible mixture
of the compounds listed in Table 1 (volatiles) and Table 2 (semi-volatiles)
was used to spike the raw wastewater influent to the pilot plant at 0.25 mg/L
of each compound for both studies.  The primary and secondary sludge from the
continuously spiked aerobic system were used as feed to the bench-scale      T
anaerobic digesters.

ANAEROBIC BENCH-SCALE DIGESTERS

    Three bench-scale digesters (see Figure 1) were constructed from
Plexiglass cylinders with quarter-inch thick walls.  Influent and effluent
ports were constructed from polyvinyl chloride pipe.  One-inch valves were
used at each port location to provide for sludge feeding and digester mixed-
liquor withdrawal. *A Plexiglass cup with approximate capacity of one liter
was threaded at the top of the inlet pipe to act as a funnel during digester
feeding.  A long thermistor probe was inserted through the top endplate to

        TABLE  1.   VOLATILE ORGANIC  COMPOUNDS  IN RAW WASTEWATER  SPIKE
      Acetone
      Atrazine
      Carbon tetrachloride
      Chloroform
      Cyclohexanone
      1,2-dichloroethane
      1,2-Dichloropropane
      2,4-Dimethylphenol
      2,4-Dinitrophenol  •
      Ethylbenzene
      Furfural
Methylene Chloride
Methyl ethyl ketone
Methyl isobutyl ketone
4-Ni tropheno1
Phenol
Tetrachloroethylene
Tetrahydrofuran
Toluene
1,1,1-Trichloroethane
1,1,2-Trichloroethane
Trichloroethylene
       TABLE  2.   SEMI-VOLATILE ORGANIC  COMPOUNDS IN RAW WASTEWATER SPIKE
      1,2-Dichlorobenzene
      1,3-Dichlorobenzene
      1,4-Dichlorobenzene
      1,2,4-Trichlorobenzene
      Nitrobenzene
      1,3-Dinitrobenzene
      2,6-Dinitrotoluene
      E-Cresol
      Hexachloroethane
      Hexachloro-1,3-butadiene
N-nitrosodiphenyl amine
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Butyl benzyl phthalate
(Bis) 2-ethylhexyl phthalate
Napthalene
4-Chloroaniline
Lindane
Dieldrin
                                      332
-------
measure digester temperature.   All fittings were sealed with silicone caulking
and each unit was pressure  tested to insure air tightness.  The gas
delivery/mixing system was  constructed from one-eighth inch stainless steel
tubing and stainless  steel  fittings.   The H-shaped sparger was placed one-half
inch above the endplate  and levelled.
    On the outside of the digester,  a two-inch long,  stainless steel tube was
inserted through the  upper  endplate to act as the digester gas outlet port.
Butyl rubber tubing was  inserted over the inlet and outlet stainless steel
tubes with hoseclamps.   The tubing ran from the gas recycle outlet port to a
gas pump and from the gas pump  to the gas recycle inlet port.  A glass tee was
inserted between the  gas recycle outlet port and the pump to allow for the
escape of digester off-gases to a Tedlar gas collection bag.  A stainless
steel tee with a septum  and cap was placed in line with the gas collection bag
for withdrawal of gas samples for analysis.
                BUTYL RUBBER
                STOPPER
                TEMPERATURE
                PROBE

                STAINLESS
                STEEL TUBING
                 SPARGER
                 BUTYL
                 RUBBER SEAL
                 (TYPICAL)

                 r BRASS
                 SVAGELDK
                 VALVE
                 (TYPICAL)
                                                        GAS RECYCLE
                                                        PUMP
                                                        TO GAS
                                                        COLLECTION
                                                        BAG
                                          61 cm
SIDE
VIEW
                                        EFFLUENT
                                          PORT
                          30.5 cm —
Figure 1.  Schematic of anaerobic digester for volatile
            and semi-volatile studies.
DIGESTER OPERATION

    During startup  for both studies,  the three 40-liter digesters were seeded
with 30 liters of anaerobic biomass from the Mill Creek plant.  The digesters
were cleaned out and reseeded between volatile and semi-volatile experiments.
Initially, all three digesters were fed one liter of unspiked primary and
secondary sludge for three  or four days to achieve start-up.  After start-up,
digester 1 continued to  receive unspiked sludge from the Mill Creek plant,
while digesters 2 and 3  were fed sludge from the continuously-spiked aerobic
                                     333
-------
 pilot plant.  The feed rate of one liter per day resulted in a 30 day solids
 retention time.  The digesters were maintained at 35 + 1°C in an environmental
 room.  Off-gases were recirculated through the sparger for 15 minutes every
 six hours to keep the contents of the digesters well mixed.   Conventional
 parameter evaluations were performed biweekly to monitor digester biomass
 acclimation and the approach of steady state.  Steady state operation was
 achieved after seven weeks for the volatiles study and five weeks for the
 semi-volatiles study, after which analysis for the spiked organics began.
     The daily draw and feed proceeded after digester biomass was mixed
 sufficiently uniformly distribute the cpntents.  One liter of mixed liquor was
 purged from the effluent port and then, reintroduced through the feed port to
 insure representative sampling.  A second one liter sample was withdrawn and
 used for the analysis of conventional parameters and organic compounds.   One
 liter of feed sludge was then added to the digester through the influent port.

 FEED SLUDGE

     For the volatile organics study,  unspiked feed sludge for digester 1 were
 gathered weekly from the Mill Creek treatment plant to act as control.
 Primary sludge was drawn from the gravity thickeners and diluted to 4% total
 solids (if necessary) with clarified primary influent.   Secondary sludge was
 collected from the return activated sludge line, thickened by means of a
 perforated bowl centrifuge and diluted with the resulting centrate to 4% total
 solids.  Both unspiked sludge were then combined in equal volumes and the
 resultant feed sludge was stored at 4°C in a large carboy until needed for
 digester feeding.
     Volatile spiked feed sludge for digesters 2 and 3 were gathered daily to
 minimize component loss.   The primary sludge was collected from the clarifier
 of the continuously-spiked aerobic pilot plant.  Thickening was achieved by
 overnight settling at 4°C in completely-filled glass containers to avoid loss
 of volatiles in the headspace.  Secondary sludge was drawn from the waste
 activated sludge line of the pilot plant and thickened by overnight settling
 in glass jars at 4°C followed by centrifugation.  After achieving 4% total
'solids, the spiked primary and secondary sludge were combined in equal volumes
 and stored in glass containers at 4°C (with no headspace)  until needed for
 digester feeding.   Slight losses of volatiles occurred during processing,
 however all feed sludge were analyzed as fed to the digesters.
     For the semi-volatile organics study,  unspiked sludge were taken weekly
 from the Mill Creek facility and spiked sludge from the T&E  pilot plant.
 Unspiked primary sludge was gathered  from the gravity thickeners and diluted
 to 4% total solids,  when necessary, with primary influent.   Secondary sludge
 was gathered from the thickened secondary sludge line (after polymer addition
 but prior to gravity-belt thickening).   Return activated sludge was used for
 dilution when necessary.   Both primary and secondary sludge  were combined in a
 1:1 volume ratio to yield the final feed sludge for digester 1.
     Semi-volatile spiked sludge for digesters 2 and 3 were gathered biweekly
 from the pilot plant, as  the danger of component loss by volatilization  was
 not significant as that for the volatile organics.   Primary  sludge was
 thickened by overnight settling at 4°C in large Nalgene carboys.   Secondary
 sludge was thickened by the perforated bowl centrifuge and diluted with  the
 centrate to 4%.   Both spiked sludge were combined in equal volume ratios prior
 to storage at 4°C in a large carboy.
                                     334
-------
DIGESTER PERFORMANCE

    Conventional parameters were measured for the sludge feed, mixed-liquor,
and gas samples from each of the three digesters in order to assess digester
performance.  Digester temperature, pH, and gas production were measured
daily.  Chemical oxygen demand, alkalinity, volatile fatty acids, total and
volatile suspended solids, total Kjeldahl nitrogen, ammonia nitrogen, and
total and organic phosphorus were measured biweekly throughout the study.
Analyses of off-gases for methane and carbon dioxide content were performed
weekly.
    For organics analysis, the primary/secondary sludge mixtures used as
digester feed were analyzed daily for the volatile organics or composited
daily and analyzed weekly for the semi-volatile organics.  Mixed liquor was
collected for analysis once a week.  A portion of the mixed liquor was
centrifuged, and the centrate was collected for analysis.  For the volatiles
study, gas samples were collected and analyzed weekly.  Gas samples for the
semi-volatiles study were analyzed near the end of the testing period.

ANALYTICAL METHODS

    Conventional parameters were evaluated by EPA Methods (7).  All analyses
were performed in duplicate, when possible, for each sample.  Volatile organic
compounds were analyzed using EPA Method 601 (8) and EPA Method 602 (9).  The
analytical system was composed of a Tracer LC2 Sample Concentrator, a Tekmar
Model ALS Automatic Laboratory Sampler, and a Tracer 585 Gas Chromatograph
with a Nelson Analytical 900 Series Interface to an IBM Personal Computer AT.
Digester off-gas samples were analyzed for specific volatile organics by
injecting a 5-cc sample directly into the purging chamber with a gas-tight
syringe.  From this point, the methods used for gas analysis were identical to
those described above for the aqueous samples.
    Semi-volatile organics analysis was performed by EPA Method 1625 (10).
Analytical equipment included a Varian gas chromatograph and an Incos 50 mass
spectrometer.  Off gas was analyzed as described above.  For those compounds
not quantifiable in the spectra, gas-phase concentrations of the semi-volatile
compounds were estimated by pure component Henry's Law constants evaluated at
35°C.  The chemical component in the gas was assumed to be in equilibrium with
the mixed-liquor centrate.
    Mass balance calculations were performed for each of the sampling events
from the sludge feed, mixed liquor, mixed-liquor centrate, and gas
concentrations, multiplying by the appropriate volumetric flows and gas
productions for each digester yielded the mass flows.  The amount of component
sorbed on solids was determined by mass difference between the mixed liquor
and the mixed-liquor centrate.  Fates of the specific organics were then
determined.

                            RESULTS AND DISCUSSION
CONVENTIONAL DIGESTER PARAMETERS

    Conventional parameters are reported in Table 3 to document operation and
performance of all three digesters during the 7-week test period for the
                                      335
-------
volatile organics study.  Parameters for digesters 2 and 3 were very similar,
and are thus presented as an average for the spiked system.  Results in Table
3 indicate that all digesters functioned within the limits of normal operation
over the course of experimentation.  Of particular interest is the methane
content of off-gases from the digesters.  As shown, the difference in methane
production (per gram VSS destroyed) between the spiked and control digesters
is only 13%.  Further, for each digester the methane produced per gram COD
reduced was above the theoretical value (11) of 0.35 liters(at STP), again
signalling normal operation.  The total gas production from each digester was
also similar, with a difference of 6% between the spiked and unspiked units.
    Conventional operating parameters for the control and spiked digesters
during the 9-week test period of the semi-volatile organics study are
presented in Table 4.  As with the volatiles study, the operational variations

  TABLE 3.  SUMMARY OF ACCLIMATED DIGESTER CONVENTIONAL OPERATING PARAMETERS
                           FOR RCRA VOLATILES STUDY
  Parameter:
                         Digester
    CONTROL
              SPIKED
                    CONTROL
                                    SPIKED
Temp. (C) 35.1 ± 0.8
pH 7.34 ± 0.07
TSS (mg/L) 29,200 ± 1,000
VSS (mg/L) 13,600 ± 1,500
COD (mg/L) 22,000 ± 3,600
TEA (mg/L) 3,560 ± 100
VFA (mg/L) <50
TKN (mg/L) 895.0 ± 40.0
TP (mg/L) 51.0 ± 4.0
GAS (Lgjp/d) 14.0 ± 3.6
CH4 (L/gVSS) 0.62 ± 0.02
CH4 (L/gCOD) 0.41 ± 0.07
35.2 ± 0.8
7.32 ± 0.12
33,200 ± 2,500
15,300 ± 2,200
24,000 ± 4,000
3,610 ± 320
<50
860.0 ± 25.0
28.0 ± 9.0
13.1 ± 3.8
0.71 ± 0.11
0.39 ± 0.17
...
	
39,400 ± 3,500
24,500 ± 2,100
42,300 ± 5,700
675 ± 80
360 ± 90
200.0 ± 45.0
42.0 ± 4.0
	
	
...
	
	
42,700 ± 4,700
24,700 ± 6,400
44,000 ± 9,500
1,360 ± 515
470 ± 195
175.0 ± 25.0
22.0 ± 5.0
	
	
— — —
  TABLE 4.  SUMMARY OF ACCLIMATED DIGESTER CONVENTIONAL OPERATING PARAMETERS
                        FOR CERCLA SEMI-VOLATILES  STUDY
                         Digester
  Parameter:     CONTROL           SPIKED
                                               Feed
                                     CONTROL
                                               SPIKED
  Temp. (C)
  PH
 TSS (mg/L)
 VSS (mg/L)
 COD (mg/L)
 TBA (mg/L)
 VFA (mg/L)
 TKN (mg/L)
  TP (mg/L)
 GAS (Lsxp/d)
 CH< (L/gVSS)
 CHA (L/gCOD)
  35.1 ±
  7.08 ±
31,000 ±
17,400 ±
30,700 ±
 4,150 ±
     <50
 709.1 ±
  87.9 ±
 12.1 ±
  0.70 ±
  0.48 ±
 0.8
 0.03
 1,400
 1,100
 9,400
 150

 123.1
 16.9
2.1
 0.12
 0.25
32
17
35.2 ± 0.8
7.07 ± 0.03
       1,600
       800
       4,300
       170
800
700
27,400 ±
 3,670 ±
     <50
 569.2 ± 93.9
  23.9 ± 5.8
 10.7 ± 1.4
  0.65 ± 0.12
  0.34 ± 0.09
43,600
30,200
55,900
   560
   890
4,800
3,000
12,600
120
220
               181.0 ± 43.1
                88.2 ± 20.6
43,500 ± 5,400
28,700 ± 3,000
49,700 ± 8,900
   450 ± 140
   570 ± 170
  90.8 ± 22.3
  19.1 ± 5.7
                                      336
-------
between the two spiked digesters were slight, therefore parameters are
tabulated as an average of both.  No major operational difference between the
spiked and control digesters is seen.  Methane production was near the
theoretical value of 0.35 l(STP)/g COD destroyed for all digesters.  Average
daily gas production from the control and test units differ by 12%.  An
increase in gas production from digester 1 near the end of the test period is
attributed to an increase in COD in the control feed, while the COD of the
spiked feed and spiked units' gas production remained relatively constant.  As
with the volatiles study the semi-volatile spike does not appear to have
hindered digester operation.

FATE OF CHEMICAL COMPONENTS

    The specific fate of each compound attaining steady-state conditions is
summarized in Table 5 according to treatment mechanism.  Volatile and semi-
volatile concentrations in samples from the two spiked digesters were averaged
together in the mass balance calculations.  For both studies,  it was found
that weekly pollutant concentrations, especially in the sludge feed, varied
considerably in some instances.  Fluctuation in the feed concentration was due
to variability in the background chemical concentrations in the raw wastewater
and the method of the feed preparation.  An additional (and unquantifiable)
complication in assessing the fate of individual components is that some of
the degradation products are also components of the spike mixture (i.e.,
tetrachloroethylene to trichloroethylene).  Finally, it should also be
emphasized that accurate analytical data for specific components in an
anaerobic digester matrix are difficult to obtain.   In spite of these
difficulties,  the experimental results provide useful information on the fate
of sorbed organic toxics in anaerobic digesters.
    For the volatile organics study, eleven of the compounds originally spiked
into the raw wastewater were consistently present in the digester feed sludge.
Other compounds in the original spike were either not present in the feed or
were present sporadically or in trace quantities.  Steady-state conditions
were achieved during the test period for seven of these volatile compounds,
shown in Table 5.  Biodegradation was the most significant removal mechanism
for all of these compounds except 1,2-dichloropropane.  As would be expected,
volatilization of the components into the digester off-gas was also a primary
removal mechanism.  Four other compounds, including chlorobenzene,  chloroform,
1,2-dichloroethane, and toluene, initially accumulated in the digesters, but
all showed evidence of degradation by the end of the testing period.
    Analytical results of the semi-volatile organic analysis showed twelve
compounds exhibiting steady-state behavior.  Degradation was the predominant
removal mechanism for lindane,  dibutyl phthalate, butyl benzyl phthalate, and
2,-cresol were almost completely degraded.  The other compounds,  while
generally showing significant biotransformation,  tended to sorb or remain
sorbed on the solid material.  Results of off gas analysis showed only the
dichlorobenzenes, trichlorobenzene and hexachloro-1,3-butadiene to be
quantifiably present.  As expected from the high Kows of the components,
solubilization was not significant except for 4-chloroaniline.  As mentioned
previously, some components may biotransform into some of the other spiked
components, thus affecting the apparent fates.  Of note here is 1,2,4-
trichlorobenzene, which has been shown to degrade to 1,2-dichlorobenzene (12).
                                     337
-------
                                    SUMMARY
    The performance of the anaerobic digesters were not hindered by either the
volatile or semi-volatile organic spiked sludge.  Gas production and methane
content for spiked and unspiked systems were nearly identical.  All parameters
were within typical operating limits.  Aliphatic volatile compounds were
significantly degraded or otherwise volatilized.  Removal of volatile organics
by sorption was generally not a significant mechanism.  The semi-volatile
compounds tended to sorb onto the mixed-liquor solids or be degraded.
Solubilization and volatilization generally did not play key roles.  For both
studies, all compounds showed some evidence of degradation.  Anaerobic
digestion is a viable method of treating toxic-laden primary and secondary
sludge, however, more research is necessary to determine specific fates of
individual compounds in actual digestion processes.

                               ACKNOWLEDGEMENT
    This work was performed jointly by the USEPA and the University of
Cincinnati under cooperative agreement no. CR812939-01.

            TABLE 5:  STEADY-STATE FATES  OF SPIKED ORGANIC  COMPOUNDS
I. VOLATILES:
Fate Mechanism Average
SOL VOL SORB DEC FEED LOAD
(%) (%) (%) (%) (mg/kg)
1,1,2-Trichloroethane
1,1,1-Trichloroethane
Trichloroethylene
Tetrachloroethylene
Methylene Chloride
Ethylbenzene
1,2-Dichloropropane

II. SEMI-VOLATILES:
0
0
0
0
0
0
5
 0
 3
 8
 8
14
26
46
 0
 2
 0
 1
 3
14
10
100
 95
 92
 91
 83
 72
 39
 39
 64
 84
 37
 33
10
14
144 ± 45
61
20
25
 61 ± 14
Lindane
Butly benzyl phthalate
Di-n-butyl phthalate
2-Cresol
4- Chloroaniline
Napthalene
1 , 3-Dichlorobenzene
Bis (2-ethylhexyl) phthalate
1,2, 4-Trichlorobenzene
Hexachloro -1,3 -butadiene
1 , 2-Dichlorobenzene
1 , 4-Dichlorobenzene
0
1
1
6
31
4
3
5
3
1
4
4
0
0
0
0
0
4
10
0
5
8
14
16
2
2
3
20
34
65
61
69
66
73
66
68
98
97
96
74
34
27
26
26
26
18
16
13
490
290
270
13
12
230
320
1100
750
1100
280
275
± 140
± 190
± 140
± 6
± 3
± 40
± 110
± 700
± 320
± 560
± 70
± 60
SOL—solubilized(amount in mixed-liquor centrate)   VOL=volatilized
SORB-sorbed on solids   DEG=degraded   FEED LOAD=mg organic/kg dry wt.  solids
                                      338
-------
                                  REFERENCES
 3.
 4.
 5.
 6.
 7.
 8.
 9.
10.
11.
12.
Hannah, S. A., and Rossman, L.  Monitoring and analysis of hazardous
organics in municipal wastewater--a study of twenty-five treatment
plants.  Paper presented at the Seminar on Hazardous Substances in
Wastewater, Toronto, Ontario.  November 3, 1982.

Hannah, S. A., Austern, B. M: , Eralp, A. E.,  and Dobbs, R. A. J. WPCF.
60(7): 1281, 1988.

Dobbs, R. A., Jelus, M., and Cheng, K.  Partitioning of toxic organic ''
compounds on municipal wastewater treatment plant solids.  EPA/600/D-
86/137, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1986.

Malz, F. Heavy metals and chlorinated hydrocarbons in sewage sludge.  In;
Proceedings of the Workshop on Removal of Chlorinated Hydrocarbons and
Heavy Metals from Wastewater by Advanced Treatment Systems.  IUPAC
Commission on Water Chemistry (VI.6), Frankfurt, FRG, 1986.

Sludge handling and disposal practices at selected municipal wastewater
treatment plants.  MCD36.  U.S.Environmental Protection Agency, Office of
Water Program Operations, Washington, DC, 1977.

Bhattacharya, S. K., et al.  Fate and effects of selected RCRA and CERCLA
compounds in activated sludge systems.  Paper presented at the 15th
Annual EPA Research Symposium, Cincinnati, OH.  April 11, 1989.
EPA methods for chemical analysis of water and wastes.
U.S. Environmental Protection Agency, 1979.
EPA-600/4-79-020.
Test methods - Methods for organic chemical analysis of municipal and
industrial wastewater:  Purgeable halocarbons - Method 601.  EPA-600/4-
82-057.  U.S. Environmental Protection Agency, 1982.

Test methods - Methods for organic chemical analysis of municipal and
industrial wastewater:  Purgeable aromatics - Method 602.  EPA-600/4-
82-057.  U.S. Environmental Protection Agency, 1982.

EPA Method 1625 - Semivolatile organic compounds by isotope dilution
GC-MS.  Federal Register. 49(209): 1984.

Parkin, G. F., and Owen, W. F. Fundamentals of anaerobic digestion of
wastewater sludge.  J. Environ. Eng.  112(5): 879, 1986.

Tsuchiaya, T. and Yamaha, T.  Reductive dechlorination of 1,2,4-
trichlorobenzene by Staphylococcus epidermidis isolated from intestinal
contents of rats.  Agric. Biol. Chem.. 48: 1080, 1984.
                                      339
-------
                  FATE AND EFFECTS OF SELECTED RCRA AND CERCLA
                     COMPOUNDS IN ACTIVATED SLUDGE SYSTEMS

                                       by

                    Sanjoy K. Bhattacharya, Rao V.R. Angara

               Department of Civil and Environmental Engineering
                            University of Cincinnati
                            Cincinnati, Ohio  45221

                    Sidney A. Hannah, Dolloff F. Bishop, Jr.
                     Richard A. Dobbs, and Barry M. Austern

                      U.S. Environmental Protection Agency
                            Cincinnati, Ohio  45268
                                    ABSTRACT
     Two separate studies were conducted to investigate the removal and fate
of 28 selected RCRA compounds (0.25'tng/l of each compound) and 19 selected
CERCLA compounds (0.5 mg/1 of each compound) in conventional activated sludge
treatment.  In each study, two pilot-scale (35 gpm) activated sludge systems
(SRT: 4 days for RCRA study and 8 days for CERCLA study) were operated in
parallel at the USEPA Test & Evaluation Facility in Cincinnati, Ohio.  One
system was spiked continuously with either RCRA or CERCLA toxics to produce
an acclimated biomass; the other was spiked intermittently with the same
toxics and sampled to determine performance under unacclimated conditions.
The selected RCRA or CERCLA compounds did not cause any adverse effects on
COD and SS removals.  The concentrations of organics (RCRA study) in the air
emissions indicated that the chlorinated aliphatic solvents were essentially
volatilized into the plant air emission stream, whereas the aromatic volatile
benzenes were substantially degraded.  Additional work is planned to attempt
to reduce the analytical variability encountered in these studies.
                                    340
-------
                                  INTRODUCTION
     A survey of publicly owned treatment works (POTW) showed that concentra-
tion of priority pollutants in the influent wastewater to many of these
plants exceeded the allowable concentrations for these chemicals (1).
Petrasek, et al. (2) studied the fate of 22 toxic organics in wastewater
treatment plants.  They reported that a typical POTW significantly (up to
90%) reduced the concentrations of most of these compounds.  However certain
compounds were present in the activated sludge effluent in relatively high
(20-30 jug/L) concentration.  Hannah, et al. (3) investigated the comparative
removal of priority pollutants by six biological and physical-chemical treat-
ment processes.  They reported that activated sludge process provided the
best results.  A further review of the literature indicated that only limited
data are available for many priority pollutants.

     In this study, the removal and fate of selected RCRA and CERCLA toxic
organic pollutants were evaluated with two pilot-scale activated sludge
systems fed municipal wastewater at the USEPA's Test and Evaluation Facility
in Cincinnati, Ohio.  The 28 RCRA (semivolatile and volatile) and 19 CERCLA
(semivolatile only) chemicals spiked into the systems are shown in Table 1.
The selected RCRA and CERCLA toxics were spiked into the raw wastewater in
two separate test periods.
                   EXPERIMENTAL SYSTEMS AND TESTING APPROACH
     The twq activated sludge systems were operated at a flow rate of 35 gpm
and a hydraulic retention time (HRT) of 7.5 hours.  An operational sludge
retention time (SRT) of 4 days was used in the RCRA study period. In the
CERCLA study period, the SRT was 8 days.  Each compound was spiked at 0.25
mg/L for the RCRA study and at 0.5 mg/L for the CERCLA study.  The operating
conditions and design characteristics for the two systems used in the study
are given in Table 2.  Both RCRA and CERCLA studies were performed with an
acclimated (continuous addition of toxicants) and an unacclimated (intermit-
tent spiking of toxicants) system.  These systems were operated in parallel.

     To sample from the air space above the primary clarifier, the units were
covered and vented through a duct to the roof.  An air sweep equivalent to a
5 kilometer per hour wind was maintained over the surface of the primary
clarifier by exhausting air at 14,000 liters/min.  The aeration basin was
fitted with an air tight cover and the off-gas was also vented to the roof.
Air flow in the aeration basins averaged 5,600 liters/min.
                                     341
-------
               TABLE  1.   RCRA AND  CERCLA TOXIC ORGANIC  POLLUTANTS
         RCRA Study  Period
     CERCLA Study Period
    acetone
    cyclohexanone
    furfural
    2-butanone
    4-methyl-2-pentanone
    tetrahydrofuran
    carbon tetrachloride
    chlorobenzene
    chloroform
    1,2-di chloroethane
    1,2-dichloropropane
    methylene chloride
    tetrachloroethylene
    trichloroethylene
    1,1,1-tri chloroethane
    1,1,2-tri chloroethane
    ethyl benzene
    toluene
    total xylenes
    bis(2-ethylhexyl) phthalate
    butyl benzyl phthalate
    1,4-dichlorobenzene
    2,4-dimethylphenol
    2,4-dinitrophenol
    naphthalene
    nitrobenzene
    4-nitrophenol
    phenol
1,2-dichlorobenzene
1,3-dichlorobenzene
1,4-dichlorobenzene
1,2,4-tri chlorobenzene
nitrobenzene
1,3-di ni trobenzene
2,6-dinitrotoluene
p-cresol
4-chloroaniline
hexachloroethane
hexachlorobutadi ene
dimethyl phthalate
diethyl phthalate
dibutyl phthalate
butyl benzyl phthalate
bis(2-ethylhexyl) phthalate
naphthalene
lindane
dieldrin
     Automated analytical procedures were used for the conventional pollu-
tants (COD, BOD, NH4-N, N03-N and TKN) and 6C/MS procedures were used for
the toxic organic compounds.  RCRA samples were analyzed by a contract
laboratory (PEI Associates Inc., Cincinnati, OH).  Air samples were collected
in stainless steel canisters, and were analyzed by GC/MS.  From these data,
masses of each RCRA compound stripped during the sampling event were calcu-
lated.  Sludge and liquid samples were also analyzed by GC/MS according to
approved USEPA methods (4).  Semi-volatile RCRA compounds were extracted from
the samples using continuous liquid-liquid extraction.  Prepared portions of
the extracts were injected into the GC/MS for analysis.  The semi-volatile
CERCLA compounds were analyzed following Method 1625 (5). Analyses were per-
formed at the EPA, RREL.  Details of the analytical procedures were reported
elsewhere (6).
                                     342
-------
          TABLE 2.   OPERATING CONDITIONS AND DESIGN  CHARACTERISTICS  OF
                    THE PILOT SYSTEMS
         I.


        II.




       III.




        IV.
Design Flow
= 2.2 Us
= 191 m3/d
Primary Clarifiers - Diameter              = 2.9 m
                     Weir Diameter         = 2.8 m
                     Surface Area          =6.8 m2
                     Surface Overflow Rate =28.0 m3/m-d

Aeration Basins - L:W:D                    = 5.4 m:3.0 m:3.6 m
                  Surface Area             = 16.3 n£
                  Volume                   = 59.7 m3
                  Hydraulic Residence Time = 7.5 hrs.

Secondary Clarifiers - Diameter              = 3.6 m
                       Surface Area          = 10.4 n£
                       Surface Overflow Rate =18.4 m3/mz.d
     Three tests  (sample collection events) were performed during the RCRA
study period.   For the CERCLA study, eleven tests on the continuously spiked
(acclimated) system and 4 tests on the intermittently spiked  (unacclimated)
system were performed.


                                    RESULTS


     The presence of the spiked toxic organics in the wastewater produced no
major adverse effects on the treatment of conventional pollutants.  Average
removals of conventional pollutants in the pilot systems during the two
studies were between 94 and 97 percent for SS and between 81 and 88 percent
for COD (Table 3).  In the RCRA study period nitrification in the activated
sludge processes produced average NH4-N reductions between 76 and 81
percent.  In the CERCLA study, the NH4-N removal was between 88 and 98
percent.  The CERCLA toxics (0.5 mg/L) did interfere with nitrification in
the acclimated system.

     Substantial variability occurred in the reported results with some toxic
compounds, especially in the RCRA study period.  Table 4 lists average
measured concentrations of the selected RCRA organics in wastewater and
sludges for the acclimated system.  The difference between the concentration
of most toxics in the spiked wastewater feed and primary effluent was very
low.  The primary sludge showed enhanced concentrations of the two phthalates
and naphthalene along with reduced concentrations in primary effluent indica-
ting adsorption onto sludge solids.  Four compounds (tetrahydrofuran, 1,2-
dichloroethane, methylene chloride and 1,1,2-trichloroethane) were present in
                                      343
-------
      TABLE  3.   AVERAGE PERCENT REMOVALS  OF  CONVENTIONAL  POLLUTANTS
                 DURING THE  RCRA AND  CERCLA STUDIES
                       Acclimated  System
                     % Removal    Standard
                     	Deviation
                         Unacclimated System
                         % Removal   Standard
                        	Deviation
   RCRA Study

      SS

      COD

      NH4-N

   CERCLA Study
97

82

76
 4

 7

19
97

81

81
 3

 8

18
SS
COD
NH4-N
95
88
88
3
4
14
94
87
98
6
10
3
the secondary effluent stream  in high concentrations  (between 95 and 140
ug/L) indicating poor removals of these organics.  Five other compounds
(cyclohexanone, furfural, 2,4-dimethyl phenol, 2,4-dinitrophenol and 4-
nitrophenol) were not evaluated due to inconsistent results.  The average
removals of the toxic compounds in the RCRA study are summarized in Table 5.
The removal of RCRA organics with primary treatment was between 3 and 44
percent.  The total removal was between 36.6 and 99.0 percent.  The calcu-
lated percent stripped for the individual volatile compounds varied from 1 to
139 percent.  No air analyses were performed for the semivolatiles.  Bio-
degradation of a compound was estimated by subtracting the measured removals
by adsorption and stripping of the compound from the total removal.  The
estimated biodegradation was between -42 and 97 percent.  A negative bio-
degradation indicated inconsistent mass balance and the problems of
estimating biodegradation using this approach.  Biodegradation appeared to be
the predominant removal mechanism for the polar solvents, and the aromatic
volatiles (e.g., toluene: 72%, xylenes: 66%, chlorobenzenes: 60%).  The
unacclimated system (e.g., biodegradation of toluene: 56%) showed similar
results and no significant advantage of acclimation was observed (6).

     Table 6 lists the average concentrations of the CERCLA organics.  Five
compounds (1,2,4-trichlorobenzene:  89 ug/L, 2,6-dinitrotoluene: 125 ug/L,
p-cresol: 156 ug/L, lindane: 198 ug/L, and dieldrin: 99 ug/L) showed high
concentration in the secondary effluent (Table 7).  The average removals of
the toxic compounds in the CERCLA study are summarized in Table 7.  The
                                      344
-------

s:
Oi
co co
0 -Zi
i — i ZD
K
CH. CO
O LU
o: co
o
o: LJ_
ll_ °
CD LU
CD
CO ==C

O LU
i — i ^*


£ «5"
•y* LU
LU 1 —
O CO
Z >-
O CO
0
Q
LU LU
CD 1 —

LU >— l
•=C 0

CO UJ
UJ
tD
a:
UJ
<;

CkC O
Q H-
3: <
I— >
to uj
o
UJ


1


g 0

C> 1—
z ^

00 UJ
o

UJ
C3
C£.
%












UJ
5
z







II
II
II
II
II
II
II
II
II
II
II
II
—1 II
-^ 11
O> II
D II
11
11
II
II
II
II
II
II
II
II
II
II
—1 II
^ II
a> u
n n
n
n
n
n
n
n
u
u
u
n
u
ii
n
n
—i u
^, n
CD II
D II
II
11
II
II
II
II
II
II
II
II
II
II
||
II
-J II
^« II
a> n
=i II
n
u
ii
n
n
n
ti
u
it
n
u
It

n
-i u
•-- II
O) II
3 II
II
II
II

II
II
II
jj
II
II
It
II
It
II
II
II
II
II
II
II
II
II
It
II
11
II
II
II
II
tl
II
II
II

OJ
«— OJ CO
•<»• %o oj
OJ «-
«— ro in
JO 0 ,0

o in ro
CM



0s *"~ *O
ro o CM
CO




CM CO CM
SCO NO
«-




ro co N-
*- CO CO

*—




N- --* co
*»j" ~ ro ro
v± «— CM













c
OJ
4-J
c
C 0-
C Q) CM
OJ C «
O Q) OJ C >-
c JT t- m j=
O O D *J -M
4-> — ' M- D 0)
. 3 « i
<£ CJ u. CM ^f


CM
1
st

co
OJ



tn
OJ





in
OJ
OJ




o-






1













C
m
L.
D
M-
o
|
0)
t_
4-*
Q)


in
1
«-

ro



o
*~




JJ





j:






1













C.
O
^1
2
4J
0)
1—
c


OJ
o


ro
s.
o^
in
OJ
»-

ro



o

O
O
.r:
0
Q
i
OJ



sf
o
CO
o
OJ
o
OJ

Nf
in



R





OJ
OJ




o
in






OJ












c
1

o
t.
o
u
Q

OJ



' R ~* -
co -o ro
ro o o
in o in
t- OJ OJ
ro sj- ro
>o in r^
«— OJ OJ
oj ro oj in

r^ o in in
o »-



r^ -o m co
*~




co sf -o ro
o- o- co ro
ro «— «— oj




i? 5 in 5






sf «- «- O
ro *— o ^
-d- oj oj oj









 • ai o
O JS — ' t-
-• 4J X 0
jr 4) .c — •
O O 4J JZ
c. v o
QJ O O "^
C — • c- t-
CJ JZ O H-
— ' U — ' <
>* (0 J= «-
JZ L. U *

01 O tl
5E 1— 1— «-


O OJ
ro
fe M
in ^~
•o o
r- OJ
r- OJ
 1C





OJ . OJ
0 O
OJ «—




Sf !J






^ t>










c
ro

o co
S. x± <— c\j . , t\j
' ^_ ]
1
«— O CD O CD *—
1
. .
1
1
>t in in in in in



»•$• o« N- ' 'CM co *o
CM ro . in ro o
T-
.



(
f
in ro in to ro CD
o ro o- ro oj o
i— «— CM




ro ^o ro *— r^- CD
ro fM ro NQ >d- co

-•

•
.
•
o CM CM *4- ON . in
OJ vO O -
ro
_.
03 .
JZ
-M
.C
Q. . OJ N C O
x ^: c .— ' t- >. O 0) C C
f-v>ni^c_c~aj-UOQ4J O-M O
1 >- 1 - 1 i .C L- ••- C
yj-+j^j--J--st O. *J Z 
-------


CO
o- —
l— l CO
2: 2:
— >
5:
OH-
— °°
^C L fJ
S1-
Q£ CO
1 1 l ^t^
o o
CO LU
—I CD
11

z: LU
it
.11. 1
-
II 0
II I—
II
II
II
!!
n
n
u
u
H
IS »
ii s
ii »_
n a:
It Q.
n
ii
n
ii

ii
11
n
n
n
n
n
it
n
it
n
it
ii
ii
n
ti
u
u
u
11
u
II
tl
n
H
"

1 1
| REMOVAL STANDARD |
(%) DEVIATION |
1 1
[REMOVAL STANDARD
(%) DEVIATION

1
| REMOVAL STANOARt
| (%) DEVIATIOI
& 1
STANDAF
DEVI ATI

_,
^
o *!

UJ

g I
«C t—
0 <

^c ^>
1— UJ
CO O

— I
o S
UJ
C£













Ul

z







II
ii in t- NO
it
it
it
ii
n
n
n
ti
n
n
ll oj -si- in
II ON ON ON
II
II
II
II
II
II
II
II *~ "~ *~
II
It
II
II
II
It
II
II
II
ll ro <— oj
it
ii
ii
ii
u
It •
tl
n *~
it
ll
it
it
u
ll
ll
It OJ T- «-
II
II
II
It
ll NO co in
u oj oj ro
u
u
u
ii
u
HO CD in
UN. in S-
jl ON ON ON

II
II
II CO T— CO
II NO «— 0
ll oj ro NO

ii
ll
it
ll
it
u
IIO 00
II NO -.
II C J= t-  JC
ii w — < H- D a!
Ii 8 S. D ? ?
It 
OJ
O
t_


o
CO

1 —

NO
i

-
in


ro
£



sf
^_
OJ





ro
ON



















O
ON






0
•J-













OJ
•M
CO
CD
JCZ
4->
•>.
C
OJ
J.3
D
CO

ro c\j vQ c\J
.


(\J CM *— T—
«— in in o


ro ro N. CD -
O O O o



*O OJ f\J *—
S- >d" N" r—






ro N- ro o
co NQ -j- ro













• O  O
i . . Ji t_ .~ C
"^ *^«. "^ (D 't- ^ .C
T— OJ OJ Z Z -J- CL
I
I
1
1
1
I
1
1
1
1
1
1
1
I
J
1
)
1
I
1
1
1
t
1
I
1
1
1
1
t
1
1

1
:
|
I
I
1
1
I
1
I
1 CD
f
l Crt
l
1 (U
1 4-»
(D
f >
I "-
t O
I CU
1 *-*
t M
1 03
I 13
t +
I
t I-

1 E
t ••-
1 f-
1 Q_
1
t
t
( 0)
! f
) — •
1 C/)
1
1 *


1
t
I
1
1
r
t
i
l
i
i
l
i
i
l
i
i
l
i
i
i
l
i
l








t
f
o
o
-------

1
u_
co
i— t CO
 CD
O &
ii ^^
i— >
A? ^j

UJ UJ
CJ 1—

o >-
o co
LU Q
CD UJ
fff |— •
Dig
UJ S
^> i — i
0

UJ
ca
r


















ii
u
n
u
n
n
n
u
ii
u
it
it
n
n
u
it
u
n
n
ii
u
n
u
n
n
u
n
n
n
n
n
ii
u
n
u
n
n
n
n
n
n
u
u
n
u
u
n
n
n
it
n
n
u
n
u
n
n
n
u
n
n
n
ii
n
n
n
n
n
n
n
n
u
u
u
n
!!
jj

j!
II

Jj

U

II

II

It
11

It
'j
II
II
ii
SECONDARY SLUDGE
PRIMARY SLUDGE |
1
EFFLUENT |
1

>-
a
 Q 11
II
II
UJ It
(D —1 II
K ~O> II
UJ 3 II
> II
 II
t— UJ II
CO O II
II
UJ II
0 _J II
ce. 01 n
UJ 3 II
> tl
< it
n
Z II
Q O II
 II
h- LU II
CO O II
II
UJ II
C3 —I II
< — II
UI 3 II
> II
< II
II '
II
II
II
II
II
II
It
II
II
II
II
LU II


CM ro
ro in
in co
o ro
ro in
- <*^ *^ ^ *~~ N- ^O
fo «— *sj~ in «— in in
^- r— N-
CM«-inNOCQ COS-O
NO-J-O>OCO OfMO
O «— in ^a- «— N- -o
«- ^- ro
in QS ON r**- ^f ^o *o ^~
«— «— rocoo mco«—
o O* o NO in co
sjO^*oOi>fs- N-O>O-
coco^ro>o ru <— **
tn^fON-NO fMN-in
T— «— m 0* o in NO
-r -j-ro-st
rONOcor^-cM CO^NQ
cviin-4-OJOj r^irorvj



rONOororo rotoo
m-ON-coco <— in ON
ONN-ONC\JON rOf\Jf>
roin*4-"J'ro -vf-^d-in



UJ

-~i
UJ -
< < _I _J Q. X
*— 31 ^ «4 UJ

CO 3= 1— H- >- -J UJ
CZ Q. D_ 2 3= UJ
O -J UJ h- -J 2
CJi— rct-_jfMj-
to

c
i
S
*




















347
-------
       II
       II
       II
       II
       II

       !!     ~
       n
       n
       n

       n
       n
       n
       n
  .
CO

0
   co
-
uj co  N

°"Q  |i
LU LU  M


^S£  "
LU I
   0
CO

e£
       It




1—

_j


z
*~*

UJ
>
g

o:




•K


<
O
1—














«I
S

CL















UJ
z:

o
E
LU
o:
g
«c
o
z

UJ
a

&?



g
i—
<

>
MJ
*£



s^



g
§

>
UJ
0


s«





g

1—



UJ
o



































II
II
11
11
11
II
II
II
II
It
It
II
II
II
II
II
11
11
II

II



11
|j

11

ii

II
1!
ii

n
n


n
n



it
ii
n
n
n
n
n
n
n
n
n

n
n
n

n
"
H
||
II


11


II

II
II
II

II


II
II


II



II
II

O







to




0.






o
CM





ro
•*





-3-
ro
o





f-
tr-







*—
in






CM





ca

m
0
a_


UJ
z
LU

Z
LU


g

3:
0

a









e




o






vO






,_
in





O
CO
CO





o
0







o
in






ro





CO

ca
0
a:
o_


UJ
z
UJ
IM
Z
UI


g


0

a

o







&




0






N.






CM
in





ro
in
o





*o
,_







ro
,„.






^3-





CO

ca
0
ee.
o_


UJ
LU
IM

UJ


g


0

Q

^







CO
•J-




CM






N.
ro





CO
1^





«-
in
CO





o
in







CM
N.
CM












LU
Z
Si
z
LU
CO
O
g


O

en
t-
1


CM

^—

ro







c>




r-






^_






<0
CM





CO
CM
O





CO
CM







O.
>o




















UI

UI
IM
Z
UJ
m
o

h-

z




























r^
ro





N.
§





co
o







o
o>
















LJJ

UI
IM
Z
UI
CO
o

1—

z
a
i
ro

^

CO
ro






r-
in




in






CM






ro
in





CM
S





in
co







ro
ro
CM















LU
z
LU
3
O
t—
O

1—

z

1


CM


CM






O
•a




,-






CM






O
CM





O
R








































_i
o
LU
C£
U

Q.

co







IS




,-






CM






,_
5Q





o
So





CM
in
ro






si-
,_
CM

















LU
_J

Z

O


o

(J

^^


*
*





o
o
^~"



JUt
*
4c





O






in
CM





«-
fe





CM
^O







ro
^.
ro

















UI
|
i—
LJJ
O

o
_J
3:
0
1


CO

o

CM CO CO ^t
v- ro CM






f^~ "^ CO *O
 co ** i
•*3" T— in i

i
i

	 i
i
i
i
i
UJ 1
h- 1

_J I
UJ - I
CL X 1
UJ 1
_J 3T 1
>- — 1 LLJ 1
rsl >- z I
Z 3= HI 1
UI 1— — J -SS. \
03 LU < LU — 1
—1 CM 1— - N-^ 3= Q _J |
t-  i-, < ^-. •-• i
CO CO Z -J O 1
1
1
1
1
1
1
1
1
1
1
i
i
i
i
i
i
i
i
i
i
t
i
i
i
i
i
i
i
i
i "
i
i
i
i
i
i o
1 O)
1 TJ
1 3
1 (/)
\
\ TJ
1  — •

1 *-• >. (D
1 - O t_ — •
i 
1 4-« O 

.^ •*- TJ
U I_
a. a. .
TI *J "c
3 O O
-^ 1—
a> *
* *
* * *
_

,
,



.



.

•

-

' ,
-

.
'

,



                                            348
-------
removal of CERCLA organics in primary treatment with the acclimated system
was between 3.5 and 79.0 percent.  The total removal varied between 55.9 and
98.1 percent.  Biodegradation was estimated by subtracting the removal by
adsorption from the total removal of a particular compound.  The extent of
biodegradation varied between 28 and 100 percent.  The unacclimated system
also exhibited similar removals.  Biodegradation values were similar for both
acclimated and unacclimated systems (e.g., naphthalene: 79% and dimethyl
phthalate: 85%).  Like the RCRA study, no significant advantage of acclima-
tion was observed for the CERCLA compounds (6).  In the CERCLA study period,
the amounts of organics found in the complex primary sludge samples were
substantially lower than the measured removals across the primary process.
Due to the analytical variability encountered in these studies, additional
work has been planned.
                                  CONCLUSIONS
     The following conclusions were drawn from this study:

 1. The polar solvents and aromatic volatiles were biodegraded to a great
   extent.  For example, toluene exhibited 72 percent and total xylene  showed
   66 percent biodegradation.

 2. A significant  amount of chlorinated aliphatic solvents may be volatilized
   from an  activated sludge  system.  The percent stripped varied between  1
   and 139.

 3. Pesticides (lindane and dieldrin) were removed by both adsorption onto
   primary  and secondary sludge and biodegradation  in the secondary tank.
                                    REFERENCES
 4.
 5.
Fate of priority pollutants in publicly owned treatment works, 1,
EPA440/1-82/303, USEPA Effluent guidelines div.. WH-552, Wash. D.C,
Sept. 1982.
Petrasek, A. C, et a!., "Fate of toxic organic compounds in wastewater
treatment plants", Journal of Hater Pollution Control Federation 1983,
1286-1296.
Hannah, S.A, et al., "Comparative removal of toxic pollutants by six
wastewater treatment processes", Journal of Water Pollution Control
Federation 1986, 27-34.
USEPA Methods for Evaluating Solid Waste, SW846, 3rd edition, Nov. 1986.
Federal Register. Guidelines establishing test procedures for the analysis
of pollutants under the clean water act; 40 CFR, 136, 49, No. 209, 1984.
Bhattacharya, S.K., et al., "Removal and fate of RCRA and CERCLA toxic
pollutants in wastewater treatment", Final Report, Contract No. 68-03-4038.
                                      349
-------
                COMPATIBILITY OF FLEXIBLE MEMBRANE LINERS AND
                       MUNICIPAL SOLID WASTE LEACHATES
                   by:  Henry E. Haxo, Jr.
                        Matrecon, Inc.
                        Alameda, CA  94501
                                 ABSTRACT

      This paper describes the results of a survey of the  open technical
 literature relating to the composition of currently produced  municipal solid
 waste (MOT)  leachate,  the compatibility of flexible membrane  liners  (FMLs)
 with  such leachate, and the results  of limited experimental work  on  the
 absorption of  organics from dilute aqueous solutions that  simulate MSW
 leachates.

      The  results of the survey revealed little information on the compati-
 bility of FMLs  currently being used  in the construction of disposal  facili-
 ties  with MSW  leachates.   Some information was available from studies per-
 formed in the  1970s.   The little information  that  was available on the
 composition  of  MSW  leachates  currently being  produced indicates (1)  that
 concentrations  of potentially polluting species in current leachates appear
 to be low and  (2) that  current leachates  may  contain more  potentially pol-
 luting organics which are less biodegradable  and which may be more aggressive
 to FMLs than leachates  generated during the 1970's.

      Because of uncertainties  regarding the compatibility  of FMLs and MSW
 leachates and the absorption  of organics  from dilute solutions, such as
 leachates, limited laboratory  experiments were performed to measure the
 partitioning of selected  organics from  dilute  aqueous solutions to various
 types of FMLs.  The results indicate that, depending on the similarity of
 the solubility parameters of the organics and  the FMLs, some of the organics
even at low concentrations can partition  from  the water in which they are
dissolved and significantly swell an FML.
                                    350
-------
               COMPATIBILITY OF FLEXIBLE MEMBRANE LINERS AND
                      MUNICIPAL SOLID WASTE LEACHATES
                               INTRODUCTION

    Under Subtitle "C" of the Resource Conservation and Recovery Act (RCRA),
the EPA requires all materials that are used in constructing hazardous waste
disposal facilities and may come in contact with leachate to be tested for
chemical resistance to the specific leachate.  At the present time, the EPA
is considering extending this requirement for chemical resistance testing to
all materials which are used in constructing municipal solid waste (MSW)
landfills and which may come in contact with leachate.

     Previous studies indicate that commercially-available flexible membrane
liners (FMLs) are chemically resistant to typical MSW leachate (1, 2).  These
studies were conducted with leachate generated in simulators containing
freshly collected and shredded MSW.  Analyses of that leachate showed in-
organic salts, volatile organic acids, and suspended solids as the only
contaminating species.  However, specific analyses of additional organics do
not appear to have been perfromed.  In addition, adding small amounts of
organics from small quantity generators and nonhazardous industrial wastes
to an MSW stream could generate a leachate containing measurable quantities
of organic compounds that could chemically attack an FML and reduce its
service life.

     Even though commercially available FMLs are generally chemically
resistant, some could be adversely affected by specific constituents of
the leachate.  The effects can depend on the type of constituent, its con-
centration in the leachate, and the specific type of FML.  Furthermore, the
effects of the constituents can be synergistic and can vary with time as the
concentrations change with the aging  of the waste.   Organic molecules  [as
indicated by such analyses as volatile acids, volatile solids, total organic
carbon  (TOG), and chemical oxygen demand (COD)] may be damaging to some FMLs.
For example, specific organics can cause some types of FMLs to swell, to
become  softer and more permeable, and to lose in mechanical properties, such
as tensile strength and tear resistance; thus, exposure to some organics can
allow FMLs to be more easily torn and damaged.  Other constituents of a
leachate, e.g. water, can also cause  some FMLs to swell.

     Determining  the  chemical resistance of  FMLs and other materials of
construction is necessary when the waste liquid or leachate is. known to
contain constituents  that are aggressive to  these materials and when the
                                     351
-------
           concentrations of these constituents are sufficiently high.  The current
           method for determining chemical resistance of FMLs to waste liquids and
           leachates for permitting purposes under RCRA is EPA Method 9090.  However,
           this method as presently performed may not be adequate for assessing FML
           resistance to MSW leachate due to the instability of the leachate during the
           required 120-day exposure at the temperatures (23° and 50°C).  To assess FML
           resistance, the FMLs under test should be exposed to leachates that are in an
           in situ condition.

                This paper describes the results of a survey of the open technical
           literature relating to information that would be appropriate to a discussion
           on extending the requirement for chemical resistance testing of materials
           used in constructing MSW landfills, particularly FMLs.  The information
           sought in the survey included data on the composition of currently generated
           MSW leachates, the compatibility of FMLs with such leachates, and basic
           information regarding dilute aqueous solutions.   Results of limited experi-
           mental work on the partitioning of selected organics from dilute aqueous
           solutions to various types of FMLs are also presented in this paper.   Because
           MSW leachates are essentially complex dilute solutions,  these results  can
           supply information on the tendency of FMLs  to absorb organics from leachates
           and, ultimately, on the compatibility between FMLs and MSW leachates.


           SURVEY OF THE SCIENTIFIC AND OPEN TECHNICAL LITERATURE

                The  information sought in the survey included the following:

                o Data on MSW leachate composition and characteristics,  particularly
                  currently generated leachate.

                o Data from the compatibility testing of FMLs  with MSW  leachates.

                o Information on the  preservation  and stabilization of MSW  leachates,
                  particularly at  temperatures  somewhat higher than 23°C  so  that  these
                  leachates can be  used  in  conventional compatibility tests,  e.g.  EPA
                  Method  9090.

                o Data on the  resistance  of  FMLs and  polymeric compositions  to dilute
                  aqueous  solutions of organics and inorganics.  Such solutions simu-
                  late MSW  leachates  in many respects.

                o  Basic information from  physical  chemistry  regarding such factors as
                  the  chemistry of dilute aqueous  solutions  of  organics,  solubility
                  parameters,  constituent activity, partitioning of dissolved organics
                  between phases, and  transport of organic species.

          The information resulting from  this literature search is  summarized in the
          following subsections.
                                               352
_
-------
Composition of MSW Leachate

     The literature search revealed that most of the available information
on the composition and characteristics of MSW leachate had been generated in
the 1970s when the disposal of MSW was a principal concern of the EPA (3, 4,
5, 6, 7, 8, 9, 10, 11, 12).  The data reflect the characteristics of leach-
ates generated in both laboratory and pilot-scale projects and in actual
full-scale MSW landfills; these data include information such as chemical
oxygen demand (COD), biological oxygen demand (BOD), total organic carbon
(TOG), hardness, pH, electrical conductivity (EC), total dissolved solids
(TDS), trace metal analysis, volatile organic acids, and dissolved inor-
ganics.  Almost no specific information on organic species was found, except
for information on the organic acids, even though many are potentially pol-
luting or aggressive to FMLs and other polymeric .construction materials.
Much of the information on leachate composition was considered in developing
the chapter on wastes for the Technical Resource Document, SW-870, "Lining of
Waste Impoundment and Disposal Facilities" and its subsequent revision (2,
13).  Overall, the information indicates the following:

     o  MSW leachate is a complex mixture of inorganics, organics, and
        bacteriological constituents usually generated in anaerobic environ-
        ments in MSW landfills.  MSW leachate is generally mildly acidic,
        and many of the constituents are biodegradable.

     o  Leachates from different MSW landfills vary widely; the precise compo-
        sition is waste- and site-specific and depends on such variables as
        type of waste, amount of infiltrating water, age of the landfill, and
        pH.  Table 1 presents typical results of analyses of MSW leachates
        that were obtained and reported in the 1970s.

     o  MSW leachate is highly oxidizable and unstable and subject to rapid
        changes in composition on removal from the in situ anaerobic environ-
        ment in which it is usually  generated.  Even when cooled and sealed
        in bottles, the composition  of leachate will change for a limited
        time (4).

     Data in the open literature on  the composition of recently generated
MSW  leachates are limited; data that are available indicate the presence of
priority pollutants, aromatic hydrocarbons, and other constituents which
may  be  absorbed by FMLs  (14, 15, 16).  Table 2 presents a statistical
analysis of data on the concentrations of organic constituents of leachates
generated between  1980 and 1985.  These data include data from 15 case
studies performed by the EPA (15); data from landfill leachate sampling by
Minnesota and Wisconsin, and data from the National Pollutant Discharge
Elimination System  (NPDES) discharge permits for leachates from landfills
in New Jersey.  Even though only a relatively small number of facilities
were surveyed,  the data  resulting from those studies were considered reli-
able by the EPA (16).  The concentrations of most of the organics are low
and, even  if  they partition to  the FMLs, the amounts may not  be sufficiently
large  to cause  significant changes in properties of the FMLs, even after long
exposures.
                                     353
-------
       TABLE 1.  COMPOSITION OF MSW LANDFILL LEACHATES GENERATED BEFORE 1980
    Concentration of Constituents. (mg/L), Except pH and Electrical Conductivity

Constituent
BOD5
COD
TOC
Total solids
Total dissolved solids
Total suspended solids
Total volatile acids as acetic acid
Acetic acid
Fropionic acid
Butyric acid
Valeric acid
Organic nitrogen as N
Ammonia nitrogen as N
Kjeldahl nitrogen as N
Total phosphorus
PH
Electrical conductivity (ftmho/cm)
Total alkalinity as CaC03
Total acidity as CaCC>3
Total hardness as CaC03
Metals and anions:
Arsenic
Boron
Cadmium
Calcium
Chloride
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Phosphate
Potassium
Silica
Sodium
Sulfate
Zinc

Reference
No. 17
• • *
42,000
• • •
36,250
• • •
• • •
• • •
• • •
• • •
• • •
• • •
• • •
950
1,240
• • •
6.2
16,000
8,965
5,060
6,700

• • •
* • •
• • •
2,300
2,260
• • •
• • •
1,185
» • •
410
58
* • •
• • •
82
1,890
• • •
1,375
1,280
67
Source
Reference
No. 18
13,400
18,100
5,000
12,500
• • •
85
9,300
5,160
2,840
1,830
1,000
107
117
• • •
• • •
5.1
• • •
2,480
3,460
5,555

• • •
• • •
• • •
1,250
180
• • •
• • •
185
• • •
260
18
• • •
• • •
1.3
500
• • •
160
...
...
of Data
Reference
No. 19
• • •
1,340
• • •
• • *
• • •
• • •
333
• • •
• • •
• • •
• • •
* • •
862
• • •
• • •
6.9
• • •
* • •
• • •
• • •

0.11
29.9
1.95
354.1
1.95
<0.1
<0.1
4.2
4.46
233
0.04
0.008
0.3
• • •
• • *
14.9
748
<0.01
18.8

Reference
No. 4
81-33,360
40-89,520
256-28,000
0-59,200
584-44,900
10-700
• •
• •
• •
• •
• •
• • •
0-1,106
• • •
0-130
3.5-8.5
2,810-16,800
0-20,850
i • • «
0-22,800

• • •
• • *
0.03-17
60-7,000
4.7-2,467
• • •
0-99
0
0-2,820
17-15,600
0.09-125
• • *
• • •
• • •
28-3,770
• • •
0-7,700
1-1,558
0-370
Based on:  Reference No. 2, p A-5.
                                        354
-------
             TABLE 2.  CONCENTRATIONS OF ORGANIC CONSTITUENTS/
                    MSW LEACHATES GENERATED AFTER 1980a
(Units in ppb)
Constituent
Acetone
Benzene
Bromome thane
1-Butanol
Carbon tetrachloride
Chlorobenzene
Chloroethane
Bi s ( 2-chloroethoxy )me thane
Chloroform
Chloromethane
Delta BHC
Dibromome thane
1 ,4-Di chlorobenzene
Dichlorodifluoromethane
1 , 1-Di chloroethane
1 ,2-Dichloroethane
Cis 1 , 2-dichloroethene
Trans 1 , 2-dichloroethene
Di chloromethane
1 ,2-Dichloropropane
Diethyl phthalate
Dimethyl phthalate
Di-n-butyl phthalate
End r in
Ethyl acetate
Ethyl benzene
Bis(2-ethyl hexyl) phthalate
Isophorene
Methyl ethyl ketone
Methyl isobutyl ketone
Naphthalene
Nitrobenzene
4-Nitrophenol
Pentachlorophenol
Phenol
2-Propanol
1,1,2, 2-Te trachloroe thane
Tetrachloroethene
Tetrahydrofuran
Toluene
Toxaphene
1,1, 1-Trichloroethane
1,1, 2-Trichloroethane
Trichloroethene
Trichlorof luoromethane
Vinyl chloride
m-Xylene
p-Xylene and o-Xylene
Minimum
140
2
10
50
2
2
5
2
2
10
0
5
2
10
2
0
4
4
2
2
2
4
4
0
5
5
6
10
110
10
4
2
17
3
10
94
7
2
5
2
0
0
2
1
4
0
21
12
Maximum
11,000
410
170
360
398
237
170
14
1,300
170
5
25
20
369
6,300
11,000
190
1,300
3,300
100
45
55
12
1
50
580
110
85
28,000
660
19
40
40
25
28,800
10,000
210
100
260
1,600
5
2,400
500
43
100
100
79
50
Median
7,500
17.
55
220
10
10
7.5
10
10
55
0
10
7.7
95
65.5
7.5
97
10
230
10
31.5
15
10
0.1
42
38
22
10
8,300
270
•8
15
25
3
257
6,900
20
40
18
166
1
10
10
3.5
12.5
10
26
18
aThis table was developed by the U.S. EPA, Office of Solid Waste,  Economic
 Analysis Branch.

Source:  Reference No. 16.
                                  355
-------
 Compatibility of FMLs with MSW Leachate

      The only data on the compatibility of FMLs with MSW leachate found in
 the literature survey had been developed by Haxo et al (1, 2, 13).  In this
 work the major types of liner materials that were available and used for
 containment in the early 1970s were evaluated on exposure to MSW leachate for
 up to 56 months.  The leachate used in these studies was generated in pilot-
 scale MSW landfill simulators filled with a shredded residential solid waste.
 As the work was part of an exploratory research program performed before
 Method 9090 was developed, no immersion testing that approximates EPA Method
 9090 or other short-term exposures were conducted.  All of the exposure
 testing was performed at 23°C or less with leachate that was continually being
 generated in the landfill simulators.  Complete analyses of the leachate were
 not performed, though several parameters, i.e. solids, pH, COD, and total
 volatile acids (TVA), were followed.  Overall, the changes in the physical
 properties of the FMLs resulting from 56 months of exposure were relatively
 minor.   It should be noted,  however, that the leachate generated for this
 study probably did not contain aromatics, chlorinated hydrocarbons,  and other
 volatile organics that are known to affect hydrocarbon FMLs.   Consequently,
 these results may not be completely applicable to leachates being generated
 currently in active landfills.


 Preservation of  MSW Leachates for Use in Compatibility Tests

      An important factor in  conducting a compatibility test is  that  the
 composition of the test  leachate  reflects  the  composition  of  in  situ
 leachate.   Thus,  the  test  leachate should  maintain a constant~~concentration
 of those constituents  that could  affect an FML during  extended  service.
 Because MSW  leachate  is  highly oxidizable  and  unstable and  is subject to
 change  almost  immediately after removal from a service environment,  which
 generally is anaerobic,  methods of preserving,  stabilizing, or sterilizing
 leachates need to  be  developed  for possible use in  performing exposure  tests
 to assess the  compatibility of  FMLs  with MSW leachate.

     For analytical purposes, refrigeration has been used to protect leach-
 ates from bacteriological changes  before they  can be analyzed.   However, EPA
 Method  9090, the standard test  for determining the  compatibility of an  FML
 with a waste liquid, is  performed  at higher temperatures (23° and 50°C);
 therefore, refrigeration can only  be used to store  the  leachate  before use
 in exposure testing.

     No  information was  located in the open literature  indicating any methods
 of long-term stabilization of MSW leachates that could  be used in the com-
 patibility testing of FMLs at 23° and 50°C.

     Analysis of a single sample of stored MSW leachate indicated that stabi-
 lization of the leachate may be possible with sterilization.  A  1-gal bottle
of MSW leachate, which had been generated in October 1976 in a pilot-scale
simulator (1) and had been stored in a sealed brown glass bottle, was found
in September 1988 to be only slightly changed in composition since 1976.
                                      356
-------
There was no indication of volatile chlorinated hydrocarbon organics in
the leachate.

     If the EPA should require compatibility testing of FMLs with MSW leach-
ates in accordance with EPA Method 9090, additional research is needed to
develop adequate means of preserving or sterilizing MSW leachates to prevent
changes in leachate composition during the test, except for those changes
resulting from absorption by the FML under test.


Applicable Information on the Chemical Resistance of FMLs

     Solubility parameters are used in polymer science and technology to
assess the solubilities of polymers in different organic solvents (20).
Data generated in a recent study (21, 22) on the solubility parameters of FMLs
are applicable to assessing the compatibility of FMLs and MSW leachates if the
composition of the leachate is well established.  In that study, the solubility
parameters of polymers used in manufacturing FMLs were determined using their
equilibrium swelling in a variety of different-organics, the solubility
parameters of which are well documented in the literature (23, 20, 24, 25).
However, the solubility parameters of an FML are only one property of an FML
that can affect the magnitude of its swelling when it is in contact with a
leachate or waste liquid; for example, crosslinking, crystallinity, and filler
content of the FML compound can significantly affect the amount of swelling.


Applicable Information from Physical Chemistry on Dilute Solutions

     When a substance such as an organic solvent is added to a two-phase
system such as a polymer and water It will, in general, distribute at equi-
librium with different concentrations in the two phases (26).  The equi-
librium concentrations are related to relationship between the solubility
parameters of the phase and those of the solvent.  The ratio of the con-
centrations at equilibrium of the solute in the two phases, also known as the
distribution coefficient, remains essentially constant over a wide range of
concentrations.  Therefore, in the case of an organic distributed between
water and a polymer, a decrease in the concentration of the organic in the
aqueous phase would eventually result in a decrease in its concentration in
the polymer phase.  This characteristic has been found to be applicable to
dilute aqueous solutions in contact with a polyethylene FML.  Limited data
have been reported on the distribution coefficients of selected organics in
dilute aqueous solutions between the water and selected polyethylene FMLs
(22, 27, 28).


EXPERIMENTAL WORK ON DISTRIBUTION COEFFICIENTS OF FMLS

     On reviewing the data obtained in the literature survey, it was obvious
that there were areas of uncertainty which should be resolved before valid
recommendations could be made as to the type of testing needed to assess the
long-term compatibility of a specific FML with a specific MSW leachate.
                                   357
-------
      Inasmuch  as  only a limited  amount  of  experimental  work could  be  per-
 formed  in  this study, it was  decided  to measure  the  absorption  of  three
 organics by  different FMLs  from  dilute  aqueous solutions.   Samples of FMLs
 based on four  different polymers,  including  a linear low-density polyethyl-
 ene  (LLDPE), polyvinyl chloride  (PVC),  chlorinated polyethylene (CPE), and
 chlorosulfonated  polyethylene (CSPE), were placed in vapor-tight cells con-
 taining a  dilute, unsaturated aqueous solution containing  three different
 organics.  The concentrations of the  organics in the aqueous solutions were
 monitored  until they  had become  relatively constant  at  which time  the cells
 were  opened  and the FML samples  analyzed to  determine the  concentration of
 the organics in the FML samples.  Analyses of both the  solutions and  the FMLs
 for the organics were performed  using gas  chromatographic  (GC)  procedures.
 Experimental details  and results are  presented in this  section.


 Gas Chromatography Procedures

      A  Perkin-Elmer Sigma Three  Series  gas chromatograph with a flame ioni-
 zation  detector was used for  the GC analyses.  The instrument was  fitted with
 an open-capillary column coated  with polyethylene glycol.   Details  of the
 GC analyses  are presented in  Table 3.

      The concentrations  of  the organics in the aqueous  solutions were
 determined by  injecting  samples  removed from the test cells  directly  into
 the GC.  These cells  featured septums through which  the aqueous  solutions
 could be sampled during  the test to determine whether equilibrium had been
 reached.

      The concentrations  of  the organics in the FML samples were  determined
 by headspace GC.  In  this procedure, an FML  specimen containing  absorbed
 organics is placed in a  small vapor-tight  can provided with  a septum
 through which  vapors  from the specimen  can be sampled.  The  can  is  placed
 in an oven at  105°C and  heated for approximately an  hour.  A sample of the
vapors  is drawn from  the  can  and injected  into the GC.  The  FML  specimen is
 removed from the sampled  can  and placed in a new can which is then heated
 in a  105°C oven for approximately  an hour.   Once again, the  vapors inside
 the can are sampled and  injected into the  GC.  The process of heating,
 sampling, and  injecting  is repeated until  no organics are detected  in the
 sampled vapors  by the  GC.  The amount for  each organic were  summed  to
represent a total for  the organic  in the sample.

      The concentrations  of the organics in the injected samples  were
calculated by  comparing  peak  height values resulting from the GC analyses
with  calibration curves.  The calibration  curves for the aqueous solutions
were  determined by injecting  1 /*L  of various solutions of known  concentra-
tions of the different organics  into the GC  column.  Injections  of each
standard were  performed  five  times to ensure reproducibility  of  injection
techniques.  The calibration  curves for the  headspace GC analyses were
prepared by analyzing  a  specific 100 /xL volume of vapor from headspace cans
filled with different  amounts of the organics.
                                      358
-------
          TABLE 3.  GAS CHROMATOGRAPHY CONDITIONS FOR LEACHATE
                           AND VAPOR ANALYSIS
            Condition
            Setting
Detector range

Injector and detector temperature

Initial temperature

Initial holding time

Final temperature

Final holding time

Temperature program rate

Detector



Column^
Chart speed

Carrier

Specimen size:
  Liquid
  Vapor from headspace

Attenuation
x 10

250°C

60°C

1 min.

120°Ca

1 min.

20°C/min.

Flame ionization:
  H2 30 cc/min.
  02 40 cc/min.

Polyethylene glycol coated
fused silica open tube
(FSOT) capillary: 0.53 mm
in diameter and 15 m in
length

30 cm/hour

Helium, 10 cc/min.
1 ML
100 /uL

32 (for 500 ppm concentration)
down to 4 (for low concentration)
aCan vary with the solvent mixture from 120°C to 260°C, the maximum
 temperature for the column.                        ,
^Trade name Superox (Altec), Megabore DB WAX (J and W).
                                   359
-------
 Selection of Volatile Organics

      For this study, three types of volatile organics representing a range
 of chemical characteristics were desired, including a volatile containing
 oxygen,  a volatile that was a chlorinated solvent, and a volatile that was
 an aromatic.  The three selected were methyl ethyl ketone (MEK), trichloro-
 ethylene (TCE),  and toluene.  All of these organics have been observed in
 MSW leachates.  Properties of these solvents are presented in Table 4.
            TABLE 4.   ORGANICS USED IN ABSORPTION EXPERIMENTS
                      WITH DILUTE AQUEOUS SOLUTIONS
Organic
Toluene
Trichloroethylene
Methyl ethyl ketone
Mole-
cular
weight
92.13
31.40
72.10
Density
at 20°C,
K cm""
0.866
1.476
0.805
Boiling
point ,
°C
110.6
87.2
79.6
Vapor
pressure
at 25°C,
mm Hg
31.96
80.30
100.0
«o
8.9
9.2
9.3
Solubility
parameters3
<5d See Reference No. 29.
 cChemical Abstract Services' number.
Immersion Test Cells

     The immersions were  performed in vapor-tight cells consisting of 8-oz
glass jars with ground  polished top edges, and Teflon-lined phenolic resin
tops (Figure  1).  Each  of these tops was fitted with a Swagelock sampling
port and a Teflon-lined silicone rubber septum for withdrawing samples for
GC analysis.  Special arrangements were made in the cells to suspend the FML
samples in the solutions.
Experimental Procedure

     The work performed  in  this  study developed out of previously reported
work (22, 27, 28).  In the  earlier  study,  samples of a polyethylene FML were
placed in the same type  of  test  cells as used in the present work.  These
cells were then filled with a  series  of dilute but saturated aqueous solu-
tions, each of which contained a single organic.  These solutions were
prepared with an excess  of  the organics to maintain saturation.   After
analyzing the aqueous solutions  and FML samples at the end of the immersions,
the distribution coefficients  (i.e. the ratio of the concentration of the
organic in the FML to its concentration in the aqueous solution)  were"cal-
culated.  It was noted,  however,  that the  amounts of organics absorbed by the
FML approached the amounts  the FML  absorbed when immersed in neat organics.
It appears that the water had  acted as a permeable medium between the FML and
                                      360
-------
the reservoir of excess organics which allowed the organics to be absorbed by
the FML until saturation of the FML was reached.  These conditions are not
representative of those in a landfill due to the improbability of an excess
of the organic being maintained until saturation was reached in the installed
FML.
              Teflon
              Septum
                  \
   Teflon-lined
   Screw cap
                            Swagelock
                            Assembly
         Washer
                          Nut
             TOP ASSEMBLY
Jar  with
ground and
polished
edge
                                                 8 OZ JAR
           Figure 1.  Schematic of the immersion test cell.
     In the experimental work conducted in the present study, the aqueous
solutions were prepared at concentrations less than saturation and the
number of FMLs in test were increased to include the four basic polymer
types.  Four immersion test cells were each filled with an aqueous solution
containing a mixture of 500 ppm (on a weight basis) each of MEK, TCE., and
toluene.  Samples of CPE, fabric-reinforced chlorosulfonated polyethylene
(CSPE-R), LLDPE, and PVC FMLs were placed in separate test cells, and the
concentrations of the volatile organics in each cell were followed by GC.
In all cases, the initial concentrations of the MEK in solution showed
little change and, thus, little distribution to the FMLs, whereas the
toluene and TCE showed drops in concentration and thus a transfer from the
water to the FMLs.  The cells were dismantled and the respective FMLs
were analyzed for the volatile organics by headspace GC.


Experimental Results

     The results of analyzing the aqueous solutions and the FMLs at the end
of the immersion are presented in Table 5.  The results show that there was
a significant increase in the weight of the FML samples; however, the dif-
ference between the weight gains of the FML samples and the total amount
of organics detected in each FML sample indicates that not all of the weight
gains could be attributed to the absorption of the organics.  Thus, consider-
able water had also been absorbed by the specimens.  In addition, it is also
                                     361
-------
TABLE 5.  ABSORPTION BY IMMERSED FMLS OF A MIXTURE OF ORGANICS3
                  FROM DILUTEb AQUEOUS SOLUTIONS
Parameter
Contents of cell at beginning
of experiment
Amount of water in cell, g
Amount of organics in cell:
MEK, mg
TCE, mg
Toluene, mg
Total
Original weight of FML
specimen, g
FML specimen at end of
immersion
Swollen weight , g
Weight gain, g
Weight gain, %
Headspace analysis of
swollen FML specimen
Amount in swollen FML
specimen:
MEK, mg
TCE, mg
Toluene , mg
Total
% of original weight
of FML
Concentration of organics
in swollen FML specimen
MEK, ppm
TCE, ppm
Toluene , ppm
PVCC
579


225.5

112.8
112.8
112.8
338.4

3.4580


3.668
0.210
6.1




1.35
54.0
55.6
110.95

3.21


370
14,720
15,200
LLDPEC
580


226.6

113.3
113.3
113.3
339.9

2.5337


2.626
0.093
3.7




0.08
36.3
44.5 .
80.88

3.19


30
13,800
16,900
CPEC
581


228.4

114.2
114.2
114.2
342.6

5.4670


6.246
0.779
14.2




1.16
63.7
65.5
130.36

2.38


190
10,200
10,490
CSPE-RC
582


225.2

112.6
112.6
112.6
337.8

6.1601


6.702
0.542
8.8




0.55
34.3
38.8
73.65

1.20


80
5,120
5,790
                                                      continued  .
                              362
-------
                            TABLE 5.   CONTINUED
Parameter
FML specimen after head-
space analysis
PVCC
579

LLDPEC
580

CPEC
581

CSPE-RC
582

 Weight.g                      3.4501     2.5310      5.4551       6.1253

 Loss in weight  (based on
    original weight),  %           0.23       0.11        0.22         0.56
at end of experiment
Amount in aqueous solution:
MEK, mg
TCE, mg
Toluene , mg
Total
Concentration of organics
in aqueous solution (CH_Q):
MEK, ppm
TCE, ppm
Toluene , ppm
Total amount of organics
in celld:
MEK, mg
TCE, mg
Toluene , mg
Total
Organic accounted for at
end of experiment , %
Distribution coefficient
(CFML/CH2o):
MEK
TCE
Toluene


112.5
48.3
39.2
200.0


500
214
174


114
102
95
311

92


0.74
68.8
87.4


110.1
87.5
80.4
278.0


486
386
355


110
124
125
359

106


0.06
35i.8
47.6


91.4
41.6
27.9
160.9


400
182
122


93
105
93
291

85


0.48
56.0
86.0


111
46.6
38.1
195.7


493
207
169


111
81
77
269

80


0.16
24.7
34.3
aMethyl ethyl ketone (MEK), trichloroethylene (TCE), and toluene.
CPVC » polyvinyl chloride; LLDPE = linear low-density polyethylene;
 CPE = chlorinated polyethylene; CSPE-R = chlorosulfonated polyethylene
 (fabric-reinforced).
^Initial concentration of each organic was 500 ppm.
     of organics in water and in the FML.
                                    363
-------
possible that some of the organics not accounted for at the end of the experi-
ment had been absorbed by the FML but were not recovered during the headspace
GC procedure.  Regardless, the results do show a large partitioning of the
organics from the water to the FMLs and, furthermore, they show that the FMLs
varied considerably in their absorption of the different organics.  The total
amounts of organics and water that were absorbed ranged from 3.7% for the
polyethylene to 14.2% for the CPE.  This experiment was repeated with similar
overall results.  The results of very limited physical testing, which was
performed on percut specimens immersed in the second experiment, consistently
indicated moderate losses in tensile property values.


CONCLUSIONS

     The results of the literature survey and the limited experimental work
that was conducted in this study indicate that:

     o  There is little information in the literature regarding the compo-
        sition and characteristics of current generation leachates and the
        compatibility of FMLs with these leachates.

     o  The limited data available indicate that MSW leachates currently
        being generated may contain low concentrations of organics having low
        biodegradability; in addition, depending on their concentration, some
        of these organics are aggressive towards FMLs.

     o  It is questionable whether EPA Method 9090, as presently designed,
        can result in a consistent and realistic assessment of the compati-
        bility of an FML with a MSW leachate without modification of the
        exposure cells to maintain anaerobic conditions and a standard means
        of assuring stabilization of the leachates so that little change in
        concentrations of the constituents during the required four months is
        ensured.

     o  The use of distribution coefficients that represent the distribution
        of dissolved organics between aquous and polymeric phases appears to
        be applicable to assessing the compatibility of FMLs  and other poly-
        meric construction materials with MSW leachates,  which are basically
        dilute aqueous solutions of organic and nonorganic solutes.  Thus, the
        amount of organics absorbed by an FML from a dilute aqueous solution
        with which it is in contact may be estimated.  For example, at low
        concentrations of organics in solution, the amount of the organics
        that an FML will absorb will diminish as the concentration in the
        leachate is lowered.

     o  The solubility parameters of an FML and the individual organic are
        important tools in assessing the level of absorption  of the organics
        by the FML and in determining the distribution coefficient of an
        organic between an aqueous solution and an FML.
                                    364
-------
RECOMMENDATIONS

     It is recommended that further experimental research be performed:

     o  To develop quantitative data regarding the distribution of dissolved
        organics between various FMLs and aqueous solutions or actual MSW
        leachates in which the FMLs are immersed.

     o  To determine threshold concentrations of various organics found
        in MSW leachates, below which levels absorption will not signif-
        icantly affect the properties of FMLs and other materials of con-
        struction used in liner systems.

     o  To assess possible interactions between organics in dilute aqueous
        solutions and the effects of the absorbed organics on permeability
        and properties of FMLs.

     o  To perform a round-robin series of exposure tests performed in
        accordance with EPA Method 9090 to establish error and bias in
        the test method.

     o  To perform more in-depth analyses of MSW leachates than have
        normally been made to determine the presence of specific organics
        that swell various FMLs.  These organics include chlorinated
        solvents, aromatic solvents, and some aliphatic solvents, all of
        which which are known to deteriorate the properties of polymeric
        compositions.
ACKNOWLEDGMENT

     The work on this project was performed by Matrecon under Work Assignment
No. 0/01 of Subcontract 771-87 of Environmental Protection Agency Contract
68-03-3413 to PEI Associates, Inc.
REFERENCES

 1.  Haxo, H. E., R. M. White, P. D. Haxo, and M. A. Fong.  1982.  Final
     Report: Evaluation of Liner Materials Exposed to Municipal Solid Waste
     Leachate.  NTIS No. PB 83-147-801.  U.S. Environmental Protection
     Agency, Cincinnati, OH.

 2.  Matrecon, Inc.  1988.  Lining of Waste Containment and Other Impoundment
     Facilities.  EPA-600/2-88-052. [SW-870, second revised edition.]  U.S.
     Environmental Protection Agency, Washington, B.C. 991 pp.

 3.  Chian, E. S. K., and F. B. DeWalle.  1976.  Analytical Methodologies for
     Leachate and Gas Analysis.  In: Proceedings of a Research Symposium on
     Gas and Leachate from Landfills: Formation, Collection, and Treatment.
     EPA-600/9-76-004 (NTIS No. PB-251-161).  U.S. Environmental Protection
     Agency, Cincinnati, OH.  pp 44-53.
                                    365
-------
 4.  Chian, E. S. K., and F. B. DeWalle.  1977.  Evaluation of Leachate
     Treatment.  2 volumes.  EPA-600/2-77-186 a,b.  U.S. Environmental
     Protection Agency, Cincinnati, OH.

 5.  Dunlap, W. J., D. C. Shew, J. M. Robertson, and C. R. Toussaint.  1976.
     Organic Pollutants Contributed to Groundwater by a Landfill.  In:
     Proceedings of a Research Symposium on Gas and Leachate from Land-
     fills: Formation, Collection, and Treatment.  EPA-600/9-76-004 (NTIS
     No. PB-251-161).  U.S. Environmental Protection Agency, Cincinnati,
     OH.  pp 96-110.

 6.  Ham, R. K.  1975.  Milled Refuse Landfill Studies at Pompano Beach, FL.
     Approx. Range, Three Cells Aged One Year.  21 pp.

 7.  Ham, R. K.  1976.  Solid Waste Degradation Due to Shredding and Sludge
     Addition.  In: Proceedings of a Research Symposium on Gas and Leachate
     from Landfills: Formation, Collection, and Treatment.  EPA-600/9-76-
     004 (NTIS No. PB-251-161).  U.S. Environmental Protection Agency,
     Cincinnati, OH.  pp 168-176.

 8.  Ham, R. K., K. Hekimian, S. Katten, W. J. Lockman, R. J. Lofty, D. E.
     McFaddin, and E. J. Daley.  1979.  Recovery, Processing, and Utili-
     zation of Gas from Sanitary Landfills.  EPA-600/2-79-001.  U.S.
     Environmental Protection Agency, Cincinnati, OH.  133 pp.

 9.  Pohland, F. G.  1975.  Sanitary Landfill Stabilization and Leachate
     Recycle and Residual Treatment.  EPA-600/2-75-043.  U.S. Environmental
     Protection Agency, Cincinnati, OH.  105 pp.

10.  Pohland, F. G.  1976.  Landfill Management with Leachate Recycle and
     Treatment: An Overview.  In: Proceedings of a Research Symposium on Gas
     and Leachate from Landfills Formation, Collection, and Treatment.
     EPA-600/9-76-004 (NTIS No. PB-251-161).  U.S. Environmental Protection
     Agency, Cincinnati, OH.  pp 159-167.

11.  Pohland, F. G., D. E. Shank, R. E. Benson, and H, H. Timmerman.  1979.
     Pilot Scale Investigations of Accelerated Landfill Stabilization with
     Leachate Recycle.  In: Municipal Solid Waste: Land Disposal.  Proc.
     5th Annual Res. Sympos.  EPA-600/9-79-023a.  U.S. Environmental Pro-
     tection Agency, Cincinnati, OH.  pp 283-295.

12.  Pohland, F. G., W. H. Cross, and J. P. Gould.  1987.  The Behavior and
     Assimilation of Organic and Inorganic Priority Pollutants Codisposed
     with Municipal Refuse - A Progress Report.  In: Proceedings of the
     Thirteenth Annual Research Symposium: Land Disposal of Hazardous Waste.
     EPA-600/9-87-015.  U.S. Environmental Protection Agency, Cincinnati, OH.
     pp 26-37.

13.  Matrecon, Inc.  1983.  Lining of Waste Impoundment and Disposal Facili-
     ties.  SW-870 Revised.  U.S. Environmental Protection Agency, Washington,
     D.C.  448 pp.
                                   366
-------
14.  Kmet, P., and P. M. McGinley.   1982.  Chemical Characteristics of Leach-
     ate from Municipal Solid Waste  Landfills in Wisconsin.  In: Proceedings
     of the 5th Annual Madison Conference of Applied Research and Practice on
     Municipal and Industrial Wastes, September 22-24, 1982.  Dept. of Eng.
     and Applied Science, University of Wisconsin Extension, Madison, WI.
     pp 225-254.

15.  EPA.  1986.  Muncipal Landfill  Case Studies (unpublished).  These
     studies were prepared by PEI, SRW, and ICF.  U.S. Environmental Pro-
     tection Agency, Office of Solid Waste, Washington, D.C.  Cited in: U.S.
     EPA.  1986.  Subtitle D Study - Phase I Report.  EPA/530-SW-86-054.
     U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
     D.C.

16.  EPA.  1986.  Subtitle D Study - Phase I Report.  EPA/530-SW-86-054.
     U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
     D.C.
17.  Wigh, R. J.  1979.  Boone County Field Site.  Interim Report, Test Cells
     2A, 2B, 2C, and 2D.  EPA-600/2-79-058.  U.S. Environmental Protection
     Agency, Cincinnati, OH.  202 pp.  (NTIS PB-299-689).
18.
19.
20.
21.
22.
23.
Breland, C. G.   1972.  Landfill Stabilization with Leachate Recir-
culation, Neutralization, and Sludge Seeding.  CE-756A6.   School of
Civil Engineering, Georgia Institute of Technology, Atlanta, GA.  80
pp.

Griffin, R. A.,  and N. F. Shimp.   1978.  Attenuation of Pollutants in
Municipal Landfill Leachate by Clay Minerals.  EPA 600/2-78-157 (NTIS
PB 287-140).  U.S. Environmental Protection Agency, Cincinnati, OH.
146 pp.

Barton, A. F. M.  1983.  Solubility Parameters and Other Cohesion
Parameters Handbook.  CRC Press, Boca Raton, FL.

Haxo, H. E., and P. J. Pick.  1986.  Determination of the  Solubility
Parameters of FMLs for Use in Assessing Resistance to Organics.  In:
Proceedings of the Twelfth Annual  Solid Waste Research Symposium: Land
Disposal, Remedial Action, Incineration and Treatment of Hazardous
Waste.  EPA/600/9-86/022.  U.S. Environmental Protection Agency,
Cincinnati, OH.  pp 132-145.

Haxo, H. E., T.  P. Lahey, and M. L. Rosenberg.  1988.  Factors in Asses-
sing the Compatibility of FMLs and Waste Liquids.  EPA/600/2-88/017
(NTIS No. PB 88-173-372/AS).  U.S. Environmental Protection Agency,
Cincinnati, OH.  143 pp.
Barton, A. F. M.
75(6):731-753.
1975.   Solubility Parameters.  Chemical Reviews
                                    367
-------
24.  Leo, A. and C. Hansch.  1970.  Linear Free-Energy Relationships Between
     Partitioning Solvents Systems.  Journal of Organic Chemistry 36(11):
     1539-1544.

25.  Leo, A., C. Hansch, and D. Elkins.  1971.  Partition Coefficients and
     Their Uses.  Chemical Reviews, Vol. 71, No. 6.  pp. 525-554.

26.  Daniels, F., and R. A. Alberty.  1961.  Physical Chemistry.  2nd
     Edition.  John Wiley and Sons, NY.

27.  Haxo, H. E.  1988.  Transport of Dissolved Organics from Dilute Aqueous
     Solutions Through Flexible Membrane Liners.  In: Proceedings of the
     Fourteenth Annual Solid Waste Research Symposium: Land Disposal,
     Remedial Action, Incineration and Treatment of Hazardous Waste, May
     9-11, 1988.  U.S. Environmental Protection Agency, Cincinnati, OH.
     21 pp.

28.  Haxo, H. E., and T. P. Lahey.  1988.  Transport of Dissolved Organics
     from Dilute Aqueous Solutions Through Flexible Membrane Liners.
     Hazardous Waste and Hazardous Materials.  5(4):275-294.

29.  Riddick, J., and W. Bunger.   1970.  Techniques of Chemistry, Volume
     II - Organic Solvents, Physical Properties and Methods of Purification.
     Wiley-Interscience, NY.
                                    368
-------
        GEOSYNTHETIC CONCERNS  IN LANDFILL LINER AND COLLECTION SYSTEMS

        by: Robert M. Koerner, Arthur E. Lord,  Jr.,  and Yick H.  Halse
           Geosynthetic Research Institute
           Drexel University
           Philadelphia, PA  19104
                                  ABSTRACT

     The  use of geosynthetic materials in landfill liner and closure  systems
is commonplace. This use includes the containment  of  all types of solid waste
materials,   e.g.,  hazardous,  municipal,   industrial,   ash,   low  level
radioactive, and heap leach, ores. The geosynthetic materials involved  include
the following;
     •  geomembranes  or flexible membrane liners (FMLs),
     •  geotextiles as filters and cushions,
     •  geonets  as drains,
     •  geogrid  reforcing elements, and
     •  geocomposites for various  uses.

While it is felt that current systems can indeed be designed and constructed
with confidence, there are  certain  aspects  in need of further investigation
and/or clarification. This paper highlights several of them. Included are an
investigation of FML behavior in anchor  trenches,  stress cracking behavior of
HOPE seams,  flow  rate  reduction in  geonets,  and  particulate  and biological
clogging of geotextile filters.  Some amount  of  data is presented  on each
topic but,   clearly,  the paper represents research-in-progress from the point
of view  of  definitive guidelines.  The  goal  of each  individual  topic is to
generate  a data  base  so  as  to   avoid long-term  problems  when  using
geosynthetics in association with waste  containment facilities.

                          INTRODUCTION AND OVERVIEW

     Current waste  containment  facilities  represent  a bevy of geosynthetic
materials.  Included in the group  are  FMLs  which  form  the primary and/or
secondary liners against escaping liquids, geonets which are commonly used as
drainage materials  in  primary and/or secondary leachate  collection systems
and  geotextiles  used as  filters to allow leachate  to enter  into  drainage
collectors.  When the waste containment facility is doubly lined,  as are many
hazardous,  municipal and industrial landfills, the cross  section of Figure 1
                                    369
-------
is common. Here one can itemize the following geosynthetics and their primary
functions (progressing from the waste downward);
     •  geotextile filter beneath the waste,
     •  geonet as a  primary  leachate collector  on side slopes,
     •  geotextile filter around perforated pipe drains,
     •  geotextile cushion above primary FML,
     •  the primary  FML (i.e.,  a geomembrane),
     •  geotextile separator between primary clay and geonet drain,
     •  geonet as secondary  leachate collector,
     •  the secondary FML  (i.e., a geomembrane), and
     •  geotextile cushion against soil subgrade

Thus as many as nine (9)  geosynthetics are  used  in  the liner system shown.
When adding the various geosynthetics placed  in the cap, or closure, of such
facilities,   it  is  easily seen that  these materials  play a key  role in the
proper functioning of the  system.  This  is tantamount  to  stating that they
must be properly evaluated, selected,  designed, and installed.

     Toward  this end we have selected to investigate  a number of areas which
are  felt  to  be of  concern.  They are = identified on  Figure  1  and will be
described in the sections to follow.

1.   FML Behavior  in Anchor Trenches - As shown in Figure 1 at Location "1",
FMLs will typically come out  of  a lined facility and terminate around the
perimeter of  the site. There  is  a short horizontal runout length and then a
vertical drop into an anchor trench.  The lengths involved can be  estimated by
available design models (1,2)  but their evaluation awaits field verification.
As an  intermediate step  we have constructed  a large  scale test facility to
model  the situation in  the  laboratory.  A  steel reinforced wooden  box of
internal size 3 ft. height by 3ft.  width by 6 ft.  length has been constructed
which  can contain various  anchor  trench configurations,  see  Figure 2. The
jacking system to which the outside of the FML is  attached  can be oriented so
as to  exert  a downward force thereby simulating an actual situation. Normal
Stress is exerted by an air-bag system to mobilize overburden pressures up to
1500  lb/ft2.  While not necessary  for the perimeter  anchor  trench shown in
Figure 1, these high normal stresses are mobilized on the interior berms of
discrete cells when a zoned landfill expands laterally and  then vertically.

     Strain  gages  are attached to  the  FML test specimens at four locations.
The initial data indicate that the largest FML stresses are in the portion of
the FML closest to the pullout force. These  stresses  rapidly dissipate with
increasing  distance  into  the test box  as seen  in  the curves  of  Figure 2
moving from  locations Gl  through G4.  Obviously,   soil friction  plays an
important role in  this  stress dissipation.  Also note that the FML in the
vertical anchor trench is essentially non-stressed, i.e. gage number G4. This
has been  typical of tests  conducted to  date.  The  planned experiments  include
30 and 60 mil HDPE,  40 mil LLDPE, 30 mil PVC and 36 mil CSPE-R.  Normal stress
will  vary from 100  lb/ft2  to 1500 lb/ft2. Results will be  compared to the
various  design models  so as to  verify,  modify  or refute  the  existing
literature.
                                     370
-------
2.   Stress Cracking Behavior  of HOPE Seams - ASTM  defines  stress cracking
as,  "an internal or external rupture in  a  plastic  caused by tensile stress
less than  its short-term mechanical strength."  Thus  any mechanism which cause
premature  failure  of  the FML  can fall  into  this  general  category,  e.g.,
environmental stress cracking,  residual stress  cracking,  scratches  and cracks
resulting  from grinding  marks,  geometrical irregularities,  and/or fatigue
failure. Many  (if not  all) of these situations are encountered at HDPE liner
seam locations.

     An evaluation  of  HDPE seams  was  initially addressed using  a modified
version of  ASTM D-2552  test  method  which immerses dumbbell-shaped  test
specimens  in a surface wetting agent at an  elevated temperature. A number of
HDPE  seam types were evaluated  under  different   constant stress  levels
resulting  in a large  number that  cracked  (3), see Tables l(a) and (b)  for
test results of 168 and 1000 hour test durations,  respectively.
     TABLE l(a)  - LABORATORY STRESS CRACKING RESULTS FOR 168 HR.  DURATION
Type of Seam
fillet extrusion
flat extrusion
hot wedge
hot air
ultrasonic
Number of
Tests
179
40
60
80
40
Elastic
74
15
25
27
28
Results
Plastic
13
19
19
22
5
Cracked
92
6
15
31
7
Percent
Cracked
51
15
27
39
25
    TABLE l(b) - LABORATORY STRESS CRACKING RESULTS FOR 1000 HR. DURATION
Type of Seam

fillet extrusion
flat extrusion
hot wedge
Number of
Tests
20
20
20

Elastic
14
11
8
Results
Plastic
1
0
0

Cracked
5
9
12
Percent
Cracked
25
45
60
To be  emphasized is  that these  laboratory tests are  index tests  wherein
stress relaxation of  the polymer cannot  occur  and that they are  conducted
under very harsh conditions.  The  extent of the problem in the  field is  not
known,  although it is known to exist at a few surface impoundments  where  the
FML was exposed to the atmosphere <4>, see Figure  1  at  Location "2".

    Work is  currently ongoing  in  evaluating  various  seaming procedures
(e.g.,  temperature,  pressure and  time),  determining  residual  stresses,
evaluating cracking initiation mechanisms, measuring crack growth  rate in
various  HDPE  formulations, and  attempting to  optimize  seam geometry.  The
ultimate goal is to  assess the magnitude of the  stress cracking  problem,  its
optimization and/or  its remediation.
                                     371
-------
3.   Flow Rate Reduction of Geonets -  The  intrusion  of a geotextile  (or any
other material) into the flow apertures of a drainage net will surely reduce
its flow rate capability. While a number of cross  sections can be envisioned,
the  situation of  a clay over  a geonet with a  geotextile  separator between
them is of great concern, see Figure 1  at Location "3".

     Flow rate tests using  a  cross  section  consisting of  a geonet between two
60 mil HDPE liners versus the same type of  geonet  between a  60 mil HOPE liner
beneath  it  and a  geotextile/clay layer above  it have  been  conducted.  The
geotextile used as the  separator was  a needle punched, continuous filament,
nonwoven  fabric of 8.0  oz/yd2 mass per unit  area.  The test  procedure was
performed in  accordance with ASTM D-4716. Results are given in Table 2 where
one  can  see  flow  rate decreases of 19%  to  41%  depending upon the hydraulic
gradient and applied normal stress. Clearly,  intrusion  of the geotextile/clay
is occurring as it spans  the open apertures  of  the geonet. Note  that the
geotextile  acts in a  membrane  reinforcement  mode and  must  do so  for the
design lifetime of the  system.  If it  fails,  the clay will immediate extrude
into the  geonet thereby blocking  all  flow.  Thus long term creep  tests are
warranted for many situations.

  TABLE 2 - FLOW RATES (IN GAL/MIN-FT)  AND  REDUCTIONS  (IN %) FOR DIFFERENT
                       GEONET DRAINAGE CROSS SECTIONS
Normal
e>4-
(Ib/ft2)
5000



10,000



15,000



Cross


HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
Hydraulic Gradient "i"

i = 0.25
1.6
JUJi
0.3
19%
1.4
•J-l
0.3
21%
1.2
ILu2
0.3
25%

i = 0.50
3.7
2^1
1.0
27%
3.4
2^3.
1.1
32%
3.2
JU&
1.4
44%

i = 1.00
7.1
±*&
2.3
32%
6.4
AJ.
2.3
36%
5.6
3.3
2.3
41%
 4.   Parfciculate  Clogging  of  Geotextile  Filters  -  Shown in  Figure 1  at
 Locations  "4"  is  a geotextile covering a primary leachate  collection  geonet
 on  the  sideslope and  drainage  gravel on  the bottom.  In both  cases  the
 geotextile must act as a filter. Hence, it must provide adequate  permeability
 (thus  open voids) and  soil retention  (thus tight  voids),  i.e., a  balanced
 void structure must be achieved. While these considerations  can be  adequately
 handled by proper design,  there  is  a third necessary condition, that  being
 prevention of long-term clogging.
                                     372
-------
    The  customary approach toward the assessment of geotextile  clogging is
by the  gradient  ratio  test,  CW-02115.  This  U.S. Army  Corps of  Engineers
developed  test,  however,  was  directed  toward  sandy  soils  and  woven,
monofilament geotextiles of relatively high open areas (5). For the very fine
sediment carried by leachate,  the authors prefer the  use  of long-term column
tests  (6,7).  Here the  actual  cross section  being  considered is  simulated
using  site-specific cover  soil,  the  candidate geotextile  and actual  (or
simulated)  leachate.  The  long  term  flow through  this  cross-section  is
measured over a period of time. As seen in Figure 3,  the  flow rate initially
decreases until a transition time is reached.  This is due to soil compaction
and an  initial  "tuning" of the  geotextile  to  the  upstream soil and  the
permeating liquid. Beyond this point, the  flow  rate  either becomes constant
(thus  equilibrium is  established),  becomes  uncertain  (which  requires
continued testing), or decreases to  zero  (hence the system is clogged).

    A  broad base study is ongoing in this  regard  using four  geotextiles
(woven  monofilament,   nonwoven  melt-bonded,  lightweight  needle-punched
nonwoven, and heavyweight needle-punched  nonwoven); four conditions above the
geotextile  (no soil, Ottawa sand, clayey silt and a  combination  of sand and
clayey silt);  and two  permeating liquids  (water  and sediment  laden water).
Upon establishing one  of the trends shown  in Figure 3,  the  system will be
epoxy-set,  properly sectioned and microscopically viewed  so as to understand
the mechanisms involved. The goal of this  study  is a laboratory test  matrix
illustrating the behavior and proper design against fine particulate clogging
of geotextile filters.

5.  Biological  Clogging  of  Geotextile Filters  -  To  be sure,  municipal
landfill leachates  have many viable (active)  bacteria present in  them,  see
Figure 4. The geotextile filters  shown  in Figure 1 at Locations "5" interface
with these bacteria. Columns of the type shown in Figure  3 have  been  set up
at six  municipal landfills under aerobic  conditions resulting;  in the flow
rate trends shown in Table 3(a). Relatively large  reductions over the elapsed
times are indicated, e.g.,  50%  reduction  (or higher)  occurred in one-third of
the test materials. A separate  phase of the study was also aimed at anaerobic
bacteria clogging of  geotextiles  which produced the results  of  Table 3 (b) .
Here the flow  rate reductions  are much less  than noted previously, but are
nonetheless quantifiable in all cases of 7 months or  longer.  See reference 9
for additional details.

    This  study is also ongoing into a second phase  where  backflushing with
water and  with biocide-treated water will be performed.  The  resulting flow
rate response will be evaluated.  If  flow  is reconstituted to its original, or
near to original,  flow  rate  the  recommendation will  be  forthcoming  to
incorporate such  procedures  into  the  design of  the  collection  system.  Such
designs are certainly possible  and,  if  justified, should be implemented.

                           SUMMARY AND CONCLUSIONS

    Numerous types  of geosynthetic materials are currently seeing widespread
use in  waste containment facilities.  Use  of these  materials  has led to  a
certain  degree of  confidence which  is augmented by the installation  of
redundant  liner  and  leachate  collection  systems.  Due  to their   rapid
                                      373
-------
 co

 in
 m
 d>
-H  r-i
 S   W
    •P
 w   to
\p   a)
 (0  4J
 0)
4J  T)
     (0
 0)   0)

^•^
 M   tn

 S  -H
 O  H
H  H

           Cu  Cn
           Cu  o
           I  rH
                                    id H
                                    *  3
                                       <4H
                                    o
                                   -H -P
                                    1-1  10
                                   X>  <1)
                                    (0 JJ
                                   <4H *
                                                                             H
                                                      fa

                                                      O
                                                      HI
                                                         co
                                                                            CO
                                                          «.
                                                                                                       I   rH
                                                                                                             df>  dP O  O
                                                                                                      IT) O  O  O

                                                                                                       I  rH  t-l  H

                                                                                                          I   I    I
                                                      dp dp dp  dp O  U
                                                      m o o  o --^  ~~.
                                                       I  rH rH  rH S  2
                                                                                   O  O


                                                                                   S3 ^
                                                                                                      dP dp  6IP o\»
                                                                                                      o in  o o
                                                                                                      I   I    I   I
                                                                                       co  to  w   to
                                                                    rH  O
                                                                    rH  rH CT>
                                                                                                                    in
                                                                                                      r- r-
                                                                                                             00 CO  00  00
                                                                                                      rH rH  in
                                                                                                                00  OO
                                                                                                                        OO
                                                                                                             eg in  vo  CN
                                                                                                     rH  H  rH
                                                                                                                    VO
                                                      rH eg  ro i1  in  vo
                                                      I   I    I   I   I    I.
                                                                                                                                a<
                                                                                                         CO
                                                                                                                                Q.
                                                                                                                               •o
                                                                                                                               o
                                                                                                                               II   M
                                                                                                                               _  <1>
                                                                                                                               B  4J
                                                                                                                                   a.
                                                                                                                                  H
                                                                                                                               IS CU
                                                               374
-------
development,  however,  there are areas where additional inquiry is warranted.
Some of these areas have been the  focus of this paper.

     •  FML  anchor trench behavior  and mobilization mechanisms will soon have
       a  meaningful  data  base via  large  scale laboratory evaluation.  A
       rigorous critique  of available design methods will result.

     •  Stress cracking of HDPE field  seams  appears to be strongly related to
       the  workmanship involved  in  making  the  seam.  Strict  construction
       quality control (10)  should alleviate the concern for most situations
       except possibly the exposed berms  of surface impoundments.  In  these
       areas   proper  design  with  sufficient  expansion  and contraction
       capability must  be  considered.(11)

     •  Flow rate  reduction of  geonets  via  intrusion of materials  above  and
       below  the  net's apertures  or openings  is very  quantifiable.  Proper
       design of  the  geotextile  or geomembrane  above  and  below the net is
       necessary. The  laboratory simulation test  is  available  to  assess  the
       adequacy of  the  design.

     •  Fine  precipitate  clogging of  geotextile  filters  associated with
       leachate  collection systems  is of concern for those  leachates with
       high particulate content. A laboratory simulation test is proposed  and
       a  broad based series of  experiments is ongoing. If such clogging is of
       serious concern, backflushing might  be required for  remediation  and
       reconstitution  of  flow.

     •  Biological  clogging  of geotextile filters associated  with  leachate
       collection  systems is  a  likely  concern for  municipal  landfills.
       Clearly,  biocide  treatment either within  the geotextiles or  in  the
       backflushing liquid as proposed above are possible remedial measures.
       Work is ongoing to evaluate these various remediation schemes.

                              ACKNOWLEDGEMENT S

     This project is funded by  the U.S. Environmental Protection Agency under
Project No. CR-813953-01. Our sincere appreciation is extended to the Agency
and in particular to the Project Officer, Robert  E. Landreth.

                                 REFERENCES

1. Koerner,  R. M.,  Designing with Geosynthetics, Prentice  Hall  Publ. Co.,
   Englewood Cliffs, NJ, 1st  Ed. 1986,  2nd Ed.  1989  (to appear).

2. Richardson, G.  N.  and Koerner, R. M., "Geosynthetic Design Guidance  for
   Hazardous Waste Landfill Cells  and Surface Impoundments," EPA Contract  No.
   68-03-3338, 1987, GRI, Drexel University, Philadelphia, PA.

3. Halse, Y.  H.,  Koerner, R. M. and Lord, A. E., Jr., "Laboratory Evaluation
   of Stress Cracking  in HDPE Geomembrane Seams," Proc. Aging and Durability
   of Geosynthetics, Dec. 1988, GRI,  Drexel University, Philadelphia, PA.
                                     375
-------
4. Peggs,  I.  and  Carlson,  D.  S.,  "Stress  Cracking  of  Polyethylene
   Geomembranes:   Field   Experience,"  Proc.  Aging  and  Durability  of
   Geosynthetics,  Dec.  1988, GRI, Drexel University,  Philadelphia,  PA.

5. Haliburton,  T.  A.  and  Wood,  P.  D.,  "Evaluation of U.S.  Army Corps of
   Engineers Gradient  Ratio Test for Geotextile  Performance,"  Proc. 2nd Int.
   Conf. on Geotextiles, Las Vegas, NV, Aug. 1-6, 1982,  IFAI,  pp.  97-101.

6. Koerner, R. M. and  Ko,  F.  K.,  "Laboratory Studies on Long-Term Drainage
   Capability of Geotextiles," Proc.  2nd.  Int.  Conf. Geotextiles,  Las Vegas,
   NV, Aug. 1-6, 1982,  IFAI, pp. 91-95.

7. Halse, Y. H., Koerner,  R. M. and Lord, A. E. Jr., "Filtration  Properties
   of Geotextiles  Under Long  Term Testing," Proc. ASCE/PennDOT Conf.  on
   Advances in Geotechnical Engineering, Hershey, PA, Apr.  1987, pp. 1-13.

8. Rios,  N.  and  Gealt,  M.  A.,  "Biological Clogging  Growth in Landfill
   Leachate  Collection   Systems,"  Proc.  Aging  and  Durability  and
   Geosynthetics,   Dec., 1988, GRI, Drexel University, Philadelphia, PA.

9. Koerner,  G.  R.  and Koerner,  R.  M.,  "Biological Clogging  in Leachate
   Collection Systems," Proc.  Aging and  Durability of Geosynthetics, Dec.
   1988, GRI, Drexel University, Philadelphia, PA.

10. Rollin,  A.,  "Factors  Influencing Geomembrane Seam  Quality,"  Proc.
   Geosynthetics '89,  San Diego, CA, IFAI.

11. Peggs,  I.  D.,  "Failure  and  Regain  of Geomembrane  Lining System,"
   Geotech. Fabrics Report, Vol. 6, No. 6, 1988,  pp.  13-16.
                                         GEOTEXTILE FILTERS
                                                           P-FML
                                                           S-FML
 Figure 1 - Cross Section of Double Lined Landfill Facility Often Used for
             Municipal/Industrial Solid Waste Disposal
                                     376
-------
                                                                  2'-0"
      D
01
n
* ^
10 •
400 -jw —
__ - J» H
200-1
nJ
'fa
!y


*• 	 ••"




> 	 o
a
a


+
o
*
p
D
J








O
a

«
!l>"J'™"

Q
D
3 	









H
a


* 	 	














u 	



*

c;
"~ET~















»

Q


































*

E
Q

















Fl


















c
















































...








                             1              2
                            TRUE STRflIN (7o)
                                                               El  G1
                                                               *  G2
                                                               a  G3

                                                               «  G4
             - Pullout of 60mil HOPE with 350 Ib/ft2 Surcharge

Figure 2  - Details of Test Facility and FML Test Results for Anchor Trench
          Pullout Experiments
                                 377
-------
                                                  TIME
  Figure 3 - Long-Term Flow Column and Usual Trend in Flow Rate  Data versus
             Time
Figure 4 - Viable  Bacterial  Titer  of Each  of  the  Leachate  Samples as
           Determined by Multiplying Viability by the Total Direct Count  TDC,
           after Rios and Gealt<8)
                                   378
-------
                ATTENUATION OF PRIORITY POLLUTANTS CODISPOSED
                       WITH MSW IN SIMULATED LANDFILLS

                        by:  Dr.  Frederick G.  Pohland
                             Department of Civil Engineering
                             University of Pittsburgh
                             Pittsburgh,  PA  15261

                                     and

                             Dr.  Wendall H. Cross, Dr.  Joseph P.
                             Gould and Ms. Debra R.  Reinhart
                             School of Civil Engineering
                             Georgia Institute of Technology
                             Atlanta, GA 30332
                                   ABSTRACT

     Organic  and  inorganic  priority  pollutants  codisposed  with  shredded
municipal solid  waste (MSW)  in  ten pilot-scale simulated landfill columns,
operated under single pass leaching or leachate  recycle, were capable of being
attenuated  by microbially-mediated  landfill stabilization  processes.   The
results of detailed investigations have indicated that inorganic heavy metals
(Cd, Cr, Hg,  Ni, Pb and  Zn)  were subject to a complex  array of attenuation
mechanisms  within the  MSW matrix,  including  precipitation,  encapsulation,
complexation, reduction,  adsorption and ion exchange.   Similarly,  the major
classes  of  organic priority  pollutants  (aromatic  hydrocarbons,  halogenated
hydrocarbons, pesticides,  phenols and phthalate  esters) were attenuated mainly
by  sorption,  volatilization and bioassimilation and  release  of identifiable
by-products within the leachate and gas transport phases.

     Collectively these in situ processes  constitute the assimilative capacity
of a landfill for priority pollutants,  the magnitude of which is dependent on
microbial viability during the sequential phases of landfill stabilization as
affected by loading intensity and contact opportunity.  Data are presented to
demonstrate  some of  the  principal  assimilative mechanisms  as  well  as  the
efficacy of in situ process control through leachate and gas management.  Based
upon these findings, a landfill management strategy is proposed for landfills
receiving inputs of organic and inorganic priority pollutants.
                                      379
-------
                                  INTRODUCTION
      There exist few more complex and challenging problems than those associated
 with the management of municipal and  industrial  solid wastes.   Because these
 wastes often contain toxic and hazardous substances,  they may become a threat
 to human  health and  the  environment unless proper  treatment and  disposal
 techniques are employed.   Of the array of options available for such treatment
 and disposal,  codisposal  in sanitary landfills is probably the most prevalent,
 considering the magnitude  of household  and  small quantity generator sources.
 Therefore,  it  is imperative that the efficacy of codisposal of hazardous wastes
 with MSW be properly assessed,  and that decisions on codisposal be approached
 in as scientifically and  technically sound a manner as possible.

      One approach to such  an assessment is  to examine the progress  of waste
 stabilization  with time and under various operational conditions, either in the
 laboratory or  in the field.  Laboratory or pilot-scale  investigations are often
 more cost effective and permit evaluations with greater operational control and
 less parametric variability.  Hence, pilot-scale simulations were chosen for a
 3-year investigation on the behavior and fate of selected inorganic and organic
 priority pollutants codisposed with  shredded MSW under the influence of single
 pass leaching  or leachate recycle.   This paper presents  a progress  report on
 some of the research results  to  date.

                            EXPERIMENTAL PROCEDURE


 Column Construction and Loading

      Since  the behavior and fate of toxic inorganic and  organic compounds within
 a  landfill  setting are controlled by the  related effects of various  mass trans-
 fer and removal mechanisms,  including solubilization,  sorption, volatilization,
 chemical transformation  and bioassimilation, it  was  necessary to develop  an
 experimental protocol sufficient to  achieve an adequate description and control
 of the waste mass and the gas and  leachate transport phases.   Accordingly,
 shredded MSW was selected  as the primary waste  matrix, and was augmented  by
heavy metal sludge  and selected classes of  organic  priority pollutants.   In
addition, to provide  a range of loadings under both single pass leaching  and
leachate recycle, 10 simulated landfill  columns were constructed and prepared
for operation  as indicated  in Figure 1 and Table 1.

      The  simulated  landfill  columns   were   constructed  from  0.3-cm  steel
cylinders,  0.9 m in diameter and 3 m in  height. The columns were lined with  30-
mil HOPE which  was supported by an underdrain  system composed of 20 cm of gravel
and 2.5 cm  of  sand.  The underdrain  system allowed collection of leachate  for
sampling,  discard  (single  pass  columns) or  recirculation (recycle columns)
through a perforated pipe distribution system  installed  after loading.  Moisture
addition (or recycle) was facilitated through this pipe distribution system.

     A leachate sight glass was also  connected to the underdrain  system  as well
as the headspace under the cover.  Gas  generated during landfill stabilization
                                       380
-------
                                 1.22m
                                  1.83m
                                  0.61m
  LEGEND
1 GAS METER
2 TEMPERATURE INDICATOR
3 GAS SAMPLING VALVE
4 GAS TRAP
5 CHECK VALVE
6 PRESSURE GUAGE
7 DISTRIBUTOR ARM
8 RECYCLE PUMP
9 FLANGE
10 THERMOCOUPLE
11 HOPE LINER
12 IN-LINE FILTER
13 STEEL
14 LEACHATE DRAIN
15 LIQUID SAMPLE PORT
16 LIQUID LEVEL CONTROL
17 GRAVEL, SAND, AND
 GEOTEXTILE LAYERS
18 GEOTEXTILE, SAND,
 GEOTEXTILE, AND GRAVEL
 LAYERS
19 110 V AC
20 110 V AC
 TO PUMP
21 110 V AC FROM
 LIQUID LEVEL CONTROL
22 VENT TO ATMOSPHERE
23 SHREDDED REFUSE
® BALL VALVE
                  RECYCLE UNIT
                                                              SINGLE PASS UNIT
Figure  1.   Construction and Operational Features of Simulated Landfill Columns


was collected from this system and measured with an automatic gas-displacement
meter.

     Each of the 10 columns  received 42 individual  9-kg batches of shredded MSW
placed  and compacted over an 8-hour period.   Columns 1 and 2,  designated control
columns (Table 1) , received only shredded MSW and an equivalent amount of  saw-
dust  also  added to  the other  test  columns because  of  its  use  to facilitate
homogeneity in the sludge-loaded columns.   This loading  strategy provided  five
pairs,  each pair  identically  loaded,  but  operated  with either  single  pass
leaching or leachate recycle.

      Batches of metal sludge were prepared by mixing with sawdust and augmenting
with metal oxides as necessary to deliver low, medium  and high loadings  (Table
1) .  The  organic priority pollutants (Table  1) were mixed together and added at
the  surface of  the  first 30 cm of compacted MSW to increase the potential for
detection during operations.  After column  loading, an 8-cm layer of washed pea
gravel  was added to the surface of the MSW,  the columns were sealed, and initial
                                       381
-------
3
I

53
M

£
             •§
             o-g

              •. a


             il
     II
                                                    *   =8   *
                                       ro   M
                     OJ   M   (Nl   c3
                                          >«.
                                                   fl)   —*   «4

                                                                       O) CD O) O)
                                                                          lit
                                                                       '1!
                                                                  -   2.
                                                                         I-S
                                                                       1'
                                                                        a. ro
                                                                          §
                                                                        8.-I
                                                                                         ITS
                                                 382
-------
moisture  (tap  water)  was introduced to  satisfy  field capacity and  encourage
leaching.   Once  leaching  commenced,  the  five single  pass  columns  received
periodic  moisture additions  equivalent  to  local  annual  rainfall  (123  cm),
whereas  the five  recycle  columns  were  subjected  to  leachate  recycle  with
moisture  additions only of  quantity necessary  to  replace  that  removed  for
sampling and analysis.

Analytical Methods

     Leachate and gas samples were analyzed routinely for both gross parameters
and  priority pollutants once sufficient quantities were  produced to  permit
analysis.   Leachate  samples were  subjected  to  immediate  analyses  for  pH,
conductivity,  chemical  oxygen demand (COD),  and  oxidation-reduction  potential
(ORP), supplemented by  biochemical  oxygen  demand (BOD5) , total organic  carbon
(TOG), volatile organic acids  (TVA),  alkalinity,  nutrients (N,P),  selected
anions (Cl~, SOA~2, S"2,  Br") and cations (Ca+2, Mg+2,  Na+, K+, Li+) ,  heavy metals
(Cd, Cr,  Fe, Pb, Hg, Ni, Zn) and organic priority pollutants (Table 1)  according
to standard techniques.  Gas  samples were analyzed for C02, 02, N2, H2, CHA and
volatile  organic priority pollutants using GC and GC-MS techniques.

                      EXPERIMENTAL RESULTS AND  DISCUSSION


     Since  the primary  purpose  of the research investigations was to evaluate
the  fate  of selected organic  and inorganic priority pollutants cpdisposed with
MSW  under the  influence of single pass leaching and leachate recycle, emphasis
here is placed on these pollutants as the  simulated landfills progressed through
acid formation and into the  methane fermentation phase of  landfill  stabiliza-
tion.    Accordingly,   experimental  data  are  presented to  reveal  potential
attenuating mechanisms  operative during these phases  and  contributing  to the
overall  assimilative  capacity of the landfill  environment.

Progress  of Landfill  Stabilization

     The  progress  of  landfill stabilization  can  be described by  certain
indicator parameters  (1).   For  purposes  here,  gas production and leachate COD
and  TVA  concentrations and pH were selected as  indicated  for the single pass
and  recycle columns in Figure 2. As illustrated  in this figure, acid formation
(acid phase) became prominent and was intentionally sustained for the majority
of   the  report period.   Reductions  in  leachate  strength  (COD  and TVA)  or
increases in gas production,  as  well as some increase in pH, occurred only after
the  onset  of  methane  fermentation (methane phase),  which was  initiated by
digested sludge seeding with neutralization and  took effect on about Day 720.
Thereafter,  leachate and gas quality changed dramatically,  particularly for the
control  columns (1 CR,  2 C) without heavy metal and organic priority pollutant
codisposal.

     The differences  in results between  the  individual  single pass and recycle
test columns  also  are  indicative  of loading  implications,  particularly with
respect  to  the heavy metals,  and the washout by  single pass  operations  as
contrasted  to  leachate  recycle.  The apparent release of leachate constituents
through washout contributed to a decrease in concentrations, but did not greatly
                                      383
-------
W  S0.n
+j
o

J  40.
3

15  30.-4
 n  20.-
   10.-
ti

I
  0.

100.
       »-o 1 C
       o-o 5 0
       *-* 70L
       *-* 9 OM
       «—f 10 OH
                                                o-o 2 C
                                                o-a 3 0
                                                A-A 4 OL
                                                «—» 5 OM
                                                *-* 8 OH
                                                                                  smr.LC
                                                                                o I C
                                                                                aao
                                                                                • 4 01
                                                                                « 5 OW
                                                                                • 8 OH
        200.    400.   600.    800.   1000.
            Time since loading, days
                                            1200.  0.
                                                      200.    400.    600.    800.   1000.
                                                          Time since loading,  days
                                                                                       1200.
 Figure
       2.   Changes  in Gas Production and Leachate  COD,  TVA  and pH during
            Simulated Landfill Investigations
                                              384
-------
enhance  the  extent of  stabilization in the  single pass  columns  where heavy
metal and  organic priority pollutants  were admixed.   With recycle, leachate
concentrations  and gas production  tended  to stabilize  initially  and then
decreased or increased, respectively, when methane fermentation ensued. There-
fore, the containment for conversion of these  leachate  constituents  to gas, by
better contact and controlled in situ treatment,  emerged as a major advantage
of recycle over single  pass operations.

Mechanisms of Change in Inorganic Priority  Pollutants (Heavy Metals)

     A principal focus of these investigations was the behavior and fate  of  the
toxic heavy metals in  terms  of their mobility in the leachate (or gas) phases
as landfill stabilization progressed.   Therefore, special emphasis was placed
on analyses of these metals and potential attenuating mechanisms.

Hydroxide Precipitation--

     Of  the heavy metals added to the sludge loaded columns (Table 1), only
Cr+3  was  subjected to significant solubility control by hydroxide.   Based upon
its extremely low  solubility  (pKso = 30.5),  precipitation of the chromic  ion as
the hydroxide, Cr(OH)3, at equilibrium concentrations of  chromium  on the order
of 1 mg/L, would be attained  even at pH levels as low as  5.0.  Therefore, even
during the acid phase  of landfill stabilization  (pH 5.5  to  6.0),  the mobility
of trivalent chromium would be limited by hydroxide precipitation.   This control
was demonstrated by leachate chromium analyses  (Figure 3), although some, initial
fluctuation  occurred  for  both  the  single  pass  and recycle  columns.   These
fluctuations  coincided  with  early  fluctuations in pH  and,  no  doubt, were
affected somewhat by the respective  column operations  and  the possibility of
short-circuiting.   Moreover,  since leachate chromium concentrations  decreased
sharply  to very low levels during  the  acid phase and  into  the methane  phase,
coupled with the absence of other important precipitants for chromium in typical
landfill  leachates,   hydroxide  precipitation  was  considered   the  primary
attenuating mechanism  for this heavy metal.
  30.
-------
Redox Processes --

      The ORP is particularly important  in defining  the chemical character of
the landfill environment.  As illustrated in Figure 4, negative  leachate redox
values provide reducing conditions  that mediate  the behavior  of many  of the
codisposed inorganic species.   The  impact  of this  condition  can be  either
directly through a modification of the nature  of  the pollutants with  a change
in  their mobility,  or  indirectly  by  reductive  generation   of a  potent
precipitant.
        •400 L
                                                                         SINGLE PASS
                                                                        ate
                                                                        030
                                                                        • 4 01
                                                                        + S OU
                                                                        • tl Oil
               800.    900.   1000.   1100.
                 Tlmo since loading, days
                                        1200.
                                             0.0
                                              700.
800.    900.    1000.   1100.
  Time since loading, days
1200.
           Figure 4.   Changes in Leachate ORP,  Sulfate and Sulfide  during
                       Simulated Landfill Investigations
                                       386
-------
     An  important  example of  the latter  indirect  case is  the  reduction of
sulfate to sulfide  and subsequent precipitation of very sparingly  soluble metal
sulfides.  This is  an  extremely important attenuating mechanism in the landfill
environment  since  many heavy  metals are  precipitated and  removed from  the
leachate, particularly when leachate recycle provides  an additional mechanism
by filtration through the  waste matrix.  However, the efficiency of attenuation
is also  a  function of the availability of sulfides, primarily provided by in
situ microbially-mediated reduction of sulfates.

     As  also  illustrated  in  Figure 4,  sulfate reduction to  sulfide tended to
coincide with  the  onset of methane fermentation, most  noticeably  in the least
loaded  recycle columns.   The  associated impact of available sulfide on  the
leachate lead  and zinc concentrations is dramatically illustrated in Figure 5.
Both  metals rapidly  decreased in concentration in  correspondence with  the
reduction of sulfate to sulf ide.  Therefore, the important attenuating mechanism
for these  heavy metals is precipitation  as PbS  (pKso  = 28.0)  and ZnS (pKso  =
24.8).

      Similarly,  a  direct  impact of reducing conditions on attenuation  of toxic
heavy metals is also  apparent with mercury (Figure 5).  Following an initial
peak, leachate mercury decreased rapidly to concentrations below 100  /zg/L and
eventually stabilized at about 10 /ig/L.   However, in spite of the  exceptionally
low solubility of  mercuric  sulfide (pKso  = 52.5),  leachate mercury concentra-
tions  tended   to  exhibit  no  additional decrease with  sulfate  reduction.
Collectively,  these observations  are consistent with reduction of mercuric ion
to metallic mercury, which has been shown to have a  water solubility of 5 to 30
jzg/L  (2).   This behavior has  also been documented for bottom sediments  (3),
where similar reducing conditions prevail.  Hence, reduction to volatile mercury
and  subsequent sorptive containment within the  waste  matrix would constitute
important attenuating mechanisms  for this heavy  metal.

Other Precipitants --

      During the acid phase,  when  significant  sulfate reduction  and sulfide
generation were not in evidence,  control  of metal solubility may have involved
such  generally  abundant  anions  as chloride,  sulfate,  carbonate  and  possibly
phosphate.  Within the landfill environment,  these potential precipitants may
have  only transient significance,  particularly where  leachate  recycle  and
promotion  of  sulfate reduction  and reducing  conditions  are provided under
controlled operating conditions.

Heterogeneous Physical-Chemical Processes --

      Codisposal of alkaline metal sludges with MSW presents the potential for
 generation  of chemical  microenvironments within  the waste matrix,   and the
 opportunity for a complex array of heterogeneous  (liquid/solid) reactions  of
major  importance  to  the overall  attenuation  of heavy metals.   These micro-
 environments  provide a  source  of alkalinity and  acid neutralizing  capacity
 which will lessen  leachate metal mobility, while  the hydroxide sludge will react
 with leachate anions such  as  sulfide,  phosphate  and sulfate  to develop  an
 encapsulating  layer capable of impeding dissolution of the  sludge  metals  into
 the  leachate.  Therefore,   these alkaline microenvironments  will contribute
                                       387
-------
                                 BtCYCtC.
                                  1 CD
                                  6 OS
                                  7 OtR
                                  9 OUR
                                 10 OHR
    0.
         200. •  400.   600.    800.   1000.
             Time  since loading, days
1200.
                                             0.
                                           2000.'
        n    700.   400.    600.   800.   1000.  1200.
                                            1500. -
                                            1000,-
                                            500.
                                              0.
             200.   400.    600.    800.   1000.
                 Time since loading, days
1200.
         Figure 5.  Changes  in Leachate Lead,  Zinc and Mercury  during
                    Simulated Landfill Investigations


effectively to in situ attenuation during the acid phase of landfill stabiliza-
tion  when  it  is most needed, whereas  this contribution  will  diminish  as
encapsulation  proceeds  and  other  attenuating  mechanisms,   such  as  sulfide
precipitation, gain prominence.

     The  MSW  itself also  provides abundant surface  area for  sorptive  inter-
actions with leachate constituents.  These processes include physical adsorption
on  the solid  surface,  ion exchange  (particularly  on soil),  chemisorption  by
                                       388
-------
complexation with insoluble  ligands  associated with  the MSW,  and physical
containment  in  transiently  stagnant  void  volume liquid.     In addition,
complexation of metals by soluble ligands, particularly with moderate  to high
molecular weight humic-like substances, may enhance sorption by  incorporating
the metal into  a relatively hydrophobic molecule.  Current analyses on leachate
aromatic hydroxyl concentrations and molecular weight distributions are being
provided  to help  determine   the  significance  of  this  latter  attenuating
mechanism.

Mechanisms of Change in Organic Priority Pollutants

     Since  the  landfill  environment  provides  numerous possibilities   for
contaminant transport and transformation, an  equally important focus of these
investigations  was  the  behavior and fate of  the organic priority pollutants
during both acid and methane  phases of  landfill  stabilization.   In this case,
mobility and possible attenuation  were  not  as clear as for the  heavy  metals.
However,  scrutiny of the  data  (Figures  6  through  8)  suggests  that  the  12
compounds added to the test columns (Table 1)  can be divided into four  general
groups in terms  of their relative mobility and reactivity.

     Four  of  the compounds,   dieldrin,  hexachlorobenzene, bis-2-ethylhexyl-
phthalate  and  7-1,2,3,4,5,6-hexachlorocyclohexane (lindane) essentially have
not  emerged in  either  the leachate  or gas  phases  although in the   case  of
lindane, three samples from over 600 analyzed  revealed its presence at about 10
to  20  /ig/L.     Three  of  the  compounds,  dibromome thane,  2-nitrophenol   and
nitrobenzene,  appeared  early  in appreciable concentrations and then decreased
to below detection  limits  (Figure  6), particularly  after methane fermentation
had  been established.    In  contrast,1,2,4-trichlorobenzene has  been detected
regularly in the leachate (Figure 7) , but  at relatively low levels  (< 1 mg/L) .
The  remaining  four  compounds, naphthalene,  1,4-dichlorobenzene,  2,4-dichloro-
phenol  and 1,1,2-trichloroethene,  appeared  in  the   leachate  in   varying
concentrations  throughout the project period  (Figure 8).

     Headspace  gas analysis has indicated the presence of only three of the test
compounds;  trichloroethene  fairly  regularly at 1 to 20  jug/m3, and naphthalene
and  dichlorobenzene sporadically  at  equally  low concentrations.  Moreover,
evidence  of biodegradation  of  at  least two of the added organic priority
pollutants has  been obtained.  Bromide ion (Figure 9) has been detected in the
leachates  from all  test columns;  slowly leaching from the single pass  columns
due  to  washout, while  remaining relatively constant in the leachates of  the
recycle  columns (-150  mg/L).   Since no leachate bromide  was detected  for  the
control columns, and appeared from the test columns coincident with  the decrease
in  dibromomethane  (Figure 6),  microbially-mediated debromination  would  be
suggested.  Likewise, vinyl chloride (as high  as  300 /*g/L) has been detected in
the  headspace  gas  of  the  test  columns   after  reductions   in   leachate
trichloroethene were observed, suggesting partial bioconversion  to this  known
intermediate  (4).

Primary Attenuating Mechanisms --

     In  addition to the suggestion of bioconversion of some of  the  organic
priority pollutants into detectable intermediates and reaction products, several
                                      389
-------
                                                                        SINGLE PASS.
                                                                       o 2 C
                                                                       0 3 O
                                                                       » 4 OL
                                                                       + 5 OU
                                                                       * B OH
     0.
  200.   400.   600.   800.   1000.
      Time  since loading, days
                                       1200. 0.
                                               200.    400.    600.    800.   1000.
                                                   Time since loading, days
1200.
Figure
6.  Changes in Leachate, Dibromomethane, 2-Nitrophenol and Nitrobenzene
    during Simulated Landfill Investigations
other potential attenuating mechanisms could be envisioned.   Inspection of the
concentration data,  as well as the mass of  each  added compound released from
the test columns (Table 2) , provides some basis for deducing possible mechanisms
controlling relative mobility in the leachate or gas.  Although £here does not
appear to be  a variation in perceptible patterns  of release of the organic
priority pollutants between the single pass and recycle columns,  the operational
procedure in  either  case would determine whether  the leached compounds would
                                      390
-------



UJ
1
c_>
LJJ
O
i
CO
o
O£.
Q

O
Ul
0
&
111
OK
01 CO
UJ 5
GO —1
CO O
O O
(O — 1
rf II
rz* s^
r— O
^ Z
	 | ^
_I _j
O
a. a
UJ
^ <
|i
a.: co
UJ
u =

3 o
0£ 0£
O "-

CO
UJ
(__
_J
13
Q
*>
o

LL
O
CO





UJ
3
i—








ujapug;
a^BjBtpqd


3T19 Z U9 q
OJOTUOBX9H



suBpufT;






euazuaq
— o aoxqoTax— -17 * 3 ' j


•[ouaqdoj^Tji-^


auazuaq

••OJt OTUO"E(T™^7 T



auazuaaoaaxM






3Ua dBN
+



d OTUOTfT-17'7






aUBUaaOJOTUOTJT






*



A^j^uapi uuinjog



II II II II II



II It II II II



ii ii ii *4~ r- ii
" O

V
^/


it t i O '"* O *•* O '•^
in *^ CM CM N. %o
o o o


t i ii ^D ^^ C3 ^^ O ^^
v^ vx v^

^D ^^ CZ3 *"* C3 *"*
^D ^— CD O* ^5 r1*
ii ii C\l *^ O • O •
» » S.X >^



O ^N 0 ^ O '^
ii ii «— ^" r^in in'O
»— • in » N- •





C5 ^% ^D ^^ O ^^ C3 ^^ CD ^^
CDO OO CD -J- O O OCM
^T ,-T ^^ ^^



O • *"* O ^N O *^
SN- O sf O *-
»- . ro •

v— vx C? ^^ iN- ^^
. . .


§x*x ^3 ^N CO ^^
CO ON- O CM

*~ ^ O. ^f O> ^



^D ^^ ^D ^^ CD ^^
ON- O CD O CO
ii ii OK) CD jTO CD CM
•S* %O -*
** to ro


0 CJ O O S
T- CM ro st in


it 11 ii ii ii



11 ii it 11 ii


^•n' " XX
II II 11 *Q «— ""d" «—
• O • O
«- O O

N.X V
^^


^D ^^ ^D ^* CD ^"* CD *"* CD ^^
%oro I'-NO -OCM oro ro«—
ro • o CM • sa- • «- •


§**\ O ^^ O *"* O '•^ O *•*»
^o co CM o «~ in ro r*- ^3
-* ~

O«— OTO OCM OCM 0*4-

»t«— »*— •>•:— -*v- ».«—



CM-4- ON- O-« OCO Oo «- in • to • •* • IO,'~

N- O^-* v— v-' v-:^^ in



o *** o ^^ o ^^ o ,^* o ^^
ot- OTO OTO o*r oin
o« in*— ^is" ^o*" r^«—

t*» T- in co ^ *O t*-
•r- v^ T- a- CON- ^CM or- OCM
^^ ^^ TO t^> ^O
in co TO «- • CM

a OL ai
CZ —1 31 £ 3=
o o o o o
-o N. co o o

























•
c
3
"o
o
~o
L.
c
o
0
c
"^
T3

UJ
CA
o
(/)
w
l_
o
• • M-
't-> "D
CO  ?S 3
01 t- >
£ ° ~
O O> 55
•M
c -c c
c 0) "ra
- O l_ •!->
fi. 0 J=
S Q. Q-
O 
-------
  3.0
 .2,0
s>

I
B9
J3

oW

o
  0.0
                          RECYCLE.
                         Q  1 CR
                         O  6 OR
                         »  7 OLR
                         +  9 OUR
                         « 10 OHR
 SINGLE PASS.
D 2 C
030
* 4 OL
+ 5 OM
« 8 OH
    0.
200.   400.    600.    800.
    Time since loading, days
                                       1200.
                                                 200.   400.    600.    800.   1000.
                                                     Time since loading,  days
      1200.
               Figure 7.   Changes in Leachate 1,2,4-Trichlorobenzene during
                         Simulated Landfill Investigations
      escape due to washout (single pass) or be more  likely  subjected to additional
      attenuation opportunity with recirculation (recycle).

           Evidence to date indicates  that  the organic priority pollutants  will be
      influenced by the equilibrium  conditions established at the point of  contact
      between the gas, liquid and solid phases.  Supplemental  studies with the test
      compounds have demonstrated that other leachate constituents  have enhanced the
      solubility of the more hydrophobic molecules, but that physical  sorption was
      a primary attenuating mechanism.  Moreover, a strong correlation of sorbability
      existed with the octanol/water partition coefficient (Figure 10).   Therefore,
      sorption of  the test compounds  on the  waste matrix  could  be  attributed to
      hydrophobic interactions and removal  from the aqueous phase onto  nonspecific
      surface sites, particularly such MSW constituents as fats, oils,  waxes, biomass,
      humic-like substances,  lignin,  plastics and leather.

           In the final analysis,  only soluble, less hydrophobic compounds would be
      expected to elute from the wastes in the  test columns.  Hence,  compounds would
      be  expected to emerge in the approximate order of increasing affinity  for the
      waste.    However,  the  degree   of MSW   stabilization  and  the  corresponding
      microbially-mediated phase  (acid or  methane) would influence  this pattern,
      i.e. ,  pH  changes  could impair  sorption  of ionizable  compounds and the more
      stable the waste,  the  greater the mobility.  Hence, upon  completion   of  the
      stabilization process through methane fermentation and into  final  maturation,
      the liquid (leachate) transport phase should be  removed  to preclude continued
      fractionation of many of the residual organic priority pollutants from the waste
      matrix.  With leachate containment and recycle, such operational control can be
      easily accommodated.

                                 SUMMARY AND CONCLUSIONS
           Simulated landfill  columns,  containing  shredded MSW and  any array  of
      inorganic  and organic  priority  pollutants  have been  shown  to  possess  a
      significant  assimilative capacity due to a variety of attenuating mechanisms.
                                           392
-------
200.
           400.   600.'   800.   1000
         Time  since loading, days
                                   1200. 0.
240.     480.     720.     960.
  Time since loading,  days
                                                                         1200.
Figure  8.   Changes  in Leachate Naphthalene, 1,4-Dichlorobenzene,
            Trichloroethene  and 2,4-Dichlorophenol during
            Simulated Landfill Investigations
                                     393
-------
   400.
   300.-

              o
              D  Soil/Sediment Data
                   (Karlckhoff, 1981)
                                                                      -b-
                 R»lu«« Beat Fit Curv* Plu* 86 Percent
                   Confidence Interval
                                     D
           1234

                     Log  Octanol/Water  Partition Coefficient

           Figure 10.   Relationship Between Log Octanol Water Partition
                        Coefficient and Log  Sorption  Partition Coefficient
                        Normalized to Carbon Content.
                                               394
-------
     Metal mobility  was minimized by microbially-mediated  physical and both
homogeneous   and   heterogeneous  chemical  processes,   including  hydroxide
precipitation, direct and indirect reductive removal, sorption, and retention
in  stagnant  intersticial  void  liquid.    Associated  specific  attenuative
mechanisms, such as encapsulation, tended to function during  operational phases
when other mechanisms, such as sulfide precipitation, were unavailable.

     Organic priority pollutants appeared to be removed primarily by sorption
and/or biodegradation, with identifiable metabolic intermediates and reaction
products being detected. The sorptive capacity of MSW results in extremely Ipng
contact  times within  the  landfill  setting,  and  enhanced  opportunity  for
acclimation and further degradation of more recalcitrant compounds.

     Leachate  recycle  offers  the advantage  of  not only  retaining leached
materials within  the landfill system, but also  providing better contact and
redistribution of components contained in the leachate, thus  facilitating and
beneficiating the overall efficiency of stabilization.

     Landfills containing MSW and operated with leachate recycle, can provide
opportunities for codisposal of organic and inorganic priority pollutants with
reasonable assurance that mobility and potential release can be controlled to
safeguard human health and the environment.
                                  REFERENCES
1.  Pohland, F. G. Dertien, J. T., and Ghosh, S. B.  Leachate and Gas Quality
    Changes During Landfill Stabilization of Municipal Refuse.  In: Proc. 3rd
    Intl. Symp. on Anaerobic Digestion, Boston,  MA, 1983.

2.  Hughes, W. L.  A Physicochemical Rationale for the Biological Activity of
    Mercury and Its Compounds.  Ann. N.Y. Acad.  Sci.  65:454, 1957.

3.  Mercury in the Environment. An Epidemiological and Toxicological Appraisal.
    L. Fribery and J. Vostal, [Eds].  CRC Press, Cleveland, OH,  1972.

4.  Wilson, B. H., Smith, G.  B.  and Reas, J.  F.  Biotransformation of Selected
    Alkylbenzenes and Halogenated Aliphatic Hydrocarbons in Methanogenic Aquifer
    Material:  A Microcosm Study,  Environmental Science and Technology 20, 997-
    1002,,1986.

5.  Karickhoff,  S. W.   Semi-Empirical Estimation  of Sorption of Hydrophobic
    Pollutants on Natural Sediments and Soils, Chemosphere, 10(8),  833,  1981.
                                     395
-------
               SITE DEMONSTRATION OF HAZCON PROCESS
                        Paul R. de Percin
             SITE Demonstration and Evaluation Branch
              Risk Reduction Engineering Laboratory
               U.S.  Environmental  Protection Agency
                         Cincinnati,  Ohio
ABSTRACT

     In October 1987 the HAZCON stabilization process was tested
and evaluated at the Douglassville, Pa. superfund site.  There
was extensive physical and chemical testing of the untreated and
treated waste samples.  Stabilized wastes were also stored
underground on site as a field durability test.  After 270 days
(July 1988) these wastes were dug up and the samples obtained
were tested for any changes in chemical and physical properties.
The wastes were reburied and will be sampled again in July 1989.

     The durability of the treated samples was determined by
comparing-the 28-day and the 270-day test results.  Test results
of the HAZCON 270-day samples indicated a slight loss in long-
term durability, but because of the data scatter this may not be
statistically important.  There was some loss in UCS strength in
the low oil and grease samples (<5%) , but the high oil and grease
samples (16-25%) the UCS strength increased.  TCLP leachate
concentrations for lead remained low (ppb) ; and the VOC, BNA and
PCB leachate concentrations appeared lower than the 28-day sample
results.  A noticeable decrease in the porosity of the samples
was observed.  Overall, there was not a major change in sample
characteristics and some durability was demonstrated.
                               396
-------
INTRODUCTION

     The HAZCON solidification/stabilization treatment process
was demonstrated at the Douglassville, Pennsylvania superfund
site under the Superfund Innovative Technology Evaluation (SITE)
program in October 1987.  Reliable field performance and cost
information were the major objectives of this demonstration.  A
secondary objective was to study the long-term durability of the
treated waste under field conditions.  The treated wastes
resulting from the HAZCON demonstration were stored on-site,
underground and isolated from the untreated waste.

     One aspect of solidification/stabilization processes is that
the waste is not destroyed, but macro- and micro-encapsulated
(trapped)  or  fixated  (reacted) in/to the stabilizer matrix.
Thus, in future years there is potential for the contaminants to
leach from the treated waste as the stabilizer degenerates.
Long-term durability or lifetime treatment effectiveness is a key
questions both regulatory agencies and responsible parties have
about stabilization.

     After being stored on-site for nine months (July 1988),
sample cores were bored from the treated wastes and the samples
were tested using physical and chemical procedures.  The sampling
and the sample testing procedures were the same as the procedures
used to obtain and test the one month samples.  One and nine-
month sample test data were compared to determine if any
degradation of the treated waste was evident.

BACKGROUND

     The Douglassville, Pa superfund site is a former oil
recovery facility covering about 50 acres near Pottstown, Pa.
There are six contaminated areas: two large lagoons referred to
as lagoon north (LAN) and lagoon south (LAS), an oily filter cake
storage area (FSA) , an oil drum storage area (DSA) , an area where
generated sludge was landfarmed (LFA), and the plant facility
area (PFA).  The major contaminants at this site are oil and
grease, semivolatile organics (BNAs) and lead; minor contaminants
were PCBs and volatile organics (VOCs).

     The HAZCON proprietary solidification process involves the
mixing of hazardous waste material and cement with a patented
nontoxic chemical called Chloranan.  Chloranan is claimed to
neutralize the inhibiting effect that organic contaminants
normally have on the crystallization of cement-based materials.
For this treatment, the wastes were immobilized and bound into a
hardened, leach-resistant concrete-like mass.  Waste from the
Douglassville superfund site was selected because of the high
levels of oil and grease (up to 25%) and lead (up to 2.2%)  in the
FSA waste.   This combination was considered very difficult to
                             397
-------
process successfully.

     During the HAZCON SITE demonstration at Douglassville, five
cubic yards of treated waste was produce for each of five wastes.
The sixth waste (LAS) generated, twenty-five cubic yards of waste
as a test for equipment reliability.  Waste was mixed with the
Chloranan at a ten to one  (10:1) ratio and with the portland
cement at a one to one (1:1) ratio.  Water was added as needed.
A 1:1 ratio of waste to cement  is high, but it was believed
necessary because of the high level of organics and heavy metals
in the FSA waste.  These mixing ratios were not changed for the
other wastes.  The treated waste was poured into 1-cu-yd molds
(as a slurry) and allowed to harden.  The molds were stripped
from the hardened waste after one to two days.

     During the waste processing, the holes created excavating
the feed waste were enlarged to accommodate burial of the treated
waste blocks.  Before the blocks for a particular waste area were
buried, the hole was lined with plastic (to prevent seepage of
contaminated water into the hole) and a one foot layer of clean
soil was deposited.  After the  blocks were placed into the
excavation hole, additional soil was added to cover the blocks.
Stakes were planted to identify the location of each block.
SAMPLING AND ANALYSIS PROCEDURES

     One month and nine months after the blocks had been buried,
samples were collected from two blocks in each area for physical
property tests and from one block for chemical tests (soil
analyses and leaching tests) and weathering tests.

     Samples of solidified waste were collected using a rotary
rig with diamond-tipped core barrel.  The outside diameters for
the cores were 7 cm for unconfined compressive strength,
permeability, and leaching tests; and 4.5 cm for wet/dry and
freeze/thaw tests.  These multiple samples of different diameters
required the boring of at least two holes in each of twelve
blocks, two blocks in each area.  Cores were removed from the
core barrel and sealed in aluminum foil, placed in glass jars
closed with a custody seal, placed in zip-loc bags, and stored
with ice packs during storage and shipment to the laboratory.


     Both standard methods from SW-846 and experimental test
methods were used to evaluate the core samples.  The test
procedures are listed in Table 1 below:
                                398
-------
                             Table 1
     Physical Tests
     Chemical Tests
     Micro Scale Tests
Moisture
Bulk Density
Unconfined Compressive Strength (UCS)
Wet/Dry Weathering
Freeze/Thaw Weathering
Permeability - Falling Head

Leaching   -   TCLP - VOCs, BNAs,  PCBs
Priority Pollutant Metals - Pb, Cr, Ni,
               Zn, Cu, Cd
Total Oil and Grease

X-Ray Diffraction
Scanning Electron Microscopy
Unconfined compressive strengths were performed on the Wet/Dry
test specimens and controls after the twelve-cycle weathering
tests, except for LAS where a permeability test was performed on
the test specimen.  Unconfined compressive strengths were
performed on the Freeze/Thaw test specimens and controls after
the twelve-cycle weathering tests, except for DSA and FSA where
permeabilities were performed on the test specimens.  The listed
metals were the only metals detected in significant quantities.


TEST RESULTS

     The results are presented in three parts; physical, chemical
and microstructural properties.  The physical properties include
moisture content, bulk density, permeability, unconfined
compressive strength, and weathering effects.  The chemical
analyses were for treated soil and TCLP leachate analyte
concentrations for volatile organic compounds (VOC), base
neutral/acid extractables (BNA), six priority pollutant metals,
and polychlorinated biphenyls (PCB).  Microstructural analysis
included X-ray diffraction and.scanning electron microscopy.
It should be noted that the 28-day (i.e., 1 month) sample data is
from the report "Technology Evaluation Report, SITE Program
Demonstration Test, HAZCON Solidification, Douglassville,
Pennsylvania - Volume I (EPA/540/5-89-001a) .
Physical Properties

     The results of the physical tests are summarized in Tables
2, 3 and 4; and are discussed below.

                                399
-------
     The moisture content ranged from 8.8% by wt for PFA-1 to
20.4% by wt for FSA-3.  All the values are approximately 1% to 2%
by wt less than for the 28-day samples, with the largest
reduction for FSA.  This may indicate that the blocks continued
to cure thus reducing the free moisture content.

     The bulk densities were essentially equivalent to those
obtained in the 28-day cores.  They ranged from 1.51 g/ml for FSA
to 2.05g/ml for PFA.  The values at LAS and LAN increased by 5%
by wt and 3% by wt, respectively.  This increase is most likely
due to sample variations.

     The permeabilities ranged from 2.2xlO"10 cm/sec at DSA-3 to
2.3X10"7 cm/sec at  DSA-1.  These values are in the same general
range as the 28-day cores but appear to be slightly larger.  The
apparent increases in permeability are larger than the decreases,
and eight of the .twelve samples increased.  This trend is not
conclusive as data on permeability variation within individual
blocks is needed.  This apparent increase could be explained by a
small breakdown in the porous structure of the blocks, which
would allow interconnection of the pores and a path for water
flow.

     The unconfined compressive strength  (UCS) ranged from 230
psi at FSA-3 to 1170 psi at PFA-3.  Comparing the results to
those of the 28-day cores, nine of twelve samples had a lower
UCS.  The largest  losses occurred for the areas with low oil and
grease content.  Averaging all of the values across the six plant
areas shows an overall decrease of about 20%.  This apparent
•reduction in UCS is consistent with an increase in permeability,
if the solid structure is deteriorating.

     The oil and grease content in the treated soil ranged from
0.48% by wt at DSA-3 to 12.1% by wt for FSA-1.  All values are
consistent with 28-day cores.  The samples are also consistent on
a material balance basis to the untreated soil oil and grease
concentrations.

     The wet/dry weathering tests weight  losses of the test
specimens were slightly greater than  for the 28-day samples, with
losses ranging from 0.7% to 1.7% by wt.  The unconfined
corapressive strength  for the  long-term and 28-day samples were
the same order of  magnitude.  A permeability was performed on the
weathered test specimen from  LAS-LM1  and was 5.5x10"  cm/sec
which agreed with  the unweathered  long-term sample but is greater
by a factor of 30  compared to the unweathered 28-day  cores.

     The freeze/thaw weathering test  weight losses of the test
specimens was nearly  double that  of the 28-day  cores, with values
ranging from 0.99% by wt at PFA-3  to  3.19% by wt at LAS-LM1.  The
loss for the controls was only  slightly greater than  for the  28-

                                400
-------
day core controls.  Therefore, unlike the wet/dry samples, the
corrected cumulative weight  loss  (wt. loss of test specimens
minus wt loss of control each divided by its dry starting weight)
for the long-term samples were larger than for the 28-day cores.
This indicates, along with UCS _and permeability results, that the
freeze/thaw cycles at Douglassville during the winter of 1987-
1988 may have damaged the internal structure of the test blocks.
The unconfined compressive strengths of the 12-cycle weathered
test specimens ranged from 430 psi at LFA-5 to 860 psi at LAN-1.
These values on average are  larger than the UCS values from the
28-day weathered cores.  However, individual UCS values are so
scattered that a comparison  of UCS values between long-term and
28-day cores shows them to be approximately equal.


Chemical Properties

     The results of the chemical  tests are summarized in Tables 5
and 6; and are discussed below.

     The total PCBs ranged from <0.64 mg/kg at DSA-3 to 5.9 mg/kg
at FSA-1.  These values are  considerably lower than anticipated
based upon untreated soil sample  analysis performed during the
Demonstration Test.  Analytical difficulties were encountered.
Analysis on 28-day core samples were not performed.  PCBs were
not detected in the TCLP leachates.

     The VOC content ranged  from  210 ug/kg at DSA-3 to 70,800
ug/kg at FSA-1.  These values are considerably lower than i the
28-day cores where the values ranged from 3,490 ug/kg at LAN-1 to
108,700 ug/kg at FSA-1.  The VOC  content in the TCLP leachate
ranged from 13 ug/1 for LAN-1 to  493 ug/1 for LAS-LM1.  A
calculation of the migration potential, which is defined as
analyte weight in the leachate divided by analyte weight in the
solid, provided erratic results,  but appear to be equivalent to
the 28-day cores.  The Demonstration Test results showed
equivalent migration potentials for the treated and untreated
soils.  Therefore, no immobilization of VOCs could be seen.

     The total metals content in  the long-term treated soil
samples was equivalent to those obtained for the 28-day cores
with lead values ranging from 980 mg/kg at DSA-3 to 8,600 mg/kg
at FSA-1.  As previously, the primary metal contaminant is lead.
Leachate results from the TCLP leach test showed total metals
content ranging from 3 ug/1  at DSA-3 to 120 ug/1 at LFA-5.   The
FSA-1 results were quite low, 39  ug/1, wit only lead detected in
the leachate.  A calculation of migration potential produced
values equivalent to those for the 28-day samples,  which showed
high immobilization of the metals.  The migration potential for
lead ranged from 2.1x10  ug  leachate/ug soil for LAN-1 to
6.3x10  ug leachate/ug soil  for LFA-5.
                              401
-------
     The quantity of BNAs for the long-term soil samples was
considerably less than reported for the 28-day cores.  However,
the detection limits used by the laboratory contractor were a
factor of 50 to 100 higher due to laboratory difficulties.  For
example, at FSA-1, no BNAs were_ detected in the long-term samples
while the 28-day cores showed a value of 210 mg/kg.  However, the
long-term and 28-day samples had TCLP leachate concentrations
that were equivalent.  The long-term leachate values ranged from
35 ug/1 for PFA-3 to 3,109 ug/1 for FSA-1, the latter being about
the same as that of the 28-day cores.  Therefore, it appears that
no further immobilization of these organics occurred.

     The oil and grease levels in the long-term sample leachates
were below 1.0 mg/1, which is less than for the 28-day samples,
where the values were primarily between 2.0 mg/1 and 4.0 mg/1.
Oil and grease was detected in the TCLP leachates only for LAN-1
and FSA-1.  The other four samples showed values below the
detection limits for oil and grease of 0.4 mg/1.  It appears that
the mobility for oil and grease is equivalent for the treated and
untreated soils.  This is an apparent improvement over the
earlier results.

     Analysis of clean soil samples taken from the backfill used
around the buried blocks in each plant area showed no
distinguishable change in its measured physical properties or
organic content.  No VOC, BNA or PCBs were detected in the soil.
Microstructural Analysis

     Analyses  of thirteen  samples were performed on a
microstructural scale.  The  samples were  studied by scanning
electron microscopy  (SEM), optical microscopy  (OM), and X-ray
diffraction  (XRD).   The type of information to be  obtained from
the tests were:

     Microscopy      -    crystal appearances, porosity,
          fractures, and the presence of  unaltered soil
                                    crystalline  structure of the


The  results  can be summarized as follows:
X-ray Diffractometry
     hydration products.
      The samples were morphologically poorly  defined material
      compared to hardened portland cement/soil mixtures which do
      not contain waste material.

      The porosity of the long-term samples were  low, having
      decreased noticeably from examinations of the  28-day
      samples.  Decreased porosity suggests that  either
      microstructures vary significantly or that  dehydration

                               402
-------
     reactions continued after the previous samples were
     analyzed.

     Mixing appears to be poor as observed previously.  Brownish
     colored aggregates and opajgue particles, likely to be
     untreated waste material, could be observed in both long-
     term samples and those of the 28-day cores.

     X-ray diffraction analyses continue to support that
     encapsulation is the major process contributing to
     stabilization of these contaminated soils.

     As in the 28-day samples, previously examined, unhydrated
     clinker was detected in the long-term samples, suggesting
     poor mixing and/or incomplete hydration.


CONCLUSIONS

     Based on the comparison of the long-term and 28-day sample
test results, the following conclusions were drawn:

     *    The priority pollutant metals were immobilized,  with
          the long-term results equivalent to the 28-day results.

     *    The VOCs and BNAs leaching did not change.

     *    The clean soil backfill appears not to have been
          contaminated by the buried blocks.

     *    A small deterioration in physical properties appears to
          have occurred.  This was concluded from a series of
          small changes, which are as follows:

               a decrease in UCS for samples of low oil and
                    grease content

               an increase in permeability of the samples

               increased weight losses of test specimens and
               controls during the wet/dry and freeze/thaw
               weathering tests

               large increase (double)  in weight loss of
               freeze/thaw test specimens.

     *    The microstructural results are similar to earlier
          results except that the porosity was low having
          increased noticeably from the earlier samples.


     The general  conclusion is that the 9-month test results were
                               403
-------
similar to the earlier result, but where they deviated the
results appear to show a small reduction in the physical
properties.

     Further long-term samples .are scheduled to be collected in
June 1989 and analyzed as described in this study.  Comparison of
three sets of data will allow a stronger conclusion to be drawn
about the long term durability of solidification/stabilization
processes.
                                404
-------











trt


S
i
u_
0

IU
ot
IU
&
£
PHYSICAL
(M
IU
•J

















1
I


§
« X

i


•—
"2 .i **•
ill
S h. ?
a s I
5 o *-
s
p^ V
15

1*6
£ -
II

?£
-. i
O ci
L

i

o. <

!>.»«« COM CMK> MN «-O M^
•<) ^T ep O> eo •- u>K. NCM ino
ineo «7in •- ro o>o KM o- S
M N» 40 *• ^" co ** N. •o in
cMo  (O O> 03 •- -O
•*in -or-- M<- •* m coo inin
•-•- ^-»- CMCM »-»- «- •-»-





MK inp RS inia SS «l8
»- -* in h- CMCM coco (O'- eo I--


OO Kl CM OO -J- O -* 1M OO
JQO •*•"* '"t KO OO OM


£ S 3 S E 5 S S S S 35

57* o •» 0 in
l^>* v*i tin to <
-------
                              TABLE 3.  WEATHERING TEST RESULTS  - WET/DRY
'
Plant
Area
DSA-3
LAH-1
FSA-1
LFA-5
PFA-3
LAS-LH-1
Average
Long-term
Ut. Loss, X
Specimen Control
1.50 1.33
0.80 0.66
0.86 0.80
1.71 1.40
0.91 0.73
1.42 1.00
1.20 0.99

UCS, psi
Specimen Control
830 470
890 500
340 430
250 230
2290 1270
(a) 830
920 622
a) Permeability performed on weathered sample resulted


Plant
Area
DSA-3
LAH-1
FSA-1
LFA-5
PFA-3
LAS-LH-1
Average
TABLE
Long-term
Wt. Lose, X
Specimen Control
1.85 1.73
1.55 0.62
1.48 1.04
2.89 1.28
0.99 0.88
3.19 1.00
1.99 1.09
28-day
Ut. Loss, X UCS,
Specimen Control Specimen
0.90 0.88 1150
0.74 0.77 180
0.93 0.75 340
1.53 1.15 1230
0.73 0.66 1170
1.05 0.84 400
0.98 0.84 739
in 5.5x10"8 cm/sec.

psi
Control
750
610
330
500
1190
450
638

4. WEATHERING TEST RESULTS - FREEZE/THAW

UCS, psi
Specimen Control
(a) 1530
860 420
(b) 520
430 870
750 620
470 770
628 788
28-day
Ut. Loss, X UCS,
Specimen Control Specimen
1.29 1.24 660
1.07 0.60 370
0.53 0.53 400
2.17 1.20 520
0.58 0.49 980
0.95 0.73 210
1.10 0.80 523

psi
Control
1020
250
210
320
350
400
425
a)  Panaeability performed on weathered sample resulted in 4.0x10'  on/sec.
b)  Permeability performed on weathered sample resulted in 7.1x10   cm/sec.

                                                       406
-------





















S
1
CD
U
OL
i
UJ
X-
_J
_J

u
X
uu
ac
o


in
ui
i

































i
00
CM




























E
4)
i














+ S ci
f — • O
£55
* 8 s
8.

a.
u

c

X
*



c

5


8
S - S
— ex:
*rf **- O
« *-. ro
&• C 4)
01 C —
— *••
x 8. S




c








c
g
>




0.
CJ
»—
c


§

te

***
Cfl
(J
OL



J
Q.

•O
O
»
5

««
3

41
**
a
u
a
±
a,

3
_^
to
|

i.

T3
O
3



^
5?
«T
a
£
S


O*


o>
^
(0
"8
*J
0
«>
u

ff

V
a
f
a
a
.

"o
(A
U
1


I

in to o o in
•". ^ S S R S
iX o o o d o





§T- &t |W
O Q O
*^ X> •"
— *





0 O 0 0 0 0
in O- o K) in ry
Kl »» h- O> to «»
in" m to K? in" ^
3 «M *-



Kl 00 tO
AJ ro o *o -^ h*
• •••••
>» 0 0 U\ 0  to o> o o to
^r •- o o^ o o








o o o o o »-
«- Q Q O O O.
CM oo oo in in in
g" S






333333






>O Kl t> •- K IO
?.* in «- to •*




Ki »- -- in to 5
• 111*1
5 3 35 2 2 S
o -J u. -J a. -J



0)
0
«5
i.
^
u
0
C
o
i

2
0
1
S 0
•v. w
! oo a
*•» u
s -
§•
DC -O
S "
— M
I I
~a » ^
_o S
1 0
f. ,C
84* •
C
i- .•«- 5
0 ° 8
0 X i
M U
0 <->
« » 2
g S —

a 01.
> JC 0
1- *1 i
ii M U
•*" CB
•0 I ^


a u ^
£ a «
•a 5 - cT
CM O0
O ** C
SO. C 0
S St.
— — u a
o.
1 £ i1 i

f «D 0 0
1 si M
s - * Is
« 33 * ~*

£ 00 •«• CD
5 s 5 ^ «^
? C C ° 0
•^ O o 01*
0 — ~- .CO
oa u u n •*.
" " • £
an 80
3 xx t i£
. * :
407
-------
                                      TABLE 6.  LONG-TERM CHEMICAL ANALYSES - 8HA AND LEAD
Plant
Are*
DSA-3
LAN-1
FSA-1
IFA-5
PFA-3
LAS-LH-1
BHAs in
treated soil, mg/kg
DO* (12.15)
BD (17.90)
BD (534.2)
1.1 (36.70)
0.91 (18.45)
BO (39.60)
BKA in TCLP
leachate, ug/l
43 (14)*
1,260 (1.409)
3,109 (4,083)
60 (169)
35 (55)
216 (689)
Migration
potential - BNA,
ug teachate
ug solid
- (•)
- (.62)
- (.15)
1.09 (-)
.077 (.53)
• (.31)
Lead in
treated
soil, mg/kg
980 (570)*
3,800 (3,010)
8,600 (10,200)
3,800 (2,700)
3,400 (4,900)
4,300 (3,850)
Lead in TCLP
leachate, ug/l
3 (7)
4 (5)
39 (400)
120 (50)
38 (11)
23 (51)
Migration potential
lead, (104)
ug leachate
ug solid
0.61 (1.69)*
0.21 (0.36)
0.91 (8.0)
.6.32 (5.40)
2.24 (0.61)
1.30 (3.13)
*  Values In parentheses are from 28 day samples as reported in the Technology Evaluation Report.




+  Below detection limits.  The detection limits were quite high due to interferences.
                                                         408
-------
   SITE DEMONSTRATION OF THE TERRA  VAC  IN SITU VACUUM  EXTRACTION TECHNOLOGY
                by:  Peter A. Michaels
                     Foster Wheeler Enviresponse, Inc.
                     Edison, NJ  08837
Mary K. Stihson
USEPA, RGB, RREL
Edison, NJ  08837
                                    ABSTRACT


    Terra Vac Inc's vacuum extraction system was the subject of a SITE
program demonstration test in Grovel and, Massachusetts.  The site chosen was
contaminated with volatile organic compounds, mainly trichloroethylene, which
were used as degreasing solvents in an operating machine shop on the site.

    The eight-week test run produced the following results:

o   extraction of 1,300 Ib of VOCs                                    •  ..

o   a steady decline in the VOC recovery rate with time

o   a marked reduction in soil VOC concentration in the test area

o   an indication that the process can remove VOCs from clay strata

    The system operation proved to be very reliable.  Upon achievement of a
steady operation, the only stoppages occurred in order to replace spent
activated carbon canisters with fresh canisters.
                                     409
-------
                                 INTRODUCTION

    This SITE program demonstration test was planned to determine the
effectiveness of Terra Vac Inc's vacuum extraction technology in the
removal of volatile organic compounds from the vadose zone.  The location
of the test was on the property of an operating machine shop.  The property
is part of a Superfund site and is contaminated by degreasing solvents,
mainly trichloroethylene.

                                  OBJECTIVES

    The main objectives of this project were:

    The quantification of the contaminants removed by the process.

    The correlation of the recovery rate of contaminants with time.
o

o

o
    The prediction of operating time required before obtaining site
    remediation.

o   The effectiveness of the process in removing contamination from
    different soil strata.

                                   APPROACH

    The objectives of the project were achieved by following a
demonstration test plan which included a sampling and analytical plan.  The
sampling and analytical plan contained a quality assurance project plan.
This QAPP assured that the data collected during the course of this project
would be of adequate quality to support the objectives.

    The sampling and analytical program for the test was split up into a
pretest period, which has been called a pretreatment period; an active
period; midtreatment; and a posttreatment period.

    The pretreatment period sampling program consisted of:

o   soil boring samples taken with split spoons

o   soil boring samples taken with Shelby tubes

o   soil gas samples taken with punch bar probes

    Soil borings taken by split spoon sampling were analyzed for volatile
organic compounds (VOCs) using headspace screening techniques, purge and
trap, GC/MS procedures, and the EPA-TCLP procedure.  Additional properties
of the soil were determined by sampling using a Shelby tube, which was
pressed hydraulically into the soil by a drill rig to a total depth of 24
feet.  These Shelby tube samples were analyzed to determine physical
characteristics of the subsurface stratigraphy such as bulk density,
particle density, porosity, pH, grain size,  and moisture.  These parameters
were used to define the basic soil characteristics.
                                     410
-------
    Shallow soil gas concentrations were collected during pre-, mid-, and
posttreatment activities.  Four shallow vacuum monitoring wells and twelve
shallow punch bar tubes were used at sample locations.  The punch bar
samples were collected from hollow stainless steel probes that had been
driven to a depth of 3 to 5 feet.  Soil gas was drawn up the punch bar,.
probes with a low-volume personal pump and tygon tubing.  Gas-tight .50-ml
syringes were used to collect the sample out of the tygon tubing.

    The active treatment period consisted of collecting samples of:

o   wellhead gas

o   separator outlet gas

o   primary carbon outlet gas

o   secondary carbon outlet gas

o   separator drain water

    All samples with the exception of the separator drain water were  r
analyzed on site.  On-site gas analysis consisted of gas chromatography
with a flame ionization detector  (FID) or an electron capture detector
(ECD).  The FID was used generally to quantify the trichloroethylene (TCE)
and trans 1,2-dichloroethylene (DCE) values, while the ECD was used to
quantify the 1,1,1-trichloroethane (TRI) and the tetrachloroethylene (PCE)
values.  The use of two detectors, FID and ECD, was necessitated by high
concentrations of TCE in the extracted well head gas.  Owing to the high
TCE concentrations, most of the  samples injected on the ECD had to ,be
diluted.  Even with dilution factors of 333 to 1, the TCE concentration on
the ECD would exceed the linear  range of the detector, thus necessitating
the use of two detectors.

    The separator drain water-was analyzed for VOC content using SW846
8010.  Moisture content of the separator inlet gas from the wells was
analyzed using  EPA Modified Method 4.  This method is good for the
two-phase flow  regime that existed in the gas emanating from the wellhead.
Table  1 lists analytical methods  used for this project.             !

    The posttreatment sampling essentially consisted  of repeating
pretreatment sampling procedures  at locations as  close as possible to the
pretreatment sampling locations.                          . ... .

    The activated carbon canisters were sampled,  as close to the center  of
the canister- as possible, and these samples were  analyzed for  VOC content
as a check on the material balance for the process.   The method  used was
P&CAM  12.7, which consisted of desorption of the carbon with CS2  and
subsequent gas  chromatographic analysis.

                             PROCESS DESCRIPTION

    The vacuum  extraction process is a technique  for  the removal and.
                                    411
-------
                         TABLE  1.   ANALYTICAL METHODS
   Parameter
Analytical method
Sample Source
Grain size
PH
Moisture (110°C)
Particle density
Oil and grease
EPA-TCLP

TOC
Headspace VOC
VOC
VOC
VOC
VOC
VOC
VOC
ASTM D422-63
SW846* 9040
ASTM D2216-80
ASTM D698-78
SW846* 9071
F.R. 11/7/86,
Vol. 51, No. 216,
SW846* 8240
SW846* 9060
SW846* 3810
GC/FID or ECD
GC/FID or ECD
SW846* 8010
SW846* 8010
Modified P&CAM 127
SW846* 8240
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings

Soil borings
Soil borings
Soil gas
Process gas
Separator liquid
Groundwater
Activated carbon
Soil borings
*Third Edition, November 1986.
                                     412
-------
venting of volatile organic constituents (VOCs) from the vadose or
unsaturated zone of soils.  Once a contaminated area is completely defined,
an extraction well or wells, depending upon the extent of contamination,
will be installed.  A vacuum pump or blower induces air flow through the
soil, stripping and volatilizing the VOCs from the soil matrix into the air
stream.  Liquid water is generally extracted along with the contamination.
The two-phase flow of contaminated air and water flows to a vapor liquid
separator where contaminated water is removed.  The contaminated air stream
then flows through activated carbon canisters arranged in a parallel-series
fashion.  Primary or main adsorbing canisters are followed by a secondary
or backup adsorber in order to insure that no contamination reaches the
atmosphere.  Figure 1 illustrates the layout of wells and equipment.

                     EQUIPMENT LAYOUT AND SPECIFICATIONS

    Specifications are given in Table 2 for the equipment used in the
initial phase of the demonstration.  This equipment was later modified when
unforeseen circumstances required a shutdown of the system.  The
vapor-liquid separator, activated carbon canisters, and vacuum pump skid
were inside the building, with the stack discharge outside the building.
The equipment was in an area of the machine shop where used cutting oils
and metal shavings had been stored.

    Four extraction wells (EW1-EW4) and four monitoring wells (MW1  - MW4)
were drilled south of the shop.  Each well was installed in two sections,
one section to just above the clay lens and one section to just below the
clay lens.  The extraction wells were screened above the clay and below the
clay.  As shown in Figure 2, the well section below the clay lens was
isolated from the section above by a bentonite portland cement grout seal.
Each section operated independently of the other.  The wells were arranged
in  a triangular configuration, with three wells on the base of the  triangle
(EW2,  EW3, EW4) and one well at the apex  (EW1).  The three wells on the
base were called  barrier wells.  Their purpose was to  intercept
contamination, from underneath the building and to the side of the
demonstration area, before this contamination reached  the main extraction
well (EW1).  The  area enclosed by the four extraction  wells defined the
area to be cleaned.

INSTALLATION OF EQUIPMENT

    Well drilling and equipment setup were begun on December 1, 1987.  A
mobile drill rig  was brought in, equipped with hollow-stem augers,  split
spoons, and Shelby tubes.  The locations  of the extraction wells and
monitoring wells  had been staked out previously based  on contaminant
concentration profiles  from a previously  conducted remedial investigation
and from bar punch probe  soil gas monitoring.

     Each well drilled was sampled at 2-foot intervals  with a split  spoon
pounded  into the  subsurface by the drill  rig  in advance of the hollow  stem
auger. The hollow stem auger would then  clear out the soil down to the
depth  of the split spoon, and the cycle would  continue in that manner  to  a
depth  of 24 feet.  The  drilling tailings  were  shoveled into 55-gallon  drums
                                     413
-------
                                   MW1
1   Schematic diagram of equipment layout.
                 414
-------
                 2//PVCPIPE
                             — BENTONITE
                           3'
                                SAND
                                SCREENING
                           12.67'
                          «	GROUT
                               " BENTONITE
                                SCREENING
                            24'
Fig.   2   Schematic diagram of an extraction well
                    415
-------
                            TABLE 2.  EQUIPMENT LIST
      Equipment
   Number required
           Description
Extraction wells

Monitoring wells

Vapor-liquid
  separator

Activated carbon
  canisters


Vacuum pump skid


Holding tank

Pump
4 (2 sections each)

4 (2 sections each)

         1
Primary:  2 units in
  parallel
Secondary:  1 unit
2" SCH 40 PVC 24' total depth

2" SCH 40 PVC 24' total depth

1000-gal capacity, steel


Canisters with 1200 Ib of carbon
  in each canister - 304 SS
4" inlet and outlet nozzles

25 HP motor - positive displace-
  ment lobe type blower - 3250 rpm

2000-gal capacity - steel

1 HP motor - centrifugal
for eventual disposal.  After the holes were sampled, the wells were
installed using 2-inch PVC pipes screened at various depths depending upon
the characteristics of the soil in the particular hole.  The deep well  was
installed first, screened from the bottom to various depths.  A layer of
sand followed by a layer of bentonite and finally a thick layer of grout
were required to seal off the section below the clay lens from the section
above the clay lens.  The grout was allowed to set overnight before the
shallow well pipe was installed at the top of the grout.  A layer of sand
bentonite and grout finished the installation.

                      VOC  REMOVAL  FROM THE VADOSE ZONE

    The permeable vadose zone at the Groveland site is divided into two
layers by a horizontal clay lens,  which is relatively impermeable. As
explained previously, each extraction well had a separate shallow and deep
section to enable VOCs to be extracted from that section of the vadose  zone
above and below the clay lens.  The quantification of VOCs removed was
achieved by measuring

    o    gas volumetric flow rate  by rotameter and wellhead gas VOC
         concentration by gas chromatography
                                     416
-------
    o    the amount of VOCs adsorbed by the activated carbon canisters by
         desorption into C$2 followed by gas chromatography.

    VOC flow rates were measured and tabulated for each well section separ-
ately.  The results of gas sampling by syringe and gas chromatographic
analysis indicate a total of 1,297 Ib of VOCs were extracted over a 56-day
period, 95% of which was trichloroethylene.  A very good check on this
total was made by the activated carbon VOC analysis, the results of which
indicated a VOC recovery of 1353 Ib; virtually the same result was obtained
by two very different methods.

    One view of the reduction in VOC concentrations in the vadose zone can
be seen from examining the three-dimensional shallow soil gas plots.  Soil
gas was collected during pretreatment, midtreatment, and posttreatment from
punch bar probes and shallow vacuum monitoring wells.  The collection
points were located on a coordinate system with extraction well 1 as the
origin (0,0).  Each collection point has an x and y coordinate, and TCE
concentrations are plotted on a "Z" scale, which gives a three-dimensional
plot across the grid.  Values of "Z" between data points and around the
grid are generated by the Kriging technique, which uses given data points
and a regional variable theory to generate values between and around sample
locations.  Kriging is the name given to the least squares prediction of
spatial processes and is used in surface fitting, trend surface analysis,
and contouring of sparse spatial data.

    A total of twelve shallow punch bar tubes were utilized along with the
four shallow vacuum monitoring wells.  The punch bars were driven to a
depth of 3 to 5 feet, and as with the vacuum wells, soil gas was drawn up
the punch bar probes with a low-volume personal pump and tygon tubing.
50-ml gas-tight syringes were used to collect the sample out of the tygon
tubing.  The gas samples were analyzed in the field trailer using gas
chromatographs with flame ionization detectors and electron capture
detectors.

    The soil gas results show a considerable reduction in concentration
over the course of the 56-day demonstration period as can be seen from
Figures 3 and 4.  This is to be expected since soil gas is the vapor halo
existing around the contamination and should be relatively easy to remove
by vacuum methods.

    A more modest reduction can be seen in the results obtained for soil
VOC concentrations by 6C/MS purge-and-trap analytical techniques.  Soil
concentrations include not only the vapor halo but also interstitial liquid
contamination that is either dissolved in the moisture in the soil or
existing as a two-phase liquid with the moisture.

    Table 3 shows the reduction of the weighted average TCE levels in the
soil during the course of the 56-day demonstration test.  The weighted
average TCE level was obtained by averaging soil concentrations obtained
every two feet by split spoon sampling methods over the entire 24-foot
depth of the wells.  The.largest reduction in soil TCE concentration.
occurred in EW4, which had the highest initial level of contamination.
                                    417
-------
                 MU2
a
a
      0
      Fig.   3    Pretreatment  shallow  soil-gas concentration
                                  418
-------
                                              MAP VIEW
    MW2
                             MU3
                                               MUJ4
Fig.  4    Posttreatment shallow soil-gas concentration
                            419
-------
           TABLE 3.  REDUCTION OF WEIGHTED AVERAGE TCE LEVELS IN SOIL
                     (TCE  Cone,  in mg/kg)
Extraction  Well
Monitoring Well

       1
       2
       3
       4
Pretreatment
Posttreatment
    1.10
   14.75
  227.31
    0.87
     0.34
     8.98
    84.50
     1.05
% Reduction
1
2
3
4
33.98
3.38
6.89
96.10
29.31
2.36
6.30
4.19
13.74
30.18 . .
8.56
95.64
   69.09
   39.12
   62.83
EW1, which was expected to achieve the greatest concentration reduction,
exhibited only a minor decrease over the course of the test.  Undoubtedly
this was because of the greater-than-expected level of contamination that
existed in the area around MW3 that was drawn into the soil .around EW1.
The decrease in the TCE level around MW3 tends to bear this out.

            EFFECTIVENESS OF THE TECHNOLOGY IN VARIOUS SOIL TYPES

    The soil strata at the Groveland site can be characterized generally as
consisting of the following types in order of increasing depth:

    o    medium to very fine silty sands

    o    stiff and wet clays

    o    sand and gravel

    Soil porosity, which is the percentage of total soil volume occupied by
pores, was relatively the same for both the clays and the sands.
Typically, porosity over the 24-foot depth of the wells would range between
40% and 50%.  Permeabilities, or more accurately hydraulic conductivities,
ranged from 10"* cm/sec for the sands to 10"° cm/sec for the clays,
with corresponding grain sizes equal to 10"1 mm to 10"3 mm.

    Pretest soil boring analyses indicated in general  that most of the
contamination was in the strata above the clay lens with a considerable
quantity perched on top of the clay lens.  This was the case for EW4, which
                                     420
-------
showed an excellent reduction of TCE concentration in the medium to fine
sandy soils existing above the clay layer,  with no TCE detected in the clay
in either the pretest or posttest borings (see Table 4).   One of the wells,
however, was an exception.  This was MW3, which contained the highest
contamination levels of any of the wells, and was exceptional in that most
of the contamination was in a wet clay stratum.  The levels of
contamination were in the 200-1600 ppm range before the test.

    After the test, analyses of the soil boring adjacent to MW3 showed
levels in the range of ND-60 ppm in the same clay stratum.  The data, as
shown in Table 5, suggest that the technology can desorb or otherwise
mobilize VOCs out of certain clays.

    From the results of this demonstration it appears that the permeability
of a soil need not be a consideration in applying the vacuum extraction
technology.  This may be explained by the fact that the porosities were
approximately the same for all soil strata, so that the total flow area for
stripping air was the same in all soil strata.  It will take a long time
for a liquid contaminant to percolate through clay with its small pore size
and consequent low permeability.  However, the much smaller air molecules
have a lower resistance in passing through the same pores.  This may
explain why contamination was generally not present in the clay strata, but
when it was, it was not difficult to remove.  Further testing should be
done in order to confirm this finding.

                 CORRELATION OF DECLINING VOC RECOVERY RATES

    The vacuum extraction of volatile organic constituents from the soil
may be viewed as an unsteady state process taking place in a nonhomogeneous
environment acted upon by the combined convective forces of induced
stripping air and by the diffusion of volatiles from a dissolved or sorbed
state.  As such  it is a very complicated process, even though the equipment
required to operate the process is very simple.

    Unsteady state diffusion processes  in general correlate well by
plotting the logarithm of the rate of diffusion versus time.  Although the
representation of the vacuum extraction process presented here might be
somewhat simplistic, the correlation obtained by plotting the logarithm of
the concentration of contaminant  in the wellhead gas versus time and
Obtaining a least squares best fit line was reasonably good.  This type of
plot, shown in Figure 5, represents the data very well and is more valid
than both a linear graph or one plotting concentration versus log time, in
which a  best fit curve would actually predict gas concentrations of zero or
less.

    Looking at the plots  for EW1,  shallow  and deep, equations are given for
the least  squares  best fit line for the data points.   If  the vacuum
extraction process is run long enough so that the detection  limit for TCE
on the  ECD, which  is 1 ppbv,  is reached, the length of time  required to
reach that concentration would be approximately 250 days  on  the  shallow
well  and approximately 300 days on the  deep well.
                                     421
-------
 TABLE  4.   EXTRACTION  WELL  4:
          TCE REDUCTION IN SOIL STRATA
Depth
ft
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
Description Permeability TCE Cone, com
of strata
Med. sand w/gravel
Lt. brown fine sand
Med. stiff It. brown fine sand
Soft dk. brown fine sand
Med. stiff brown sand
V. stiff It. brown med. sand
V. stiff brown fine sand w/silt
M. stiff grn-brn clay w/silt
Soft wet clay
Soft wet clay
V. stiff brn med-coarse sand
V. stiff brn med-coarse w/gravel
cm/sec
1°1
10-J
10-5
10";
lo-J
lo-J
lO-J
IQ-f
10"°
10"!
lo-J
lO'3
pre
2.94
29.90
260.0
303.0
351.0
195.0
3.14
ND
ND
ND
ND
6.71
post
ND
ND
39
9
ND
ND
2.3
ND
ND
ND
ND
ND
TABLE 5.  MONITORING WELL 3:
          TCE  REDUCTION  IN SOIL STRATA
Depth
ft
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
Description
of strata
M. stiff brn. fine sand
M. stiff grey fine sand
Soft It. brn. fine sand
Lt. brn. fine sand
Stiff V. fine brn. silty sand
Silty sand
Soft brown silt
Wet green -brown silty clay
Wet green -brown silty clay
Wet green-brown silty clay
Silt, gravel, and rock frag.
M. stiff It. brn. med. sand
Permeability
cm/sec
1Q-5
10"?
10"!
10-J
lo-J
10-J
10-J
10 Q
1° "I
10"?
10-J
ID'4
TCE Cone.
pre
10.30
8.33
80.0
160.0
ND
NR
316.0
195.0
218.0
1570.0
106.0
64.1
ppm
post
ND
800
84
ND
63
2.3
ND
ND
62
2.4
ND
ND
          422
-------
CROVELAND/TERRA-VAC DEMONSTRATION


 _ 1000
 £
 CL
O
cr
 LU
 CJ

 O
 C>

 Ld
 O

    °'1
   0.0.1
                              EXTRACTION WELL #1
                              SHALLOW
                                                     CURVE COEFFICIENT- R2=0 62.
        0       20       40       60       80       100
             DAY  OF  ACTIVE  TREATMENT
          Fig.  5    Wellhead TCE concentration vs. time
                             423
-------
               PREDICTION OF TIME REQUIRED FOR SITE REMEDIATION

    The soil concentration  that would be  calculated from the wellhead gas
concentration using Henry's Law is  included  in the last column of Table 6.
Calculations for the predicted soil concentrations were made assuming a
bulk density of the soil of 1761  kg/nr3, a total porosity of 50%, and a
moisture content of 20%.  The calculated  air filled porosity of the soil is
approximately 15%.  Henry's constant was  taken to be 0.492 KPa/m3-gmol at
40°F.

    Given the nonhomogeneous nature of the subsurface contamination and
interactions of TCE with organic  matter in the soil, it was not possible to
obtain a good correlation between VOC concentrations in wellhead gas and
soil in order to predict site remediation times.  Henry's Law constants
were used to calculate soil  concentrations from wellhead gas concentrations
and the calculated values obtained, correcting for air filled porosity,
were lower than actual soil  concentrations by at least an order of
magnitude (see Table 6).

                   TABLE 6.   COMPARISON  OF WELLHEAD GAS VOC
                              CONCENTRATION AND SOIL VOC CONCENTRATION
Extraction Well
TCE concentration
 in wellhead gas
      ppmv
TCE concentration
  in soil ppmw
Predicted by
Henry's Law
    ppmw
IS
ID
2S
2D
3S
3D
4S
9.7
5.6
16.4
14.4
125.0
58.7
1095.6
54.5
7.2
ND
20.4
20.9
18.0
9.1
0.11
0.07
0.20
0.17
1.5.3
0.74
12.49
    Before one can attempt to make a rough estimation of the remediation
time, a target value for the particular contaminant in the remediated soil
must be calculated.  This target concentration is calculated by using two
mathematical models, the Vertical and Horizontal Spread ModelW and the
Organic Leachate Model^a'.  The mathematical models allow the use of a
regulatory standard for drinking water in order to arrive at a target soil
concentration.

    The VHS model is expressed as the following equation:

    Cy = C0 erf (Z/(2(azY)°-5)) erf (X/(atY)°-5)
                                     424
-------
    where:

    Cy   =    concentration of VOC at compliance point (mg/1)
    CQ   =    concentration of VOC in leachate (mg/1)
    erf  =    error function (dimensionless)
    Z    =    penetration depth of leachate into the aquifer
    Y    =    distance from site to compliance point (m)
    X    =    length of site measured perpendicular to the direction of
              ground water flow (m)
    at   =    lateral transverse dispersivity (m)
    az   =    vertical dispersivity (m)

    A simplified version of the VHS model is most often used,  which reduces
the above equation to:

    cy = cocf

    where:

    Cf   =    erf (Z/(2(azY)°-5)) erf (X/(atY)°-5), which is
              reduced to a conversion factor corresponding to the amount of
              contaminated soil

    The Organic Leachate Model (OLM) is written as:

    C0 = 0.00211 cs°'678S°-373
    where:
              concentration of VOC in leachate (mg/1)
              concentration of VOC in soil (mg/1)
              solubility of VOC in water  (mg/1)
                                                                   This
    The regulatory standard for TCE in drinking water is 3.2 ppb.
regulatory limit is used in the VHS model as the compliance point
concentration in order to solve for a value of the leachate concentration.
This value of leachate concentration is then used in the OLM model to solve
for the target soil concentration.

    Once the target soil concentration is determined, a rough estimation of
the remediation time can be made by taking the ratio of soil concentration
to wellhead gas concentration and extrapolating in order to arrive at a
wellhead gas concentration at the target soil concentration.  The
calculated target soil concentration for this site is 500 ppbw.  This
corresponds to an approximate wellhead gas concentration of 89 ppb for
EW1S.  The equation correlating wellhead gas concentration with time (see
Figure 5) is then solved to give 150 days running time.

    After 150 days the vacuum extraction system can be run intermittently
to see if significant increases in gas concentrations occur upon
restarting, after at least a two day stoppage.  If there are no appreciable
increases in gas concentration, the soil has reached its residual
                                    425
-------
equilibrium contaminant concentration and the system may be stopped and
soil borings taken and analyzed.

(a) EPA Draft Guidelines for Petitioning Waste Generated by the Petroleum
    Refinery Industry, June 12, 1987.

                               ACKNOWLEDGMENTS

    The authors wish to thank Mr. James S. Ciriello, formerly of the U.S.
Environmental Protection Agency, Region I, Boston, Massachusetts at the
time of the project, for his efforts during the course of this project.  A
special note of gratitude is to be given to Mr. Thomas Quinlan of the
Valley Manufactured Products Company, Inc. for his special  support and
cooperation that helped make this project a successful one.
                                    426
-------
              THE OFFICE OF RESEARCH & DEVELOPMENT
                          WRITE PROGRAM
                          Ivars J.  Licis
                     Research Program Manager
              Risk Reduction Engineering Laboratory

                               and

                           M.  Lynn  Apel
                Chief,  Process Engineering  Section
              Risk Reduction Engineering  Laboratory,
              U.S.  Environmental Protection Agency
                      Cincinnati,  OH  45268
     Passage of the 1984 Hazardous and Solid Waste Amendments
(HSWA) to the Resource Conservation and Recovery Act (RCRA) of
1976 marked a strong change in the policies of the United States
concerning the generation of hazardous and nonhazardous wastes.
In addition to authorizing very stringent treatment and disposal
regulations, the Amendments also indicated, as the Nation's top
waste management priority a redirection toward "waste
minimization" as a preferential strategy for encouraging
improvement in environmental quality.

     To carry out the intent of the Amendments to reduce the
generation of waste, the U.S. Environmental Protection Agency
(EPA) has developed a comprehensive pollution prevention program
addressing the release and transport of hazardous, toxic, and
nonhazardous materials through air, water, and solid media.  The
EPA pollution prevention program includes information gathering,
research and development, demonstration, support of state and
local government waste minimization and pollution prevention
programs, training and education, technology transfer activities,
waste minimization assessments, and extensive communication with
universities, state and local governments, and the general
public.

     This paper describes one of the major pollution prevention
research programs being undertaken by EPA, the Waste Reduction
Innovative Technology Evaluation [WRITE] Program.  The WRITE
Program has been designed to identify, evaluate, and demonstrate
new ideas and technologies that lead to waste reduction.  The
WRITE Program involves the cooperative efforts of the EPA, state
and local governments, private industry, universities,  and other
organizations to encourage the research and development of
                               427
-------
effective techniques  and technologies  for multimedia pollution
prevention.
BACKGROUND

     During the last decade, waste management and disposal
regulations have changed drastically.  During the 1970's, the
seriousness of the hazardous waste problem prompted the Congress' *
to enact the Resource Conservation and Recovery Act in 1976.  By
the early 1980's, it had become apparent that even well-regulated
land disposal could cause significant environmental damage.  The
Amendments to RCRA resulted in the banning of the land disposal
of over 400 chemicals and hazardous wastes.  During the 1980's,
significant effort was devoted to research, development, and
commercialization of hazardous waste treatment technologies.
Although some efforts were successful in substantially mitigating
the problem of existing waste, it became apparent that treatment
alone would not solve all of the problems associated with the
increasing amounts of waste being generated.

     Congress recognized this fact while drafting the national
policy on hazardous wastes and directed EPA to report on the
feasibility and desirability of developing mandatory requirements
to compel the adoption of pollution prevention techniques.  The
1984 HSWAs to RCRA define the national policy on hazardous waste
management as follows:

          "The Congress hereby declares it to be the
          national policy of the United States that,
          wherever feasible, the generation of
          hazardous waste is to be reduced or
          eliminated as expeditiously as possible.
          Waste that is nevertheless generated should
          be treated, stored, or disposed of so as to
          minimize the present and future threat to
          human health and the environment." [1]

In 1986, EPA responded to the Congressional request  with a
Report to Congress on the Minimization of Hazardous Waste [2].
In this report, the Agency defined waste minimization as:

          "The reduction, to the extent feasible, of
          hazardous waste that is generated or
          subsequently treated, stored or disposed of.
          It includes any source reduction or
          recycling activity undertaken by a generator
          that results in either (1) the reduction of
          total volume or quantity of hazardous waste
          or (2) the reduction of toxicity of
          hazardous waste, or both, so long as the
          reduction is consistent with the goal of
                                428
-------
          minimization of present and  future threats
          to human health and environment."
     In January 1989, the Agency issued a Pollution Prevention
Policy Statement [3].  This policy encourages organizations,
facilities and individuals to fully utilize source reduction and
recycling practices and procedures to reduce risk to public
health, safety and the environment.  Source reduction is
emphasized as the preferred approach since it eliminates or
significantly reduces the quantity of pollutants generated,
thereby significantly reducing the potential risk to health and
the environment.  As a second choice, recycling reduces the
amount of waste to be treated or sent for disposal.
Additionally, many of the waste reduction efforts have proven to
be both more economical, in the long run, as well as beneficial
to the environment producing a win-win situation.  On this basis,
source reduction and recycling are the primary two elements in
the Waste Management Hierarchy  preceding treatment and disposal.
In this context, source reduction has been defined as the
reduction or elimination of waste at the source, usually within a
process.  Source reduction measures include process
modifications, feedstock substitutions, improvements in feedstock
purity, housekeeping and management procedural changes, increases
in the efficiency of equipment, and recycling within a process.
Likewise, recycling has been defined as the use or reuse of a
waste material as an effective substitute for a commercial
product or as an ingredient or feedstock in an industrial
process.  It includes the reclamation of useful constituent
fractions within a waste material or the removal of contaminants
from a waste to allow it to be reused [4].

EPA'S POLLUTION PREVENTION PROGRAM

     In its Report to Congress, the Agency explored various
technical, economic, and policy issues relevant to the reduction
and recycling of hazardous and nonhazardous wastes, and concluded
that it would be counterproductive for EPA to establish a
mandatory program for waste minimization at this time.  The
Report to Congress stressed that the most constructive role
government could assume is to promote voluntary waste
minimization by providing information, technology transfer, and
assistance to waste generators.  The Agency proposed a waste
minimization program to encourage industry to accelerate efforts
to reduce the generation of wastes through implementation of
process changes and/or the incorporation of recycling methods.

     In its efforts to pursue the objectives set forth by
Congress, EPA has established a national program to effect waste
minimization, or what has since become included in the term
"pollution prevention".  Pollution prevention is a term that has
been used more frequently within the last year by several
organizations including EPA to describe techniques, practices, or
procedures implemented by the private and public sectors to
                               429
-------
prevent the generation of pollutants.  As it is used today,
"pollution prevention" has replaced the term "waste minimization"
which was generally applied to reducing the generation of
hazardous wastes.  Through elimination of this term, that may be
perceived as closely tied to RCRA, EPA is emphasizing that it's
pollution prevention policy has applicability beyond the RCRA
hazardous waste context.

     EPA's pollution prevention program is based on several
elements including: (1) promoting waste reduction by transferring
technical information to firms and state and local government
technical assistance officers; (2) encouraging the adoption of
waste reduction by identifying, evaluating, and demonstrating
appropriate technologies and by promoting the use of waste audits
or assessments; (3) keeping Congress advised on national progress
of waste reduction techniques and practices; (4) fostering
development of state and local government pollution prevention
programs; (5) developing outreach and communication programs with
the goal of raising awareness of the benefits of waste reduction
practices within government, industry, and the public.

     To accomplish some of these objectives and encourage the
identification, development, and demonstration of processes and
technologies that result in less waste being generated, EPA's
Office of Research and Development (ORD) has initiated several
multimedia pollution prevention research programs.  These
programs are designed to be the cornerstone of the Agency's
pollution prevention research and demonstration effort and will
provide much needed data on new technologies to the generating
sectors. Addressed under these programs are hazardous,
nonhazardous, industrial and municipal wastes.  Design and
implementation of these programs is being undertaken by the
Pollution Prevention Research Branch  (formerly the Waste
Minimization Branch) of the Risk Reduction Engineering Laboratory
of ORD in Cincinnati, Ohio.

     The Agency has also established a Pollution Prevention
Office (PPO) within the Office of Policy, Planning and Evaluation
in Washington, D.C.  The Pollution Prevention Office is
responsible for award and oversight of the Source Reduction and
Recycling Technical Assistance grants, establishing a National
Pollution Prevention Information Clearinghouse, and publishing a
quarterly newsletter on pollution prevention topics.  PPO
together with ORD and the Agency's media-specific offices  (i.e.,
Office of Solid Waste, Office of Toxic Substances, etc.) will
develop and implement EPA's multi-media pollution prevention
program.


THE WRITE PROGRAM

     Reducing the generation of industrial and other wastes can
be achieved in many ways.  Process chemistry can be changed;
potential waste streams can be recycled within a manufacturing
                               430
-------
process or back into the process; process technology and/or
equipment can be modified to produce products more efficiently,
resulting in less waste; plant operations, i.e., "housekeeping"
methods, can be changed or controlled to produce fewer and
smaller waste streams or less waste in general; changes in raw
materials (feedstocks) can lead to fewer waste streams or
less-hazardous waste streams; finally, changes in the end
products from manufacturing operations can, in some instances, be
made so as to affect the types and quantities of wastes emitted.
The EPA WRITE Program outlined below is designed to identify,
evaluate and enhance the application of technologies and
practices containing these attributes.

     The WRITE Program is a research program designed to
identify, evaluate, and/or demonstrate the use of innovative
engineering and scientific technologies to reduce the volume
and/or toxicity of wastes produced from the manufacture,
processing,  and use of materials.  The WRITE Program is broad in
technical scope and addresses the reduction of pollutants across
all environmental media: air, land, surface water, and
groundwater.  Attention is directed toward methodologies with the
potential for reducing the quantity and/or toxicity of waste
produced at the source of generation, or to achieve practicable
onsite reuse or recycling of waste materials.  Strong
consideration is given to the applicability of a technique on an
industry-wide basis and across industries.  Industries of primary
interest under the WRITE Program include chemical, fabricated
metal, electronic, printing and publishing, lumber, petroleum,
transportation, food, and textile [5].

     The objectives of the WRITE Program are:

        0  To establish reliable performance and cost
           information on pollution prevention techniques by
           conducting  evaluations or demonstrations of the more
           promising innovative technologies.

        0  To accomplish an early introduction of waste
           reduction techniques into broad commercial practice.

        0  To encourage active participation of small- and
           medium-sized companies in evaluating and adopting
           pollution prevention  concepts by providing support to
           these companies through State and local government
           agencies.

           To encourage the transfer of knowledge and technology
           concerning pollution prevention practices between
           large, medium-sized, and small industries.

        0  To provide solutions to important chemical-,
           wastestream-, and industry-specific pollution
           prevention  research needs.
                              431
-------
     For the purposes of this program, innovative waste
minimization technologies are defined as those technologies that
represent an alternative approach to conventional waste
management methods  (i.e., alternatives to incineration, land
disposal, treatment, etc.).  This approach involves substantially
reducing the volume or toxicity of waste generated at the source
and/or recycling or reusing the waste materials.  In some cases,
the technologies may incorporate novel modifications to a
component of an existing process or operation, or the adaptation
of an existing technology to a new process.

        To date the WRITE Program encompasses approximately 12
research and demonstration projects totalling over $4.0 million.
Under these projects, approximately 30 waste reduction
technologies will be evaluated throughout the next three years,
and several long-term waste- and industry-specific research
studies will be undertaken.  These activities are conducted under
three subprograms of the WRITE Program; the WRITE Pilot Program
with State and Local Governments, the WRITE Program With
Industry, and the WRITE Research Program [see WRITE Program
Chart).


WRITE PILOT PROGRAM WITH STATE AND LOCAL GOVERNMENTS

        At the moment, the "WRITE Pilot Program with state and
Local Governments" is the largest subprogram under WRITE and
addresses immediate information transfer needs between government
and industry.  Through the joint efforts of EPA and various state
and local governments, technical and economic evaluations of
source reduction and recycling technologies are being conducted
of manufacturing and  processing operations across approximately
twenty industries.  This joint approach was chosen because State
and local government officials are often more familiar with local
industrial practices and regional manufacturing and economic
interests that can affect the potential success and widespread
applicability of proposed pollution prevention technologies.
States currently participating in this program include
California, Connecticut, Illinois, Minnesota, New Jersey, and
Washington.

     Under this program, $100,000 per year is provided by EPA to
each participating state/local government.  The state/local
government also contributes additional matching funds ranging
from 33% to 50% of the cost of the research.  An average of five
waste reduction technologies will be evaluated by each
state/local government during the 3-year period of this pilot
program.  Waste reduction technologies evaluated under this
program are based on several selection criteria.  These include:
(1) type of waste minimization technology, (2) status of
development, (3) unique nature of the technology, (4)
application, (5) source reduction performance capability, (6)
extent of process modification,  (7) cost effectiveness of the
technology, (8) process safety and health considerations, (9)
                               432
-------
cost to the EPA and state/local government, and  (10)
legal/contractual issues.  A worth assessment model is then used
as one of the decision tools to evaluate and rank potential waste
minimization technologies.

     The technical and economic evaluations conducted for each
technology include an in-depth study of the process, a literature
review of comparable processes, material and energy balance
computations, a field demonstration, and determination of cost
estimation parameters including itemization of capital and
operational costs, calculation of the payback period and return
on investment.  A summary of the types of information collected
during a technical and economic evaluation of a waste
minimization technology under this program is shown in Table 1.
The example concerns the modification of a cold solvent cleaning
process.  In the cold cleaning of ball bearings with solvents,
using a two-step countercurrent cleaning sequence can increase
the cleaning efficiency.  It can also substantially reduce the
solvent requirement and, hence, the waste generation.  This
process does not involve substantial equipment modification.
Material balance calculations indicate a waste reduction of 50
percent and a 33 percent reduction in fresh solvent requirements
[6].


WRITE PROGRAM WITH INDUSTRY

     The "WRITE Program With Industry" focuses on  evaluations of
waste reduction technologies currently in use or under
development by large industries.  One of the  objectives of this
program is to encourage the transfer of knowledge and technology
concerning pollution prevention practices between large,
mid-size, and small industries.  Under the "WRITE Program With
Industry", evaluations of waste reduction technologies are
performed directly with industrial firms or through industrial
trade associations and/or technical societies.


WRITE RESEARCH PROGRAM

     In addition to evaluation programs, the WRITE Program  has a
research subprogram.  This subprogram focuses on pollution
prevention research needs, i.e., the generation of data to allow
the future demonstration of emerging new pollution prevention
techniques.  Projects under this component of the WRITE  Program
address various technical obstacles to waste  reduction and to
chemical-, wastestream-, and industry- specific waste
minimization issues.  These research efforts are conducted with
industrial firms,  universities, other government agencies,
technical societies, and industrial trade organizations.

        While emphasis to date has been on industrial processes
the area of focus is being enlarged to include  techniques for
pollution prevention for any source and especially those that are
identified with posing significant risk in terms of high toxicity
                               433
-------
or high threat or those producing large amounts of pollution.


CONCLUSIONS

        The WRITE Program will continue to expand and address
pollution prevention research needs within government and
industry.  A number of challenging areas have been identified and
are receiving attention.  An effort is being launched to provide
a way to prioritize needed work.  This is especially significant
with the change of focus from the industrial and hazardous waste
area to all sources of pollution.  Another area deals with the
means by which technologies can be best evaluated in terms waste
reduction and cost when the differences between the old and the
new technology involve different pollutants at differing
toxicities, concentrations and exposure pathways.
  Additional information concerning this and other EPA pollution
prevention research programs can be obtained from the  Pollution
Prevention Research Branch  of the Risk Reduction Engineering
Laboratory, U.S. EPA, Cincinnati, Ohio.
REFERENCES
1.
4.
6.
U.S. Congress.  Hazardous and Solid Waste Amendments.
Washington, D.C. 1984.

U.S. Environmental Protection Agency.  Report to
Congress, Minimization of Hazardous Wastes.  U.S.   EPA,
Office of      Solid Waste, Washington, DC., EPA/530-SW-
86-033. 1986.

U.S. Environmental Protection Agency.  Pollution
Prevention Policy Statement.  Federal Register, Vol,. 54,
No. 16, January 26, 1989, p. 3845.

Freeman, H.M., and J. Lounsbury.  The U.S. EPA Hazardous
Waste Minimization Program.  Proceedings of American
Institute of  Chemical Engineers Annual Meeting,
Washington, D.C.  1988.

Apel, M.L., H.M. Freeman, M.F. Szabo, S.H. Ambekar.
Guidance Document for the WRITE Pilot Program With State
and Local Governments.  U.S. EPA, Risk Reduction
Engineering Laboratory,  Cincinnati, OH. 1988.

Jacobs Engineering.  Waste Minimization Audit Report:
Case Studies  of Minimization of Solvent Waste from
Parts Cleaning and From Electronic Capacitor
Manufacturing Operations.  U.S. EPA, Risk Reduction
Engineering Laboratory, Cincinnati, OH,
EPA/600/52-87/057.  1987.
                                434
-------
  TABLE 1.  SUMMARY OF ENGINEERING EVALUATION FOR COLD SOLVENT
                    CLEANING PROCESS MODIFICATION [4]
Type of
Application

Stage of
Development

Unique Nature of
Technology

Applications of
cold solvent
Performance
Need for
Modification
Process modification
Demonstration
First-of-a-kind demonstration
Reduces hazardous waste generation in the
cleaning operations of the parts cleaning
industry, which  is a medium/small-scale
operation.

Achieves 50% waste reduction by reducing the
fresh  solvent requirement by 33%

Requires essentially only minor equipment
modification.
Cost Effectiveness  Added capital costs = $ 600
of Technology       Net operating savings = $ 380 per year
                    Payback period = 1.6 years
                    Note: Net operating savings include savings
                    resulting  from  reduced waste  disposal,
                    reduced  solvent requirement,  and operation
                    and  maintenance expense.
Safety & Health
 Properly designed  system  is considered  safe.
 Metal  cleaning  systems are routinely used  in
 industry without any  safety or health
 problems.
                               435
-------
CC
O
O
cc
CL
111

5E
UJ


UJ
DC
cn
CL
CQ

CO
                 CO
   CO


o<^



ZDOO

CQ
UJZ>
   CO
                                          O
                                          rr
                                          "3 «3
                                          < ^
                                          LU<
                                          CODC
                                          cc
                                                 «
                                                       i=Ol-
                                       w
DCt±
CD CO
OZ>
DCQ


[r
DC
                                  CM
                                         ; Q ^ LJLJ uj
                                         • ^ {/J O ft"
                                       3 pc N o S
                                                          : LU
                                                   ZZQi
                                                                           CO

                                                                           O)
                                                                           LLJ


                                                                           DC
                                                                           CD
                                                                         CO
                                                                        'a
                                                                                       Q.
                                          into
                              DC
                              CD

                              S:
                                            cc
                                            UJ

                                           ,
                                            Q
                                                     cc
                                                    2 Q
                                                    > y
                                                    o!
                                        Q O
                                            436
-------
                      Solidification/Stabilization as a
                 Best Demonstrated Available Technology for
                Resource Conservation and Recovery Act Wastes

                                     by

                  R. Mark Bricka and M. John Cullinane, Jr.

                          Environmental Laboratory
                           Department of the Army
              Waterways Experiment Station, Corps of Engineers
                                P.O. Box 631
                      Vicksburg, Mississippi 39180-0631

                                  ABSTRACT

     Early in 1987, the U.S. Army Engineers Waterways Experiment Station
(WES) was tasked by the U.S. Environmental Protection Agency (USEPA) to
investigate solidification/stabilization (S/S) as a Best Demonstrated Avail-
able Technology (BOAT).  Under this program EPA supplied WES with a number
of Resource Conservation and Recovery Act (RCRA) listed hazardous wastes for
evaluation.

     Nine listed wastes were processed using three generic S/S processes:
Portland cement, kiln dust, and lime/flyash.  The physical and leaching
characteristics of the treated materials were evaluated.  Physical charac-
teristics were evaluated using the unconfined compressive-strength (UCS)
test.  The leaching characteristics of the wastes were evaluated using the
Toxicity Characteristic Leaching Procedure (TCLP).  The TCLP extracts were
analyzed for a total of 22 metals.  No attempt was made to analyze for semi-
volatile and volatile compounds in the TCLP extracts, due to the absence of
these compounds in the raw wastes.  The effectiveness of the S/S processes
in immobilizing the metals was evaluated by comparing the TCLP extract
result of the raw (unsolidified/unstabilized) waste to the TCLP extract
result of the solidified/stabilized waste.

     Preliminary results of the physical tests indicate that the solidified/
stabilized wastes developed strengths ranging from not measurable to
4400 psi.  The results of the TCLP extractions indicate that mercury and
lead concentrations as high as 26.8 and 1.16 mg/1 respectively, were leached
from the solidified/stabilized wastes.  The effectiveness of the S/S pro-
cesses appears to be dependent on the specific waste and metal of interest.
                                     437
-------
            INTRODUCTION
Background
     Amendments to the Resource Con-
servation and Recovery Act  (RCRA)
enacted through the Hazardous and
Solid Waste Amendments of 1984
(HSWA), impose substantial new
responsibilities on handlers of haz-
ardous waste.  In particular, these
amendments prohibit the continued
land disposal of untreated hazardous
waste beyond specified dates,"
...unless the Administrator deter-
mines that the prohibition...is not
required in order to protect human
health and the environment for as
long as the waste remain haz-
ardous..." (RCRA sections 3004 (d)
(1), (e) (1), (g) (5), 42USC 6924 (d)
(1), (e) (1), and (g) (5) ).

     To expedite the development of
treatment standards, various dead-
lines have been established for
Agency action.  Land disposal of a
particular group of hazardous wastes
is prohibited after certain deadlines
if the USEPA has not set treatment
standards under RCRA section 3004 (m)
for such wastes or determined, based
on a case by case specific petition,
that there will be no migration of
hazardous constituents from the
wastes for as long as the wastes
remain hazardous.  Additional dead-
lines result in conditional restric-
tions on land disposal taking effect
if treatment standards have not been
promulgated or if a petition has not
been granted.

     Treatment standards will be
established based on Best Demon-
strated Available Technology (BOAT)
and developed in accordance with RCRA
section 3004 (m).  USEPA (1986a)
defined a technology as best, demon-
strated and available as follows:
     a.  Best—if several technologies
are available for treating the same
(or similar) waste(s), the waste
treatment method which reduces the
concentration and/or the migration of
contaminants most effectively is con-
sidered best.

     b.  Demonstrated—for a waste-
treatment technology to be considered
demonstrated, a full scale facility
must be known to be in operation for
treating the waste.

     £.  Available—for a waste-
treatment technology to be considered
available, it must:  (a) be able to be
purchased or licensed from the propri-
etor if a technology is a proprietary
or patented process; and (b) provide
substantial treatment.

Solidification/stabilization (S/S) is
one technology that meets the demon-
strated and available criteria (USEPA
1986b).  This process has been pro-
posed as a hazardous waste treatment
method that has the potential to sub-
stantially reduce the likelihood of
contaminant migration at land disposal
sites.

Purpose and Scope

     The purpose of this study was to
determine if S/S techniques could be
applied to specific wastes and to
determine the effect of S/S on the
wastes.  The physical and chemical
properties of the treated wastes were
evaluated to determine if S/S tech-
niques substantially reduced the
amount of hazardous contaminants in
the leachate and improved the physical
handling properties of the wastes.
The study described in this paper was
initiated by the USEPA to evaluate S/S
technology as a BDAT and to develop
data to support the establishment of
treatment standards.
                                     43G
-------
     This paper is not intended to,
nor does it, make any attempt to
determine whether S/S is a BOAT for
any specific waste.  This determina-
tion will be made by the USEPA in
accordance with their regulatory
procedures.

        MATERIALS AND METHODS

Wastes of Interest

     Nine waste types listed under the
first third scheduled wastes (USEPA
1988) were evaluated in this study.
Table 1 lists a description of these
wastes, their appropriate USEPA waste
identification numbers, and a single
contaminant of interest.  While many
contaminants for each waste were mea-
sured, only a single contaminant is
discussed to simplify this paper.

General Description
of the S/S processes

     Three S/S processes were used to
treat the wastes and are differen-
tiated by the type of binder material
used in the process.  The three pro-
cesses included:  Portland cement,
kiln dust, and lime/flyash.  The S/S
processes involved the addition of
water and binder material to a waste
followed by a mixing and curing
period.  A schematic diagram of S/S
processing performed during this
study is shown in Figure 1.

Initial Screening Test

     The moisture content of the
wastes varied from almost zero per-
cent to 95 percent water (water/
weight basis).  The first step was to
determine if enough water was avail-
able in the waste for binder hydra-
tion.  An initial screening test was
used to determine the appropriate
water-to-waste/binder ratio for each
S/S process and also to narrow the
range of binder-to-waste ratios used
in the detailed evaluations.  For the
purpose of initial screening, the Cone
Index Test (CI) (Bricka et al. 1989)
was used to measure the strengths of
various water—to-waste/binder mixtures
after they had cured for 48 hours.  By
preparing several water-to-waste/
binder mixtures and subjecting each
mixture to the CI test, the water-to-
waste/binder mixture that developed
substantial strength could be
selected.  An example of the screening
matrix for the Portland cement binder
is given in Table 2.

Preparation of Specimens
for Detailed Evaluation

     Based on the results of the ini-
tial screening test, one water ratio
for each binder was selected for the
evaluation of each waste.  A minimum
of 4 and a maximum of 9 binder-to-
waste ratios were selected for addi-
tional study.  Batches of the water-
to-waste/binder mixtures were prepared
and either poured or compacted into
2 by 2 in. brass molds and cured at a
temperature of 23° C and 98-percent
relative humidity for a minimum of
24 hr (Bricka et al. 1989).  Specimens
were removed from the molds when they
developed sufficient strength to be
free standing and were cured at the
same conditions until used for further
testing.

Physical and Con-
taminant Release Testing

     Unconfined Compressive Strength.
Unconfined Compressive Strength (UCS)
was used to define and characterize
the effects of S/S on the physical
characteristics of the wastes.  The
UCS of each waste specimen was deter-
mined using ASTM method C 109-86 (ASTM
1986).  The only deviation from this
method was the vibration or compaction
of the specimens.  UCS testing was
performed on the cubes after 7, 14,
21, and 28 days of curing.
                                     439
-------
      Contaminant Mobility Testing.
Only  one binder-to-waste ratio was
selected for TCLP  analysis  for each
binder.  The binder-to-waste  ratio
selected for evaluation was the ratio
that  exhibited an  UCS value closest
to but greater that 50 psi.   A UCS of
50 psi was chosen  based on  informa-
tion  found in the  USEPA Office of
Solid Waste and Emergency Response
Policy Directive 9487.00-2A (USEPA
1986c).  TCLP extractions were per-
formed in triplicate for the  selected
binder-to-waste ratio.  A total of
nine  TCLP extractions were  performed
for each waste, i.e. triplicate
extractions for the three binders.

        DISCUSSION OF RESULTS

      The following discussions are
presented on the UCS and TCLP tests.

Unconfined Compressive Strength

      Generally the UCS data,  for each
waste, is presented as UCS versus
curing time at the various binder-to-
waste ratios.  A typical example of
the relationship between UCS and cure
time  is illustrated by the K048/K051
ash waste for the  cement binder as
presented in Figure 2.  The shape of
the UCS binder ratio curves gives an
indication as to the degree of ulti-
mate  strength development that will
be achieved by a S/S waste.  Figure 2
illustrates that the 0.8 cement-to-
waste ratio (binder to raw waste
basis) has achieved the majority of
its strength by a  cure time of
28 days.

      The binder ratio selected for
TCLP  evaluation was based on the 50
psi criteria.  The ratios selected
for each binder and each waste are
presented in Table 3.  The 28 day UCS
for these ratios are presented in
Figure 3.  This figure illustrates
that  for each waste, except for K061,
the cement treated wastes developed
 substantially  higher  strengths than
 the  wastes  treated with  other binders.
 Figure  3  also  illustrates  that many of
 the  binder  ratios selected for TCLP
 evaluation  are in excess of the 50 psi
 criteria.   This can be attributed to
 the  fact  that  it is difficult to
 predict the 28 day UCS and that for
 some wastes it was difficult to add
 binder  and  not exceed the  50 psi
 criteria.

 TCLP Results

     Results for the  raw waste com-
 posite  analysis, the  TCLP  for the raw
 waste,  and  the TCLP for  the treated
 waste are presented in Table 4.  Data
 are  presented  for the contaminant of
 interest for each waste.   All TCLP
 data for the F006 waste  are below the
 detection limit.  In  addition, the
 data for the K031 waste  are not pre-
 sented  because the data  are not avail-
 able to date.   The data  in Table 4
 indicate generally, that there is a
 reduction in the concentration of con-
 taminant measured in  the TCLP extract
 of treated wastes.  Only the cement
 and  lime/flyash binders  for  the K106
waste have a higher concentration of
 contaminant in the TCLP extract of the
 treated waste  than that measured in
 the  TCLP extract of the raw waste.

     Due to dilution by the binder,
 the  TCLP data  for the raw  and treated
wastes  in Table 4 cannot be directly
 compared.  The leaching data can be
more accurately evaluated  if the data
are normalized to the dry raw waste
concentration and presented as the
percent of treatment effectiveness.
Treatment effectiveness was calculated
using the following equation:
            Cdr
                    Cr
                  Wr * Mr
(1)
                                     440
-------
where
     Cdr
      Cr
      Wr
      Mr
     TCLP contaminant  concen-
     tration mass/dry  weight
     untreated waste,  mg/g
     untreated waste TCLP mass
     for the contaminant of
     interest, mg.  (Calculated
     as:  TCLP contaminant con-
     centration mg/1 * TCLP
     extraction solution
     volume, 1)
     wet weight of  the waste
     extracted, g
     Solids content of the
     untreated waste used in
     the extraction
     expressed as a decimal
         Cdt
                    Ct
               Wt * Mt * Bt
                             (2)
where
     Cdt = TCLP contaminant
           concentration/dry weight
           treated waste after S/S,
           mg/g
     Ct  = S/S waste TCLP mass, for
           the contaminant of inter-
           est, mg.  (Calculated as:
           TCLP contaminant concen-
           tration, mg/1 x TCLP
           extraction solution vol-
           ume, 1)
     Wt  = weight of wet S/S waste, g
     Mt  = solids content of the S/S
           waste used in the extrac-
           tion, expressed as a
           decimal
     Bt  = weight fraction of waste
           in stabilized/solidified
           waste calculated as
           follows :
n  - weight of raw waste
   ~ weight of raw waste +
       weight of binder
PT
                    100
                                   *_,.
                                   (4)
                                   Figure 4 presents the normalized
                              data as the percent treatment.  As
                              illustrated by Figure 4 most of the
                              contaminants were effectively immobi-
                              lized by S/S.  The mobility of chrome
                              in the K048/K051-Ash waste and the
                              mobility of mercury for the K106 waste
                              were increased after S/S for several
                              of the binders.

                                           CONCLUSIONS

                                   Several conclusions can be drawn
                              that characterize the effects of the
                              S/S processes on the various wastes
                              investigated in this project.

                                   1.  Small quantities of the Port-
                              land cement binding agent produced
                              materials with an UCS well above the
                              50 psi criterion.

                                   2.  The binders were easily mixed
                              with the wastes, improving the
                              physical handling properties of the
                              wastes.

                                   3.  Typically, S/S resulted in
                              reduced mobility of most of the con-
                              taminants studied.

                                           REFERENCES

                              1.  American Society for Testing and
                              Materials.  1986.  "Construction;
                              Cement; Lime; Gypsum," Vol 0401,
                              Annual Bank of ASTM Standards, Phila-
                              delphia, PA.

                              2.  Bricka, R. M., Holmes, T. T.,
                              Cullinane, M. J., Jr.  1989.  "An
                              Evaluation of Stabilization/
                              Solidification of Fluidized Bed
                              Incineration Ash (K048 AND K051)",  (in
                              preparation), U.S. Army Engineer
                              Waterways Experiment Station,
                              Vicksburg, MS.
 where PT = percent  of  contaminant  not
 leached due to S/S.
                                     441
-------
3.  U.S. Environmental Protection
Agency.  1986a (Nov).  "Best Demon-
strated Available Technology (BDAT)
Background Document for F001-F005
Spent Solvents,"  EPA 1530-SW-86-056,
Vol I, Office of Solid Waste,
Washington, DC.

4.  U.S. Environmental Protection
Agency.  1986b (Jun).  Handbook for
Stabilization/Solidification of
Hazardous Wastes, Hazardous Waste
Engineering Research Laboratory,
Cincinnati, OH.

5.  U.S. Environmental Protection
Agency.  1986c.  "Prohibition on the
Placement of Bulk Liquid Hazardous
Waste in Landfills; Statutory Inter-
pretive Guidance," EPA 530 SW-86-016,
OSWER Policy Directive 9487.00-24,
Office of Solid Waste and Emergency
Response, Washington, DC.
6.  U.S. Environmental Protection
Agency.  1988 (Aug. 7).  Federal
Register, "Land Disposal Restrictions
for First Third Scheduled Wastes;
Final Rule', Vol. 53, No. 159, Office
of Solid Waste, Washington, DC.

ACKNOWLEDGEMENT:  The test described
and the resulting data herein, unless
otherwise noted, were obtained from
research conducted by the U.S. Army
Engineer Waterways Experiment Station
and sponsored by the U.S. Environ-
mental Protection Agency, Risk Reduc-
tion Engineering Laboratory (RREL),
Cincinnati, Ohio, under Interagency
Agreement DW96930146-01-5.
Mr. Carlton Wiles of RREL was the EPA
project officer.  Permission to pub-
lished this information was granted by
the Chief of Engineers and the U.S.
Environmental Protection Agency.
                                     442
-------
  ITER BINDER     WATER BINDER
WASTE
TO BE 	 *-
STABILIZED
V
WATER TO WASTE
AND
BINDER TO WASTE
RATIO
SELECTION


J V
BATCH
PREPARATION


CURING



DETERMINATION OF
UNCONFINED COMPRESSIVE
STRENGTH AT
7. 14. 21. AND 28 DAYS


TOXICITY
CHARACTERISTIC
LEACHING
PROCEDURE
(AFTER 28
DAY CURE)


ANALYSIS
OF
LEACHATE
J V
INITIAL UCS TESTING TCLP
SCREEN TESTING
TESTING ' •-- • •
Figure 1.  Schematic  Flow Chart for Solidification/Stabilization
           Processing
  4.500
                               CURE TIME. DAYS
Figure 2.  An example of UCS  verses Cure Time  at Different Binder
           Ratios
                                443
-------
2.500
2.000
1.500
IOflft

Q. t-/v*\
. 500
CO
o
400
300
zoo
100
••
-i— t
J*£
\i
•wp*
±
^r
=
=
	
~""

—
:•:•:
LEGEND
^^ CEMENT
ty-yj KILNDUST
^?3 LIME/FLYASH
I

0 = '•'•'-• ^ — *:-£S3
0 0
DE: K048/K051 K106

— — — —
liL 1^ |


^q
CvlTtCM CMTfCM IOWU)
ooo odd °. °. °.
CM CM ° ° ?
8
o
F006 K004 K061
E
I




CM




-H
JJJ
R

b





r~
= D
lit II

•»N loom ^ *-. in
" O Or^O WW1^ OO°
S S ° °' 2 §
o o u> o
o o
o'
K044 K046 K031 K031
Figure 3.  Twenty-eight day UCS for Each Waste Studied at the
Binder-to-Waste Ratios Utilized in the TCLP Evaluations
                             444
-------
1 1U
100
90
80
70
60
50
40
t- 30
z.
HI 20
OC 10
Ul
0. 0
en -10
Ul -20
LU
>
H- -7.0000
ill
£ 10,0000
ill
j_ -13,0000
Ul -16,0000
< -19,0000
H -22,0000
-25,0000
-28,0000
-31,0000
-34,0000
-37,0000
-40,0000
CONTAMINEN1
WASTE CODE:

« ^— ^—



_
-

A)"
rn




::•:•:•
X;X

1






^^
LEGEND
t=4 CEMENT
Ixwl KILNDUST
^^ UIME/FLYASH
i i i i t ii i
' i i 1 1
1": Cr Hg Pb Pb Zn
K048/KO51 K106 K004 K061 K044
IHiiiMiiHiiiimiij
m

i
Pb
K046
Figure 4.  The Percent Treatment Effectiveness for the Waste and
Binder Ratios Studied.
                              445
-------
                 TABLE 1.  DESCRIPTION OF RCRA WASTE EVALUATED

No.
1.

2.
3.
4.
5.

6.


7.

8.
9.
EPA Waste
Code
K048/K051-Ash

K106
F006
K004
K061

K044


K046

K031-Liquid
K031-Solid

Waste Description
Ash from the Incineration
of API Sludge
Mercury Cell Sludge
Electroplating Sludge
Chrome Pigment Sludge
Emission Control Dust from
a Steel Electric Furnace
Wastewater Treatment Sludge
from an Explosive Manu-
facturing Process
Lead Based Initiating
Compounds
Arsenic Salts
Arsenic Salts
Contaminant
of Interest

Cr(III)
Hg
Cr(III)
Pb

Pb


Zn

Pb
As
As
       TABLE 2.   INITIAL SCREENING MATRIX FOR THE PORTLAND CEMENT BINDER
 Binder-to-Waste
	Ratios
      0.1
      0.7
           Number of Specimens at
      Indicated Water-to-Waste Ratios
0.2                 07J077
 1                   1                   1
 1                   1                   1
                                    446
-------
        TABLE  3.   BINDER-TO-WASTE  RATIOS  SELECTED  FOR TCLP EVALUATIONS

Waste Code
K048/K051
K106
F006
K004
K061
K044
K046
K031-Liquid
K031-Solid

Cement
0.2
0.1
0.2
0.2
0.05
1.2
1.0
0.065
0.1
Binder-to-Waste Ratios
Kiln dust
0.2
0.5
0.4
0.4
0.05
1.4
1.4
0.09
0.1

Lime/Flyash
0.2/0.2
0.1/0.1
0.2/0.2
0.2/0.2
0.05/0.05
0.7/0.7
0.7/0.7
0.045/0.045
0.05/0.05
           TABLE 4.   WASTE ANALYSES FOR THE CONTAMINANT OF INTEREST


Waste
ID
K048/
K051-Ash
K106
F006
K004
K061
K044
K046


Contaminant
of Interest
Cr(III)

Hg
Cr(III)
Pb
Pb
Zn
Pb
Raw
Composite
Concen-
tration
(mg/1)
1520

25,900
DL
666,820
20,300
2.61
967
Raw TCLP
Concen-
tration
(mg/D
2.2

0.1
DL
1410
45.1
0.236
103
Average TCLP for


Cement
2.13

28.6
DL
0.54
1.16
0.02
0.078
the Treated
(mg/1)
Kiln dust
1.86

0.0139
DL
7.2
0.787
0.02
1.0
Waste

Lime/Flyash
1.14

6.65
DL
31.2
0.095
0.28
0.4
DL = Compound measured below the detection limit.
                                     447
-------
                     VOLATILE EMISSIONS FROM STABILIZED WASTE

                                           By

                             Leo Weitzman, Lawrence E. Hamel
                                    Acurex Corporation
                                Research Triangle Park, NO

                                 Paul dePercin, Ben Blaney
                           Risk Reduction Engineering Laboratory
                                      Cincinnati, OH
        The 1984 Resource Conservation and Recovery Act (RCRA) amendments, which limited
 the free liquid content of wastes placed in hazardous waste landfills, resulted in significant
 quantities of hazardous waste being solidified prior to disposal.  Recent "Land Ban" restrictions
 requiring that waste be pretreated to the standards of Best Demonstrated Available Technology
 (BOAT) prior to land disposal are also resulting an increased level of solidification/stabilization
 (SS). While the mechanical strength and leaching characteristics of stabilized wastes have been
 the subjects of investigation, there is very little data available on the emissions of organics from
 the S/S process and from the treated waste. The present program was initiated to address this
 data gap. This program measured the organic emissions from synthetic hazardous wastes during
 the mixing and curing stages of the S/S process.

 APPROACH

        The tests were conducted on two synthetic wastes (1) an inorganic "waste" made of soil
 and water and (2) an organic "waste" made of soil, water and latex paint sludge.  Each of these
 wastes was doped with either an equal weight mixture of trichloroethylene  (TCE) and acetone,
 representing a volatile organic, or N-pentanol (amyl alcohol) representing a semivolatile organic.
 Two stabilization agents were used, (1) a 50%, by weight, mixture  of portland cement (Type A)
 and flyash or (2) a 50%, by weight, mixture of lime kiln dust and flyash

        The program evaluated the emissions from 64 samples of  material. First an initial
 program was conducted on 14 samples. In this program,    both  the organic and inorganic
 "wastes" were doped with two levels (2 and 4 percent each) of acetone and TCE. The emissions
 from these mixes were measured during mixing, and at 7 and 14 days. On the basis of this
 preliminary program, a more detailed program was developed involving 50 samples. Of these 50
 tests, 30 were conducted on synthetic "wastes" contaminated with  TCE and acetone and
 20 additional tests on the same type of "wastes" contaminated with n-pentanol. Three levels of
 contaminant (1,2, and 4  percent) in the waste were evaluated as part of this program. In
 addition, the sampling of the emissions was continued for 30 days  rather than the 14 days of the
 preliminary work.

       The organic dopants used in this study were acetone and trichloroethylene, representing,
 respectively, a water soluble and a relatively insoluble organic compound.  Phenol was also
 examined as a potential candidate, but it posed major analytical difficulties  as it elution time
through the GC column is much longer than for the other compounds.  More importantly, its vapor
pressure is so low that it would be difficult to detect. Furthermore, phenol will react with highly
alkaline materials found in stabilization agents to form nonvolatile phenoxide. This would have
further reduced its vapor pressure to the point of nondetectability.
                                         448
-------
       Acetone was chosen because it is a polar, volatile organic material with a very high vapor
pressure.  It is soluble in water and the presence of water reduces its vapor pressure sharply.
Acetone reacts with alkaline materials relatively slowly, forming a salt. Over a period of time, this
reaction could actually tie up the acetone. Conversely, it is conceivable that the S/S process
could actually increase the acetone emissions over leaving the waste as a liquid.  The increase
would be due to the binder removing the free water but leaving the acetone unchanged. The
removal of the water could, conceivably increase the evaporation rate of the remaining acetone,
thereby increasing its emissions. It was hoped that testing over a 30 day period could identify
such a phenomena and help identify the important parameters for predicting organic emissions.

       The TCE was selected to represent a water insoluble, volatile organic material. Based on
discussions with WES researchers, it appeared that the heat associated with the fixation
processes using lime could result in rapid release of TCE during the mixing of the sludge with the
fixating agents.

       The n-pentanol was selected to represent a semivolatile compound.  It is a relatively
nontoxic material, whose vapor pressure is high enough to be measured with the same system as
used for the TCE and acetone, but sufficiently different from those to identify whether and how
the vapor pressure of the compound impacts its emissions.

       The following pieces of equipment were developed and used for the  emission
measurements.

             1.     "Wind Tunnel" System

             2.     "Modified Headspace" Sampling System

             3.     "Sample Venting" System

        The "Wind Tunnel" system (shown in Figure 1) consisted of a Lundberg mixer mounted in
an 8" x 18" (height x width) cross section rectangular duct. The top plate of the duct was tightly
fitted around the mixer head to minimize air infiltration.  A hole was cut directly below the mixer
head, and a clamping mechanism was constructed to hold a 5 gallon pail tightly against the duct
and mixer. A hinged door was installed next to the mixer to allow addition of materials while the
system was in operation. The end of the rectangular duct (at the far right on Figure 1) was open
to the air. The other end transitioned into a long 6" circular duct containing an orifice plate,
sampling ports leading to a GC, and  pitot tubes. The duct was heat taped and insulated to
maintain the gas temperature 80°-90 °F (26°-32 °C). The organic emission rates were measured
by measuring the airflow through the wind tunnel with the orifice plate and the concentration of
the analyte with the GC sampling system.

        The second system, the "Modified Headspace" sampling system, shown in Figure 2, was
used to measure the low organic releases from the  solidified "wastes".  It consisted of the same
pail with the lid on it, attached to a pump that pulled air at a measured rate over the surface of the
solidified wastes. The rate of flow of the air over the surface of the pail was  measured by a
calibrated rotometer and the concentration of the compound in question was measured with the
 GC. This allowed determination of the emission rate of the organic constituents over a period of
time.  The GC used here was equipped with a 10-port gas sampling valve which was controlled
 by the GC computer system.  This system switched the GC sampling loop from pail to pail so that
 emissions from up to 10 pails of material could be measured automatically (one at a time) over a
 period of several hours) without moving sampling lines or shifting pails.

        The third system, the "Sample Venting System", was like the modified headspace system
 of Figure 2, but without the GC and the associated gas sampling system.  The "Sample Venting
                                           449
-------
 System"was used to maintain a constant air flow over the curing "wastes" while they were being
 stored in between measurements. In this way, airflow was maintained at a measured rate over
 the surface of the solidified/stabilized material throughout the curing period of 30 days.

        The tests were conducted by putting a weighed quantity of sludge into a five gallon pail
 which could be (but was not at this point) tightly sealed with a lid. The pail was placed in the
 "wind tunnel apparatus and the air flow was turned on. The dopant, and binder were then added
 and the mixer was turned on.  Mixing was continued for 60 to 90 minutes while the gas flow rate
 and concentration were measured at 3-10 minute intervals. The time between measurements
 was fixed by the GC elation time for the constituent being measured, approximately 3 minutes for
 the acetone and TCE, approximately 10 minutes for the amyl alcohol.

        At the end of the mixing period, the mixer was stopped and the mixing arm removed.
 The pail was then covered and placed on a pallet of ten pails, each containing a sample which
 had been mixed in a similar manner. Each pail in the pallet was then  hooked up to the "Sample
 Venting" system for the duration of the testing.  Every three to four days, the emissions from that
 sample were measured by removing that pallet from the sample venting system and placing it
 onto the "modified headspace" system. The automated sampling and analytical system
 measured the emissions from each pail on a rotating basis (one every approximately 3 minutes)
 over a period of three to four hours. In this way, each sample was cured for 30 days under a
 constant airflow, with the organic emission rate measured over an extended period every 3 to 4
 days.

        Early in this program, a concern arose that the results from the "Headspace" and
 "Modified Headspace" measurements might be a function of the system, rather than an indication
 of the actual rate of release of the organic compound from the solidified "wastes." The concern
 was that the air flowing over the surface of the "waste" in the pail might pump or purge the
 volatiles out of the solid.

        To determine whether this was the case, a test was set up whereby a sample of the
 "waste," contaminated with TCE and acetone, solidified and placed in  the "Modified Headspace"
 system.  The flow of air in the headspace of the pail was varied in a random manner between 0.2
 to 5.0 liters per minute. After each change of air flow the system was allowed to reach steady
 state and the emission rate of the acetone and TCE were measured. The results are given in
 Rgure 3.

       As can be seen, the organics emission rate  is a function of the flow rate up to
 approximately 1 liter per minute.  Above that flow rate the emissions are limited by the diffusion
through the solid phase. Functionally, this means that above 1 l/m, the loss of the organics from
the solidified mass is limited by diffusion through the solid rather than through the boundary layer
between the solid and the bulk gas phase. On the basis of this test sequence, the air flow rate
through the pail was selected to be in the three liters per minute range. This is well above one
liter per minute, but still low enough so that the air does not dilute the organic compound being
measured to below the level of detectability.

RESULTS

       Including blanks and replicates, this program resulted in approximately 70 sets of
emission curves. Figure 4 shows a typical emission curve generated for a sample. The  left hand
curve shows the emissions (in g/min) during the mixing, and the right-hand curve shows the
emissions from the solidified material during curing.  As can be seen, the time-scales for  the two
parts of these curves are different, the left portion gives the time in  minutes, the right portion in
days.  Figure 5 presents the results for the same sample in a cumulative format. These curves
                                           450
-------
show the cumulative percentage of organic constituent released during the mixing and during the
curing.

       As discussed in the "Conclusions", below, a statistical analysis of the data indicated that
only the type of binder used and the type of organic constituent influenced the level of emissions
during the mixing step to a statistically significant extent. Six average emission curves one for
each of the three organics tested and for each of the two binders were, therefore calculated by
averaging the emissions for all tests at each of these conditions. The results are given in
Figures 6 through 8 which show the cumulative percent emissions from the "wastes" for each of
the three organic compounds. Each of these figures shows the cumulative emissions for the
wastes during mixing, when the LF and PF solidification agents were used, and for the blanks—
without solidification agent. The curves were obtained by averaging the results of all the runs for
the respective conditions over concentration. These curves can be interpreted as being a first
approximation to an emission factor—the percent of the given component released during the
mixing as a function of time.

       The tests showed that only the type of organic compound being emitted was significant
during the curing portion of the tests. Figure 9 was obtained (for the curing materials) in a similar
manner to Figures 6 through 8 (for the mixing) by averaging the emission measurements at the
specific times.  Note that for all of these cases, "blank" refers to samples of synthetic waste
containing appropriate levels of organic which were mixed and stored in the same way as the
other samples but which were not mixed with any binder. By comparing the  results of each of the
samples to those of the blanks, it is possible to determine what portion of the emissions are due
to the mechanical mixing and handling and what is due to the S/S process.

       The percentages shown in Figure 9 are based on the amount of organic constituent
remaining in the waste after mixing. For example, if 90 percent of the organic constituent was
emitted during mixing, the percentage given in Figure  9 is of the remaining 10 percent. The
curves of Figures 6 through 8 can be interpreted as showing the cumulative percent (the amount
emitted from the waste during mixing) of each organic constituent released from the
solidified/stabilized waste and from the quiescent blank, which had not been mixed with binder
but otherwise handled in the same way.

CONCLUSIONS

       The experimental program described herein was designed to determine the following:

       1.   Can a laboratory system be fabricated to readily measure organic emissions from all
            phases of the stabilization process and from the stabilized wastes?

       2.   Does  stabilization reduce the emissions of organic compounds  from sludges? If not,
            how does it affect them?

       3.  Do nonvolatile organic constituents affect the emissions of organic compounds?

       4.  Does  the type of solidification agents used affect the emissions or organic
            compounds?

       5.  Are the emissions of the different organic compounds proportional to their
            concentration or does another relationship apply?

        6.  How significant is the vapor pressure of the component on its emission?

        7.  How significant is the solubility of the component on its emission?
                                          451
-------
         The first question was addressed by evaluating the overall performance of the systems
 developed to measure the organic emissions from the waste solidification process and from the
 solidified wastes themselves. The systems developed performed very well.  Measurements were
 performed continuously for a period of more than two months without any loss of data. The
 systems proved to be flexible, and were able to measure organic emissions ranging from less
 than one mg/minute to tens of grams per minute—three orders of magnitude.

         Questions 2 through 7 were answered by performing an analysis of variance (ANOVA) on
 the results of the tests.  The tests were conducted by grouping the experiments into categories
 with common parameters of interest, and then comparing the variation between the categories to
 that within each category. This analysis allowed estimation of the probability that the two
 categories are, or are not, of the same population. This is a standard method of testing
 hypotheses.

        The act of solidifying wastes had its greatest impact on the organic emissions during
 stabilization, when all of the components were being mixed. At that time, stabilization with lime
 kiln dust often raised the temperature of the "waste," resulting in emissions that were significantly
 higher than those from wastes processed in a similar manner, but without the addition of
 solidifying agents—blanks. The tests indicated that there was a 99.99% probability that the
 emissions during mixing are higher when lime kiln dust and flyash solidifying agents are used
 than when no solidifying agent is used.

        Emissions during mixing of the components did not differ significantly between the blanks
 and those samples which used portland cement as the solidifying agent. The statistical means of
 the emissions during mixing of the blanks and of the "wastes" solidified with portland cement
 were equivalent respectively. The emissions from the blanks (i.e., unsolidified wastes) were a
 affected by the agitation during the tests, which was expected, although this affect was relatively
 small. The organic compounds were released from the liquid at a very low rate until the liquid
 was disturbed by processing.

        The organic emissions after mixing showed no significant difference between any of the
 samples for each of the three compounds (acetone, TCE, and n-pentanol). The results indicated
 a greater than 90% probability that the method used for solidification does not affect these
 emissions from the solidified wastes.  In fact, they indicated a greater than 90% probability that
 the organic emissions from a solidified waste are the same as those from the quiescent liquid
 waste.  It should be noted, however, that solidification does prevent the emissions from being
 Increased by environmental factors, such as wind or activities which would disturb a pool of
 standing liquid and increase emissions from it.

        The effect of nonvolatile organics in the wastes on the emissions was tested by
 comparing the emissions from the "organics wastes" (those prepared  with latex paint) to those   k
 from the corresponding "inorganic wastes."  The analysis showed that there was a small
 difference between the two populations during mixing, but no significant difference in the
 emissions from the stabilized wastes.

        The proportionality of emissions was tested to determine whether the emissions of each
 compound are proportional to the amount in the waste, or whether another dependency occurs.
 If the emissions are proportional to the concentration then they are governed mainly by
 equilibrium relationships.  If, on the other hand, the relationship is more complex, then diffusion
 and possibly other relationships would have to be included in subsequent modeling. This
 analysis can assist further research programs.

        The results showed no clear trends  in the emissions. Various runs tended to show
differences which  could not be correlated over groups of tests. The results seem to indicate that
                                          452
-------
the organic emissions are proportional to the amount in the "waste." That is, the emissions of
each component from the wastes containing 2% at the start are approximately twice those
containing 1%, and half of those from the wastes containing 4%.

       This analysis addressed whether the vapor pressure of each of the components was a
significant influence on the emissions. It was performed by comparing the TCE emissions
against the n-pentanol emissions for the same parameters. Both TCE and ri-pentanol are
insoluble in water and have widely different vapor pressures.

       The intent of this analysis was actually to see whether their emissions were proportional
to the vapor pressure, and hence, if the governing mechanisms are equilibrium related. The
results indicated that the emissions of each of the three components tested were different during
the mixing step, although the difference was by no means as great as the differences in vapor
pressure would indicate, even considering the impact of dilution on the vapor pressure. The
emissions from the solidified waste (during curing) were also similar The n-pentanol emissions,
while lower than those of the volatile compounds, were of the same order of magnitude. This
reinforces the conclusion that mechanisms other than straight volatilization are involved in all
cases.

       The solubility of the compound in water was tested by comparing the acetone emissions
to the corresponding TCE emissions. This must be classified as a very tentative comparison, as
TCE and acetone have different vapor pressures in the pure state and, clearly, the vapor
pressure of the acetone dissolved in the water is different from the TCE which remains in the pure
state.

       The emissions were significantly different for these two compounds throughout the tests;
however, it was not possible to  ascertain whether the difference was due to solubility or to vapor
pressure. The partial pressure  of acetone drops rapidly when it is mixed with water.  This
dependency on the water content of the system was not noted during these tests.  The
significance of the emission difference was, therefore, tentatively attributed to the solubility
variation.

        In summary, a system has been developed which can rapidly and with reasonable
accuracy measure the emissions of organic compounds from mixing processes  such as S/S.
This equipment was used to measure the emissions of organic constituents from synthetic wastes
and it showed that under the conditions of these tests, significant percentages of the organic
constituents were released to the atmosphere. The experiments resulted in the determination of
emission factors which could be used to estimate the likely levels of emissions from S/S
processes.

        The equipment developed here can also be used as an adjunct to a remediation  effort.
S/S is often used to as part of a cleanup of a contaminated site, to solidify/stabilize a siudge
pond, in place, for example. The air emissions from such an operation have to be considered in
evaluating the applicability of the technology.  The equipment developed under this program can
be used to determine the emissions from this application. Samples of the waste and binder can
be tested in the laboratory and the emissions measured, thus giving a reliable and defensible
 estimate of the environmental impact associated with the remediation.
                                           453
-------
                               ACKNOWLEDGEMENT

       The work described herein was performed under contract to the Environmental Protection
Agency Contract Number 68-02-3993, WD 32 and 37 to Acurex Corporation. This paper has not
been subject to the Agency's Peer Review Process butdoes not necessarily represent the views
of the Agency.  Mention of trade names or commercial products does not constitute endorsement
or recommendation for use. The final report, of the same title, is available from the Risk
Reduction Engineering Laboratory, Cincinnati, Ohio.
               ROOF LINE

        FAN
r
                                        if=TOGC
                                        JlSAMPLING LOOP
                                                                        HINGED PLATE
                  \/ PILOT PORTS SAMPLING PORT ORIFICE PLATE TRANSITION
               DAMPERS
                   Figure 1. Schematic of "Wind Tunnel" Treatment System.
                                        454
-------
          TO VENT
                                S SAMPLING
                               .OOP AND GC/
                  10 PORT STREAM
                  SELECTOR VALV
               PALLET OF 10 PAILS OF SOLIDIFIED-WASTE
                                                      SOLIDIFIED
                                                        WASTE
                                                          TO VENT
  Figure 2. "Modified headspace" sampling system used in WP 37.
0.9  -


O.8  -


O.7  -


O.6  -


O.5  -


O.4  -






O.2  -
                            O      O
                                                     a     a
02                       +

                           FLOW (lUen/mlnuto)
      O    ACETONE            +   TWGHLOROETHYUME


                         Figure 3. Results of air flow tests.
                                 455
-------
                     DOPANT  EMISSION  RATE
1
                        UNUTB
                        O   ACfTONC          •  TOC

                         Figure 4.  Dopant emission rate.
                           IL1B % EMISSION
                           (CUKUXIKC: ACC7QNC M 1CD
    i
                 20
                        UNUttS
                         o   Accrotc
                        Figure 5. IL1B percent emission.

                                  456
-------
in
3?
                  AVERAGE ACETONE EMISSIONS
                  DURING MIXING (SOLIDIFICATION AGENTS)
                              MINUTES
                 Figure 6. Cumulative acetone emissions.
1
in
                     AVERAGE TCE EMISSIONS
                  DURING MIXING (SOLIDIFICATION AGENTS)
                                  30
                              MINUTES
                   Figure 7. TCE cumulative emissions.

                               457
-------
114
5?
                 AVERAGE 1-PENTANOL EMISSIONS
                  DURING MIXING (SOLIDIFICATION AGENTS)
                                 30
                             MINUTES
                 Figure 8. 1-pentanol cumulative emissions.
60
                    AVERAGE VOC EMISSIONS
                            AFTER MIXING
                              DAYS
                  Figure 9. Total VOC emission by VOC.

                              458
-------
  TECHNOLOGIES APPLICABLE  FOR  THE  REMEDIATION  OF  SUPERFUND  RADIATION  SITES

              by:     Ramjee Raghavan
                      Foster Wheeler Enviresponse, Inc.
                      GSA Raritan Depot
                      Woodbridge Ave.
                      Edison, NJ  08837

                      Darlene Williams
                      Releases Control Branch
                      USEPA RREL
                      GSA Raritan Depot
                      Woodbridge Ave.
                      Edison, NJ  08837
                                   ABSTRACT
    This paper identifies technologies that may be useful in removing or
stabilizing radioactive contamination at uncontrolled hazardous waste
(Superfund) sites containing radioactive material.  The radioactive
materials at some Superfund sites consist primarily of waste from radium,
uranium, and thorium processing.  Twenty-five existing Superfund sites are
known to contain radionuclides.

    Sites contaminated with radioactive material pose a unique problem
because, unlike organic wastes,' it cannot be destroyed by physical or
chemical means; it can only decay at its natural rate.  Alteration of the
radioactive decay process thereby changing the fundamental hazard is not
possible.  Several technologies have potential for removing or stabilizing
radioactive material at Superfund sites.  These fall into the categories of
disposal, on-site treatment, chemical extraction, physical separation, and
soil washing.  Applicability of these technologies is controlled by
site-specific factors, and their feasibility must be determined on a
site-by-site basis.
                                     459
-------
                                 INTRODUCTION
    The United States Environmental Protection Agency (EPA) has identified
twenty-five Superfund sites in the country that are radioactively
contaminated.  These Superfund sites are briefly described in the Appendix.

    The radioactive materials at many Superfund sites are by-products of
uranium, thorium, and radium processing in the form of tailings,
contaminated buildings and equipment, and stream sediments.  These sites,
located across the United States, vary greatly in size and may involve
radiation exposure to people who reside on and around them.

    The primary public health threat is from contact with radioactive
materials through a) external whole body exposure to gamma radiation, b)
ingestion of radionuclides in food and water, and c) inhalation of radon
and radon progeny.  Radon and radon progeny are continuously produced
through the decay and decomposition of uranium, thorium, and radium.
Radioactive material cannot be destroyed by physical or chemical means;
they can only decay at their natural rate.  The health risks will persist
throughout the entire decay period if no remedial action is taken.
Possible effects on human health include the increased risk of cancers and
increased risk of genetic damage that may cause inheritable defects in
future generations.

    Technologies that have potential to remediate structures (buildings)
and groundwater are of interest at some Superfund sites, but these are
beyond the scope of this paper.  The soil remediation technologies
discussed in this paper fall into the categories of disposal, on-site
treatment, chemical extraction, physical separation, and soil washing.
Applicability of these technologies to Superfund sites is controlled by
site-specific factors; therefore, their usefulness must be determined on a
site-by-site basis.

                                   DISPOSAL
    The disposal technologies can be in one of two categories:  on-site
disposal and off-site disposal.

ON-SITE DISPOSAL

    Several technologies are available for on-site disposal.  These may be
applied in situ.  These technologies include capping and vertical barrier
wal 1 s.

Capping

    This concept involves covering the contaminated site with a barrier
sufficiently thick and impermeable to minimize the diffusion of radon gas .
(1).  Barrier materials can be either natural low-permeability soils (e.g.,
                                     460
-------
clay) or  synthetic membrane  liners,  or both.

Application:  Appropriate  for  large  discrete contaminated  areas, or  several
  smaller areas that  are close together.
Advantages:   Low  cost,  easily  applied, well-known, and a proven technology.
Disadvantages:  Limits  further use of the site.  The cap must be maintained
as
  long as the contaminant  exists  at  the site.  Also, horizontal migration
  of the  radioactive  material  in  groundwater could still occur.
Experiences:  Exists  for radioactive contaminated soils and tailings  (1,

Cost:  $13-200 per m3 (3).   Low cost for cap only, high cost for
excavation,
  transportation, legal assistance,  and cap.

Vertical  barrier  walls

    Vertical  barrier  walls may be installed around the contaminated zone to
help confine  the  material  and  any contaminated groundwater that might
otherwise flow from the site.  The barrier walls, which might be in the
form of slurry walls  or grout  curtains (4), would have to reach down to an
impermeable natural horizontal barrier, such as a clay zone, in order to be
effective in  impeding groundwater flow.

Application:  Could be considered for large discrete waste material or
  around  several  smaller areas that  are close together.
Advantages:  Simple to install, and  applicable to a variety of soil
  conditions.
Disadvantages:  Restricts  further use of the site, possible deterioration
  of the  barrier  walls resulting from the chemicals contained in the waste,
  would not stop  vertical  contamination to groundwater below.
Experiences:  Exists  for hazardous wastes and not for radioactive wastes
  (4).
Cost:  $33-377 per m2 of vertical face (4).

OFF-SITE  DISPOSAL

    Four  off-site disposal methods are briefly described in this paper:
land encapsulation, land spreading,   underground mine disposal,  and ocean
disposal

Land encapsulation

    Land  encapsulation has been the disposal method most used to this point
in time for low-level  radioactive waste materials.  Land encapsulation can
be as simple as excavating the contaminated material  and,  without further
treatment, hauling it to a secure site.

Application:  Appropriate for wastes that have not been treated,  as well as
  for radionuclides extracted from a soil  or other type of matrix.
Advantages:  Low cost, proven,  workable technology for the disposal of
                                    461
-------
  low-level radioactive wastes.
Disadvantages:  Finding a site is politically and socially difficult.
  Transportation of large volumes also carries certain costs and risks.
  Longevity is a consideration not fully addressed by this disposal method.
Experiences:  Exists for radioactively contaminated soils (3).
Cost:  $276-895 per m3 of contaminated soil (3)

Land spreading

    This technology involves excavation of the contaminated material,
transporting it to a suitable site, and spreading it on unused land,
assuring that radioactivity levels approach the natural background level
for these materials when the operation is completed.

Application:  Appropriate for dry, granular tailings and soils with very
  low level radioactivity.
Advantages:  Simple and relatively inexpensive.
Disadvantages:  Selecting a site is both politically and socially
  difficult.  Also, it could contribute to a non-point source pollution
  problem.
Experiences:  Very limited (1).
Cost:  Not available.

Underground mine disposal

    Underground mine disposal could provide secure and remote containment.
The radioactively contaminated wastes could be excavated and transported
without treatment to the mine site, pretreated for volume reduction, or
solidified to facilitate transport and placement, thus reducing associated
costs.  Movement of radionuclides into groundwater must be investigated and
prevented.

Application:  Appropriate for variety of radionuclides and matrix types.
Advantages:  Would provide a very secure and remote containment.
Disadvantages:  Expensive.  Transportation costs and associated risks need
  to be researched further.
Experiences:  Very limited (1).
Cost:  $399-942 m3 of contaminated soil (1)

Ocean disposal

    A possible alternative to land-based disposal options is ocean
disposal.  This alternative should only be evaluated for naturally
occurring mill tailing wastes and not considered for enhanced radioactive
materials.

Application:  Appropriate for low level radioactive waste.
Advantages:  Offers extreme isolation of low-level radioactive waste.
Disadvantages:  Stringent permit requirements.
Experiences:  Exists for radioactive wastes (1).
Cost:  $332-400 per nr of contaminated soil (2).
                                     462
-------
                              ON-SITE TREATMENT
    Two on-site treatment technologies are briefly described;
stabilization/solidification and vitrification.

Stabilization/solidification

    This method immobilizes radionuclides by trapping them in an impervious
matrix (4).  The solidification agent (i.e., Portland cement, silica grout,
or chemical grout) can either be injected in situ, or the waste can be
excavated, mixed, and returned.

Application:  Can be applied to buried and/or capped material.
Advantages:  Solidification may be able to reduce the release of radon and
  associated radioactivity to acceptable levels.  Solidification also may
  make it easier to transport and dispose of the waste material off site.
Disadvantages:  Long-term effects are not known.  There can be undesired
  reaction between the additives and other types of hazardous waste.
Experiences:  Exists for hazardous wastes and not for radioactive wastes
  (4).
Cost:  $33-248 per ton of contaminated soil (4).

Vitrification

    This technology immobilizes radioactive contaminants by trapping them
in an impervious matrix.  The in situ process melts the waste materials
between two or more electrodes, using a large amount of electricity.  The
melted material then cools to a glassy mass in which the radionuclides are
trapped.

Application:  Applicable for low-level radioactive wastes.
Advantages:  Minimal site preparation required.
Disadvantages:  Many substances volatilize, requiring gas collection
  system.  Radium may volatilize, therefore extra precautions are required.
Experiences:  Limited (1).
Cost:  $161-600 per m3 of contaminated soil (4)

                             CHEMICAL EXTRACTION


    Chemical extraction generates two soil  fractions.  One fraction
contains the concentrated radioactive contaminants and may require
disposal; the remaining material is analyzed for residual contamination and
evaluated for replacement at the point of origin or at suitable alternative
sites.  The various applicable chemical  extraction techniques include
extraction with water, inorganic salts,  mineral acids, and complexing
agents.
                                     463
-------
EXTRACTION WITH INORGANIC SALTS

    Radioactive contaminants can be extracted by thoroughly mixing soil and
mill tailings in an inorganic salt solution.  The slurry is filtered,
separating the extractant from the solids.  The radioactive contaminant is
separated from the extractant by ion exchange, coprecipitation, or membrane
filtration.

Application:  Should not be used as a pretreatment to an acid extraction
 •process.  The presence of sulfate and hydroxide in soils and tailings
  will negatively impact the efficiency of radium and thorium removals.
Advantages:  A high percentage of radium and thorium removal is possible.
  Simple extraction vessels are needed.  Recycling of salts is possible.
Disadvantages:  Large amounts of salts may be required.  Some salts such as
  chloride may be environmentally undesirable.
Experiences:  Limited laboratory and bench-scale testing (5,6).
Cost:  $50-150 per ton of soil.

EXTRACTION WITH MINERAL ACID

    In these processes, the ore is ground to 28 mesh and mixed with water
to form a slurry.  The slurry is pumped into an acid leach circuit,
maintaining a pulp consistency of 50 percent solids.  The solids are
separated from the leach liquid by physical methods.  The radioactive
material is removed from the leach solution by ion exchange, solvent
extraction, or precipitation. (1,7)

Application:  Removes most of the metals, both radioactive and
  nonradioactive.
Advantages:  High percentage of radium removal is possible.  Uranium and
  other metals would also be removed.
Disadvantages:  Increased operating and capital costs due to expensive
  reagents, higher operating temperatures, and the corosion resistant
  material required.  The resulting chemically leached material may create
  a harmful waste stream.
Experiences:  Extraction of radionuclides from ores (1,7,8).
Cost:  $50-150 per ton of soil.

EXTRACTION WITH COMPLEXING AGENTS

    This process differs from acid extraction in that complexing agents
like EDTA (ethylenediaminetetraacetic acid) are used instead of mineral
acids.

Application:  Radium from soils with low concentrations of thorium.
Advantages:  High percentage of radium removal.  Low reagent concentrations
  required, reagent can be recycled, reducing operating costs.  The process
  works at ambient temperatures, and many of the reagents are innocuous.
Disadvantages:  Reagents are expensive, process would not remove thorium.
Experiences:  Used in extraction of uranium from ores (1,9,10).
Cost:  $50-150 per ton of soil.
                                     464
-------
                             PHYSICAL SEPARATION
    The radioactive contaminants in soils and tailings in many cases are
associated with the finer fractions (9,11,12).  Thus, size separation may
be used to reduce the volume of concentrated material for disposal, leaving
a cleaner fraction.  Physical separation may be used with extraction to
further reduce contaminant volume.  Four physical separation technologies
are screening—both wet and dry, classification, flotation, and gravity
concentration.

SCREENING

    Screening separates soil (or soil-like material) on the basis of size
(13).  It normally is applied to particles greater than 250 microns.  The
process can be done dry or by washing water through the screen.  Screening
is not efficient with damp materials,  which quickly blind the screen.  It
may be particularly effective as a first operation to remove the largest
particles, followed by other methods.

Application:  Appropriate for all soils, can separate fractions as low as
  50 micron in size.
Advantages:  Simple and inexpensive method.
Disadvantages:  Noisy.  Dry screening requires dust control.  Wet screening
  will require separation of contaminants from water.
Experiences:  Mature technology (1,13,14).
Cost:  Equipment cost ranges from $2-5 per Ib/hr capacity.

CLASSIFICATION

    Classification separates particles according to their settling rate in
a fluid.  Several hydraulic, mechanical, and nonmechanical configurations
are available.  Generally, heavier and coarser particles go to the bottom,
and lighter, smaller particles' (slimes) are removed from the top.
Theoretically, classifiers could be used to separate the smaller particle
fractions, which may contain much of the radioactive contamination in waste
sites.  Classifiers could be used with chemical extraction in a volume
reduction process.

Application:  Effective for sandy soil with low clay and humus content.
Advantages:  Low cost, reliable, high continuous processing capabilities.
Disadvantages:  Humus and clay soil are hard to separate by classification.
Experiences:  Extensive use in industry (1,13,14).
Cost:  Equipment cost ranges from $0.10 to $0.50 per Ib/hr capacity.

FLOTATION

    Flotation is a liquid-froth separation process often applied to
separate specific minerals (particularly sulfides) from ores.  The process
depends more on physical and chemical  attraction phenomena between the ore
and the frothing agents, and on particle size, than on material density.
                                    465
-------
Ordinarily, flotation is applied to fine materials; the process often is
preceded by grinding to reduce particle size.

Application:  Effective for extraction of radium from uranium mill
  tailings (12).
Advantages:  If the particle fraction containing the contaminants can be
  collected by the froth, then flotation is a very effective tool.
Disadvantages:  New additives may have to be developed to permit successful
  flotation separation for radioactively contaminated materials.
Experiences:  Extraction of metals from complex ores (1,12,14).
Cost:  Equipment cost varies from $160 to $830 per gallons/minute capacity.

GRAVITY CONCENTRATION

    Gravity concentration is an old and proven technology that takes
advantage of the difference in material densities to separate the materials
into layers of dense and light minerals (14,15).  Separation is influenced
by particle size, density, shape, and weight.  Shaking and other motions
are employed to keep the particles apart and in motion.  Gravity separation
can be used in conjunction with chemical extraction.

Application:  Limited to those soils in which the contaminants are
  relatively coarse and capable of resisting breakage and sliming.
Advantages:  Highly efficient and proven process for a wide range of
  applications.  It gives a high-grade concentrate over a wide range of
  particle sizes and functions well with most soil types.
Disadvantages:  Low capacity.  High throughput requires multiple decks,
  clean water.  Must ensure there is no slime buildup in recycle water.
Experiences:  Extensive use in industry and ore processing (13,14,15).
Cost:  Equipment cost varies from $65 to $75 per gallons/minute handling
  capacity.

                                 SOIL WASHING
    Soil washing uses a combination of physical separation and chemical
extraction technologies.  Contaminated soil or tailings are mixed with
water and/or extraction reagents.  The clean coarse particle sizes are
separated from the liquid containing the fines and radioactive material by
a combination of physical separation methods.  The radioactive material
would then be extracted from the liquid by standard water treatment
processes such as filtration, carbon treatment, ion exchange, chemical
treatment, and membrane separation.

Application:  Depends on chemical reagents used.
Advantages:  High percentage of contaminant removal is possible.  Recycling
  of reagents is possible.
Disadvantages:  Expensive reagents. Chemically leached material may create
  a harmful waste stream.
Experiences:  Only laboratory and bench-scale testing (1).
                                    466
-------
Cost:  $50-100 per ton of soil.
                                 CONCLUSIONS
      The 25 Superfund Sites have radiologically contaminated soil spread
      over a total of 9500 acres and have several contaminated groundwater
      wells.

      Alteration of the radioactive decay process thus changing the
      fundamental hazard is not possible.

      Any choice of remediation technologies for radioactive waste at
      Superfund sites would have to be site specific.  Extensive site soil
      characterization studies would be required prior to development and
      application of most of the technologies.

      Since none of the chemical extraction and physical separation
      technologies has been used in a site remediation situation, their
      application must be approached cautiously.  The same holds true for
      solidification and stabilization processes.  Essentially, only land
      encapsulation has been used to remediate similar sites; ocean disposal
      has been used for low-level radioactive wastes.

      Various remediation technologies have potential to reduce the volume
      of the contaminated waste with an associated increase in concentration
      of the radioactive material.

      Every site remediation involving radioactive materials must involve a
      final environmentally safe disposal site for the radioactive
      materials.

      Even if it proves feasible at a particular site to lower the
      concentration of radionuclides in the soil by physical separation
      and/or chemical extraction to some acceptable level, the "clean"
      fraction is likely to contain traces of radionuclides.  Therefore,
      adequate attention must be given to whether the "clean" fraction may
      be returned to the original site or an unrestricted location or must
      be sent to a disposal site.

                               ACKNOWLEDGMENTS


  The authors wish to express their gratitude for the contribution of F.
Freestone, P. Shapiro, R. Hartley, W. Gunter, and G.  Snodgrass of USEPA; and
G. Gupta of Foster Wheeler Enviresponse, Inc.

                                  REFERENCES
1.    U. S. Environmental Protection Agency.  Technological  Approaches to
                                    467
-------
      the Cleanup of Radiologically Contaminated Superfund Sites, USEPA.
      Report EPA/540/2-88/002, August 1988.
2.    Camp, Dresser & McKee et al.  Draft. Final Feasibility Study for the
      Montclair/West Orange and Glen Ridge, New Jersey Radium Sites, Volume
      1.  USEPA Contract 68-01-6939, 1985.
3.    U. S. Department of Energy.  Long Term Management of the Existing
      Radioactive Wastes and Residues at the Niagara Falls Storage Site,
      DOE/EIS-0109D, Washington, DC, 1984.
4.    U. S. Environmental Protection Agency.  Handbook -- Remedial Action at
      Waste Disposal Sites (Revised).  EPA-625/6-806, Hazardous Waste
      Engineering Research Laboratory, Cincinnati, OH, 1985.
5.    Landa, E. R. Leaching of Radionuclides from Uranium Ore and Mill
      Tailings.  Uranium, 1:53-64, 1982.
6.    Organization for Economic Cooperation and Development (OECD).  Uranium
      Extraction Technology - Current Practice and New Development in Use
      Processing.  OECD, Paris, 1983.
7.    Clark, D. A.  State of the Art:  Uranium Mining, Milling and Refining
      Industry.  USEPA/60/2-74-038m 1974.
8.    Ryon, A. D., F. J. Hurst, and F. 6. Seeley.  Nitric Acid Leaching of
      Radium and Other Significant Radioncuclides from Uranium Ores and
      Tailings.  ORNL/TM-5944, Oak Ridge National Laboratory, Oak Ridge,
      Tennessee, 1977.
9.    Borrowman, S. R., and P. T. Brooks.  Radium Removal from Uranium Ores
      and Mill Tailings.  RI-8099, U.S. Bureau of Mines, Salt Lake City
      Research Center, Salt Lake City, Utah, 1975.
10.   Taskayev, A. I., V. Ya. Ovchenkov, R. M. Altkaskhin, and I. I.
      Shuktomova.  Effect of pH and Liquid Phase Cation Composition on the
      Extraction of 226 Ra from Soils.  Pochvovedeniye, 12:46-50, 1976.
11.   Garnett, John, et al.  Initial Testing of Pilot Plant Scale Equipment
      for Soil Decontamination.  U.S. Dept. of Energy, RFP 3022, 1980.
12.   Raicevic, D. Decontamination of Elliot Lake Uranium Tailing.  CIM
      Bulletin, 1970.
13.   Kelly, E. G., and D. J. Spottiswood.  Introduction to Mineral
      Processing.  John Wiley, New York, 1982.
14.   Wills, B. A. Mineral Processing Technology.
      1985.
15.   O'Burt, Richard.  Gravity Concentration Technology.  Elsevier, New
      York, 1984.
Pergamon Press, New York,
                                    468
-------
APPENDIX:  RADIOACTIVELY CONTAMINATED SUPERFUND SITES
Site Name
1.
2.

3.
4.
5.
6.
7.
8.
9.
-••'
10:

t1.
12.

13.


14.
15.
16.
17.

18.

19.
20.

21.
22.
23-
24.
25.
Shpack/ALI
Haywood Chemical Co./
Sears Property
U.S. Radium Corp.
W. R. Grace & Co.
Montclair, West Orange,
Glen Ridge Radium Site
Lodi Municipal Well
Lansdowne Property
Maxey Flats Nuclear
Disposal Site
West Chicago Sewage
Treatment Plant
Reed-Keppler Park
. Kerr-McGee Off -Site
Properties
Kerr-McGee Kress Creek/
West Branch of Dupage
River
The Homestake Mining Co.
United Nuclear Corp.
We I don Spring Quarry
Monticello Radioactivity-
Contaminated Properties
Denver Radium Superfund
Sites
Lincoln Park
U.S. DOE Rocky Flats
Plant
Uravan Uranium Project
Teledyne Wah Chang
Hanford 200-area (USDOE)
Hanford 300-area (USDOE)
Hanford 100-area (USDOE)
City/County State/EPA Region
Norton/Attleboro
Maywood/Bergen Co.

Orange, Essex Co.
Wayne/Passaic Co.
Essex Co.
Essex Co.
Lodi, Bergen Co.
Lansdowne
Fleming City/Hi I Isboro
- . . .-..••••
West Chicago :
, '•„'
West Chicago
West Chicago

West Chicago


Cibola Co.
Church Rock
St. Charles City
Monticello
San Juan, Co.
Denver

Canon City
Golden

Montrose City/Uravan
Albany
Bentorij cb
Bentoh, CO
Benton, CO
MA/ 1
NJ/II

NJ/II
NJ/II
NJ/II
NJ/II
NJ/II
PA/III
KY/IV

IL/V

IL/V
IL/V

IL/V


NM/VI
NM/VI
MO/VII
UT/VI 1 1

CO/VI 1 1

CO/VI 1 1
CO/VI I I

CO/VI 1 1
OR/X
WA/X
WA/X
WA/X
Acres
31
42

1
6.5

127.0
wells
1.9
25.0-
,: 	
25.0

0.25
--

--


245
170
220
;

40

900
6,550

900
--
--
--
--
Cu yds

270,000

10,000
120,000

350,000
_ _
2,000
178,000
.
40,000

20,000
61,000

..


16,500,000
4,700,000
780,000
182,000

106,000

1,900,000
--

10,000,000
..
1,000,000,000
27,000,000
4,300,000,000
                          469
-------
             RRELevant OA DIAGNOSTICS: MAJOR HNDINGS IN FY 88

                 by:  GuyF. Simes and Ronald D. Hill
                      Risk Reduction Engineering Laboratory
                      U.S. Environmental Protection Agency, Cincinnati, Ohio 45268
                      John R. Wallace, S-CUBED, San Diego, California 92121
                                   ABSTRACT

    This paper evaluates the RREL QA Program by considering its impact on specific projects
undertaken during FY 88. During this time, RREL reviewed 108 QA Project Plans (the major
document describing QA requirements for data generation). Typical problems identified during
these reviews included poorly defined objectives, incorrect sampling and analytical methods,
incomplete or inappropriate methods for measuring precision or accuracy, and sampling strat-
egies which did not match the project objectives. Before experimental efforts began, these con-
cerns were resolved, thereby avoiding potentially serious experimental flaws. A total of 61 on-
site technical audits have been carried out during sampling and analytical activities. At times,
these Technical Systems Audits have resulted in major changes, such as discontinuation of work
at a laboratory. More commonly, problems were corrected promptly by the analytical laboratory
or field team before the integrity of the project was compromised. Audits of Data Quality
(ADQs) were carried out less frequently during FY 88, but nevertheless resulted in substantial
impacts. For example, an ADQ of one project revealed that much of the required QA data had
not been collected, and consequently much of the analytical data were of limited usefulness.
This finding was valuable for EPA policy makers as well as for the management of follow-on
projects. The last component of the RREL extramural QA Program is review of final reports,
which provide constructive feedback to RREL regarding the success (or failure) of the project.
Typical  problem areas identified at this stage were insufficient presentation of QA data or overly
generalized conclusions.
                                        470
-------
                                   INTRODUCTION
     The formal Quality Assurance Program at the Risk Reduction Engineering Laboratory
 (RREL) originated in the late 1970s in response to the need to coordinate the data requirements
 in complex, mterdisciplinary projects. For the purposes of achieving project objectives, it was
 not sufficient that engineers, chemists, economists, regulators, and statisticians each performed
 well in their own areas; rather, it was also necessary that the data requirements and data gener-
 ation activities of all experts be coordinated and communicated in the project planning stage.

     In response to these needs, RREL (then the IERL laboratory) appointed the first Quality
 Assurance Officer in 1978.  Quality Assurance Programs originally emphasized the development
 of improved analytical and sampling methods. In the early 1980s, RREL began requiring that
 data generating projects include a formal quality assurance plan.  Also at this time, most of the
 formal quality assurance functions were initiated.  The period of 1981 to present has been char-
 acterized by a gradual evolution in intensity and approach.

     The primary purpose of the Quality Assurance Program at RREL is to assure that all data
 are adequate to  meet project objectives.  A corollary to this goal is that data be of known quality
 in terms of accuracy, precision, and detectability, and that samples be obtained in a meaningful
 and representative manner.  These goals apply to all measurements (including process measure-
 ments) that may be needed to complete technical, economic, and regulatory analyses.

     Quality assurance involvement in a particular project begins in the planning stage with the
 development of data quality objectives (DQOs).  These are not, as the name might suggest,
 objectives for precision and accuracy, but rather are statements of overall project objectives. For
 example, a data quality objective for an incineration process might be "to demonstrate a destruc-
 tion efficiency of 99.99 percent for a given pesticide at a confidence level of 95 percent." The
 development of DQOs particularly requires the interaction of various managers and disparate
 experts in order to assure that the end product meets the customer's needs.  The second stage of
 quality assurance involvement in RREL projects is the development and review of a Quality
 Assurance (QA) Project Plan which describes all procedures for data generation. This plan must
 be approved by  RREL before data generation can begin. The third stage of quality assurance
 consists of various on-site technical reviews (technical audits) that take place during sample col-
 lection, chemical analysis, or data reduction.  Such on-site inspections evaluate sampling and
 data generation  vis-a-vis the QA Project Plan. Fourth, final reports are reviewed with respect to
 the credibility of data and the validity of conclusions.

    Once data quality objectives are defined, the other QA activities (described above)  can and
 do affect the outcome of RREL projects.  The remainder of this paper presents case studies from
 RREL's experience with QA Project Plans, various types of technical audits, and final report re-
 views, and discusses their impacts on overall project performance. The final section of this
paper attempts to summarize what RREL has learned from these activities and the impact this
experience may have on future quality assurance activities.
                                         471
-------
                 QUALITY ASSURANCE PROJECT PLAN REVIEWS
    The purpose of the QA Project Plan is to relate the overall project objectives to the specific
measurements that are required. Emphasis is placed on sampling strategy and methods, analyt-
ical procedures, measurements of process conditions and requirements, and data treatment, since
these elements are crucial in evaluating the candidate treatment technologies. QA Project Plans
are reviewed to determine whether the proposed data generation procedures are adequate to
achieve project goals. Common deficiencies exhibited by the first drafts of QA Project Plans
include incorrect sampling and analytical methods; poorly defined objectives; incomplete or
inappropriate procedures for measuring precision, accuracy, or detection limits; or a sampling
strategy that does not match project objectives.  Examples of some of these problems are con-
tained in the following case studies.

Case No. 1: A review of a QA Project Plan for the chemical treatment ofvolatiles in water re-
vealed several problems that would have made much of the data  useless. These problems
included sample preservation methods that were inconsistent with the intended analytical pro-
cedures and the use of untested analytical methods that were of questionable validity. These
concerns were resolved in a face-to-face meeting between EPA and the contractors before field
operations began. In the process, approximately $80,000 in costs for unnecessary analytical
measurements were eliminated from the project budget.

Case No. 2: The QA Project Plan for this project, which pertained to the combustion of a herbi-
cide, originally showed several important deficiencies with respect to sample preparation and
analysis:

    (a)  The intended analytical method for the target herbicide was inappropriate and
         likely would have given erroneous results.  This error was especially important
         because this measurement was intended for the calculation of the Destruction
         Removal Efficiency (DRE).

    (b)  It provided for a single stack sample for several types of analytes, each requiring
         mutually exclusive sample preparation procedures.  Had sampling proceeded as
         planned, the required analyses would have been impossible to perform.

    (c)  The GCIMS target analyte list included labile compounds which are easily lost
         during sample preparation and analysis; in contrast, all of the intended matrix
         spike compounds were stable and easily analyzed, and were thus not represen-
         tative of the target analytes. Had these matrix spike compounds been used, the
         accuracy and precision of the key analyses would have been unknown, and the
         validity of any resulting conclusions would have been questionable.

    These three concerns were addressed to a  large extent in the next two drafts of the QA
Project Plan. While normally the second draft of a QA Project Plan is approved, in this case a
third draft was required, and a fourth was prepared that addressed several recommendations
that had been made.

Case No. 3: Another QA Project Plan for an incineration project exhibited serious problems in
the areas of analytical methods and quality control procedures:
                                         472
-------
     (a)  The analytical method for the most important compound, the Principal Organic
         Hazardous Constituent (POHC), had not been specified, although various alter-
         native, but unvalidated, methods were suggested. Under these conditions the
         probability of successfully achieving the necessary analyses was very small.

     (b)  Methods for assessing the accuracy of measurements for target compounds were
         inadequate. The only method for measuring the accuracy of some samples was
         surrogate spiking ofnontarget compounds at different points in sample recovery,
         but matrix spikes with the target compounds were needed. Data obtained under
         these conditions would not have been defensible.

     (c)  The QC procedures were not adequate, and for some analyses, including the
         most critical one, there was no discussion of QC procedures at all.  If analyses
         had proceeded as indicated, data would have been of questionable utility.

     These and other deficiencies were addressed in the second draft of the QA Project Plan,
providing a document that was complete and comprehensive enough that the objectives for the
project could be met.

Case No. 4: Another QA Project Plan which was reviewed during the last year pertained to the
chemical treatment process for PCBs in soil. Problems exhibited in this plan included an inap-
propriate sampling strategy, unclear QA objectives, and the use ofunproven methods. In addi-
tion, this plan indicated an excessive number ofGCIMS analyses that were not needed to
achieve project objectives. These concerns were corrected before the beginning of the project.
Regarding the  large number ofGCIMS analyses, it was recommended that analysis proceed to a
given level ofconfidence rather than to a fixed number of'samples.

     It is safe to say that the value of these and other projects would have been severely compro-
mised, had they proceeded as originally planned. Reviews of QA Project Plans has thus allowed
substantive concerns to be corrected before data generation begins.

           TECHNICAL SYSTEMS AUDITS DURING FIELD OPERATIONS
     During the field sampling and analysis stage of a project, a representative of the EPA fre-
quently visits the site to carry out a Technical Systems Audit (TS A). The purpose of this TS A is
to assure that sampling is being carried out as defined in the QA Project Plan and according to
scientifically sound procedures. Every effort is made to perform this TS A early enough in the
project to permit corrective action, if required, before data quality is seriously compromised.
The results of two field TS As are summarized in the case studies below.

Case No. 5: The first example of afield Technical Systems Audit (TSA) stems from an investi-
gation of the biodegradation rate of organic compounds in an aerated wastewater treatment
system.  This investigation sought to compare biodegradation rates in a full-scale treatment
plant with rates obtained from a bench-scale, in-vitro experiment. During the conduct of the
TSA, it was observed that the in-vitro test apparatus was housed in an air conditioned labora-
tory with an air temperature of 20 ° C, while the treatment plant itself was operating at 38 ° C.
Since biological reaction rates are temperature dependent, it was important that both tests be
conducted at the same temperature.

                   .                     473
-------
     Under the conditions described above, the in-vitro reaction rate could have been low by a
factor of four in comparison to the actual treatment plant results. (Biological reaction rates
decrease 50% for each 10° C decrease in temperature.) Based on the TSA observation, it was
decided to move the test apparatus from the air conditioned laboratory into an environment of
the correct temperature. Again, the problem was detected early enough in the program to pre-
vent significant loss of data. However, had this error in the basic experimental design not been
detected promptly, it would have seriously impacted the validity of the conclusions, and may
have required that the work be repeated.

    The findings of this TSA illustrate a common problem in which the field team is not
entirely informed regarding sampling and analytical requirements. It is unlikely that the field
team would have made the error described above had they been aware of the overall project
objectives.

    While some of the problems revealed during field TSAs are significant enough that they
might compromise the overall validity of the investigation, other less significant though impor-
tant problems are also found. One common problem encountered during field TSAs is that pri-
mary process wastes such as stack gases are sampled in a complete and thorough manner, while
secondary wastes or process feeds are sampled in a less rigorous manner,  as is illustrated by the
following case study.

Case No. 6: During an incinerator test burn it was observed that composite samples of the
feeds and wastes, such as scrubber water blowdown and bottom ash, were being collected by
combining aliquots of unequal volumes.  This procedure would have resulted in biased samples.
Since a primary objective of this project was to measure mass closure, representative samples
were essential, and the sampling procedure was promptly corrected as a result of this TSA.

                   LABORATORY TECHNICAL SYSTEM AUDITS
     Much like the field TSA, the laboratory TSA consists of an intensive on-site review of ana-
lytical and quality assurance procedures at a participating laboratory. Typically, such reviews
are carried out early during the laboratory phase of a project in order to make corrective action
on a timely basis. Most typically, a laboratory TSA reveals concerns or problems which the
laboratory corrects promptly. Perhaps the most common types of errors observed are associated
with a failure in communication between the various project participants, such as laboratory and
sampling teams.  The following case study provides an example of such an occurrence.

Case No. 7: During this TSA, the laboratory itself was seen to be performing satisfactorily. In
fact, the GCIMS analysis had been carefully optimized for the main compound of interest, a pes-
ticide, permitting reliable determination Of the target compound at concentrations below its
normal detection limit.  However, examination of the shipping records at the laboratory for the
project in question brought to light related problems with sample collection and shipment. Two
important concerns were noted:

     (a) Some samples which were specified in the QA Project Plan were absent from the
         shipping log because they were being discarded by the sampling personnel. The
         importance of this circumstance was somewhat mitigated by the fact that the
                                         474
-------
         samples in question were condensate waters from the volatile organic sampling
         train, and volatile organics were not critical compounds to the project.

     (b)  Some samples for VOA analysis were shipped in the same ice chest with samples
         containing high levels of organic compounds. No trip blank was included, unfor-
         tunately, and consequently it could not be determined whether any contamination
         tookplace. Again, the volatiles were not critical compounds.

     In this case, sampling requirements had not been entirely communicated to the sampling
group by those team members responsible for planning and documenting the test. Although the
major objectives of the project were not affected because VOAs were not a critical measurement,
the attainment of secondary objectives was questionable.

     In contrast to the latter case study, laboratory TSAs occasionally reveal much more severe
problems, as is illustrated by the following case study.

Case No. 8: A Technical Systems Audit of a general analytical service laboratory was per-
formed in which a number of concerns were noted, most notably in the areas of GCIMS and
trace metals analysis.

     In the GCIMS area, some of the findings were as follows:

     (a)  Tuning and calibration requirements for the GCIMS had not been met, and blank
         levels had not been checked, resulting in data of unknown quality.

     (b)  Documentation was inadequate in the GCIMS area, making it nearly impossible
         to reconstruct the analytical run sequences. It was thus nearly impossible to
         identify and isolate problems such as sample contamination that may have occur-
         red during a run, and data users could not realistically recheck their results or
         diagnose problems.

     (c)  In one case, a sample run was repeated on two consecutive days and gave results
         for some compounds differing by more than a factor of two. Although these data
         failed precision objectives, and were therefore of known poor quality, they were
         reported anyway.

     The area of trace metals analysis also exhibited some severe problems.

     (a)  An interference check sample had never been run for the ICP analysis; thus, the
         reported data was of suspect quality.                           .

     (b)  In one ICP run sequence, the calibration check failed, yet no recalibration was
         performed, and the suspect data were reported.

     (c)  A continuing calibration standard, which was run in the middle of a sequence
         during an  atomic absorption analysis, resulted in a measurement that was in
         error by a factor often. However, data were processed with no indication of the
         failed standard, meaning that the data from the run sequence were of unknown
         quality.

                                        475
-------
    As a result of this TSA, the GCIMS and trace metals analysis areas were deemed inade-
quate, andRREL suspended activities at this laboratory in these areas. Substantial corrective
actions have been made subsequently and the laboratory can now analyze RREL project
samples.

    This audit took place before the laboratory began running critical RREL samples. Had this
TSA not been carried out, it is safe to say that data quality would have been severely compro-
mised, unbeknownst to the principal engineers requiring these analytical services.

                            AUDITS OF DATA QUALITY
    Audits of Data Quality (ADQs) are performed during the latter stages of data interpretation
or early stages of report preparation. ADQs consist of a retrospective review of supporting data
beginning with raw data and proceeding to values included in the Final Report. ADQs have
been employed less frequently than other types of audits, but have nevertheless led to the iden-
tification of serious problems as well as recommendations for improving data quality.

    While it is impossible to affect the experimental part of a project at this stage, ADQs are
helpful in avoiding misinterpretation of results, as is illustrated by the following case study.

Case No. 9:  During the last year, anADQ was performed in support of a preliminary study of
an extraction technology for treating soils contaminated with volatile organic compounds. This
study required the long-term measurements of volatile compounds in gases, soil, and ground-
water.

    Review of the data from this study turned up several serious problems:

    (a)  Replicate analyses were not performed as required for obtaining an assessment
         of the precision of the methods.  Since it was important to the project to deter-
         mine differences in analyte concentration from the beginning to the end of the
         study, the lack of an adequate estimate of precision was a concern.

    (b)  An independent QC gas standard was not used as a check of the analytical
         instrumentation.  As a result, the data could not be compared from one cali-
         bration period to the next over the entire sampling period.

    (c)  Several sample concentrations fell outside the calibration range of the instru-
         ment.  For this reason, these data had to be regarded as estimates only, limiting
         their usefulness.

    (d)  Method blanks were not run on a daily basis. This means that it could not be
         determined whether interferences and contaminants were below the required
         levels.

    (e)  The holding times for VOST samples were exceeded, which could have affected
         the accuracy of these data.
                                        476
-------
     Because there were several problems found during the audit that could not be corrected,
some of the data from this preliminary study had to be considered of unknown accuracy and of
limited usefulness. While it was not possible to re-analyze the questionable samples, recommen-
dations resulting from this audit should be of benefit to similar projects in the future. It was
noted that many of the problems discovered during the audit were the result of lack of adherence
to the QA project plan.

     This example of an ADQ is perhaps extreme in its consequence, but nevertheless illustrates
that such ADQs are helpful in avoiding misinterpretation of data and potentially costly policy
decisions that might result from unjustified faith in poor data. Other ADQs completed during
the previous year have tended to support report conclusions and have been useful clarifying
ambiguous data sets.

                              FINAL REPORT REVIEW
     The primary objective of the Final Report Review is to determine how well the data support
the conclusions of the investigation. Toward this purpose, this review addresses the credibility
of the approach, sampling strategy and procedures, analytical methods, the implications of qual-
ity control results (e.g., spikes and replicates),  and the editorial quality of the reports.

     Perhaps the most common problem encountered in these documents is the presentation and
discussion of analytical results without discussions of the quality control results. These two
components are directly related, since it is impossible to judge the validity of the conclusions
without having a measure of data quality.  Final Report reviews consistently point out that the
presentation and discussion of QA data is crucial to assessing the validity of project conclusions.
This situation was illustrated by Case Study No.  10:

Case No. 10: This project dealt with the performance of a wastewater treatment system for
cyanide-bearing waste. While the report provided an excellent discussion  of the process itself, it
did not discuss the analytical results in sufficient detail. Crucial analytical data (TCLP results)
were not presented, nor was adequate discussion of the QA/QC analyses provided. From the
data presented it appeared that there was a serious problem with the cyanide analyses, yet no
explanation was provided. Based on the limited information provided in this report, it was not
possible to assess the credibility of the data presented, and the overall credibility of the report
was thus compromised. The review recommended that the complete analytical data set be pre-
sented, including the results of the QA/QC analyses. Given this information, it would be possible
to judge the overall effectiveness of the treatment process described.

     Another problem revealed less commonly by the Final Report review is conclusions that
are overly generalized and not entirely supported by the data. This situation is illustrated by the
following case study pertaining to a solidification process.

Case No 11: This report on a solidification project discussed the location and immobilization
of the hazardous trace metals within the solid matrix based on electron microscopy.  While this
technique can visualize some major constituents, it is not able to detect these metals at the con-
centrations of interest to this project. By identifying which conclusions were valid and by
recommending the removal of those conclusions which were not, the review enhanced the
reliability of this report for EPA's decision-making process.
                                          477
-------
     Another important long-term benefit of the final report review is that it provides "feedback"
 to the QA staff regarding any continuing problems, especially those that may affect future and
 ongoing projects. A case in point is the review of a Final Report for a PCB incineration project
 which brought to light several fundamental problems with implications for future and current
 projects:

 Case No. 12:  The first problem made apparent by this report was that the analytical methods
 were not sensitive enough to allow calculation of the required DRE. Referring back to the
 original QA project plan, it became apparent that project objectives had never been stated in
 numerical terms, and as a consequence, inadequately sensitive methods were selected for key
 analytes. These findings have sensitized the RREL QA staff to the importance of clearly stated,
 numerical project objectives.

     Another problem of broad interest that became apparent during this review was the dif-
 ficulty in interpreting PCB measurements made by GC/MS according to Method 680. These
 difficulties include, among others, the inability to detect PCBs at sufficiently low levels or to
 even define meaningful detection limits. This is a severe limitation for incineration processes,
 which must demonstrate a DRE of 99.9999%. In subsequent projects, RREL QA staff generally
 have recommended that pre-test samples be prepared to simulate the treatment process, and that
 these samples be analyzed and interpreted just as real treatment samples prior to the start of the
 test proper. This step helps to elucidate potential analytical and data interpretation problems
 prior to incurring the cost of a complete field test.

                                LESSONS LEARNED
     Perhaps one method of evaluating the vigor of the RREL Quality Assurance Program is to
ask whether it has had any beneficial effect on RREL projects. As is illustrated by the above
discussions and case studies, it is obvious that the in-depth objective reviews provided by the
QA Program have indeed been instrumental in avoiding numerous errors ranging from the use of
•improper analytical procedures to unsupported conclusions. The review process has also been
useful in facilitating communication between diverse interested parties, such as regulators, engi-
neers, chemists, statisticians, and developers. For this reason, the independent, objective review
has been the mainstay of the RREL Quality Assurance Program and will continue to be so in the
near future.

     Experience during FY 88 has also illustrated areas that will require additional attention in
the near future. One such problem area is that project objectives are often not as well defined as
is preferred. It is thus expected that additional effort will be directed towards the development
of (preferably numerical) data quality objectives. Another area of uncertainty is the selection of
best methods for measuring PCBs for incineration or other destructive processes, as discussed
above. Finally, experience during FY 88 has indicated that unneeded delays sometimes occur
simply because contractors are unaware of what is expected in the QA Project Plan.  For this
reason, it is hoped that additional instructional material will be prepared and disseminated during
the following fiscal year.

     Finally, it must be emphasized that the external review process afforded by the RREL
Quality Assurance Program is no substitute for the commitment of knowledgeable project staff,
                                         478
-------
who are directly responsible for the quality of work performed. However, unlike a project in-
volving a single principal investigator at a given location, RREL projects tend to require the
coordinated efforts of regulators, managers, engineers, chemists, statisticians, and economists,
and under these circumstances it is not surprising that many of the problems revealed by the
review process  arise not from lack of dedication on the part of project members, but rather from
communication problems among the diverse parties involved. Chemical analysts may be un-
aware of regulatory or engineering objectives, engineers may be misinformed regarding analyt-
ical limitations  and possibilities, the sampling team may lack instructions regarding sampling
requirements, and managers may have unrealistic expectations regarding the readiness of the
project team. It is precisely because of the complicated nature of RREL projects that such
potential problems arise, and it is for this reason that the quality assurance process has become
an integral part of RREL projects.
                                          479
-------
               EFFECT OF FEED CHARACTERISTICS ON THE PERFORMANCE
                      OF ERA'S MOBILE  INCINERATION SYSTEM
       by:  James P. Stumbar
            Robert H. Sawyer
            Gopal D. Gupta
            Foster Wheeler Enviresponse, Inc.
            Edison, NJ  08837
Joyce M. Perdek
Frank J. Freestone
Releases Control Branch, USEPA
Edison, NJ  08837
                                   ABSTRACT

    During the past four years, the EPA Mobile Incineration System (MIS),
has processed a wide variety of feeds.  Besides  ->  incinerating the
hazardous materials for which the MIS was designed, the unit has also
incinerated contaminated debris including wood pallets, steel and fiber
drums, and plastics.  This paper identifies significant physical and
chemical characteristics of various feed materials and their relationship
to MIS performance.  The paper also correlates the effect of these feed
characteristics on specific MIS components.  Corrective actions taken to
mitigate several problem characteristics are presented.  The operating
experience with the MIS has provided valuable data on the limits of
incineration capacity as well as reliability of the unit in relation to
various feed stocks.  This information is also discussed.  The information
contained in this paper is directly applicable to field use of mobile and
transportable incinerators at Superfund and other industrial cleanup sites,
                                    480
-------
INTRODUCTION

    Under the sponsorship of the Office of Research and Development of the
U.S. Environmental Protection Agency (EPA), the Mobile Incineration System
(MIS) was designed and constructed to demonstrate high-temperature
incineration of hazardous wastes (1).  The system essentially consisted of
a refractory-lined rotary kiln, a secondary combustion chamber (SCC), and
an air pollution control system.  These three components are mounted on
three separate heavy-duty semi-trailers.  Monitoring equipment is carried
by a fourth trailer.

    The MIS was rigorously tested in Edison, New Jersey, during 1982 and
1983 with PCB-contaminated and other chlorinated organic liquids (2).  The
system was transported to the Denney Farm site in McDowell, Missouri, in
December 1984 for a dioxin trial burn and field demonstration (3,4).  A
total of 900,000 kg of solid and 81,600 kg of liquid dioxin-contaminated
materials was incinerated between July 1985 and February 1986.  During
1987, the MIS was modified to double its capacity and to improve its
reliability.  A second trial burn was conducted on both solids and liquids
contaminated with chlorinated organic compounds and PCBs during August and
September of 1987 (5).  Since 1987, an additional 3,200,000 kg of solids
and 31,000 kg of liquids have been decontaminated or destroyed.

    Over the lifetime of the MIS, a wide variety of feed materials have
been processed.  These materials exhibited differences in characteristics
that affected the MIS in various ways.  Often a particular characteristic
or a combination of characteristics would affect the MIS performance
adversely.  The experiences gained from field operations of the MIS during
the past four years have increased the understanding of the interplay of
feed characteristics with hardware.

    This paper describes the effects of feed characteristics on the MIS
performance; correlates various feed characteristics with affected parts of
the system; describes actions taken to mitigate the resulting problems; and
discusses the limits imposed on*capacity and reliability by the various
feed characteristics.

FEED CHARACTERISTICS

    Both the physical and the chemical properties of the feed determine
incineration system performance (6).  Important physical properties
include: heating value, morphology, density, rheology, ash particle-size
distribution and fusion characteristics.  Important chemical properties
include the composition of the feed as shown by:  organic content, organic
hazardous constituents, acid forming elements such as sulfur and the
halogens, moisture content, and inorganic ash components.  These properties
can affect the operating parameters, the capacity, and the reliability of
the incineration system.  Many of these properties are interdependent as
far as their effect on the incinerator performance.  The manner in which
these properties affect the performance of incineration systems, based on
the experience with the MIS, is summarized in Table 1.
                                    481
-------

















s
zc
cf
OC
a
X
111

c/>
>-

g
i
o
a:
u
=s
§
u.
u.
M
UJ
i

-
UJ
_J
s
i—





















01

f~

u
o
CA
"8
0)
u.


01
i
t-
o
L.
a.
s
4->
U
>s
15





•§

o
"co
fl>
g
g.
g
ca
8.
C>


I
1
**-
ca
o
c.
cn
3:
(0
as
















1
CL
cu tu ••
O 4J 0)L.(U CU (U O
» _n> C/J •• D) DO) _O) •* P) _O)
,f •. iff CO •» "O 4-*^ T» — ' i? JCT3
co to 3 * o co D X 3 3—3 COD
C04-'— 'C0"~ 4-> — J , .E c -c •" '£ •£

M "O E QJ D) "O E E co E 32 E lr— 6 E co E
— f O C- 4-> -r- O 0) t- tic. 5 L. L.L.Q) — ' £-
a. IE ca crt t- n > ca H- CD s: co LU co > o. co
1 . . • o> S
(U CO (U CU* E* C (1)
O) 3 O) C  o> 4->
O) O CO — ' CO U> W
A) tO 3 (U t- "D
3 C co D - cu -
UU O O U. CO 4J • C 08
• p. 4-» (0 — '
*. 4_» * -O3; "-CO
>* Q. >*>s>s >s3C0^4J
• P. t_ IP- «r- IP- >p- CO C H- CU
CO C COCO CO CO O) U C. C-MO
T3 *n T3 T3 "O "O ^^ * 1— O) •'~ "~
m O (Del) CU u> DO*-H COD
flj 
S >: « S b
o> C4-» co *O4-* «
_/ .P- .P- O •— CO CO
% H>o CE&UJ
*. CO 4-* CD 1— •*
cu "n^c-^ CJCDCU cu
t. — i O D £ «- 4-» » t-
DO CU«- OO «p- 3 CO — ' D
+^CJ JSO> •*- 4-» 4-» — • O 4-*
CO CO C*O=C4-> CODE- CO
t, — » 'p- CO Q. (- CUC-O4-' C-
CUC » CO ••- -P- S- X H- O
^ O O O 4-» > C 01 ^i CO O COL.
c c o co 3: co ^ t- — « t_4-j
>. 5 4-fCUC-OO O)>-73COO TI C
co»2 "SlrcQ.* 4-' co
4-»C/l .P- CU — ' E O D4J— 'OC — •
OO CDL..P-3— » — • O •— X 0 -r-CNJ
fv» r T ~y ii y^ Q^ m LU Ctf *sjf f t f O ^ O
CA
S - ^ « - c" 3
CT ^ .i:-<0 Socx ° S
s: i- 3C>-0)C
<: u «4-'>i- c/) ca^^ocjo
ECCO .4J.r- ^.^
•. >. *. CUOCU •• CJCAC_4-> »O4-*
C E C EC4JOT3 S CJOOCO CCCO
^ S •'^4J''^>»C4JC4J 4JCOC"^3C

tT W [^ "^"SSS-lo " Suj-S'S tTuTS
CO "D COT3CO C — ^ "- t- "U — ' O4J I04J4-*
.M CD 4^ CU4-> 35 — ' 3 CU CU O4^CUCA 4-*CJCO
O CU O 0) O 3 O V Q- CU >s fl) t. C O 13 C
EV* ii rw n r^ Qf Q. 0) C3 t^ O 2E CL *~* O£ O "^
CO
CO
4-i C  CO •
c o o o at
* ir-
n u> o .a 4-<
. H- ii- CD W
"5 CO CU C
§M 01 .C C 0)
N O •£ 4-»

CO >* C W) C t_ CO

O >« CUOI ii-4-»CO
n> — • >.6)C "- — uia>j=
C 0 iVocu 30 DQO
•^ ^ ••- — • O) 4-> — f- »^
4J ^. « p 0 0) 4->
CO I- CCU — ' IP-C- JZ
cu o cujcca oco to
33 s: ac^sc z:a. «t
482
-------
EFFECT OF HEATING VALUE

    Heating value of the feed material affects both feed capacity and fuel
usage of the incinerator.  As the heating value of feed material increases,
kiln temperature can increase and sometimes become uncontrollable.  The
kiln also requires greater amounts of oxidant to complete combustion and
greater quantities of inert material to control kiln temperature.  This
temperature increase can limit feed capacity.  The MIS reached its capacity
limit at 1.33 to 1.61 megawatts (MW) heat input to the kiln.  Feed
materials, such as plastics, trash, wooden pallets, and brominated sludge,
had capacity constraints caused by high calorific values.  Maximum feed
capacities for these materials ranged from 90 kg/hr for pure plastics to
859 kg/hr for brominated sludge as shown in Table 2.

    Solid materials with high calorific values cause transient behaviors
that sometimes further limit feed capacity.  Plastics, trash, and wood
ignite almost immediately after they are fed to the kiln.  Gases evolved
from these materials burn rapidly producing a sharp increase in kiln
temperature and a sharp decrease in excess oxygen.  Prior to the 1987
modifications, the MIS was extremely sensitive to these transients, which
caused many feed stoppages.

    After the addition of the LINDER Oxygen Combustion System (OCS), the
MIS response to the above transient behavior was improved, and. feed
stoppages due to low oxygen were virtually eliminated (7).  However, there
were still many feed cut-offs caused by excessive kiln temperatures.  These
were minimized by operating the kiln at the lower end (790°C) of the
temperature range, allowed by the RCRA permit, and by using water injection
to control kiln temperatures.

    For brominated sludge, the behavior of the MIS was somewhat different.
Large oscillations of the kiln temperature and excess oxygen level occurred
even when the kiln was operated at 790°C.  The resulting
over-temperatures (greater than 1040 °C) caused feed cut-offs and loss of
the kiln burners.  Loss of the burners increased the length of the feed
cut-off period.  The operating changes required to alleviate this
phenomenon are described below.

    To reduce the amplitude of the temperature excursions, an automatic
feed cut-off was introduced into the kiln control system.  This stopped
feed whenever the kiln temperature exceeded 945°C or the oxygen level
dropped below 4% (wet).  This action minimized overtemperature  incidents,
but the oscillation frequency was still large.  As shown in Figure 1, about
four oscillations occurred per hour.  Feed was cut off for approximately
eight minutes during each oscillation.  Feed rate was limited to 450 kg/hr
under these operating conditions.

    Observations of the  kiln during the oscillations showed that the sludge
was not igniting rapidly.  Several batches of  sludge would be fed by the
ram before ignition occurred.  After  ignition, flame would fill the kiln
and oxygen flow and temperature would increase rapidly.  After the sludge
                                     483
-------











tu
z


g
u.
s

CO
X
g
CO
o
UJ
Ul
u.
CO
s
a:
u.


0
UJ
u.
u.
UJ


1
CO
u
CO


Ul

1


u
s
UJ
u.


CM
UJ

CD

fr—






















0.
•5 c
4-» (_
2 g
d§


o
H-
g

g
4->
S

]i
J3

o
g 4->
1 2 C
CO 09 CD

•- C
c o> ^

t» c O
c5 8

0)
3

U

1




_e 

co ca co
a> > o
3= **r

CA CO
G> C

U c.
••— O
4-> .=
u E
CO ^^
a. o jc
*o V S







"to

fe
4-»


1
U.


1
O£
«4-
O
i
—3
E
CO
a:

.1
4-*
CO
cu
u

u
u
c/>



o
o
o

f*.
CM
o
1
CO
ro
*";
o





in

CM





o
CM






f".
si-








co





X
CO
1


o
CO
E
c.
u.


0)

3 X
O (D

01 O
0 ~
0 t
% §2
*:
7 IT
ro t-






4-f
'o

a.
CO
o
















































ca
T3
o
0) X
L. 4-»
O "-
H- CO
Q 0)
T3
CO U
0. 0

I CD
8 j*
*-* v*





S

£^ g
— ' CO
0)
•t- T3
O CO
z !a
— • c.
1- O
C. 4->
CU U
> 0
o a
x -g

"° ~0 E
in XI o
N- C. sr
CM O «-
CM O «-

in
co

1

_ _














o
= h-
A





















ca

i
C-

4^
CO

18 ^

«- c/>
1
8 9
^^ £
1
CO
cu
o
c
X
•8
1
ro
c

• r-
***

c 'i x
S g. 1 -§
o -S ^ ro
g. e

4-> o:
c -
s .i
4J
ro D • X
CO. CO 4-*
•^ C 0) .ii
CD — t- 0
O) CO
ro D cj a.
— • I— U CO
co ca » u



o
o
o

N-
S!
1
i
















o
CM
















^j.

m
I-O

0)

4-» O
'« CU
t L.
— • O
• p. H-
O O
CO ^3

E MD
L. CO
CO O
U. T—
1
II
Ul NJ*


g.
CO
13
in
1


C

H-
o
8
H-
L.

°
>s
t.

1C
RJ

in
CO
i
CO
•R

j.








S





























C/l

i
L.
0)

H-
CD

§
O

1
8
\_x





M


._,
-s
CO
^ 2
"cu o

t. u
o o
u a
T.,

g
o






=














o

A



























'5
CO




s:

i
(G
O
0 g.
CO
_c -a
•— o
7 >
ro cu
c &
_j (D

^ ^

(tJ
^ g g.
O CD "-
4-* CD
8 CO 13
— • H-
a co ca



o
CO






_








5












O
o
in
















CO
u



J2
Q.




E

O
C
CO
_fl)
u
c
X
•8
1
ro
m
c

e
CO
0)
CO
JD
c_
o

u
o
• a


o
CM
T
o
o

tn
CO
I
g
r—
in
i
o




o
0


o
in



o


o

0
0
in
o
i
o
o
 r>
CD t~
co m
CO »


^ e
^ 4J























































u
o
•v,
O>

CM

CM
1
O
o
484
-------



























^
T3
V
c
4J
C
O
o

<\J
UJ

CO







































g
to
1

13
tu
4-
c
o
c
o
to
4-»
I

i

'x


c

s
c
Q)
0
"to
3:



o
'c
to
CO
t_
O
01
3
CO

o
ss

O)
c
4-f
to
tu
=c
CO
tu
"o
4J
t.
to
a.
H-
o
a*



to


tu

i

1

u-








co
concern








a>
eg

"g £
tU D)

*-*
Q. ^N

4-> --^
4-f
C

0 *"*
0




g ~
£ S
O
o








O D)
"to 'to
> 0
sx
CO
c
E
o
'i
O J=
V v^
- - •


















«^


CO
o
£ CO
•*_ 
£•
•*-
!5
8. x
ID 4->
O •*-
O
c. to
tu a.
•a to
"8
t- B
w ex

in
CJ

& s

in tn
CO CO
i i




o o






0 0




o !






CD O





0 0




to

tu tu

c. »-

"° 0

11 e
tu t-.
4J tU
CO >

c - c
CD CO 1 C O
tu >. — • at •—
— » to 4J ••- C 4-»
o -a u !M! ~^ (D _,

ftM T3 C 3 C O
«— s C. 4J •. t- O
CO tU O) tU CJ 4-* O
> c J5 cj co
>. tU -r- CO C 4->
T3 4J 3 t- -n "~ to
I 3 O O C •«- L.
ro o H- 3 to o -a
tA
TO
!S •
o tu
Co E

, i g i
~ -§ .£
.- 8 2
;*: t- ca
0
S3"
i
o
in

^

o
§'



«-





ro
i
in



t\j
i
in
ro
o
o

ro
i
*~




s.


o £
f3
4->
_, .X

CO E

•D C
tu t
4^ rj
to .a

o •
1
o
§


•o
CM





in
i



CO
t>






i





i










o
CO



ca
tu
0
o
c.
a.


4-*
instabili
water IDS

o
i
4J 0)
CD <-
O
O t-
tO CD
CO «-
O »-
o
o
i
o


=

=


-o
<>






o
o




o


o
co

^^
i
CD
CO
t>
ro




£






at


3
_cr
















v
B.
C
*
'il
o
1





















































485
-------
had burned several minutes, the flames would extinguish and the oxygen flow
and temperature would decrease rapidly to complete the cycle of
oscillation.  It became apparent that steady ignition of the material was
required to prevent the oscillations.

    The kiln temperature was increased from 790°C to 900°C to provide
the necessary energy to evaporate the water and volatile organics so that
ignition could be sustained.  This operating change was successful in
reducing the oscillations to a minimum as shown in Figure 2.  Maximum feed
rate was increased to 900 kg/hr by the above changes.

    For feeds with high heat content such as brominated sludge, the
capacity of the MIS is increased by water injection.  Due to its high heat
capacity, water provides a very effective heat sink.  Consequently, when it
is used to control kiln temperature, moisture increases the SCC residence
time as compared to the use of excess air.  Figure 3 shows that the use of
water injection can increase capacity by about 20% over the use of excess
air at a given SCC residence time.  The use of oxygen in the kiln enhances
the effect of water injection by allowing further capacity increases.  At
an enrichment to 40% 0? in the combustion air, capacity can increase by
60%.

EFFECT OF MORPHOLOGY

    The morphology of the feed material affects the feed system by causing
periodic jams.  Most of these problems are caused by materials that are
poorly prepared by the shredder due to their morphological
characteristics.  Problem materials consist of wooden pallets, metal drum
closure rings, plastics, trash, clothing, and mud (8,9).

    The feeding of these materials restricts the MIS capacity.  As shown in
Table 2, relatively dry soils can be fed at rates up to 2275 kg/hr but the
presence of plastics and mud reduces the feed rate to 680 kg/hr.

    The shredder is used to prepare the solid feed materials for
incineration.  For most materials, the shredder works extremely well.
However, wooden pallets and metal drum closure rings often cause feed
blockages when shredded in the present equipment.  While the shredder
breaks most of the wooden pallet into 5-cm wood chips, an occasional board
will position itself to go through the shredder as a 5-cm wide by 1.3-m
long sliver.  The same is true for the drum lid rings.  The shredder
sometimes drags a ring through, straightening it but not cutting it.  In
each case, plugging of the conveyor, weigh scale, or ram follows.  The best
solution has been to manually separate and prepare these feed materials by
cutting them into small pieces (about 24 cm in length) prior to shredding.

    The shredder is unable to handle pipe or thick metal pieces.  These
must be manually classified, cut to proper size, and placed on the main
feed conveyor downstream of the shredder.

    The shredder also performs poorly on materials such as plastic,
clothing, trash, and mud.  These poorly shredded materials often jammed at
                                    436
-------
                                                                        Excess
                                                                        Oxygen
                                                                        % Wet
                   -• [-l__ Excess Oxygen -±~ j
          NOTES:
          - Brominated Sludge
          - September 2,1988
          - Before Operating Changes
           Figure 1.   Oscillations In  temperature and
                     excess oxygen In rotary kiln.
Temp.
  °C

 1095
  955
  815
  675
Excess
Oxygen
% Wet

  12.5


  10.0


   7.5


   5.0


   2.5

    0
         NOTES:
         - Brominated Sludge
         - September 8,1988
         - After Operating Changes
         Figure 2.  K11n temperature and excess oxygen
                   after operating change.
                                      487
-------
the doctor blade or restricting dam that was originally used to level the
granular material on the conveyor belt as the belt exits the shredder
hopper.  The doctor blade worked quite well for granular material but it
created large blockages when materials such as shredded plastic, clothing,
trash, metal, or mud were fed.  A roller has been installed to replace the
doctor blade, and this has reduced the number of jamming incidents.

    In one instance, the combination of metal and mud damaged the main
conveyor belt.  The belt was slit for 21 of its 26-m length by a metal
shard which was embedded in mud.  The clump of mud containing the shard had
stuck to the belt, passed under the top-end belt wiper, and lodged in the
underside rollers.  The top-end wiper was repositioned to provide a more
positive wiping action.

    Finely granulated material affected the operation of the ram feeder.
It would bypass the ram head and collect on the backside.  The material on
the backside would periodically prevent the ram from fully retracting.  A
small chain-plug conveyor was installed and timed to convey the bypassed
material to the front side of the ram.  This solution has worked quite
well.

EFFECT OF DENSITY

    The density of the feed determines capacity for many feed materials.
For feeds of typical densities (1.5 g/cc), such as soils, the maximum feed
rates were 2275 kg/hr.  The maximum feed rate was obtained at a kiln
revolution rate of 1.6 rpm, which gives a typical solids residence time of
30 minutes.  For low density materials such as vermiculite (0.096 g/cc),
feed rates up to 364 kg/hr were feasible.

    The density of the feed determines capacity for many feed materials,
because the density of a material is inversely proportional to the volume
it occupies, and the ultimate feed rate for a given material  is limited by
the volume capacity of the kiln.  The volume capacity of the kiln is
determined by the amount of material that can be held by the kiln without
overflowing the corbel.  The corbel is an annular lip on the front end of
the kiln, which rotates with the kiln.  There is a space between the corbel
and the front plate of the kiln.  When material  gets into this space, the
seals of the kiln are damaged.  The maximum volume that the kiln can hold
without spilling over the corbel is about 15% of its total  volume.

EFFECT OF RHEOLOGY

    The rheology of the material affects either the feed system or the
decontamination behavior.  Muddy soils, fed to the MIS, formed clumps of
material, which were caught by the doctor blade and also would stick to the
conveyor belt, weigh scale and ram trough.  The resulting buildup would
periodically plug various parts of the feed system thus reducing overall
feed rates to about 1140 kg/hr.  This is approximately 50% of the maximum
rates achievable with dry soils.  Addition of vermiculite has eliminated
the sticking of the muddy material  while adding only a small  amount to the
throughput weight.
                                     488
-------
    The brominated sludge that was fed had a tendency to form balls up to 8
cm in diameter.  Since the time required for burnout of a sphere is
proportional to the square of its diameter, the large balls require a much
longer residence time in the kiln for decontamination.  At the normal 1.6
rpm revolution speed of the kiln for soil, small smoking particles would
exit the kiln with the kiln ash.  This required limiting feed rate to about
450 kg/hr.  However, when the residence time of the sludge was increased by
reducing the revolution speed to 0.8 rpm, feed rates up to 900 kg/hr were
achievable without smoking.

EFFECT OF HALOGEN AND SULFUR CONTENT

    Incineration of brominated and chlorinated wastes generate the acid
gases hydrogen bromide (HBr) and hydrogen chloride (HC1).  These acid gases
affect the capacity, the blowdown rates that control total dissolved solids
(TDS) in the process water system and the particulate emissions.

    A capacity limit of 115 kg/hr of acid-forming organic chlorides was
encountered during the 1987 trial burn (5).  The capacity limit was caused
by pump cavitation in the quench system, which cools the gases exiting from
the SCC to about 95°C.  The cavitation reduced the quench water flow rate
which activates the protective instrumentation resulting in cut-off of the
feed and shutdown of the burners.  This cavitation was produced by
excessive chlorinated waste feed rate as follows:  The quench water is
treated with caustic solution to neutralize acid gases.  Reaction between
HC1 in the combustion gases and sodium hydroxide (NaOH) in the quench sump
produces effervescence.  The amount of effervescence increases to violent
levels as HC1 flow rate increases.  The violent effervescence reduces the
available net positive suction head (NPSH) of the pump, which causes
cavitation at very high HC1 loads.

    High organic chloride loads also affect particulate emissions through
the phenomenon of mist carry-over into the stack.  The amount of carry-over
is determined by both the HC1 .loading of the flue gas and the TDS of the
process water (5,10).

    In tests performed prior to the trial burn, particulate emissions were
found to exceed the allowable emissions (180 mg/dscm) by as much as a
factor of three.  The relationship between particulate emissions and
organic chloride loading is shown in Figure 4.  Table 3 presents the
analysis of the Method 5 particulate filter cakes, which shows that the
major portion of the particulate was sodium chloride (NaCl) and sodium
hydroxide (NaOH).  The emissions were brought into compliance after a mist
eliminator was installed.

    However, data taken during the 1987 trial burn showed that TDS of the
process water also affected particulate emissions.  As shown in Figure 5,
particulate emissions were proportional to TDS during the trial burn
tests.  The data shows that operation with TDS at 20,000 ppm adequately
controls particulate levels at high chlorine loadings.

    The TDS is controlled by adjusting the blowdown rates as follows:  For
                                    489
-------
                     TABLE 3.  RESULTS OF ANALYSES OF METHOD 5
                         PARTICULATE FILTER CAKES FROM  HIGH
                       ORGANIC CHLORIDE LOADING TESTS OF MIS
   Test Description
   Date  (1987)
High Chloride
    without
mist eliminator

   6/18-19
High Chloride
     with
mist eliminator

     7/20
   Total particulates

        Iron  (Fe)
        Chromium  (Cr)
        Sodium (Na)
        Chloride  (Cl)
        Aluminum  (Al)

   As Sodium Chloride  (Nad)
   Remaining Na  as Sodium Hydroxide
   (NaOH)  (Excess caustic)
(weight in grams)

   0.3412

   0.0202
   0.0022
   0.1560
   0.1367
(weight in grams)

   0.0524

   0.0057
   0.2253
   0.1174
   0.0175
   0.0200
   0.0022

   0.0329
   0.0079
      Organic chloride loading » 71.5 kg/hr
      Organic chloride loading = 83.4 kg/hr
                 TABLE  4.   RESULTS OF ASH DEPOSIT ANALYSIS
                         FROM MIS CYCLONE RISER DUCT
Analyte
Silicon
Aluminum
Titanium
Iron
Calcium
Magnesium
Sodium
Potassium
Sulfur
Phosphorous

(Si)
(Al)
(Ti)
(Fe)
(Ca)
(Hg)
(Na)
W
(S)
(P)
Reported as
Silicon Dioxide
Aluminum Dioxide
Titanium Dioxide
Ferric Oxide
Calcium Oxide
Magnesium Oxide
Sodium Oxide
Potassium Oxide
Sulfur Trioxide
Phos. Pentoxide
Amount
(wt. %)
25.6
8.5
0.3
5.0
36.7
1.5
0.4
0.4
21.0
0.5
Analytical
technique
X-ray
X-ray
X-ray
X-ray
X-ray
AA
AA
X-ray
X-ray
X-ray
AA » Atomic Absorption Spectroscopy
                                       490
-------
      Ill
      H
      Ul
      IU
      a
             5.
            4 .
            3.
            2 . .
            1 -
                        \
                           Permit Limit
                    0.2
                         0.4
                               —i—
                               0.6
—1—
 0.8
T—r—
   1.2
                                                       1.4
                                                             1.6
—I—
 1.0
                                  FEED RATE (kg/hr)
               NOTES:
               Solid Feed Heating Value    1.50 Kcal/g
               Kiln Temperature           925°C
               SCC Temperature           1200°C
               A - Water injection using 40% oxygen-enriched air for
                   combustion
               B - Water injection using air for combustion
               C - Air cooled using either air or 40% oxygen-enriched
                   air for combustion

               Figure  3.  Effect of cooling media on SCC residence time.
VI
en
CO
Ul
I
     200 .
     100 .
                    20          40          60
                    ORGANIC CHLORIDE LOADING (kg/hr)

         NOTES:
         + Before Addition of Mist Eliminator
         0 After Addition of Mist Eliminator

      Figure 4.  Effect of organic chloride loading
                on particulate  emissions.

                                  491
                              100
-------
a-chlorinated waste 1.65 kg NaCl is formed per kg Cl in the waste.  To
maintain the IDS at 20,000 ppm, 82.4 kg of process water must be drawn from
the system for every kilogram of Cl processed.  This illustrates that the
acid content determines the required process water blowdown rate.

EFFECT OF MOISTURE

    Moisture content affects incinerator performance and can adversely
affect rheological behavior as described above.  Depending upon the heat
content of the waste, moisture can either improve or impede incinerator
performance.

    When using feeds with high heat content moisture acts as a heat sink to
control kiln temperatures.

    Conversely, when using feeds with low heat content, moisture increases
auxiliary fuel requirements and decreases SCC residence time causing a
capacity debit.  The effect of moisture on SCC residence time is shown in
Figure 6.

EFFECT OF PARTICLE SIZE DISTRIBUTION

    The particle size distribution of the ash generated from the waste
determines the amount of particulate carry-over from the rotary kiln to the
rest of the system.  The importance of this characteristic can be
demonstrated by the MIS experience with Denney Farm soil and Erwin Farm
soil.  As shown in Table 2, Denney Farm soil is much coarser than the Erwin
Farm soil.  Up to 25% of the Erwin Farm soil would carry over from the kiln
to the SCC,  This caused a rapid buildup of solids in the SCC.  The solids
buildup necessitated a 70-hr shutdown for removal of the slag after each 96
to 120 hours of operation (45,000 kg of soil processed).  The behavior of
the silt caused the unit to be unavailable for operation an average of 40%
of the time due to the need to clean out the SCC.  On the other hand, the
unit could process the coarser Denney Farm soil for about 600 hours
(270,000 kg of soil processed) before a shutdown for slag removal from the
SCC was required.  The unit was unavailable for 10% of the time due to SCC
clean-outs with the coarser Denney Farm soil.  In both cases, the buildup
of solids in the SCC significantly reduced the availability of the
incinerator.

    The problem was mitigated in 1987 when a cyclone was added, between the
kiln and the SCC, to remove the fines carried over from the kiln.  The
system operated over a three-month period and processed over 500,000 kg of
solid material without requiring a shutdown for slag removal from the SCC.

    Although the cyclone has alleviated the solids buildup in the SCC, fine
particulates still have caused problems with the operating instruments.
The large number of fine particulates associated with brominated sludge
fouled the kiln oxygen meter and the SCC thermocouple about once every
eight hours.  This increased the number of over-temperature incidents in
the rotary kiln, caused the incinerator feed cutoff to actuate due to a
false SCC low temperature measurement, and increased the fuel flow to SCC
                                     492
-------
     1
     55
     0)
     UJ
     3
     I
                                 20
                          TOTAL DISSOLVED SOLIDS - TDS (ppm)
               NOTES:
               Solids Feed Rate           1800 kg/hr
               Liquids Feed Rate          60-160 kg/hr
               Organic Chloride Feed Rate  50-74 kg/hr

         Figure 5.  Effect of TDS on particulate emissions.
ui
P
in
o
UI
Q
DC
8
CO
                               30
                             MOISTURE (%)

         NOTES:
         Solid Feed Rate         1820 kg/hr
         Kiln Temperature        925°C
         SCC Temperature       1200°C
         30% Oxygen-Enriched Air
       Figure 6.
Effect  of feed moisture content
on SCC  residence time.
                                  493
-------
       Sample 1
   QUENCH ELBOW
1.5x
   Sample 2
CYCLONE RISER
DARK SURFACE
Figure 7.  Macrophotograph of samples from Quench Elbow (1)
         and Cyclone Riser (2).
                            494
-------
Figure 8.   SEM photomicrographs of the cross section
           of the deposit from the Cyclone Riser.
                                   495
-------
burners.  The thermocouple problems have been eased by changing the
thermocouple location and using a thermowell rather than an aspirating
thermocouple.  No satisfactory solution has been found for the kiln oxygen
measurement.

EFFECT OF ASH FUSION CHARACTERISTICS

    The ash fusion characteristics of some feed materials caused the
formation of hard slag deposits in the kiln; consolidated deposits in the
ductwork between the kiln and cyclone, between the cyclone and SCC, and in
the cyclone exit tube; and consolidated deposits in the SCC exit venturi.
The ash fusion characteristics are determined by the chemical composition
of the ash.

    Ashes containing elements such as sodium (Na) and potassium (K) have
low slagging temperatures.  Ashes from plastics, glasses, wood, and other
components of trash are  rich in these compounds.  The increased slagging
tendency of trash was experienced in the rotary kiln, which required a
system shutdown about every ten days to remove the slag build up caused by
incineration of trash.

    Ashes containing significant quantities of calcium  (Ca), iron  (Fe),
sulfur  (S), or phosphorous  (P) have moderate slagging temperatures.
Although brominated sludge, containing both Ca and S did not slag  the kiln,
the ash produced a consolidated deposit, shown in Figure 7, which  fouled
the ductwork between the kiln and the SCC.  A consolidated deposit also
occasionally formed in the  quench elbow upstream of the quench nozzles.
This  fouling necessitated a system shutdown about every twelve days to
remove the deposits.

    Samples of the deposits were analyzed  to determine  the mechanism of
-deposition.  More details on this topic are provided in reference  11.  The
deposit mechanism was found to be similar  to those operative in boilers
fired with subbituminous coal.  The deposits were formed by  sintering of
calcium sulfate  (CaS04)  in  the temperature range of 870 and  980°C  in an
ash containing 14% CaS04, 23% calcium oxide  (CaO), and  about 2.5%  sodium
oxide (Na20).  The formation of fused calcium silicates as a result of
the decomposition of CaS04  in the presence of quartz and aluminosilicates
between 870-980°C was also  an important factor in the mechanism of
deposition.  A photomicrograph showing the sintering is presented  in Figure
8.  Table  4 gives the bulk  composition of  the deposit from the cyclone
riser.

    Most  ashes consist mainly of  aluminum  (Al) and silicon  (Si), which
generally  have good fusion  characteristics (fusion temperatures above
1650°C).   Both the Denney Farm and  Erwin soils were composed mainly of
SiOo  and AloOo.  The lack of  slagging  and  of troublesome deposits
experienced with the MIS when processing these materials demonstrates these
good  fusion  characteristics of Al  and  Si.
                                     496
-------
CONCLUSIONS

    The paper has  shown how various feed characteristics affected the MIS
performance.  Concerns stemming from these effects have been discussed and
are summarized below:

    o     Increased heating value of the feed often reduces the fuel
          requirements of the unit.  However, as heating value increases
          control of kiln temperature becomes important and feed capacity
          can be restricted.  Water injection can increase kiln capacity
          under these conditions.  The experience with the MIS also shows
          that operating conditions must be adjusted to insure rapid
          ignition of solids that have high heating value.

    o     Proper feed preparation is very important for maximizing
          throughput.  Problem materials should be sorted out and specially
          prepared prior to feeding.

    o     Rheology can affect either the feed system or the decontamination
          behavior.

    o     Increased halogen content increases mist formation and can
          increase particulate emissions.  TDS of process water can also be
          important in controlling particulate emissions.

    o    Moisture content increases fuel requirements and can create
         feeding problems for muddy feeds.

    o    Materials containing large quantities of micron-sized particles
         can foul critical instruments and decrease reliability of the
         system.

    o    Ash fusion characteristics can cause formation of slag or other
         deposits in various parts of the system.   This also decreases
         reliability.   The processing of pure trash produced slagging
         problems in the rotary kiln.

Most of this experience is directly applicable to  other mobile or
transportable incinerator systems.
                                    497
-------
                                 REFERENCES
1. Yezzi, J.J., Jr. et  al.   Results  of the  Initial Trial  Burn  of the
   EPA-ORD Mobile  Incineration Systems.   In:  Proceedings  of the 1984
   National Waste  Processing Conference,  ASME, pp. 514-534.

2. Yezzi, J.J., Jr. et  al.   The  EPA  Mobile  Incineration Systems  In:
   Proceedings of  the 1982  National  Waste Processing  Conference, ASME,  pp.
   199-212.

3. Lovell, R.J., et al.  Trial  Burn Testing  of the  EPA-ORD Mobile
   Incineration System.   EPA-600/D-84-054,  Municipal  Environmental
   Research Laboratory,  Cincinnati,  Ohio, 1984.

4  Mortensen, H. et al.   Destruction of  Dioxin-Contaminated Solids  and
   Liquids by Mobile  Incineration.   EPA  Contract 68-03-3255,  Hazardous
   Waste  Engineering  Research Laboratory, Cincinnati,  Ohio, 1987.

5. King,  G., and Stumbar, J.  Demonstration Test Report for Rotary  Kiln
   Mobile Incinerator System at  the  James Denney Farm Site, McDowell,
   Missouri.   EPA  Hazardous Waste Engineering Research Laboratory,
   Cincinnati, Ohio,  1988.

6. Brunner, C.   Incineration Systems Selection and Design. Van  Nostrand
   Reinhold Company,  New York,  1984.

7. Ho,  M.,  and Ding,  M. G.  Field Testing and Computer Modeling of  an
   Oxygen Combustion  System at the EPA Mobile Incinerator. JAPCA,  Vol.
   38,  No.  9,  September 1988.

8. Gupta, G.D.,  et al.   Operating Experiences with EPA's  Mobile
    Incineration  System,   In:  Proceedings  of the  International  Symposium
   on Incineration of Hazardous, Municipal, and Other Wastes.  American
   Flame Research  Committee, Palm Springs,  CA, 1987.

9.  Freestone,  F.J.,  et al.  Evaluation of On-site Incineration for  Cleanup
   of Dioxin-contaminated Materials.  Nuclear and  Chemical Waste
   Management,  Vol 7, pp 3-20, 1987.

10. Gupta, G.D.,  et al.   MIS Modifications Trial  Burn and  Operations
    February 1986 to September 1987  Draft Report EPA Contract 69-03-3255,
    Risk Reduction  Engineering Laboratory, Edison,  New Jersey, 1988.

11.  Bryers,  R.W.   Deposit Analysis:  Cyclone Riser/Quench  Elbow - Denney
    Farm Site,  Foster Wheeler Development Corp., Livingston,  New Jersey,
    1988.
                                    498
-------
           LONG-TERM FIELD  DEMONSTRATION OF  THE  LINDE®OXYGEN
       COMBUSTION SYSTEM INSTALLED  ON  THE EPA MOBILE  INCINERATOR

     Min-Da Ho                            Robert H. Sawyer, James P. Stumbar
     Linde Division                      F.  W.  Enviresponse,  Inc.
     Union Carbide  Industrial Gases Inc.  Edison, NJ  08837
     Tarrytown,  New York 10591

                           Joyce M. Perdek
                           U.S. EPA, Risk Reduction Engineering Laboratory
                           Edison,  NJ  08837

                                   ABSTRACT

     This  paper  summarizes the various  system performance tests and  the
long term  operating experience of  the LINDE®Oxygen Combustion System (OCS)
installed  on the EPA Mobile Incineration  System (MIS).  The LINDE OCS was
installed  on the MIS as part of a major modification program in 1987.  The
modified system  was demonstrated for  three months in 1987 (June-September).
During this period  various system performance tests were conducted.  Later,
the system resumed  operation in February  1988 to continue the incineration
of dioxin-contaminated materials from sites  in  southwestern Missouri.  Since
the implementation  of the  modifications,  over seven million pounds of
dioxin-contaminated material including  soil, lagoon sludge, plastics, trash,
grass, protective clothing, wood, etc.  have been processed.

     The Mobile  Incinerator's capacity  for treating contaminated soil more
than doubled because of the LINDE OCS and modifications to the feed system.
Test burns of the unit show destruction and removal efficiencies (DREs)
surpassing both  RCRA and TSCA standards.  Recently, the Mobile Incinerator
has been used successfully to burn brominated sludge materials with high
heating value.  Water was used to absorb excess heat release.  The LINDE OCS
and other modifications have been instrumental for maximizing the throughput
of  this brominated  sludge material.

     The LINDE OCS has shown excellent dynamic response to varying oxygen
demand.  This ability has  significantly reduced the frequency of waste feed
shutdowns due to  low oxygen and high carbon monoxide conditions resulting
from variations  in the BTU content of the waste feed.  It was also shown
that nitrogen oxides emissions from the oxygen enriched operation compare
favorably with the previous air-based operation due to the unique design of
the LINDE burner.  In addition to the oxygen burner,  a microprocessor-based
control system has been used in the OCS.  This control system enhanced the
MIS performance in processing high BTU-content waste materials.   Some common
considerations about oxygen combustion and experiences learned during this
operation will also be discussed.
                                    499
-------
                                INTRODUCTION

     The EPA Mobile Incineration System (MIS) was designed and constructed
to demonstrate high-temperature incineration of hazardous wastes[l].  The
original system consisted of a refractory-lined rotary kiln, a secondary
combustion chamber (SCO), and an air pollution control system.  These three
components are mounted on three separate heavy-duty semi-trailers.
Monitoring equipment is carried by a fourth trailer.  During 1987, the MIS
was modified to increase its capacity and reliability.  The addition of the
LINDE Oxygen Combustion System (OCS) and modifications to the feed system
were needed to increase the capacity of the MIS.  The cyclone and wet
electrostatic precipitator (WEP) were installed to increase the reliability
of the MIS.                                                          ,

     This paper summarizes the various system performance tests and  the
long term operating experience of the LINDE Oxygen Combustion System (OCS)
with the MIS.  The LINDE System was installed on the MIS as part of a major
modification program in 1987.  The modified system was demonstrated for
three months in 1987 (June-September).  During this period various system
performance tests were conducted in addition to the normal incineration
operation.  The capacity of the MIS for treating contaminated soil more than
doubled because of the addition of the oxygen system and modifications to
the feed system.  Test burns of the unit showed DREs surpassing EPA
standards.  All incinerator effluents such as kiln ash and scrubber water
passed the stringent delisting criteria established by the EPA.

     The system resumed operation in February 1988 to continue the
incineration of dioxin-contaminated materials from sites in southwestern
Missouri.  Since the implementation of the 1987 modifications, over seven
million pounds of dioxin-contaminated material including soil, lagoon
sludge, plastics, trash, grass, protective clothing, wood, etc. have been
processed.

        TECHNICAL CONSIDERATIONS IN THE USE OF OXYGEN IN INCINERATION

Oxygen Enrichment

     The use of pure oxygen (O_) or of oxygen-enriched air reduces the
volume of both the oxidant and the flue gas.  As the oxygen replaces part or
all of the combustion air, the nitrogen portion (which is 79% by volume of
air) is reduced or eliminated which results in decreased volumes of the
oxidant and flue gases per unit of soil processed.  This has the direct
benefit of reducing particulate carryover from the kiln, reducing the
pressure differentials through the system, increasing the gas residence time
for destruction, and reducing fuel consumption.   While various levels of O^
enrichment were considered, it was decided that the highest level of oxygen
enrichment be used in the kiln to attain the lowest possible gas flow, which
would minimize particulate carryover and, as a result, cause less downtime
of the system.                                             ,

     The use of the pure oxygen system produced a 60% reduction in gas
velocity through the kiln which helped to reduce the particulate carryover
                                    500
-------
problems.  When the MIS processed 4000 Ib/hr of solids at 20 wt% water with
the pure oxygen system, the gas velocity at the kiln exit was approximately
3.3 feet per second (fps) compared with 8.0 fps when it processed only 2000
Ib/hr of dry soil using the air system.(2)

Oxygen Burner Criteria

     The major considerations in the selection of an oxygen burner were the
burner's flame temperature and mixing characteristics.  The use of oxygen
normally increases flame temperatures significantly.  Previous studies have
shown that even an additional 300-400°F in flame temperature would have an
unacceptable effect on the NO  levels and would also increase the kiln shell
temperature.  This would increase the potential for kiln distortion, and
could also increase the maintenance on the seals, refractory, etc.  It could
also increase the tendency for slagging in the kiln with materials that have
low ash fusion temperatures.

     To minimize local hot spots and ensure good combustion, good mixing of
fuel, oxygen, and recirculated combustion products is needed.  High momentum
is needed to produce good mixing.  This may be obtained through either high
mass or high velocity or a combination of both.  Due to the lack of nitrogen
mass, an oxygen burner must rely primarily on high velocity jets to achieve
good mixing.

LINDE Oxygen Combustion System

     The LINDE Oxygen Combustion System, with its patented "A" Burner, was
selected based on Union Carbide's data which had substantiated low NO
levels and flame temperatures lower than other oxygen burners.  This Burner,
which uses 100% oxygen, replaced one of the conventional air burners in the
rotary kiln.  The remaining air burner has been used as a backup.  A
schematic of the "A" Burner installation on the rotary kiln is shown in
Figure 1.  A unique feature of this burner is that it produces a flame
having a temperature comparable to the conventional air burner flame
temperature.  This is accomplished by aspirating the furnace gases into the
oxidant jets prior to mixing with the fuel.  Referring to Figure 2, air
atomized fuel oil (or fuel gas) is supplied at the burner axis as a
relatively low velocity jet.  Most of the oxygen is supplied as a ring of
high velocity jets surrounding the center fuel stream.  The purpose of the
oxygen annulus around the fuel stream is to enhance flame stability.  The
aspiration of furnace gases into the oxygen jets prior to mixing with the
fuel is indicated in Figure 2 by the small arrows.  By maintaining
sufficient distance between the oxygen jets and fuel supply, enough of the
furnace gases are aspirated into the oxygen jets prior to their mixing with
the fuel so that the resulting flame temperature is reduced to values
substantially below the theoretical flame temperature.  At the point of
combustion, both conventional air burners and the "A" Burner using 100%
oxygen have an oxidant with a low concentration of oxygen.  For the air
burner, the dilutent is nitrogen while for the "A" Burner using pure oxygen,
the furnace gases make up the dilutent.  These high velocity oxygen jets
also cause vigorous recirculation of gases within the furnace which produces
a more uniform temperature distribution within the furnace.
                                    501
-------
     Flame  shape measurements were made with a suction pyrometer to confirm
 the  low  flame  temperature.   In Figure  3 is the result obtained when straight
 oxygen nozzles were used on  a natural  gas "A" Burner.  The peak temperature
 is about 2800°F which is only 800°F higher than the average furnace
 temperature.
 [2-5].
More details about the "A" Burner can be found in references
      In  addition to  the  "A" Burner,  the LINDE Oxygen Combustion System
 included flow control piping  skids and a microprocessor-based control
 system.   This control system  was  designed with  safety as a prime
 consideration.   This system is  able  to control  transient emissions  ("puffs")
 better than the  previous MIS  combustion control system because of its fast
 reaction time.

 NO Emissions
 —x	
      In  the field demonstration of the MIS  at the Denney Farm, the  NO^ level
 was continuously monitored and  recorded.  Due to the air infiltration, the
 effective O  enrichment  levels  were  40 to 50% O in the rotary kiln.  The
 mean  NO   emission levels reported in the 1987 trial burns[6] were between
 54.6  toX138.3 ppm at 15% C02  (or  0.07 to 0.18 Ib NOx/MMBTU)  (lb/MMBTU=pound
 per million BTU). This  compares  favorably  with the previous air system
 levels of 126 to 166 PPM at 11% CO2  (or 0.19 to 0.235 Ib NOx/MMBTU) which
 were  obtained in the 1985  trial burns.[1]   It should be noted, however, that
 the NO   emission level from the LINDE system was more sensitive to  the
 operating conditions than  the air system.   It is anticipated that additional
 NO reduction can be achieved by  (1) reducing air leakage into the  kiln, (2)
 using steam instead  of air as an  oil atomization flui'd for the oxygen burner
 and (3)  reducing the excess oxygen level in the rotary kiln.

                   SYSTEM  PERMITTING AT INCREASED CAPACITY

      Due to the  increase in solids feed capacity to 4000 Ib/hr and  the
•installation of  the  cyclone and WEP, operating  permit modifications for both
 the RCRA and air permits were filed  jointly with the Missouri Department of
 Natural  Resources (MDNR) and  EPA  Region VII.[7]  Since MDNR had not
 received final authorization  from EPA to regulate dioxins, EPA and  MDNR
 jointly  issued the final RCRA permit[8], and MDNR issued the revised air
 permit.[9]

      The modified permits  required compliance testing of the MIS for
 particulate emissions at the  new  feed rates.  Verification tests to show
 that  the byproduct residues passed the delisting criteria also were
 required.  However,  further Destruction and Removal Efficiency (ORE) testing
 was not  mandated by  the  regulatory agencies because the critical operating
 parameters remained  unchanged.

      All requirements for  compliance testing were met.  The  stack
 particulates during  these  tests were well below the regulatory limit of 0.08
 gr/dscf.  All effluent streams  (ash  solids  and  scrubber liquid) also met the
 delisting requirements,  and hence could be  discharged as non-hazardous
 wastes.
                                     502
-------
                         CONTAMINATED SOIL OPERATION

     The maximum contaminated soil throughput of the Mobile Incineration
System during its operation with air burners had been 2000 Ib/hr.  However,
this maximum rate was not sustainable.  For example, the "average throughput
rate of four test runs in the Spring of 1985 had been only 1478 Ib/hr of dry
soil (<10% moisture).  With the addition of LINDE Oxygen Combustion System
and other modifications, the MIS achieved a sustainable soil throughput rate
of 4000 Ib/hr even with moisture content up to 20%, as confirmed by
certified verification tests[11].

     As detailed in a complete report on subsequent TSCA/RCRA trial burn
tests conducted in the summer of 1987[6], the MIS was able to achieve the
required destruction and removal efficiencies (DREs) for all the principal
organic hazardous constituents (POHCs).  DRE for PCBs exceeded 99.9999%.
The DRE for carbon tetrachloride (CCl ), hexachloroethane (C Cl ), and
trichlorobenzene (TCB) exceeded 99.99%.

     With normal and relatively dry (about 20% moisture) soil, it was quite
easy to feed up to 4,000 Ib/hr.  Although the feed system had trouble
processing muddy soil, the kiln and SCC operated well with it.  Due to the
addition of a cyclone downstream of the kiln and the significant reduction
in the kiln combustion gas velocity, a dust carryover problem in the SCC,
experienced before the modifications!7], was greatly reduced.  The
particulate carryover was less than 15% of the feed rate for most feeds and
was easily removed by the cyclone.

     It should be noted that the current capacity limitation of the MIS is
about 4000 Ib/hr of soil, mainly due to the mechanical limitations of the
rotary kiln (which could be upgraded).  The capacity limitation is not due
to heat transfer or flue gas volume limitations.  Therefore, oxygen
enrichment of the SCC to obtain additional throughput is not currently
warranted.  The kiln limitation can be solved by using one or more larger
rotary kilns to feed common downstream equipment.  It is estimated that up
to 10 tons per hour (TPH) of soil could be handled by such an upgraded
system fired with LINDE oxygen burners in both the kiln and SCC.  Such a
throughput increase would make mobile incineration technology much more
economically attractive.

                         BROMINATED SLUDGE OPERATION

     In 1988 and 1989, the MIS processed over 2,150,000 Ibs. of
dioxin-contaminated brominated sludge from a site in southwest Missouri.
The sludge had characteristics which caused processing difficulties in the
MIS.  The sludge was a very heterogeneous material.  Analyses showed that
measured bromine content varied between 1 and 12%; moisture varied between
35% and 42%; and heating value ranged from 3200 to 4600 BTU/lb.  The sludge
required considerable testing before it could be processed as described
below.

     Previous testing!1] in 1985 had shown that this material was difficult
to process in the MIS partially due to the high moisture content and the
                                   503
-------
non-uniform heat content.  The two runs yielded throughput of less than 1000
Ib/hr.  The throughput was mainly restricted by the required gas residence
time in the SCC.  Also, 1000 ppm of organic bromide remained in the ash and
300 ppm was detected in the scrubber water.

     During the summer of 1937, a number of feasibility tests for
incinerating this sludge were run.  The program examined the effect of
blending contaminated soil, sludge and sodium sulfate (Na SO ) in different
ratios!. 11].  The three tests produced mixed results.  Mecnanzcal performance
of the unit was generally good, but chemical performance, ascertained by
analyzing the residues, ranged from poor to good.  Samples of purge water,
kiln ash, cyclone ash and separator sludges were analyzed for selected
constituents.  The analyses showed that the byproduct streams slightly
exceeded the stringent delisting criteria proposed for these residues
(Federal Register, Vol. 52, No. 171, Thursday, September 3, 1987).
Therefore, further testing was required to establish proper processing
conditions.

     Since it appeared that the unsatisfactory effluent quality was due
mainly to insufficient heating of the kiln ash, the residence time for
processing the solids was increased by decreasing the kiln rotation speed
and by adding tests using pure sludge.  The Na SO  was eliminated to permit
higher processing temperatures.  The new tests showed that straight sludge
could be processed successfully and provided the operating basis for the
commercial run which occurred between September 1988 and January 1989.

     At the beginning of the commercial run, a series of delisting tests
were conducted to demonstrate that the effluents of the incinerator were
delistable.  It was determined that oxygen enrichment, with the LINDE OCS,
should be continued for increased throughput and better controllability.
Since the sludge had substantial heating value, water was used to absorb the
excess heat.  Such a practice is more advantageous than using air injection
to absorb excess heat.  This is because water vapor contains approximately
twice the enthalpy per unit volume as nitrogen at 1800°F, so that for a
given heat load, the flue gas volume is smaller when using water spray than
when using excess air.

     Sludge throughput rates between 1600 to 1800 Ib/hr were achieved during
the delisting tests.  The kiln ash and other effluents met the delisting
criteria.

     The oxygen injection rate responded dynamically to control the kiln
oxygen level.  Instantaneous flow rates in the delisting test runs ranged
between 3000 and 12000 standard cubic feet per hour (scfh).  The average
oxygen flow rate was about 5000 scfh.  All of the excess pure oxygen was
also injected through the LINDE burner, and a small amount of diesel fuel
was also fired through the burner to stabilize the combustion.  An estimated
25000 scfh of air enters the kiln mainly due to infiltration.
                                   504
-------
          OTHER OPERATING EXPERIENCES AND BENEFITS OF THE LINDE PCS

Fuel and Cost Savings

     When treating contaminated soil, supplemental fuel is reguired to
provide the heat required to operate the rotary kiln at 1500-1600 F and the
SCC at about 2100 F since the contaminated soil does not have a sufficient
heating value to sustain self-combustion.  Even with high BTU waste
material, auxiliary fuel is still reguired in the SCC.

     Specific fuel savings of over 60% (7.2 MMBTU/ton soil) were achieved
during operation of the EPA/MIS with the LINDE system processing
contaminated soil.  This result can also be expressed as 50 MMBTU saved per
ton of oxygen used.[2]

     The economics of using oxygen to save fuel, of course, depend on the
relative cost of fuel and oxygen.  With No. 2 fuel oil costing $0.70 per
gallon (or $5.50 per million BTU) and a fuel savings of about 50 million BTU
for every ton of oxygen consumed, the breakeven oxygen cost is about $275
per ton of oxygen.  The cost of oxygen depends on methods of oxygen
generation, size of plant, and location.  For example, it ranges from about
$50 per ton of oxygen produced by a large on-site facility to about $120 per
ton for delivered liquid.

     The principal economic benefit from oxygen combustion is derived from
the very significant throughput improvement.  The large fixed portion of
daily incinerator operating costs is spread over a much larger quantity of
waste processed.  For example, for mobile/transportable incinerators a
doubled throughput can reduce the allocated incineration cost of
contaminated soil by typically $100 to $500 per ton of waste, while the cost
of oxygen reguired is typically between $10 to $50 per ton of waste
incinerated.

Flame Stability and Operational Flexibility

     During operation of the EPA/MIS, good flame stability and operational
flexibility were achieved with the LINDE OCS.  A number of system stability
tests were conducted with various feed materials at different feed rates.
By cycling liquid waste feed and water spray on and off, disturbances were
generated to test the dynamic response of the burner system.  Satisfactory
system response and flame stability were demonstrated in all the tests
conducted.[2]  It has also been demonstrated that very light vermiculite can
be processed through the system with modest particulate carryover (about
17%) from the kiln.  During the entire operation period, no significant MIS
downtime was attributed to the LINDE System.

Transient Emissions Control

     When high-BTU wastes are fed into rotary kiln incinerators in an
intermittent mode (typical of ram feed systems), the transient combustion
behavior of these materials creates unsteady releases of combustible gases
which may momentarily deplete the oxygen content of kiln gases.  These
                                    505
-------
temporary oxygen-deficient conditions could cause the release of products of
incomplete combustion (PICs).  These phenomena have raised public concern
recently and have been the subject of research projects sponsored by the
EPA.[12-14]

     In the field operation of the EPA/MIS, large quantities of high BTU
materials were burned periodically.  These materials were rain-fed into the
rotary kiln every 30 to 45 seconds.  To respond to the transient oxygen
demand resulting from the burning of these materials, a unique oxygen
feedforward-feedback control logic was designed into the LINDE System.  ,

     With the normal functioning of the oxygen control feature, the
transient upset conditions associated with the release of the combustible
gases were virtually eliminated in the operation of the MIS.  As, an example,
data taken during a delisting test are shown in Figure 4.  The oxygen level
of the gas entering the SCC (kiln O %) was controlled to be above 5% (wet),
while the O_ level at the SCC is maintained at about 7.5% (dry) except for a
short perioa when the kiln oxygen analyzer was plugged, as indicated in
Figure 4.  Note that due to safety consideration, the oxygen control system
does not allow the stoichiometric ratio at the burner to drop below a preset
minimum, even when the oxygen level may be higher than set point due to,
e.g., air leakage.  Carbon monoxide levels were consistently below 50 ppm.
Note also in Figure 4, the oxygen flow rate responded promptly to the
transient oxygen demand.  This can be attributed to the fast response of the
LINDE Oxygen System and the Thermox®WDG in-situ 0  analyzer.  In addition,
the high-momentum oxygen jets in the LINDE burner enhanced mixing in the
kiln to reduce any pockets of unburned combustibles.
     The transient puffs of the incinerator are minimized with this control
system for two reasons.  First, the oxygen feed dynamically follows the
trend of the demand, so that the oxygen level is sufficient at all times.
Second, a consistent oxygen level helps to maintain a stable and smooth
combustion, so that the flame fluctuation and rapid expansions due to
unstable combustion are minimized.  Consequently the release rate of
combustible gases is more consistent.  It is also important to maintain a
high turbulence in the combustion chamber and to maintain the entire chamber
above the auto-ignition temperature of the combustible gases.

Corrosion Resistance

     The presence of hydrogen chlorides and hydrogen bromides in the
combustion gases had caused some concern about potential corrosion problems
of the burner equipment.  As a precaution, the front end of the water-cooled
burner was coated with a layer of protective ceramic material (LC-5) by
Linde Coatings Service Co.  Hastelloy C-22 material was used for the oxygen
nozzles and for the patented oil atomizer.  Experience in the MIS was that
no noticeable corrosion of the burner occurred during more than a year of
operation.  The atomization nozzle showed signs of minor erosion and
significant coke buildup after more than one year of operation.  The coke
buildup was readily removed', and the nozzle then became fully functional
again.  In comparison, the separate waste oil nozzle made of stainless steel
                                   506
-------
had severe metal loss and had to be replaced within three months of
operation.

Slagging Problems

     Due to the varying characteristics of waste feeds in hazardous waste
incinerators, the control of temperature distribution is both challenging
and important.  Local hot spots can cause refractory damage, and more
frequently, clinker formation or slag buildup (slagging), which would
eventually require shutdown of the furnace for slag removal.

     Since the use of oxygen-enrichment was historically associated with
very high flame temperature, there was considerable concern that the use of
the LINDE OCS in a rotary kiln incinerator would create serious slagging
problems.  During the field operation at Denney Farm, there was no evidence
that the slagging tendency of the rotary kiln was aggravated by the use of
the LINDE OCS.  Experience has shown that when the system is operated with a
good understanding of the process and waste feed characteristics, the
occurrence of slagging is manageable.

     During the incineration of contaminated soil, slagging was normally not
a problem.  The heat release of the "A" burner was well controlled, and the
thermal mass of the soil absorbed the heat and dampened temperature
fluctuations.  However, the combustion of plastic materials tended to be
more challenging. [ 15]                                             .,

The principal causes of slagging formations noted were:

     (1)  Temperature excursions due to fluctuations in the feed rate of
          high BTU waste would sometimes cause rapid slagging because the
          kiln temperature rose above the fusion temperature of the ash.

     (2)  Occasional false readings of the control thermocouple due to
          abnormal air leakage through the ash gates caused slagging.  In
          this situation, the kiln was running at a temperature at least
          150 F higher than the temperature reading of the control
          thermocouple.

     (3)  The presence of a small fraction of low-melting point ash in the
          feed materials may cause the formation of eutectic solid solutions
          with low melting point.  The presence of glass or sodium sulfate,
          for example, is conducive to slag formation.

     (4)  The flame characteristics of the burner could also have a
          significant impact on the slagging problem.  It was found earlier
          that the mid-section of the kiln was somewhat hotter than the
          other areas when there was no heat load in the kiln.  This was
          possibly due to the flame shape created by the divergent angles of
          the original nozzles.  A set of nozzles generating higher momentum
          and a narrower flame, which were installed later, proved to be
          effective in producing a more uniform temperature distribution.
                                   507
-------
     Occasionally, there was a gradual slag buildup on the burner wall which
would distort the flame pattern.  In the future, retractable burner mounts
should be considered to allow easy slag removal from the burner area.   ;

                                 CONCLUSION

     The long-term reliable operation of the LINDE Oxygen Combustion System
has demonstrated the significant advantages of this technology on the MIS.
Much valuable experience has also been gained in the field application of
this technology to handle a large variety of waste streams.  It is believed
that similar advantages can be realized in the application of this
technology to other thermal treatment systems.

                               ACKNOWLEDGEMENT

     The authors would like to thank Mr. Frank J. Freestone of U.S. EPA and
Dr. G.D. Gupta of Foster Wheeler Enviresponse, Inc. for their support and
guidance during this project.
                                   503
-------
                                 REFERENCES

(1]   Mortensen,  H.,  et al,  "Destruction of Dloxin-Contaminated Solids and
     Liquids by Mobile Incineration," USEPA report,  EPA Contract
     #68-03-3255, Hazardous Waste Engineering Research Laboratory,
     Cincinnati, Ohio, April, 1987.

[2]   Ho, Min-Da and Ding, M. G., "Field Testing and Computer Modeling of an
     Oxygen Combustion System at the EPA Mobile Incinerator," JAPCA, Vol.
     38, No. 9,  September 1988.

[3]   Anderson, J. E., U. S. Patent Nos. 4,378,205 and 4,541,796, "Oxygen
     Aspirator Burner and Process for Firing a Furnace," March 29,  1983,
     September 17, 1985.

[4]   Anderson, J.E., "A Low NO , Low Temperature Oxygen-Fuel Burner,"
     Proceedings of the American Society of Metals, 1986 Symposium on
     Industrial Combustion Technologies, Chicago, Illinois, April 29, 1986.

[5]   Ho, Min~Da and Ding, M. G., "Proposed Innovative Oxygen Combustion
     System for the Incineration of Hazardous Waste," Hazardous Materials
     Management Conference & Exhibition/West, December 3-5, 1986, Long
     Beach, CA.

[6]   King, G. and J. Stumbar, "Demonstration Test Report for Rotary Kiln
     Mobile Incinerator System at the James Denney Farm Site, McDowell,
     Missouri," Enviresponse, Inc., Edison, New Jersey, January 1988.

[7]   Sherman, A., et al., "Final Permit Application USEPA Mobile
     Incineration System at the James Denney Farm Site, McDowell, Missouri,"
     USEPA, Hazardous Waste Engineering Research Laboratory, Releases
     Control Branch, Edison, New Jersey, April 3, 1987.

[8]   Brurmer, F.A., and Wagoner, D.A., "Hazardous Waste Facility Permit"
     State of Missouri, Department of Natural Resources, Jefferson City, MO,
     May 6, 1987.

[9]  Letter from F. A. Brunner and M. Nikkila to F. J. Freestone, State of
     Missouri, Department of Natural Resources, Jefferson City, MO, April
     12, 1987.

[10] Gupta, G.D., et  al, "Operating Experiences with EPA's Mobile
     Incineration System,"  Int'l Symposium on Hazardous and Municipal Waste
     Incineration, AFRC, Nov.  2-4, Palm Springs, CA.

[11] G. D. Gupta, et  al, "MIS  Modifications Trial Burns and Operations Feb.
     1986  to  Sept.  1987," EPA  Contract 68-03-3255, US EPA HWERL, Cincinnati,
     Ohio  45268.

[12] Linak, W.  P.,  J. D. Kilgroe,  J. A. McSorley, J.O.L. Wendt  and  J. E.
     Durn,  "On  the  Occurrence  of Transient Puffs in a Rotary Kiln
     Incinerator  Simulator,  Part I," JAPCA, VOL. 37, No. 1, Jan. 1987.
                                    509
-------
13]  Linak, W. P., J. D. Kilgroe, J. A. McSorley, J.O.L. Wendt and J.  E.
     Durn, "On the Occurrence of Transient Puffs in a Rotary Kiln
     Incinerator Simulator, Part II," JAPCA, VOL. 37, No. 8, Aug. 1987.

[14] Linak, W. P., et_al, "Rotary Kiln Incineration:  The Effect of Oxygen
     Enrichment on Formation of Transient Puffs During Batch Introduction of
     Hazardous Wastes," Proceedings of the 14th Annual Research Symposium on
     Land Disposal, Remedial Action, Incineration and Treatment Hazardous
     Waste, USEPA, Cincinnati, Ohio, May, 1988.

[15] Stumbar,  J.P., Sawyer, R., Gupta, G.D., Perdek, J. and Freestone, F.,
     "Effect of Feed Characteristics on the Performance of EPA's Mobile
     Incineration System," 15th Annual EPA Research Symposium,  USEPA,
     Cincinnati, Ohio, April, 1989.
                                  510
-------
  _l
  5

  cc
u.z
00

°Z
E2
   
   flC
   111
   z
   cc

   CD
                                511
-------
                        Figure 2
         SCHEMATIC SKETCH OF THE A - BURNER
                                FUEL TUBE
                                   ANNULUS OXYGEN
                                       OXYGEN
                                       NOZZLES
        OXYGEN
           JET
OXYGEN TO
  NOZZLES
                              OXYGEN
                              ENVELOPE
  FUEL
                            ^iW//* FURNACE
                            sife GASES
                                r^l
OXYGEN TO
ANNULUS
                        512
-------
    s  .
U_
o
 »\

 £



I
    CM


    O
    CM
I

8
CO
    10


    CO
                                    CO

                                    *

                                    N
                                Q.L.
                                               O
                                               N.
i  if;
LL  2 U. 3
                     O>

                     I

                     8
                     CO
                     c
                         513
-------
            20OO

  KILN TEMP. 1500
             5OO
  KILN O2%
20
15
10

 5
    KILN
 O2BURNER
FLOW(SCFM)  200
             1OO
 SCC TEMP
   SCC 02%
            2500
            15OO
              20
              1O
               0

             1OO

SCC CO (PPM)  50
    WEP
 GAS FLOW  5OOO
   (ACFM)
                   14
                                         Figure 4
                             MOBILE INCINERATOR OPERATION
                               DELISTING TEST No.4, 9/12/88
                                     XV	»v.
                                                   KILN
                                                O2 ANALYZER
                                                 PLUGGED
                                CALIBRATION
                                              RESULTED FROM
                                              KILN O2 ANALYZER
                                                 PLUGGAGE
            15
16
17
18
19
20
21
                                    HOURS OF THE DAY
                                    514
-------
   IN PLACE TREATMENT OF CONTAMINATED SOIL AT SUPERFUND SITES: A REVIEW

         M. Roulier, J. Ryan, J. Houthoofd, H. Pahren, and F. Custer
                    Risk Reduction Engineering Laboratory
                     U.S.  Environmental  Protection Agency
                           Cincinnati, Ohio  45268
                                   ABSTRACT

      The USEPA's Superfund research program is developing methods for in
place (in situ) removal of contaminants from soils and for in place treatment
of contaminated soils.  This work is motivated by the high cost for managing
large volumes of soil with low levels of contamination and because of the need
to comply with provisions of the Superfund Amendments and Reauthorization Act
(SARA) and the Resource Conservation and Recovery Act (RCRA).  This paper
summarizes available information sources and improvements in technology since
1984 when USEPA issued a summary report on in place soil treatment.  There
have been only a few instances of in place treatment based on aqueous solution
chemistry and these have involved primarily organic contaminants.
Biodegradation has been successful for some organic compounds and
stabilization/solidification is increasingly successful  for inorganics and
some organics.  In place vapor-phase removal processes such as vacuum
extraction, steam stripping, and microwave heating appear most promising for
low solubility, low boiling point organic compounds.  Improvements are needed
in methods for delivering and mixing treatment materials in soil and in
methods for recovering unreacted materials and reaction products.

                                 INTRODUCTION

      Work on in situ  (in place) treatment of contaminated soils at Superfund
sites is motivated by the need to comply with current regulations and by the
high costs of treating large volumes of contaminated soil often encountered at
sites.  The provisions of the Comprehensive Environmental Response
Compensation and Liability Act (CERCLA) as amended by SARA require the maximum
possible treatment of wastes and contaminated soils at Superfund sites and
restrict on-site containment and off-site disposal.  The provisions of RCRA as
amended by the Hazardous and Solid Waste Amendments (HSWA) establish treatment
standards and disposal limitations that further restrict the disposition and
encourage treatment of contaminated materials removed from Superfund sites.
      A number of accepted technologies (e.g. incineration,
stabilization/solidification, chemical treatment) are available or are being
tested for the broad range of Superfund wastes and soils (1,2,3,4).  These
                                     515 '
-------
technologies  are  cost-effective when  applied to concentrated wastes and highly
contaminated  soils  that  have  been  excavated.  They have not been shown to be
cost effective  for  the large  amounts  of  slightly contaminated soil that are
often encountered after  removal of wastes or near-surface soils.  Volumes of
soil remaining  are  usually  large;  the cost of moving such soils for treatment
1s excessive.   The  alternatives, for  reducing these costs, are to remove
contaminants  without moving the  soils (in situ removal) or to treat the
contaminants  in place (in situ treatment).
      The USEPA's first  report (5) in 1984 on in place treatment described a
large number  of chemical and  physical processes (e.g. oxidation, reduction,
precipitation)  that could potentially be used in situ to immobilize or
detoxify contaminants in soils.  The  authors noted that the majority of these
were conceptual or  had been tested only  in the laboratory.  Another USEPA
report  (6) published two years later  evaluated in situ (in place) methods for
stabilizing (treating) waste  deposits.   This report considered treatment
materials (reactants), methods for delivering the treatment material, and
methods for recovering the  products of the reaction.  The report noted that ".
. . the combination of injection,  reaction and recovery as a system for in
situ remediation  has scarcely been practiced and is in its infancy as an
integrated technology."
      Most of these proposed  treatments  involved aqueous solution chemistry
and it was hoped  that workable in  place  treatment technologies would be
developed along these lines.  Instead, the major in place developments since
that time have  been in biodegradation, stabilization/solidification, and
removal of contaminants.  This trend  is  likely to continue with contaminants
being increasingly  recovered  for treatment above ground rather than being
treated and left  in place.

                                 RECENT  WORK

BIOLOGICAL AND  PHYSICAL  TREATMENT

      Biodegradation (1,2,3,7,8,9,10,11,12,13,14) and stabilization/
solidification  (1,3,4,10,15,16,17,18,19,20,21) are rapidly developing
technologies that are distinguished, within the Superfund research program,  as
work areas separate from in place treatment; only a few of the many recent
references are  listed.   These technologies are being applied both in place and
in above-ground reactors and  batch plants.  Biodegradation is effective only
for organic contaminants; stabilization/solidification (S/S) processes are
most effective  for  inorganic materials but are being developed to stabilize
wastes containing organic contaminants.   Experience with biodegradation and
S/S suggests that these  technologies will achieve expanded coverage of waste
types and improved  performance.
      Both these technologies are limited by the problem of delivering
materials to subsurface  soils and achieving uniform mixing.   Two commercial  in
place mixing technologies are available  (15,18,22,23);  the cost and
performance of these are being evaluated (3,4) in the Superfund Innovative
Technology Evaluation (SITE) program.   In place mixing technologies will  also
have application for in  place vapor-phase removal  of organic compounds.
                                     516
-------
CHEMICAL TREATMENT

      There have been only a few examples of in place treatment of
contaminated soil by chemical processes.  Successful projects include
treatment of polymer waste in an impoundment (24), a spill of acrylonitrile
(25), various pesticides  (26,27,28) and arsenic compounds in groundwater  (29).
The limited amount of in  place treatment appears to be due to difficulty  in
applying treatment materials uniformly and recovering unreacted materials and
reaction products.  There is also a reluctance to apply treatment materials
that are dangerous substances (e.g. hydroxides and hypochlorites).  The
contaminants treated have generally been low-solubility organic chemicals.
Organic chemicals with high aqueous solubility or high vapor pressure are
transported out of soil by natural processes or are easily removed in water or
in the vapor phase for treatment above ground.  Most inorganic contaminants
are not treatable (cannot be truly degraded).  Stabilization/solidification
has been used for most of these but there is controversy whether S/S processes
are chemical treatments that form new low-solubility compounds or merely
occlude the contaminants  through physical processes and retard their release.

DELIVERY AND RECOVERY

      The limited success of in place treatment has led to an increase in work
on processes for adding treatment materials and distributing them in soil, and
on processes for removing contaminants for treatment above ground without
disturbing the soil (in place removal).  The Superfund research program
describes these collectively as innovative delivery and recovery processes.  A
recently completed review (30) identified 17 processes, proposed or being used
in other industries, that could be used for delivery or recovery in
remediation of Superfund  sites.   The authors noted that several  of the
technologies were commercially available in industries such as petroleum
production but none were  fully proven for use in waste site remediation.
Vapor (vacuum) extraction is being used to decontaminate soils affected by
leaking underground storage tanks and has successfully removed a variety of
volatile organic compounds from soil.

LIQUID PHASE REMOVAL

Withdrawal of contaminated groundwater for treatment (pump and treat) was the
first in place removal process used at Superfund sites.  The Superfund
research program is attempting to improve the pump and treat process through
the use of intermittent (pulse)  pumping for minimizing the amount of water
removed per unit of contaminant and for avoiding hydrogeologic "dead zones"
observed in many pump and treat operations.   The engineering component of this
research is being conducted by the USEPA's Risk Reduction Engineering
Laboratory (RREL) in Cincinnati, OH (31) and the evaluation/interpretation
component by the USEPA's  Robert S.  Kerr Environmental  Research Laboratory in
Ada, OK (32).   These efforts are likely to improve removal of some inorganics
and soluble organic compounds but not most inorganic cations or low-solubility
organic compounds that are highly partitioned onto the solid phase in soil-
water systems.
                                     517
-------
      The  aqueous  liquid  phase  with  chemicals  added  has  also  been  used  as a
displacing solution  to  extract  contaminants  from  soils;  this,  is termed  soil
flushing.   A  pilot test for removal  of  organics (33) was less successful than
full  scale operations for removal  of cadmium from contaminated soils  (34) and
mixed metals  and volatile fatty acids from a sludge  lagoon  (35).   Except in
confined or well-controlled hydrogeologic settings,  soil  flushing  will  be
limited for the same reasons as chemical treatment - difficulty in applying
and recovering materials  and reluctance to bear responsibility for adding
dangerous  chemicals  to  soil.  Several commercial  processes  are available and
are being  tested under  the SITE program (3,4).

LIQUID CONTROL

      Two  control  procedures are being  tested  for use with  liquid  phase
removal.   The movement  of cations  and soil water  in  response  to an applied
direct current gradient,  called electro-kinetics  or  electro-osmosis,  has been
adapted from  applications in the mining industry  and is  being considered as a
means for  directing  the movement of  water and  contaminants  during
removal/treatment  in low-permeability soils  (2,36).  The RREL sponsored one
field test (37) and  the SITE Emerging Technologies Program  is testing the
process when  used  in conjunction with an acoustic field  (4).   In place
freezing techniques  used  in the construction industry are being considered for
temporarily making soil impermeable  or  for concentrating  contaminants in front
of a  slowly moving freezing front  (38,39).   Both  of these control  procedures
are in the early stages of testing to determine their usefulness at Superfund
sites.

GAS PHASE  REMOVAL

      For  many low-solubility organic compounds,  equilibrium  thermodynamics
indicates  that the aqueous  liquid phase will contain greater  amounts of
organic contaminants than  an equal volume of the  gas phase.   However, the
rates of transfer  from  the  solid phase  to the  liquid phase  are often slow
relative to rates  of transfer into the  gas phase.   Additionally,  it is  easier,
because of relative  densities and viscosities, to move large  volumes of gas
through the subsurface  than  it  is to  move equal volumes of  liquid.  This
combination of greater  convective flow  and more rapid phase transfer makes
gas-phase  removal  an efficient  process  for many organic compounds,
particularly  in vadose  (partially saturated) zones where  it is difficult to
remove water.
      Gas  phase removal,  called  "in  situ vacuum extraction, air stripping,
volatilization, vapor extraction", or "forced air venting," is a commercially
available  process  that  has  been  used most extensively in decontaminating soils
affected by leaking  underground  storage tanks  (21,40,41,42,43,44); these often
contain fuels and  other volatile substances.  A combination of air injection
and vacuum  venting wells  are used to  induce air flow through the contaminated
soil.   Performance data for this technology have been summarized recently (45)
and detailed cost  analyses were  conducted for several typical  systems (46).
Several  gas phase  removal  processes are being tested under the SITE program
(3,4,22,23)  Although gas phase  removal  is used almost exclusively in vadose
zone soils, there  is at least one site  in Europe  (10) and one in  the U.S.  (47)
                                     518
-------
where it has been used to simultaneously remove organics from groundwater and
soil; air was injected in the saturated zone and the organics stripped from
the water were captured by the vacuum system in the vadose zone.

SOIL HEATING

      Simple gas phase removal is most effective for organic compounds with
relatively low boiling points.  Several methods are being tested for heating
contaminated soil to increase vapor pressure and allow gas phase removal of
higher boiling point compounds.  Radio frequency (microwave) heating has been
used successfully in a U.S. Air Force pilot-scale field test for removal of
fuel oil (48,49); a full scale test is being planned.  The RREL is sponsoring
a laboratory test of radio frequency heating for removing the higher boiling
compounds creosote and pentachlorophenol from soils contaminated with wood
treating fluids.  Steam injection, developed for petroleum recovery (30), is
being adapted for use with vacuum extraction at hazardous waste sites (50).
It has been tested at a small scale in the field for solvent removal (51) and
will be examined at a larger scale in a RREL field study that is being
initiated.  A combination of steam and hot air are used to heat soil for
several of the gas phase removal technologies being tested under the SITE
program (3,4,22,23,).  Another soil heating technology melts (vitrifies) the
soil (52,53) to destroy most of the organic compounds and capture the
remainder with a vacuum hood on the surface.  This technology had been
proposed for use at one Superfund site and is being tested under the SITE
program (3,4) for treatment efficiency and for ability to control the loss of
volatiles to surrounding soils.

                                   SUMMARY

      Since EPA's first report on in place treatment, the technologies for in
place removal, particularly in the gas phase, have developed much more rapidly
than those for in place chemical treatment.  Contaminants will be increasingly
recovered for treatment above ground rather than being left in place after
treatment.  Biodegradation and stabilization/solidification can successfully
treat some organic and inorganic contaminants in place and these two
technologies will continue to improve their ability in this regard.  Further
work on development of methods for delivering and mixing treatment materials
in soil and for recovering unreacted materials and reaction products would
expedite improvements in all in place treatment and removal processes.

                                  REFERENCES

1.    U.S. Environmental Protection Agency.  1985.  Handbook:  Remedial  Action
      at Waste Disposal Sites (Revised)  EPA/625/6-85/006.  USEPA Hazardous
      Waste Engineering Research Laboratory, Cincinnati, OH and USEPA Office
      of Emergency and Remedial Response, Washington, DC.

2.    Sanning, D. E., and R. F. Lewis.  January 1987.  Technologies for In
      Situ Treatment of Hazardous Wastes.  USEPA  Hazardous Waste Engineering
      Research Laboratory, Cincinnati OH.  EPA/600/D-87/014, NTIS-PB87-146007.
      24p.
                                     519
-------
3.    U.S Environmental Protection Agency.  1988.  Technology Screening Guide
      for Treatment of CERCLA Soils and Sludges, EPA/540/2-88/004.  USEPA
      Office of Emergency and Remedial Response, Washington. D.C.  127p.

4.    U.S. Environmental Protection Agency.  1988.  The Superfund  Innovative
      Technology Evaluation Program: Technology Profiles EPA/540/5-88/003.
      USEPA Risk Reduction Engineering Laboratory, Cincinnati, OH.  83p.

5.    U.S. Environmental Protection Agency.  1984.  Review of In place
      Treatment Techniques for Contaminated Surface Soils (Volume  1:
      Technical Evaluation)  EPA/540/2-84/003a.  USEPA Municipal Environmental
      Research Laboratory, Cincinnati, OH and USEPA Office of Solid Waste and
      Emergency Response, Washington, D.C.   163p.

6.    U.S. Environmental Protection Agency.  1986.  Systems to Accelerate In
      Situ Stabilization Of Waste Deposits, EPA/540/2-86/002.  USEPA Hazardous
      Waste Engineering Research Laboratory, Cincinnati, OH.  286p.

7.    Lee, M. D., J. M. Thomas, R. C. Borden, P. B. Bedient, J. T. Wilson, and
      C. H. Ward.  1988.  Biorestoration of Aquifers Contaminated  with Organic
      Compounds.  CRC Critical Reviews in Environmental Control, Vol. 18,
      Issue 1, pp. 29-89.

8.    Soczo, E. R., and J. J .M. Staps.  1988.  Review of Biological Soil
      Treatment Techniques In The Netherlands.  IN: . Contaminated  Soil '88
      Volume 1, Proceedings of the Second International TNO/BMFT Conference on
      Contaminated Soil, 11-15 April 1988, Hamburg, Federal Republic of
      Germany, K. Wolf, J. van den Brink, and F. J. Colon, eds., Kluwer
      Academic Publishers, The Netherlands, pp. 663-670.

9.    Mischgofsky, F. H. and R. Kabos.  1988  General Survey of Site Clearance
      Techniques: Trend Towards In Situ Treatment.  IN:  Contaminated Soil '88
      Volume 1, Proceedings of the Second International TNO/BMFT Conference on
      Contaminated Soil, April 11-15, 1988, Hamburg, Federal Republic of
      Germany, K. Wolf, J. van den Brink, and F. J. Colon, eds., Kluwer
      Academic Publishers, The Netherlands, pp. 523-533.

10.   U.S. Environmental Protection Agency.  1988.  Assessment of
      International Technologies for Superfund Applications:  Technology
      Review and Trip Report Results, EPA/540/2-88/003.  USEPA Office of Solid
      Waste and Emergency Response, Washington, D.C.  50p.

11.   Wetzel, R. S., D. J. Sarno, C. M. Durst, B. C. Vickers, P. A. Spooner,
      J. R. Payne, W. D. Ellis, M. S. Floyd, and Z. A. Saleem.  1986.  In Situ
      Biological Degradation Test at Kelly Air Force Base, Volume  I:  Site
      Characterization, Laboratory Studies and Treatment System Design and
      Installation.  (ESL-TR-85-52-I)  U.S. Air Force Engineering  & Services
      Center, Tyndall Air Force Base, FL  32403  291p.
                                     520
-------
12.
13.
14.
15.
16.
17.
18.
19.
20.
Wetzel, R. S., C. M. Durst, D. H. Davidson, and D. J. Sarno.  1987.  In
Situ Biological Treatment Test at Kelly Air Force Base, Volume II:
Field Test Results and Cost Model.  (ESL-TR-85-52-II)  U.S. Air Force
Engineering & Services Center, Tyndall Air Force Base, FL  32403  170p.

Wetzel, R. S., C. M. Durst, D. H. Davidson, and D. J. Sarno.  1987.  In
Situ Biological Treatment Test at Kelly Air Force Base, Volume III:
Appendices.  (ESL-TR-85-52-III)  U.S. Air Force Engineering & Services
Center, Tyndall Air Force Base, FL  32403  215p.

Olfenbuttel, R. F., E. Heyse, D. C. Downey, and T. L. Stoddart.  1988.
Lessons Learned: A Basis For Future Success.  IN: Contaminated Soil '88
Volume 1, Proceedings of the Second International TNO/BMFT Conference on
Contaminated Soil, April 11-15, 1988, Hamburg, Federal Republic of
Germany, K. Wolf, J. van den Brink, and F. J. Colon, eds., Kluwer
Academic Publishers, The Netherlands, pp. 471-479.
Ghassemi, M.  1988. Innovative In Situ Treatment Technologies for
Cleanup of Contaminated Sites.  Journal of Hazardous Materials 17:
206.
                                                                         189-
Puglionesi, P. S., J. Kesari, M. H. Corbin, and E. B. Mangel and.  1987.
Heavy Metals-Contaminated Soils Treatment.  IN:  Superfund '87,
Proceedings of the 8th National Conference.  The Hazardous Materials
Control Research Institute, Silver Spring, MD.  pp. 380-384.

Smith, R. L., D. T. Musser and T. J. DeGrood.  1985.  In Situ
Solidification/Fixation of Industrial Wastes.  IN:  Management of
Uncontrolled Hazardous Wastes Sites.  The Hazardous Materials Control
Research Institute, Silver Spring, MD. pp. 231-233.

Stinson, M. K., and S. Sawyer.  1988.  In Situ Treatment of PCB-
Contaminated Soil.  IN:  Superfund '88, Proceedings of the 9th National
Conference.  The Hazardous Materials Control Research Institute, Silver
Spring, MD.  pp. 504-507.

Wiles, C. C. and H. K Howard.  1988.  U.S. EPA Research in
Solidification/Stabilization of Waste Materials.  IN: Land Disposal,
Remedial Action, Incineration and Treatment of Hazardous Waste -
Proceedings of the Fourteenth Annual Research Symposium.  EPA/600/9-
88/021.  USEPA Risk Reduction Engineering Laboratory, Cincinnati, OH.
pp. 126-135.

U.S. Environmental Protection Agency.  1986.  Handbook for
Stabilization/Solidification of Hazardous Wastes.  EPA/540/2-86/001
USEPA Hazardous Waste Engineering Research Laboratory, Cincinnati, OH.
156 p.
                                       521
-------
21.   White, D. C., J. R. Dunckel and T. D. Van Epp.  1986.  Applying
      Alternative Technologies at Superfund Sites.  IN:  Management of
      Uncontrolled Hazardous Wastes Sites.  The Hazardous Materials Control
      Research Institute, Silver Spring, MD.  pp. 361-364.

22.   La Mori, P. N.  1987.  In Situ Treatment Process for Removal of Volatile
      Hydrocarbons From Soils:  Results of Prototype Test.  IN:  Proceedings,
      Second International Conference on New Frontiers for Hazardous Waste
      Management.  USEPA Hazardous Waste Engineering Research Laboratory,
      Cincinnati, OH  pp. 503-509.

23.   Van Tassel, R. and P. N. LaMori.  1987.  Innovative In Situ Soil
      Decontamination System.  IN:  Superfund '87, Proceedings of the 8th
      National Conference.  The Hazardous Materials Control Research
      Institute, Silver Spring, MD.  pp. 396-402.

24.   Bort, M. R.  1985.  In Situ Stabilization of Viscoelastic Polymer Waste.
      IN:  Management of Uncontrolled Hazardous Wastes Sites.  The Hazardous
      Materials Control Research Institute, Silver Spring, MD.  pp. 152-156.

25.   Harsh, K. M.  1978.  In Situ Neutralization of an Acrylonitrile Spill.
      IN:  Proceeding of 1978 Conf. on Control of Haz. Materials Spills, April
      11-13, Miami, FL, Information Transfer Inc., Rockville, Maryland pp.
      187-189.

26.   King, J., T. Tinto, and M. Ridosh.  1985.  In Situ Treatment of
      Pesticide Contaminated Soils.  IN:  Management of Uncontrolled Hazardous
      Wastes Sites.  The Hazardous Materials Control Institute, Silver Spring,
      MD.  pp. 243-248.

27.   Paulson, D. L., R. Honeycutt, Jr., J. Lebaron, and V. Seim.  1984.
      Degradation of High Concentration of Diazinon in Soil by Parathion
      Hydrolase, in 1984 Hazardous Material Spills Conference Proc.-
      Prevention, Behavior, Control, and Cleanup of Spills and Waste Sites,
      Ludwigson, J., ED., Government Institutes, Inc., Rockville, MD.  pp. 92-
      97.

28.   Ryckman, M. D.  1984.  Detoxification of Soils,  Water and Burn Residues
      from a Major Agricultural Chemical Warehouse Fire.  IN: Proceedings of
      the 5th National Conference on Management of Uncontrolled Hazardous
      Waste Sites.  Hazardous Materials Control Research Institute, Silver
      Spring, MD.  pp. 420-426.

29.   Stief, K. 1984.  Remedial Action for Groundwater Protection Case Studies
      Within the Federal Republic of Germany.  IN: Proceedings of the 5th
      National Conference on Management of Uncontrolled Hazardous Waste Sites.
      Hazardous Materials Control Research Institute,  Silver Spring, MD.  pp.
      565-568.   '
                                     522
-------
30.   Murdoch, L., B. Patterson, G. Losonsky, and W. Harrar.  1989.  A Review
      of Innovative Technologies of Delivery or Recovery for the Remediation
      of Hazardous Waste Sites.  Final report to the USEPA Risk Reduction
      Engineering Laboratory for work assignment 1-11, Contract 68-03-3379.
      Center Hill Research Facility, Cincinnati, OH  45221.  96p.

31.   Beljin, M. S. and J. F. Keeley.  1989.  In Situ Delivery and Recovery:
      Pulse (Cyclic) Pumping, Phase I.  Contract 68-03-4038 with the USEPA
      Risk Reduction Engineering Laboratory, Cincinnati, OH., Work Assignment
      2-WM32.0, University of Cincinnati.

32.   Keeley, J. F.  1988.  Remedial Performance Evaluation for Pump and
      Treat.  Cooperative Agreement CR812808 between the USEPA Robert S. Kerr
      Environmental Research Laboratory, Ada, OK. and the National Center for
      Groundwater Research, Rice University.

33.   Nash, J. H.  1988.  Project Summary:  Field Studies of In Situ Soil
      Washing.  EPA/600/S2-87/110, February 1988.  USEPA Hazardous Waste
      Engineering Research Laboratory, Cincinnati, OH.

34.   Urlings, L. G. C. M., V. P. Ackerman, J. C. V. Woundenberg, P.P. v.d.
      Pijl, and J.J. Gaastra.  1988  In Situ Cadmium Removal - Full-Scale
      Remedial Action of Contaminated Soil.  IN:  Contaminated Soil '88 Volume
      1, Proceedings of the Second International TNO/BMFT Conference on
      Contaminated Soil, 11-15 April 1988, Hamburg, Federal Republic of
      Germany, K. Wolf, J. van den Brink, F. J. Colon, eds., Kluser Academic
      Publishers, The Netherlands, pp. 911-920.

35.   Legiec, I. A., and D. S. Kosson.  1988.  In Situ Extraction of
      Industrial Sludges.  Environmental Progress 7(4):270-278.

36.   Herrmann, J. G.  1989.  Status Report on Electro-Kinetics Technology.
      January  23, 1989 memorandum from J. G. Herrmann, USEPA Risk Reduction
      Engineering Laboratory, Cincinnati, Ohio to J. Kingscott, USEPA Office
      of Solid Waste and Emergency Response, Washington, D.C.  3p.

37.   Horng, J. and S. Banerjee.  1987.  Evaluating Electro-Kinetics as a
      Remedial Action Technique.  IN: Proceedings, Second International
      Conference on New Frontiers for Hazardous Waste Management  EPA/600/9-
      87/018F.  USEPA Hazardous Waste Engineering Research Laboratory,
      Cincinnati, OH.  pp. 65-77.

38.   Iskandar, I. K. and J. M. Houthoofd.  1985.  Effect of Freezing on the
      Level of Contaminants in Uncontrolled Hazardous Waste Sites.  Part 1.
      Literature Review and Concepts.  IN:  Proceedings, Eleventh Annual
      Research Symposium (EPA/600/9-85/013) Cincinnati, OH, April 29 - May 1,
      1985.  pp. 122-129.
                                     523
-------
39.   Iskandar, I. K., and T.  F. Jenkins   1985.  Potential Use of Artificial
      Ground Freezing For Contaminant  Immobilization.   IN:   Proceedings,
      International Conference on New  Frontiers for Hazardous Wastes
      Management, EPA/600/9-85/025.  USEPA Hazardous Wastes  Engineering
      Research Laboratory, Cincinnati, OH.  pp. 128-137.

40.   Anastos, G. J. and P. J. Marks,  M. H. Corbin and  M. F. Coia.  1985   In
      Situ Air Stripping of Soils Pilot Study, U.S. Army Toxic and Hazardous
      Materials Agency, Report AMXTH-TE-TR-85026, Aberdeen Proving Ground
      (Edgewood Area), MD.  92p.

41.   Electric Power Research  Institute.   1987.  Remedial Technologies for
      Leaking Underground Storage Tanks (EPRI CS-5261).  Electric Power
      Research Institute, Palo Alto, CA.   288p.

42.   U.S. Environmental Protection Agency.  1987.  Underground Storage Tank
      Corrective Action Technologies   EPA/625/6-87/015  USEPA Office of Solid
      Waste and Emergency Response, Washington, D.C.  203p.

43.   Glynn, W. and C. Fan.  1988.  Underground Storage Tank Corrective
      Action: Application and  Field Evaluation of Vacuum Extraction
      Technology.  IN: Land Disposal,  Remedial Action,  Incineration and
      Treatment of Hazardous Waste - Proceedings of the Fourteenth Annual
      Research Symposium.  EPA/600/9-88/021.  USEPA Risk Reduction Engineering
      Laboratory, Cincinnati,  OH.  pp. 12-24.

44.   U.S. Environmental Protection Agency.  1988.  Cleanup of Releases from
      Petroleum USTs: Selected Technologies (EPA/530/UST-88/001)  USEPA Office
      of Underground Storage Tanks, Washington, D.C.

45.   Hutzler, N. J., B. E. Murphy and J.  S. Gierke.  1989.  State of
      Technology Review: Soil  Vapor Extraction Systems.  Interim Final Report
      to the USEPA Risk Reduction Engineering Laboratory for Cooperative
      Agreement CR-814319, Michigan Technological University, Houghton,
      Michigan  49931.  94p.

46.   Metzer, N., M. Corbin, and S. Cullinan.  1987.  In Situ Volatilization
      (ISV) Remedial System Cost Analysis.  (AMXTH-TE-CR-87123) U.S. Army
      Toxic and Hazardous Materials Agency, Edgewood Area, Aberdeen Proving
      Ground, MD 21010.  36p.

47.   Webster, K. T.  1988.  Aquadetox/VES Integrated System.  AWD
      Technologies, Fairfield, NJ.  (personal communication)

48.   Dev, H. P.  Condorelli, J. Bridges, C. Rogers, and D. Downey.  1986.  In
      Situ Radio Frequency Heating Process for Decontamination of Soil.  IN:
      Solving Hazardous Waste  Problems Learning from Dioxins.  Division of
      Environmental Chemistry at the 191 Meeting of the American Chemical
      Society, New York, NY. April 13-18.   pp. 329-339.
                                     524
-------
49.   Dev, H., G. C. Sresty, J. E. Bridges, and D. Downey.   1988.  Field Test
      of the Radio Frequency In Situ Soil Decontamination Process.   IN:
      Superfund '88, Proceedings of the 9th National Conference.  The
      Hazardous Materials Control Research Institute, Silver Spring, MD.  pp.
      498-502.

50.   Lord, A. E., Jr., R. M. Koerner and V. P. Murphy.  1988.  Laboratory
      Studies of Vacuum-Assisted Steam Stripping of Organic Contaminants from
      Soil.  IN:  Land Disposal, Remedial Action, Incineration and Treatment
      of Hazardous Waste, Proceedings of the Fourteenth Annual Research
      Symposium, EPA/600/9-88/021.  USEPA Risk Reduction Engineering
      Laboratory, Cincinnati, OH.  pp.65-92.

51.   Udell, K. S.  1988.  Results of a Pilot Study of Vacuum Extraction and
      Steam Injection at Solvent Service, Inc., San Jose, California.
      University of California, Berkeley, CA. (personal communication).

52.   FitzPatrick, V. F., C. L. Timmerman and J. L. Buelt.  1987.  In Situ
      Vitrification - An Innovative Thermal Treatment Technology.  IN:
      Proceedings, Second International Conference on New Frontiers for
      Hazardous Waste Engineering Research Laboratory, Cincinnati, OH  pp.
      305-322.

53.   FitzPatrick, V. F.  1988  In Situ Vitrification an Innovative Melting
      Technology for the Remediation of Contaminated Soil.  IN: Contaminated
      Soil '88 Volume 1, Proceedings of the Second International  TNO/BMFT
      Conference on Contaminated Soil, April  11-15, 1988, Hamburg, Federal
      Republic of Germany, K. Wolf,  J. van den Brink, and F. J. Colon, eds.,
      Kluwer Academic Publishers,  The Netherlands,  pp. 857-859.
                                    525
-------
                         RREL EXPERT SYSTEMS PROJECT;
               DEVELOPING TOOLS FOR HAZARDOUS HASTE  MANAGEMENT
                by:   Jay E. Clements and Daniel G. Greathouse
                                   ABSTRACT

This paper describes  the ongoing work of the  RREL  Expert Systems Development
Project.  Due  to  recent increases in the number  of environmental regulations
and the level  of  compliance  required,  EPA regulatory personnel  must monitor a
greater number of sites and  monitor these sites with  more  scrutiny than ever
before.   In  an attempt to reduce this workload and  improve the response time
of EPA Regional Offices  and  authorized  States, the  Risk Reduction Engineering
Laboratory  of  EPA's   Office   of  Research   and   Development  is  developing
microcomputer-based  decision   support   tools   (expert  systems)   to  address
commonly encountered waste management issues.

The  authors  explore  the  basic  concepts  of  expert  system technology,  the
potential  advantages  of  applying  this  technology,  and  several  specific
applications in the area of hazardous waste management.
                                   526
-------
 INTRODUCTION

 Protection  of human health and the environment  from  the risks of exposure to
 hazardous waste  involves consideration of numerous complex issues.   Decision
 makers  of  the U.S. Environmental  Protection  Agency,  bound by Congressionally
 enacted  regulations,   must oversee  the  handling  and  disposal   of  hazardous
 substances  on a  daily  basis.    In  addition  to  the  legal  implications,  EPA
 project  managers   must  be  aware  of  a  myriad  of  scientific,  technical,
 statistical,  and  economic factors  which ultimately  influence  the  decision
 making  process.   With the  preservation  of our communities  and  the  health of
 their people  at  stake,  these  decisions  are important  public issues.   Yet with
 many EPA regional  offices  characterized by high workloads and tight deadlines,
 industrial  hazardous waste management programs are in need of tools to improve
 productivity  and the quality of decision making.

 Recognizing this  problem,  the Risk Reduction Engineering  Laboratory  of EPA's
 Office of Research and  Development embarked several years ago on an effort to
 develop computer  based utilities  to  assist  regional  decision makers.   These
 utilities   will    allow   those   charged   with   mission   critical   decision
 responsibilities   to   become   more  productive,   better  informed,   and  more
 consistent  in the  execution of their duties.  In some cases, these systems can
 also be used  as training devices  for  new  or inexperienced employees.   In this
 way the Agency can use  computers,  which  are becoming  increasingly  common in
 the workplace, to  combat huge workloads.  The technology used to develop these
 decision aides is  commonly known  as expert systems.

 DESCRIPTION OF EXPERT SYSTEMS

 Expert systems are computer programs  (software)  designed to provide  advice in
 a narrowly  defined domain.  The  expert  system is an automated  process  which
 incorporates  the judgement, experience, rules of thumb,  and intuition  used by
 a human  specialist to  emulate that  specialist's  problem  solving  abilities.
 The  objective  of  these  systems  is  to  furnish  the  same  recommendations
 concerning particular decision scenarios as would an authority in the field.

 Expert  systems  are one  product  of artificial intelligence  research  programs
which have  been  ongoing  for thirty years in  university  and computer  industry
 labs.   Shortly after  the  invention  of  the modern digital  computer,  it  was
 realized that such  a machine was capable not only of mathematical calculations
 but  also  of  executing  a  series  of  logical  calculations  which simulates  a
 subset of the human thought process.  Although sound in  theory, development of
 such systems  proved to  be an  overwhelming  task.   Early  advisory  programs,
 completed  in  the  mid  to  late 1970's,  required scores  of skilled  software
 professionals  to  develop  and  expensive  dedicated  computer hardware to  run.
These restrictions  placed  expert  systems  technology out of  the  reach  of most
commercial   enterprises.  Therefore computers  have traditionally  been  used in
 "number crunching"  type applications; accounting, statistics, and mathematical
modelling for example.
                                   527
-------
Two recent  advances  in the computer field  have  paved the way  for  the expert
system's transition from research to application.  First, knowledge processing
software which enables  complex heuristic models to be coded  quickly  and with
minimal  low  level   symbolic   programming   requirements  has  become  widely
available.   Second,  breakthroughs in microelectronics design and fabrication
have increased the computing power to  price ratio of digital  machinery beyond
even the most optimistic  predictions of a decade  ago.   Consequently, serious
expert system applications can now  be  fielded economically  enough to render
the potential of swift return of the development costs.  Additionally, many of
these systems can run on computer hardware that is currently  in place for more
common engineering and data processing tasks.   Finally,  the  proliferation of
personal  computer  systems  has provided more workers than   ever before with
access to computational resources.

Knowledge based systems have several  characteristics which differentiate them
from conventional  software.   Whereas  conventional computer  programs rely on
numeric  algorithms  as the  basis  for  their operation,  knowledge  systems are
characterized by an emphasis on symbolic processing,  logical  inferencing, and
pattern matching.  Most knowledge based systems contain facilities which allow
the user  to obtain  an explanation concerning why  a  question  was asked and/or
to obtain  clarification for a  particular  question if the user requires more
information.   In  addition, the  user  is  often  supplied  with  the   reasoning
employed in reaching  a conclusion.  This allows the user to follow the logical
path from input data to  conclusion  when a justification for  the  results is
desired.    In  short,   expert  systems  are   interactive;  they  do  not   run  and
calculate so much as  they aid thought and offer advice.

Expert systems  have  their roots  in  traditional  data base management systems.
The standard data base however,  simply presents  relevant information to the
user and often only at the  expense of learning a cryptic query language.  The
expert system automates the process  of, not merely retrieving the appropriate
information,  but  also  deducing the  implications  of this information  in the
current  context.    It is  this application  of  information that  is  generally
referred to  as intelligence.

The substance of an expert  system is found  in its  knowledge base.  This  is the
collection  of facts and suppositions which when logically evaluated attempt to
reproduce the results of human  expertise.  This evaluation process is known as
inference  and software which performs  this  function  is  known as an  inference
engine.   Not long ago development of the  inference  engine was the first  step
in expert system development.   Today programs known  as expert system  shells
provide  completed  inferencing mechanisms  ready  to  accept  problem   specific
knowledge.   Programming  skills  are  still   necessary  for the  development of
expert systems  however these  skills  revolve around collection  and structuring
of  knowledge concerning  the  problem  domain.   These are  for  the most  part
unrelated   to   the   type   of  programming  expertise  necessary  to  build   a
generalized inference engine.
                                   528
-------
The  knowledge  base  is  really  just  a  sophisticated database  of  conditional
statements  or  rules.    For  example,  a  soil  mechanics  expert  system might
contain the following  rule:

IF the universal  soil  classification  is OL
THEN the soil type is  organic clay.

Statements following the IF are  known as the premise.  The premise  may  contain
a  combination of conjunctions  (two or more  statements connected  by the AND
operator)  or  disjunctions   (two  or  more  statements  connected  by  the  OR
operator).   In this way,  complex,  multi-conditional rules  can be developed.
The premise as  a whole evaluates to either true  or  false.   If the premise is
found to be true  then  the  statements following the THEN (the conclusion of the
rule) are executed.  When  this happens the rule is said to have fired.  In the
simple  example  above,  if it   is  known  that  the   variable   containing  the
universal soil  classification contains  the value OL  then  a  variable known as
soil type  is  given the  value "organic  clay".   If another rule is added, the
system can make a logical  deduction.

IF the soil type  is organic clay
THEN the soil is  not suitable for use in a liner system
BECAUSE organic  clays  may form  pores  and  thus increase  in  permeability with
time.

This rule makes a deduction about the  suitability of a  specific soil  type for
use in a  soil liner system.   If the goal  of  the system were  to make  such a
recommendation, this  rule would be  targeted  since the inference engine  sees
that  the conclusion   of  this  rule  can  provide  us with  a  recommendation.
However,  before  the  recommendation  is presented,  it must be proven  that the
premise  of  this  rule  is  true;  therefore a  determination of  the soil  type
becomes the new temporary  goal of the  system.  The knowledge  base is searched
for any fact  or rule which gives a  value to  the  soil  type variable.  Since no
facts concerning  soil  type have been recorded,  our first rule  is targeted.
This rule in  turn requires information on the universal  soil  classification.
Again the knowledge base is searched.  This time, neither a fact nor a rule is
located.   Usually when this occurs,  the user  is  asked to  supply the necessary
information although other options may be available.   If the user is asked and
indicates that  the  universal soil   code  is  OL  then  both rules fire and  a
recommendation  is formed.    The text  following  the  BECAUSE can be used  to
justify the recommendation if the user so requests.  If a soil  type other than
OL  is  specified  the  first  rule fails  and  the  knowledge  base   is  searched
further for rules which might lead  to  a value  for  soil  type.    This process
continues until   a  recommendation   is  found  or  until  all  possibilities  are
exhausted.

This  example   is  admittedly   trivial  however  it  illustrates  an  inference
strategy known as backward chaining.  Backward chaining  is only one of several
inference strategies  which  may be employed  in  an  expert  system.    It  is
represented here  because it is  the  most commonly used  technique  in knowledge
                                   529
-------
base construction.  It should also be noted that the rules here are written in
English  and therefore  would not  currently  be meaningful  to any  inference
engine other  than the human  mind.   Each  expert  system shell has  a specific
syntax.  This  syntax  defines the format that rules  and variables must assume
if the computer is to interpret them properly.

The process of collecting the rules  and  placing them in the shell is known as
knowledge  engineering.    In  general,  knowledge  engineering   is  the  most
important part of any expert system development endeavor.   Regardless of the
attractiveness and  features  in  a system,  if the program  does  not accurately
emulate the decision  process of the experts,  it  will  not  be of  value to the
end users.  The  knowledge engineering component  differentiates  expert system
development from conventional software engineering in several ways.  The first
difference is problem selection.  Not all problems are amenable to solution by
an expert system.   In order  for  an  expert  system  to be feasible there must be
recognized experts  in  the field that the  knowledge  engineer can  access.   The
problem  should not be  too  simple  and  it's solution  must represent  a  high
enough payoff  to  justify  the time and money  spent  developing  the system.   On
the other  hand,  the problem must not  be too complex or  the development  task
will quickly  become overwhelming.   In addition, a  successful  solution to the
problem  under consideration  must  be based upon  domain  specific  intellect
rather than common  sense or physical ability.

Once a problem is selected there is the matter  of shell selection.  Each shell
has a  unique  user  interface  and a  unique  knowledge representation structure.
A shell  should be selected which simplifies the knowledge engineering task as
much as  possible since this  is  a  very expensive aspect  of development.   The
user interface provided by the  shell  is  also of concern.   If users are unable
to efficiently communicate with the system they will be reluctant to use it at
all.

Finally,   there   are  the   concepts  of  rapid  prototyping   and  iterative
development.   Since outside  parties (the domain  experts)  are  involved in the
development  process  every  bit  as  much   as  the   system  designers,  it  is
imperative  that  close contact  be  maintained between  experts  and developers.
Rapid  prototyping  dictates   that  a  small,   naive  system  based  upon common
knowledge  about  the problem  be  developed  before any  experts  are approached.
This   information   can  be   collected  from  trade   magazines  and  reference
documents.   By demonstrating the  rapid  prototype,  the experts  are shown the
form and objectives  of the  system  at  the first interview.  The  experts  will
undoubtedly find many flaws  in  the prototype.   These  insights  are the basis
for  further development  of  the system.   After the  interview,  the knowledge
engineer implements additions and changes  suggested  by  the  experts and returns
to  them  for another demonstration  and  interview.   This  iterative development
process  insures  that  the  knowledge engineer does not misinterpret information
gained  through the interviews.   After  several iterations  the   system should
evolve to  the point where the  problem is  solved to  the  satisfaction of both
the domain  experts  and the system developers.
                                    530
-------
BENEFITS OF  EXPERT  SYSTEMS

The  benefits  of applying  expert  systems  technology  depend on  the  problem
domain to be addressed.   In most  cases, expert  systems  are  implemented  in  hope
of decreasing  operational  costs.   An expert's time can be a  costly  commodity.
Obviously,  if  an automated  process can successfully  distribute expertise to
many problem instances, this  system can yield considerable  savings.

Note that the  expert system  is not an  attempt  to replace domain specialists.
Currently, knowledge representation technology  is far from the stage where the
human thought  process can be  cloned.  Rather, a good expert system can  imitate
a human expert's responses in a  limited problem  scope  so as  to reach  similar
conclusions  in routine or monotonous cases.  This liberates the  expert  so that
time can be  spent on more complex or research oriented  problems.

Conservation of  expert  resources is not  the sole benefit  of expert systems.
By using an  automated system, every solution alternative  can  be  systematically
considered.  Expert systems do not  reach conclusions without  first considering
all important  factors and outcomes.  In addition, these factors  can  usually be
examined by  a  machine much  faster than  they could manually.  An expert system
may solve  a problem  in  minutes  that could require hours  or even  days  of a
human expert's time.  Expert  systems are also potentially more  available than
human  experts.  Domain  specialists are generally  in  high  demand.    In some
cases  there may be  a  significant time  interval  before  an expert   becomes
available to begin work on a  problem.

Another advantage of the  automated  system  is that machines are not  subject to
time,  pressure,  or  personal  constraints.    A  computer  will always  deliver
consistent responses  given consistent  input information.  In  an environment
where there  are  many decision makers and a need for uniformity among them, an
expert  system  can  simultaneously  increase the  quality,   consistency,  and
response time of decisions.   Furthermore, an expert system can often be geared
to act as a  training tool  for inexperienced employees.   Lastly,  creation of a
prototype knowledge  base can be  used as  a storage media  and  organizational
mechanism for  information in  a very dynamic field  of study.   A knowledge base
can depict   more  complex informational  relationships   than  a database while
maintaining  greater  accessibility to the  information  than  standard  technical
documentation.

RREL EXPERT  SYSTEM PROJECTS

For several  years  the EPA's  Risk  Reduction Engineering  Laboratory  (formerly
Hazardous  Waste  Engineering  Research   Laboratory)   has   been  active   in
development  of expert systems to  aid hazardous waste managers.  As a result of
this project, two tools have been developed, more are scheduled for completion
this year,   and  others  are in  the  beginning  stages of  development.   These
systems  will  run  on  standard  IBM  compatible  microcomputers.    Although
dedicated workstations   would  increase the performance of many  of  these
systems,  the laboratory has decided to develop for the PC platform in order to
                                   531
-------
minimize hardware  requirements  thus  making these systems available to as many
users as possible.  Not  all of  these programs are expert systems in the strict
sense  of the  term.   Each however  has the  potential  to  make environmental
decision  makers  either  better informed,  faster,   or  more  consistent  in  a
specific  area  of  concern.   The following  paragraphs  will  briefly  describe
these systems.

FLEX

FLEX is an  advisory system to assist in the analysis of chemical compatibility
data for flexible  membrane liners.   EPA specifies that geosynthetics which are
used  in hazardous waste  containment  applications must  first be  tested for
compatibility  with leachate from the contained  substances.  The test specified
for  this  determination  is  EPA Method   9090.    FLEX  assists  the   user   in
interpretation of  the  test results of  Method  9090.  This program was completed
in November of 1988 and  is currently  undergoing beta  testing in selected EPA
offices prior  to formal  release.

GM

The  GM system is   a  geosynthetic  modelling  advisor.   It  is not  so  much  an
expert  system  as a technology transfer tool.   It contains  knowledge concerning
the  appropriate engineering  calculations  used  in  geotechnical  designs  which
incorporate geosynthetic components.  It  is  used as a quick reference to the
equations  used in determining   safety  factors  for  a  variety of  geotechnical
structures.  This  system is available.

WAPRA

The  WAPRA system  assists  in the review  of the Waste Analysis Plan Section  of
RCRA  Part  B  facility  permit  applications.     Specifically,   this   program
automates  two  functions of waste  analysis  plan  review.    First,  the  system
"checks all wastes  which will  be processed  in  the same treatment  or  disposal
unit to  identify   possible chemical  interactions  which might  pose  a  health
threat to  unit operators  or demean  the  performance of the waste  processing
unit.   Second,  sampling methods are  analyzed  and judged  for  appropriateness
based  on guidelines in EPA Reference SW-846.   WAPRA is  better classified,as  an
 intelligent database than as an expert system.   It differs from a conventional
database in that  it  is  mostly  menu  driven and thus does not  require  the user
to learn a query  language.  In  addition,  WAPRA has  the  capability to  classify
wastes according  to  chemical  characteristics  and  decide  when two classes  of
 chemicals are  incompatible.  This  system is scheduled for completion  in June
 1989.

 FINAL COVER ADVISOR

 The final  cover  advisor makes recommendations  concerning standard  RCRA  cap
 design for hazardous  waste land disposal  facilities.   Increasing restrictions
 on  the land  disposal  of hazardous  waste  may force  closure  of many such
                                     532
-------
 facilities.   This  system assesses the  long term integrity  of a final  cover
 system based on  information provided in  the facility's  closure  plan    This
 system is  scheduled for completion in  June 1989.

 LEACHATE COLLECTION SYSTEM ADVISOR

 The  leachate collection and removal  (LCR)  system  analyzes  the suitability of  a
 leachate collection system based upon various site characteristics and design
 parameters.    It  makes  judgements on  operational  specifics  such  as   flow
 capacities,  pipe  strengths, and  clogging  potential.    Development  of   this
 system is  ongoing with  a tentative completion date of June  1989.

 VEGETATIVE COVER ADVISOR

 The  VEGCOVER system  contains knowledge concerning  the suitability  of  plant
 species proposed for  use in  the  vegetative cover  layer of landfill and surface
 impoundment   caps.     This  program   contains   information  on  the  growing
 characteristics  of  over  60 types  of  vegetation  commonly  used in  capping
 applications.   This  is compared  to  soil   type  and climate  designations for
 areas  within  the 50 United  States.  Development of this system  is ongoing with
 a tentative  completion  date  of June 1989.

 TECHSCRN

 TECHSCRN advises the  user of the technical feasibility of various remediation
 alternatives  for   Superfund  sites.     The  system   prompts  the  user  for
 characteristics  of the site  and   the  contaminants.    It  then  searches its
 internal  database  of  35  remediation  technologies   and   screens  out  those
 technologies  which  are  not appropriate for the cleanup task.   This  system is
 currently  a  rapid prototype.   Plans have been made  to  continue development of
 this system  but  a  completion date has not been  identified.   Future versions
 may  incorporate cost  analysis algorithms or a treatment train optimizer.

 CONSTRUCTION DESIGN REVIEWER

 This system will conduct a prefinal  review of design plans and specifications
 for  remedial  action at  a Superfund site.   The  objective of  the  system  is to
 assure  that  these  designs  are  biddable,  constructive,   operable,   and
 consistent with applicable environmental  regulations,   this system is still in
 the planning stage.

 CONSTRUCTION CLAIMS ADVISOR

This system will assist  the  site engineer  when a  contractor proposes  a change
 in  work   activity  at   a   Superfund   site.    The  system  will   assess   the
appropriateness  of  the  proposal,  suggest  responses   to  a   variety  of
unpredictable site  events  (eg.,  delays,  additional  work,  contract disputes),
and help determine whether there is a  legal  basis for  recovering  the  costs of
a change order.   This system is still  in  the planning stage.
                                  533
-------
RISK ASSESSMENT

This  system  helps  assess  health  risks  associated  with  hazardous  waste
exposure.   It contains information  on  maximum contaminant  levels  (MCLs)  and
alternate  contaminant levels  (ACLs)  and selects  the  appropriate model  or
methodology for predicting risk  levels.   A rapid  prototype has been developed
and work on this system is continuing.

QA/QC SAMPLING PLAN ADVISOR

This  system assists  the  user  in  the  analysis  of  QA/QC  sampling  plans  for
Superfund sites.  The objective of the system is to check for conformance with
Agency guidelines and procedures.   In  addition, data is examined to assure it
is sufficiently  accurate  to support  regulatory  and/or cleanup  decisions.   A
rapid prototype has been developed and work on this system is continuing.

LOOKING TO THE FUTURE

In general,  response to this project  has been enthusiastic.   It seems that,
given the  volume of applications that  must be processed  by EPA's regulatory
arm, reviewers are  eager  to  employ  methods which  might increase productivity.
Of  course,  as  with  any  new  technology,  quality  assurance  is  a  concern.
However, by  involving the prospective users as well  as domain experts in the
iterative  development process,  those  responsible  for  decision  quality  are
contributing  to  the shape of the product.   This  does  a lot to alleviate the
skepticism with which expert systems are  often met.

If  this  project is  successful,  individual  expert  systems  will  soon  make
employees  more  informed  and  more efficient  by  placing volumes of reference
materials  and expertise  at their  fingertips.    In  the  future  all  computer
software will be expected to incorporate  the cognitive  features  distinctive of
expert  systems.  When this  happens,  a network of advanced technology transfer
systems  will  allow each environmental  decision maker to  perform  as does the
best.   In this respect, RREL is working toward the future  today.

For  current availability information  regarding  any  of the  systems described
herein  contact:  USEPA, Risk Reduction  Engineering Laboratory, Municipal Solid
Waste and Residuals Management Branch  (MSWRMB), Cincinnati,  Ohio 45268.
                                   534
-------
                ASSESSMENT  OF  CHEMICAL AND  PHYSICAL METHODS  FOR
                      DECONTAMINATING BUILDINGS AND DEBRIS~
                               AT SUPERFUND SITES

                by:  Michael L. Taylor, Ph.D.,
                     Majid  A.  Dosani,
                     John A. Wentz,
                     Roxanne B. Sukol
                     Timothy L. Kling,
                     Jack S. Greber,
                     PEI Associates, Inc.
                     Cincinnati, Ohio

                     Naomi  P.  Barkley
                     Risk Reduction Engineering Laboratory
                     U.S. Environmental Protection Agency
                     Cincinnati, Ohio
                                  ABSTRACT

     Many Superfund sites contain buildings, building materials, and debris
that are contaminated with one or more toxic organic and/or inorganic chemi-
cals.  To date, no generally applicable decontamination technique has been
developed for the removal of organic contaminants such as polychlorinated
biphenyls (PCB's) from the various materials included in a modern-day struc-
ture.  The objective of this study was to evaluate chemical and physical
methods for decontaminating buildings and debris at Superfund sites.

     For evaluation of techniques designed to remove PCB's from concrete
floors in buildings, concentrations of PCB's in the top 1/2 inch of a con-
crete floor in a building located at a Superfund site were determined before
and after treatment by analyzing cores obtained from selected locations in
the floor.

     An innovative system for decontaminating debris was also designed, as-
sembled, and evaluated.  After bench-scale experiments were performed to de-
termine an optimal solution for cleaning PCB-contaminated, corroded, metallic
components, a 300-gallon, pilot-scale module was designed and field-tested at
a Superfund site.

     The results obtained during this study were very promising, and the
techniques evaluated showed a great deal of potential for removing PCB's from
concrete flooring and from the surface of the contaminated debris.  This
paper discusses the procedures developed and the results obtained by imple-
menting the chemical and physical decontamination techniques evaluated in
this study.
                                      535
-------
                              DISCLAIMER NOTICE
     This paper was prepared as an account of work sponsored by an agency of
the United States Government.  Neither the United States nor any of their em-
ployees, nor any of the contractors, subcontractors, or their employees make
any warranty, expressed or implied, or assume any legal liability or respon-
sibility for any third party's use or the results of such of any information,
apparatus, product, or process disclosed in this paper or represents that its
use by such third party would not infringe on privately owned rights.  The
views and conclusions contained in this document are those of the author and
should not be interpreted as necessarily representing the official policies
or recommendations of the U.S. Environmental Protection Agency or of the U.S.
Government.
                                      536
-------
KEY-WORD INDEX

Remediation
Decontamination
Super-fund
PCB
Concrete
Debris
Chemical Reagent
Shotblasting
                                    537
-------
                ASSESSMENT OF CHEMICAL AND PHYSICAL METHODS FOR
                      DECONTAMINATING BUILDINGS AND DEBRIS
                               AT SUPERFUND SITES

                                INTRODUCTION

     A large number of sites (more than 20,000) in the United States are con-
taminated with hazardous waste, and cleanup of these waste sites is the top
environmental priority of the decade.  As of this writing, about 1200 sites
are listed on the National Priorities List (NPL), and many more sites have
been proposed for inclusion on the list.  In the Superfund Amendments and Re-
authorization Act (SARA), a mandatory cleanup schedule is being considered
that would require EPA to clean up at least 375 of these sites over the next
5 years.

     Many Superfund sites contain buildings (e.g., office buildings, manufac-
turing facilities), building materials (e.g., glass, concrete, mortar, brick,
stone), and debris (e.g., scrap metal, pieces of wood, equipment or furni-
ture) that are contaminated with one or more toxic organic and/or inorganic
chemicals.  Decontamination of these items is important in preventing the
spread of contamination off site and in reducing exposure levels to future
users of the buildings or equipment.  To date, no generally applicable decon-
tamination technique has been developed for the removal of organic contami-
nants, such as polychlorinated biphenyls (PCB's), from the various materials
included in a modern-day structure.  At present, equipment is usually steam-
cleaned, and buildings and structures are frequently torn down instead of
being decontaminated.

     A recent U.S. Environmental Protection Agency publication (1) discusses
various methods including scarification, hydroblasting, and a variety of
chemical treatments for cleaning the surfaces of concrete and similar mate-
rials.  Many of these methods produce large amounts of liquid residues that
have to be collected and treated.  Moreover, the overall effectiveness of
these methods has not been carefully verified.  Reliable data are lacking
that provide an indication of the efficacy of either established or emerging
methods of decontaminating intact structures and structural components.

     Therefore, the major objective of this study was to evaluate the effi-
cacy of currently available chemical and physical decontamination methods for
removing contamination from intact buildings and restoring these structures
to a usable condition.  Another goal was the investigation of methods that
show promise for removing toxic contaminants from debris.  A successful
decontamination method can offset the high costs of dismantling and disposing
of contaminated structures, while at the same time salvaging or increasing
the value of the reconditioned buildings, equipment, or property.
                                     538
-------
      Phase  I  of  this  study was  directed  toward  locating an actual contami-
 nated site  or sites that would  be  suitable for  demonstrating various cleanup
 or  site  remediation technologies.   Several sites  (some NPL-listed and some
 non-NPL-listed)  were  visited and evaluated.  Ultimately, two PCB-contaminated
 Superfund sites  were  selected for  Phase  II of this study:  the Pioneer Equip-
 ment  site and the Carter Industrial  site, which are less than 1 mile apart in
 Detroit, Michigan.

      In  Phase II of this building  and debris decontamination study, the
 efficacy of selected  decontamination technologies for removing PCB's has been
 evaluated during field studies  performed at the two contaminated sites in
 Detroit.  Methods for removing  PCB's on the surface and in the top i inch of
 a concrete  floor at the Pioneer site were assessed in a comparative fashion.

      Also during Phase II, a system for decontaminating debris was designed
 and assembled.   Bench-scale studies were conducted to determine the optimal
 solution for  hydromechanically  cleaning contaminated debris.  Based upon the
 outcome of  bench-scale studies, a  pilot-scale version of the debris cleaning
 system was  designed and demonstrated at the Carter site.


      BUILDING DECONTAMINATION  DEMONSTRATIONS AT PIONEER EQUIPMENT SITE

      At the Pioneer Equipment Site, two decontamination techniques for re-
moving PCB's were demonstrated.  These two techniques were 1) a method for in
 situ  degradation of PCB's that entails application of an alkali metal/poly-~
 ethylene glycolate mixture directly to the concrete surface, and 2) a shot-
 blasting technique that entails use of steel  shot to cut away concrete sur-
 faces.  The decontamination tests were conducted in a statistically valid
manner in an attempt  to minimize both the effects of point-to-point varia-
 tions in PCB concentrations in or on the floor and the subtle effects of
variations  in concrete composition itself on the comparison of decontamina-
tion methods.

SAMPLING PROCEDURE

     Prior to implementation of the two decontamination technologies,  the
concrete floor (which is located in an abandoned building)  was divided into
sections, and each section subdivided to form test plots.   Within these test
plots, sampling locations were identified for baseline (pretreatment)  and
posttreatment sampling.   More specifically,  the floor was  sectioned off into
four 20-foot x 10-foot rectangular plots and  each was then  bisected to create
a total of eight 10-foot x 10-foot plots.  These plots were all  in the north-
west quadrant of the north room of the building and had a  north/south  longi-
tudinal axis.  Five sampling locations were  identified in  each of the  eight
test plots.   Two adjacent points were marked  off at each  of the five sampling
locations, and the first of the two adjacent  points was sampled as part of
the pretreatment sampling program.   The second  point was  sampled  subsequent
to treatment of the test plot.
                                      539
-------
     Concrete core samples were obtained by use of a 2-inch coring tool at
each of the five sampling locations in each test plot.  Each concrete core
had a diameter of approximately 1.75 inches.  From each of these eight plots,
five core samples were obtained for a total of 40 samples.  The top i inch of
24 of the 40 samples (i.e., 3 per plot) was analyzed for PCB content.  The
remaining 16 core samples (2 from each plot) were each divided into two sec-
tions (the top i inch and the next i inch, thus creating 32 samples) and each
section was subsequently analyzed for PCB's.  Thus, a total of 56 sample
analyses were completed before the two decontamination procedures were imple-
mented.

     After completion of the two decontamination technologies, another 40,
core samples were obtained from locations on the floor that were as close as
possible to the original sampling locations.  The samples were sectioned and
analyzed as described for the pretreatment samples, resulting in another set
of 56 analyses.  The locations of the top i inch and the next i-inch core
samples are shown in Figures 1 and 2, respectively.

DEMONSTRATION OF THE IT/SEA MARCONI REAGENT:  METHODOLOGY AND RESULTS

     The chemical reagent, also called the IT/SEA Marconi reagent, is a poly-
ethylene glycol-based mixture.  During the technology demonstration at the
Pioneer Equipment Site, cans of IT Marconi reagent were heated in a waterbath
to a temperature of 180° to 190°F.  When the reagent reached the correct
temperature, a small amount was poured onto the concrete floor within desig-
nated test plots and was spread evenly by a roller.  This process of applying
the reagent was repeated two additional times at 24 and 48 hours following
the initial application of reagent.  The reagent was then allowed to dry for
2 weeks.

     After 2 weeks, 20 posttreatment core samples were collected from the
areas that had been treated with the Marconi reagent and were analyzed for
PCB's.

     The concentrations of PCB's in the top i inch of concrete before and
after the application of IT/Sea Marconi reagent are summarized in Table 1.
The reduction in PCB concentration ranges from 11 to 97 percent (average
reduction of 73 percent), and these results seem to indicate that the reagent
was able to penetrate into the concrete floor and react with PCB's.  The
penetration of the reagent into concrete and reaction with PCB's is supported
by the results obtained for the next i inch of concrete, as shown in Table 2.
In this latter table, the percentage reduction in PCB concentration ranges
from 66 to 99 percent (average reduction of 91 percent).  In previous studies,
it has been found that PCB-containing oils can penetrate deeply into concrete.
For example, at the Dayton Tire Site (an NPL site in Dayton, Ohio), a poured
concrete slab that supported large transformers was sampled for PCB content.
PCB's in the concrete ranged in concentration from 47,400 yg/g at a depth of
0.25 to 0.5 in. to 33.700 yg/g at 3.25 to 3.50 in.  In an assessment of
another PCB-contaminated site for a confidential client, a similar finding
was made—PCB's spilled on concrete penetrated into and even through an 8- to
                                      540
-------
                           O
                               r-
                               co
                                       CO
                                       CO
                               CO
                               CO
o>
ffi
CO
SHOTBLASTING
SB11 SB12 I
•«•••




• •
m
CO
eo
m —
CO
5
CO
•^••M



MARCONI REAGENT
MR11 MR12





m
«
!
t
10
E
•t
••"
"^l
L— ^



    co
                  CO
CO
                  CO
                  CO
                          ID
                                  ffi
                                                                                                                        CO
                                                                                                                       .CD
                                                                                                                        Q.


                                                                                                                        05
                                                                                                                        CO
                                                                                                                        o
                                                                                                                        o

                                                                                                                        o
                                                                                                                        a.
                                                                                                                        o
                                                                                                                       *-
                                                                                                                        CD
                                                                                                                       O
                                                                                                                       CO

                                                                                                                       g


                                                                                                                       I

                                                                                                                       O)

                                                                                                                       "5.

                                                                                                                       03
                                                                                                                       CO
                                                                                                                       §
                                                                                                                       D)
                                                                                                                       U_
                                                          541
-------
                            S3
                            en
                            co
                       O

                       CO
       111

       i

       I
              to
              CM
              m
              co
                            CD
                            CO
                       I
                       g
                       5
(O
   o
       1
              to
a
                                                                                                                        (A
                                                                                                                       .0)

                                                                                                                        Q.

                                                                                                                        E
                                                                                                                        OS
                                                                                                                        CO

                                                                                                                        £

                                                                                                                        8
                                                                                                                        c=

                                                                                                                        0)
                                                                                                        CO

                                                                                                       JO

                                                                                                       "ca
                                                                                                        O)


                                                                                                       "o.
                                                                                                        E
                                                                                                        (0
                                                                                                       CO
                                                                                                                       CM
                                                                                                                        C5)
                                                       542
-------
   TABLE 1.  CONCENTRATION OF PCB's IN A SURFACE (TOP i INCH)  OF
CONCRETE BEFORE AND AFTER APPLICATION OF SEA MARCONI  REAGENT (ppm)
Sample No.
MR1
MR2
MR3
MR4
MR5
MR6
MR7
MRS
MR9
MR10
MR11
MR12
MR13
MR14
MR15
MR16
MR17
MR18
MR19
MR20
Pretreatment
52.0
6.5
22.0
5.7
8.3
6.5
9.2
42.4
22.1
23.3
33.2
21.7
50.0
19.8
39.8
4.6
8.1
60.0
10.2
12.3
Posttreatment
3.67
2.46
0.64
1.58
2.79
2.35
1.96
1.43
1.94
1.58
5.03
3.25
5.45
5.63
6.22
3.04
4.12
5.82
9.10
5.48
% Reduction
93
62
97
72
66
64
79
97
91
93
85
85
89
72
84
34
49
90
11
55
 TABLE 2.  CONCENTRATION OF PCB's IN A SUBSURFACE (NEXT i INCH)  OF
CONCRETE BEFORE AND AFTER APPLICATION OF SEA MARCONI REAGENT (ppm)
Sample No.
MR21
MR22
MR23
MR24
MR25
MR26
MR27
MR28
Pretreatment
3.5
5.0
6.3
13.0
4.1
1.8
2.7
3.0
Posttreatment
0.23
0.23
0.21
0.10
1.41
0.35
0.10
0.01
% Reduction
93
95
97
99
66
80
96
99
                               543
-------
12-in.-thick poured concrete slab.  Of course, the permeability of the concrete
is influenced by many factors, such as the density, moisture content, and
cracking.                           ,                      j

     On the basis of this rather limited study, it appears that the IT/SEA
Marconi reagent is effective in reacting with PCB's that a're on the surface
of or perhaps contained within the upper i inch of a concrete floor.  Addi-
tional studies involving analyses of concrete cores prior to and subsequent
to application of the reagent should be performed to corroborate these
initial findings.  Furthermore, additional work is needed 'to determine the
optimum reaction conditions for the IT/SEA Marconi reagent.

DEMONSTRATION OF THE SHOTBLASTIN6 PROCESS:  METHODOLOGY AN'D RESULTS

     Shotblasting is a destructive procedure that may be used to remove
surface layers of contaminated concrete.  By selecting the shot size and the
rate at which the machine traverses ;the concrete surface, one can control the
amount of concrete removed from the .surface of a concrete floor.  Typically,
a 1/16- to 1/8-inch layer of concrete can be removed by the shotblaster.
(Note:  The shot blasting machine used in this study was Blastrac Model 1-10D
with a Model 554-DC dust collector.)  The machine is equipped with a HEPA-
filtered vacuum system that captures nearly all of the paryticulate generated
during shotblasting.  The captured dust is periodically removed from the
vacuum system of the shotblaster and transferred to barrets for subsequent
disposal..  Despite the HEPA-filtered vacuum system, use of the shotblasting
process may generate airborne debris, which has the potential for cross-
contamination of test plots.  Therefore, the shotblasting evaluation at the
Pioneer Equipment Site was commenced after all activities related to the
evaluation of the IT/Marconi reagent were completed.      [

     Designated test plots on the co'ncrete floor (see Figures 1 and 2) were
shotblasted and a minimum of 0.125 to 0.250 inch of concrete was removed.
Fugitive dust generated by the process that was not captured by the vacuum
system was gathered and vacuumed separately, and disposed of as contaminated
concrete material.  Following shotblasting, 20 posttreatment core samples
were collected.
                                                          i
     The concentrations of PCB's in the top i inch surface of concrete before
and after the shotblasting process are summarized in Table 3.  The percentage
reduction in PCB concentration after1 shotblasting ranges fjrom 19 to 96 per-
cent (average reduction of 68 percen't), which indicates that the technique is
fairly effective in removing PCB's from the surface of the concrete floor.
However, two samples (SB2 and SB11) Ishowed higher PCB concentration after the
shotblasting treatment.  These latter results can be explained on the basis
that the distribution of PCB within ,the concrete is undoubtedly nonuniform,
and therefore it is probable that thje two posttreatment samples were obtained
from locations in the concrete floor where the initial concentrations of PCB
were much higher than the concentrations of PCB found in the SB2 and SB11
pretreatment samples.               <
                                      544
-------
TABLE 3. CONCENTRATION OF PCB's IN A CONCRETE SURFACE (TOP i INCH)
BEFORE AND AFTER SHOTBLASTING (ppm)
Sample ID
SB1
SB2
SB3
SB4
SB5
SB6
SB7
SB8
SB9
SB10
SB11
SB12
SB13
SB14
SB15
SB16
SB17
SB18
SB19
SB20
Pretreatment
4.4
0.13
65.0
7.7
26.0
17.2
11.8
35.7
13.4
4.6
1.6
4.4
4.7
23.5
3.2
25.0
22.8
21.4
40.9
42.5
Post-treatment
3.0
1.5
2.52
3.17
2.11
6.92
5.44
6.3
3.4
3.91
1.86
2.75
1.97
1.56
1.41
4.84
1.28
5.38
8.15
7.78
% Reduction
32
-1054
96
59
92
60
54
82
75
15
-16
38
58
93
56
81
94
75
80
82
545
-------
     These results indicate that in the case of the PCB-contaminated floor at
the Pioneer site, PCB's in the top i to i inch of the concrete floor are ef-
fectively removed by shotblasting.  However, further studies should be per-
formed to determine the depth of penetration of PCB's into the concrete.  As
discussed previously in this paper, PCB-laden oil can, at least in some
cases, penetrate deeply into concrete.  It may be necessary, therefore, to
repeatedly shotblast the concrete surface to remove the contaminated layers
of concrete.

     Only a small amount of fugitive dusts were generated by the shotblasting
process.  An estimated 95 percent of the concrete dust was captured by the
vacuum system, which is an integral part of the machine.  In total, a 1000-
square-foot area of concrete was shotblasted, removing approximately a 1/8-
to 1/4-inch layer of concrete.  The total quantity of dust collected amounted
to two barrels.

COST ESTIMATION OF MARCONI AND SHOTBLASTING TECHNOLOGIES

     On the basis of the experience at the Pioneer site, a cost analysis was
performed to determine the costs of large-scale implementation of each of the
two building decontamination techniques.  The cost of implementation of any
cleanup technology at a hazardous waste site can vary substantially depending
on the site's location, nature, types of contaminants, availability of facil-
ities, and many other variables that can unexpectedly increase or decrease
the estimated cost.  Therefore, to establish a base in the cost estimation,
several assumptions have to be made.

     A 10,000-square-foot plot was taken as a baseline area on which the cost
estimation was performed.  These costs are based on the following assump-
tions:  1) the site is within 50 miles of the contractor's facility, 2) no
pretreatment cleaning of the plot is required, 3) the plot is free of
obstructions such as debris, equipment, or machinery, and 4) the site has
electrical power and outlets.  The costs for the large-scale implementation
of the IT/SEA Marconi reagent and shotblasting techniques that were
calculated on the basis of these assumptions are summarized in Tables 4 and
5, respectively.

     As shown in these tables, the cost of Marconi reagent is 85<£ per square
foot, compared with $2.19 per square foot for shotblasting.  The shotblasting
technique is labor intensive and generates a significant quantity of contami-
nated waste, whereas the Marconi reagent technique requires minimal labor and
the reagent does not generate wastes.  On the basis of limited experience in
these studies, it appears that the cost of shotblasting is almost 3 times
higher than that of the Marconi technique.  These costs do not include  sam-
pling and analytical costs.

      DEMONSTRATION OF DEBRIS DECONTAMINATION:   BENCH-SCALE EXPERIMENTS

      In designing the debris decontamination system,  the goal was to produce
a portable, self-contained module  in which debris could be washed by a  non-
toxic cleaning solution.  The debris decontamination  system was designed to
                                      546
-------
        TABLE 4.  ESTIMATED COST OF IMPLEMENTATION OF MARCONI REAGENT
    TECHNIQUE BASED ON THREE APPLICATIONS TO A 10,000-FT2 CONCRETE PLOT
                       Description
  Cost
1.  Fixed rate labor (includes cleanup technicians),
     70 hours at $50/h
2.  Fixed rate equipment (includes reagent heating system,
     applicator)
3.  Other direct costs
      Expendables (includes Marconi reagent, protective
       clothing, empty drums, miscellaneous)
      Nonexpendables
      Travel (@ $50 per diem)
4.  Disposal (approximately 2 drums @ $125/drum)
5.  General  and administrative (calculated on Item 3 @ 16.1%)
6.  Fee (on  Items 3 and 5 @ 10.0%)
        Total estimated cost for 10,000 ft2
        Cost per square foot
$ 3,500

    500


  2,900

      0
    450
    250
    539
    389
$ 8,528
$  0.85
                                     547
-------
    TABLE 5.  ESTIMATED COST OF IMPLEMENTATION OF SHOTBLASTIN6 TECHNIQUE
    BASED ON THE REMOVAL OF TOP 1/4 INCH OF A 10,000-FT2 PLOT OF CONCRETE
                         Description
Cost
1.  Fixed rate labor (includes foreman, cleanup technician, and     $  9,000
     equipment operator), 180 hours at $50/h
2.  Fixed rate equipment (includes rental of shotblaster,              4,612
     HEPA vacuum, air compressor)
3.  Other direct cost
      Expendables (includes vacuum bags and filters, micro-            2,090
       traps, protective clothing, miscellaneous)
      Nonexpendables                                                       0
      Travel (@ $50 per diem)                                          1,500
4.  Disposal (approximately 30 drums of contaminated concrete          3,750
     @ $125/drum)
5.  General and administrative (calculated on Item 3 @ 16.1%)          ,  578
6.  Fee (on Items 3 and 5 @ 10%)                                         417
        Total estimated cost for 10,000 ft2                         $ 21,947
        Cost per square foot                                        $   2.19
                                     548
-------
include a solvent reclamation system that would permit the cleaning solution
to be reclaimed and reused, thereby minimizing the volume of contaminated
liquid produced during the debris washing process.  In order to test some of
these concepts, the bench-scale experiments described below were performed.

     A bench-scale version of a Turbo-Washer (Bowden Industries) served as
the debris washer for the initial studies.  This unit includes an axial flow
pump, a propeller shaft, a propeller, and a pressure chamber, all housed
within a heated tank also equipped with a rotating disc for removing oil that
rises to the surface of a protected (minimal turbulence) segment of the
cleaning tank.  During operation the Turbo-Washer's pump vigorously mixes the
cleaning solution by continuously recirculating the cleaning fluid in and out
of the pump.  The items to be cleaned are placed into a wire-mesh basket that
is lowered into the heated tank where the items are exposed to the turbulent
cleaning solution.

     The bench-scale version of the Turbo-Washer was used to evaluate four
cleaning solutions including tap water, 10 percent sulfuric acid, and aqueous
dilutions of two proprietary cleaning solutions:  BB-100 (Bowden Industries)
and Power Clean (Penetone Corporation).  The experimental procedure involved
the application to rusted iron parts of measured quantities of used motor
oil, grease, and soil to simulate the kind of grime likely to be encountered
on oily, PCB-contaminated metal parts and debris in the field.  Three tests
were performed with each cleaning solution, and a fresh set of oil/grease
contaminated metal parts was employed for each test.  Each set of contam-
inated parts was matched closely with regard to size, shape, and type of
metal.  The parts were arranged in the same order in the parts washer basket
during washing.

     At the completion of each of the bench-scale tests, two aliquots of
cleaning solution were collected from the heated tank, one aliquot was sub-
mitted for oil and grease analysis, and another for total suspended solids
analysis.  In addition, two surface wipe samples were collected from the
surfaces of the cleaned metal parts, and these were also analyzed for oil and
grease to determine the level of oil/grease remaining on the metal surfaces
after treatment in the debris washer.  The skimmer oil collected during each
of the three runs was combined for oil and grease analysis.  The amount of
oil and grease in each of the samples was determined by standard EPA methods
(2).

     The results of the oil/grease and total suspended solids analyses are
summarized in Table 6.  The analytical results of the wipe samples indicate
that> after cleaning, the amount of oil and grease on the metal surfaces was
significantly higher in the case of water or sulfuric acid and comparatively
lower for BB-100 and Power-Clean.  This indicates poor cleaning performance
of water and sulfuric acid.  Moreover, the handling of 10 percent sulfuric
acid was difficult, and it also had a severe corroding effect on the debris
washing equipment.  Hence, it was concluded that neither water nor sulfuric
acid should be considered as a potential cleaning solution for. oily PCB-con-
taminated debris.

     On the basis of the results of the surface wipe testing listed in Table
6, it was concluded that BB-100 solution is a more effective cleaning solution
                                      549
-------










CO
CO
1
^
CO
o
1— <
1
o
CO

0
Ul
0
i»^|
UJ
Ou
CO

co

^
0
I*"" *
o
§
UJ
CO
UJ
CO
^**»
P"j
o

C£
o
Li-

CO
1
1

co
UJ


o

5^»
cc

3IZ
=>
CO


•
(O

UJ
_J
CO
J—
























c
o
•r"
3
"o
CO

cn
c
c
to
at
o



























































o
C 4J
re re
CU S~
010
CU
(X C
o
u
>
="
1 rH
CO
CO •
§
u





"o
•o >

u •
re t>
3r
u
•r- ••
$- fc^
3O
CO U
c
0
u

1_
CU
.[ %
re





CO
co

r—
re
c







re
cu
Q_

**
CU
'S.
E
re
CO







re
+•>

CU O
E: 2E
'£ C
CU 3
OL S-
X
UJ





O 0 0
rH rH CM





"^ CO rH
rH CO














rH CO CO
ID «J- CO
rH rH rH





CM rH rH
rH CM



CU V CU
CO CO CO
to re re
 T3 4J
K-r- C •!- C •!-
re r-~ re r^ re r-~

r— COr— COr— CD



C C C
O O O
•r- 'r- •!-•
^ ^3 ^
r— r— r—
000
CO CO CO

C C C

C C C
co re re
01 CU 01

CJ O t_>









rH CM CO




1




CO CO "G}*
o r** co






C3 «4- O
o o o
rH













co in co
CM If) «3-
rH CM rH





cn r-.. co


i~ i. i-
01 01 01
01 'r- CU "r- CU *r-
•O i— -O r- X> i—
c •»-. c >^. c >^
O) en oj co o> cn
Q. E D» E o. E
CO CO CO
3 • 3 «3 »
CO CO CO CO CO CO
•o -o -o
r— • *r- r— «r- r~ »r"
re ^^ re r^ re r*~
•P O 4-> O 4J O
o to o to o to



c c c
o o o

4J +J 4J
r— r— r—
O O O
CO CO CO

C C C
•r~ «r- *r-
C C C

<_>









rH CM CO









O CTl CO CO «* rH
CO *tf- ^}- «3* CO CO
O O O O O O





CM ID CO CM CO CO
CO CM rH «3- CM CO
• •••••
O O 0 O O O












CO IT) rH r~-

rH rH O ^ CO CO





r-» CM •* o co
f» oo cn «i- **• i
rH rH O «* CM



CU Ol Ol fll 01 CU
CO CO CO CO tO CO
re re re re re • re
01 Ol Ol Ol O) Ol
cn cn cn cn cn cn

T3N "ON "ON T3N "OM T3W
C £E C £ C E C E c F" ci E
reureureureureureo

r— cnr— cnr— cnr— cnr— cnr— cn








rH CM rH CM rH CM

O O O O O O


01 01 01 Ol 01 0)
O. CL O- O. Q. O.

3: !I !2E ^ 3! 3









rH rH CM CM CO CO









O
§
VI
CO





§
CO
CO












o

rH





1




Ol
CO
re
cn s-

"O 4J
C -r-
fO r-

if




s_
01

•r—
CO

§

4-

r—

O






CO
<*-
o oa
r— •>
re CM

o •
r- rH

















































CO
01
Q.
re
CO

+•>
C
01
B
re
£

-i i
to
o
a.

£
re
to
01

"o.

re
to
r_

cC

irj
550
-------
than a Power Clean solution.  This is shown graphically in Figure 3, which
plots the results obtained for the wipe samples (in milligrams of oil and
grease/square centimeter) from each test.  Table 6 results also show that
cleaning with BB-100 removed solids from metal surfaces more effectively than
Power-Clean.  The data also show that at the completion of the third run, the
BB-100 solution still had more cleaning capacity to remove dirt from metal
parts than did Power-Clean.  Hence, of the four cleaning solutions tried,
BB-100 was selected as the cleaning solution best suited for cleaning oily
PCB-contaminated metal parts and debris in the field.

     Throughout these cleaning solution evaluations, the Bowden Turbo-Washer
performed well.  Good agitation of the cleaning solution was attained and the
unit performed reliably.  Therefore, it was concluded that a large-scale
version of the Turbo-Washer would be used in subsequent field tests.
DEMONSTRATION OF DEBRIS DECONTAMINATION:
CARTER INDUSTRIAL
RESULTS OF PILOT-SCALE TESTS AT
     An Experimental Debris Decontamination Module (EDDM) was designed and
assembled on the basis of the bench-scale results.  A 300-gallon-capacity
Turbo-Washer was installed on a 48-foot semitrailer and the Turbo-Washer was
modified by incorporating a particulate removal system and oil/water separa-
tor.  Also mounted on the trailer was a carbon sorption system for removing
PCB's from the cleaning solution.  Figure 4 represents a flow-diagram of the
pilot-scale module.  The trailer-mounted module was transported to the Carter
site for field-testing.

     At the Carter site, two 200-1b batches of metallic debris were cleaned
by using the EDDM.  A solution of BB-100 surfactant was used as the cleaning
solution.  Prior to the cleaning process, five individual pieces of metal
from each batch were sampled for PCB's by a surface wipe technique.  The
debris items were placed into a basket and transferred into the EDDM and the
cleaning process was instituted.  Each batch of debris was cleaned for a
total of 2 hours.  During the cleaning process, a portion of the cleaning
solution contained in the Turbo-Washer was pumped through a closed-loop
system in which the oil/PCB-contaminated wash solution was passed through the
particulate filter and into the oil/water separator.  The effluent from the
oil/water separator was then recycled into the module.  At the completion of
the cleaning process, five additional wipe samples were obtained from the
same pieces of metallic debris to assess the postdecontamination level of
PCB's.  The surface wiping procedure was carried out as described in the
"Field Manual for Grid Sampling of PCB Spill Sites to Verify Cleanup" (EPA
560/5-86/017, May 1986, pg. 33), which entails use of the hexane-soaked
cotton gauze pad to wipe a 100-cm2 area on the surface of the object being
sampled.  In the case of the metallic debris sampled in this study, the
posttreatment wipe sample was obtained from a location adjacent to that of
the pretreatment sample.  The surface wipe samples were analyzed for PCB's at
Hayden Environmental, Inc. (formerly PCS, Inc.) in Dayton, Ohio.
                                     551
-------
    0.6 -
u

en
UJ
C_3
CO


1
2=
o
UJ

I
CD
    0.5 •-
                                                  POWER CLEAN
    0.4 --
     0.3 - -
                                                    BB-100
     0.2
     0.1
       (0,0)
                                         RUN NO.


            Figure  3.   Amount of  oil  and grease  on surfaces  of ;metal
                        parts after completion  of cleaning cycle.
                                          552
-------
                                s
                                9
                                ^>
                                5
                                o
                                <;=
                                OS
                                TO
                                8
                                e
                                Q.
                                J>
                                8
                                •93
                               EL
553
-------
     The quantity of PCB's on  the  surface of each piece of metal before and
after  cleaning  is summarized in Table 7.  The percentage reduction of PCB's
achieved during cleaning ranges from 33 to 87 percent  (average reduction of
58  percent) for Batch  1 and from 66 to 99 percent (average reduction of 81
percent) for Batch 2.  In the  case of Sample 1  in Batch 2, however, the PCB
analysis gave a higher concentration after the  cleaning process probably
because the posttreatment wipe sample was obtained from a location on the
debris surface  that was initially  more heavily  contaminated with PCB's.
     TABLE 7.  CONCENTRATION OF PCB's FOUND IN SURFACE WIPES AND BLANKS
Sample No.
    Pretreatment,
     pg/100 cm2
Posttreatment,
 ug/100 cm2
% Reduction
Batch 1

      1
      2
      3
      4
      5
Batch 2

      1
      2
      3
      4
      5
         134
         490
        1280
          73
         203

Blank:  <1.0 yg/100 cm2
         8.0
      6090
       374
        96
      1690

Blank:  1.0 ug/100 cm2
     50
    178
    856
     43
     23
     13.0
   1800
    128
     10
     18
      63
      64
      33
      41
      87
                                                 Average % reduction:  58
     -63
      70
      66
      90
      99
                                                 Average % reduction:  81
     The results also indicate that the quantity of PCB's removed during
cleaning of Batch 2 was greater than that of Batch 1.  The reason for better
cleaning results for Batch 2 could be due to the following:  In the case of
Batch 2, after 1 hour of cleaning the basket containing the debris was re-
moved from the washer and the parts were manually rearranged so that all
sides of the debris were exposed to the cleaning solution with the same force
of the Turbo-Washer.  The basket was then lowered back into the washer and
cleaning was continued for 1 more hour.  In the case of Batch 1, however, the
cleaning process was continued for 2 hours without the debris in the basket
being rearranged.
                                      554
-------
     The surfactant solution in the Turbo-Washer was sampled twice during the
actual cleaning process; the concentrations of PCB's found were 928 and 420
yg/liter.  Following completion of the debris washing experiment, the clean-
ing solution was pumped through a series of particulate filters and finally
through activated carbon.  The PCB concentration was reduced to 5.4 yg/liter
during this treatment.  Most municipalities allow water containing _
-------
                            REFERENCES

Esposito, M. P., McArdle, J. L., Crone,  A. H., Greber, J.  S.,  Clark,
R., Brown, S., Hallowell, 0. B., Langham, A., and McCandlish,  C.  D.
Guide for Decontaminating Buildings, Structures, and Equipment at Super-
fund Sites.  EPA Report No. 600/1-85/028, U.S. Environmental  Protection
Agency, Cincinnati, Ohio, January 1985.

Test Methods for Evaluating Solid Waste.  Volume 1C, SW 846,  3rd  ed.,
November 1986.Office of Solid Waste and Emergency Response,  Washing-
ton, D.C.
                                 556
-------
      IMMOBILIZATION  MECHANISMS IN  SOLIDIFICATION/STABILIZATION
                   USING CEMENT/SILICATE FIXING AGENTS

               by:  L.  G. Butler, F.  K.  Cartledge,  D.  Chalasani
                       H.  C.  Eaton,  F.  Frey,  M.  E.  Tittlebaum,
                       S.  L.  Yang
                   Hazardous  Waste Research Center
                   Louisiana  State University
                   Baton Rouge, LA 70803
                              ABSTRACT

      Lime treatment of aqueous solutions of heavy metal  ions prior
to solidification/stabilization produces a sludge which  is or-
dinarily assumed to consist of a mixture of the corresponding metal
hydroxides.  The behaviors of Cd- and Pb-containing sludges toward
cement solidification  is quite different.  We have investigated
this behavior for both sludges and soluble Cd and Pb salts using
TCLP leaching tests, conduction calorimetry and solid-state nuclear
magnetic resonance spectroscopy as a function of time.   Concentra-
tions of Cd in leachates are very low, while Pb concentrations are
considerably higher and would represent a serious threat to ground-
water.  We believe that the Teachability differences arise for the
following reasons.  The Cd/cement system involves Cd(OH)2, which
provides nucleation sites for precipitation of CSH and calcium
hydroxide, resulting in Cd being in the form of the insoluble
hydroxide with a very impervious coating.  On the other  hand, the
Pb/cement system is much more complicated in terms of the precipit-
ation reactions.  In the presence of hydroxide, sulfate and (in the
case of the soluble Pb salt) nitrate, mixed salts are precipitated
which contain all three anions.   These salts retard cement hydra-
tion reactions by forming an impervious coating around cement
clinker grains.   However,  as pH in the cement pore waters undergoes
fluctuations during the progress of hydration,  the Pb salts undergo
solubilization and reprecipitation on surfaces of the cement
matrix.
                             557
-------
               WASTE REDUCTION EVALUATIONS AT FEDERAL SITES

                by:   James S. Bridges
                      Risk Reduction Engineering Laboratory
                      U.S. Environmental Protection Agency
                      Cincinnati, Ohio 45268
                                 ABSTRACT

    The Waste Reduction Evaluations at  Federal Sites  (WREAFS) Program
focuses on waste minimization research opportunities and technical
information transfer within the Federal community.  Many Federal
activities are implementing waste reduction plans and programs to reduce
the generation of wastes.  For example, an overall goal within the
Department of Defense is to reduce hazardous waste generation a total of
50% by 1992 compared to 1985.  There are several operation and process
opportunities for pollution prevention practices at Federal sites which
will reduce wastes and result in providing cost avoidance or cost
reduction as well as increased productivity, and  environmental and human
health benefits.  Waste minimization technologies can be applied
throughout the Federal community in fabrication, production and/or
maintenance processes depending on the waste generating function.
Industrial activities which support the Federal community through in-house
or contract efforts, whether on-site or off-site, are generally the same
activities practiced within the private sector.  The application of waste
minimization research may vary depending on the production or consumption
function; however, waste reduction techniques and technologies will work
in both the Federal or private sectors.  WREAFS Program objectives
include:  conducting waste minimization workshops, developing technical
information transfer opportunities between industry and government;
providing EPA research information to the Federal community; performing
waste minimization opportunity assessments at selected Federal sites;
demonstrating waste minimization techniques or technologies at Federal
facilities; and promoting waste minimization within all portions of the
Federal community.
                                    558
-------
        DEMONSTRATE COMPUTER ASSISTED ENGINEERING (CAE)
           TECHNIQUES FOR REMEDIAL ACTION ASSESSMENT
          by:  P.R. Cluxton, W.G. Harrar, L.C. Murdoch,
               and M. Beljin
               University of Cincinnati
               Department of Civil & Environmental Engineering
               Cincinnati, Ohio
                            ABSTRACT
     A  computer  workstation   dedicated   to  remedial  action
assessment  is being  developed.   The  system  is  composed  of
several  off-the-shelf  software  and  hardware  modules,  with
software  development  limited  to   the   creation   of  utility
programs  used to  transfer data  from one  software module  to
another.   The component  modules  include a Geographic  Infor-
mation System, a  Database Management System,  a Computer Aided
Design and Drafting System, and a Groundwater  Modeling System.
The completed system will  be an example of a Computer Assisted
Engineering (CAE)  type system.

     The  purpose  of  the  project is  to  demonstrate how  the
remedial  action   evaluation   process  can  be  improved  and
expedited  thru  use  of  the  CAE  system.    The   CAE  system
capabilities include the production  of maps and cross-sections
showing the  geology,  hydrology,  and  distribution of  contam-
inants; the  calculation of volumes  or masses  of  contaminated
material;   the modeling  of  groundwater   flow  and  contaminant
transport; and the modeling of remedial actions such as cut-off
walls or pumping.   Several site studies have been completed for
the regional offices,  including the  Queen City Farms  Superfund
site and  adjoining Cedar Hill Landfill in Seattle,  Washington,
a  soil gas  survey in Croydon,  Pennsylvania  and  groundwater
contamination  mapping  in  the  vicinity  of  the  Taylor  Road
Landfill in Hillsborough County, Florida.
                               559
-------
             Vi.VilAULIC MECHANISMS OF A MULTIPLE SOIL LAYER COVER
    by: Ati hot (s)
       Ric ard C. Warner
       University of Kentucky
       Lex ngton, KY 40546-0075
James A. Harned
Greiner Engineering Services
Tampa, FL 33609-3416
Nathaniel Peters, III
Law Environmental
Louisville, KY 40220
                                       ABSTRACT

    Invest  ations of the hydraulic  mechanisms of a field scale cover, with  a  clay barrier,
were collided under both stable and subsidence conditions.  Temporal water movement as
influenced by precipitation and  evapotranspiration were documented.  These investigations
encompassed:  11) runoff, drainage,  leachate and moisture migration throughout  the complete
cover,  (2j  a  variety  of  field   scale  hydraulic   conductivity  testing  procedures,  (3)
determination of tracer breakthrough times, (4) modeling of preferential flow paths, and (5)
determination of clay cap subsidence as a function of a cavity formed beneath the cap.
    The research was conducted  on  a 27.4 m long by 6.1 m wide cover consisting of a topsoil
layer,  sand  drainage  layer,  and  compacted  clay  barrier layer.    The  clay  barrier  was
constructed in 5 lifts at  a moisture  content 2 to 3% above optimum with regard to Standard
Proctor.  The hydraulic  conductivity of the clay cap was examined using  sealed double  ring
infiltrator (SDRI) and borehole methods.  The SDRI tests yield consistent values  of 7.9 x
10"   and  8.9 x 10"  cm/s with an average  standard error  p^  9.0  x   10"   cm/|.    The
borehole   lethod  yield  hydraulic   conductivities  of  2  x  10   and  8.0  x  10    cm/s,
respective1:    A  ;00 day  breakthrough experiment was conducted using a conservative anionic
tracer ami  >ui;i were compared to the SOILINER model.  After 90 days  the tracer initially
appeared   ,-ve  Background levels.  This time frame is analogous to the tracer replacing the
pore voiuir s of the largest macropores prior to being discharged.  Over  the next 110 days
the effluent increased to within 80% of the initial tracer concentration.
    Througi a continuous stirred tank reactors (CSTR) modeling analysis it  was concluded
that prefer ntial flow paths, representing only 6.2% of the total pore volume, conveyed 95%
of the totai  volume of tracer through the barrier layer.   CSTR modeling determined that
complete  tr cer breakthrough would take 620 days whereas the SOILINER Model predicted 5.76
years.
    Research  on the influence of a  known cavity size created  directly beneath  the multiple
soil  layer eover is  continuing  this  spring.   Preliminary results  indicate that  a clay cap,
constructed 1 to  3% above optimum Proctor moisture content and subjected to an overburden of
water and saturated sand, collapsed above a 1.57 m diameter cavity.
    Previoii, centrifuge experiments concluded that potential subsidence would decrease  due
to increased clay cap thickness, construction at moisture contents near Proctor optimum,  and
increased dry density.   Thus the selection of the  water content, used in  compaction of the
clay  layer,  appears  to  present   a tradeoff   between  decreasing  clay cap's  hydraulic
conductivity  and the structural  capability  of  the clay layer  to  span a  given size cavity
created beneath the cap.
                                          560
-------
                  Evaluation of Solidification/Stabilization
                         Treatability Studies at the
                Uni'ted States  Environmental  Protection Agency
                             Center Hill  Facility

                                      by

                             Edwin  F.  Barth,  P.E.
                United States  Environmental  Protection Agency
                    Risk Reduction Engineering Laboratory
                                Cincinnati,  OH

                                   Abstract
 Solidification/Stabilization  is  being  considered  as  either  a  treatment or

 post treatment  option  for  several  uncontrolled  hazardous waste  sites

 containing contaminated  soil.  The up  front  prediction  of the feasibility of

 this treatment  process is  complicated  because of  the many possible chemical

 interactions  between the soil, site contaminant,  and binder.  RREL has been

 active  in performing solidification/stabilization treatability  studies at

 the Center Hill  Facility as well as reviewing treatability  studies from

 other sources.   The following general  technical issue areas have been common

 in the  studies  evaluated:

                 * feasibility of stabilizing organic waste
                 * long term durability testing
                 * placability of material
                 * volatile emissions during mixing and curing
                 * ability  to mix in field
                 * construction quality assurance

A preliminary testing evaluation process has been developed for determining

the feasibility of a solidification/stabilization process for contaminated

soil.   Results of the treatability studies reviewed have been favorable for

certain contaminated matrices and less favorable for others.
                                   561
-------
  THE EPA MANUAL FOR WASTE MINIMIZATION OPPORTUNITY ASSESSMENTS
             by:  Mary Ann Curran, Harry Freeman,
                  Kenneth R. Stone
                  U.S. EPA, RREL
                  Cincinnati, Ohio 45268
                             ABSTRACT

     A waste minimization assessment (WMA) is a systematic
planned procedure with the objective of identifying ways to
reduce or eliminate waste.  The assessment consists of a careful
review of a plant's operations and waste streams, and the
selection of specific areas to assess.  After a specific waste
stream or area is established as the WMA focus, a number of
options with the potential to minimize waste are developed and
screened.  The technical and economic feasibility of the selected
options is then evaluated.  Finally, the most promising options
are selected for implementation.

     Waste minimization is fast gaining recognition as a means of
contending with the nation's hazardous waste problem and other
forms of environmental pollution.  Opportunities exist for waste
minimization throughout industry and government.  The waste
minimization assessment procedure described^in the manual offers
a means of determining a facility's waste situation and
identifying and evaluating potential viable options for reducing
waste.

     The waste minimization  assessment offers opportunities to
reduce operating costs, reduce potential liability, and improve
the environment, while improving regulatory compliance.  The WMA
results in a careful review  of a plant's operations toward
reducing wastes.
                                562
-------
                   THE U.S. EPA COMBUSTION RESEARCH FACILITY
                   by:   Johannes W. Lee, Larry R. Waterland
                         Acurex Corporation
                         Jefferson, Arkansas  72079

                         Robert C. Thurnau
                         U.S. Environmental Protection Agency
                         Cincinnati, Ohio  45268
                                   ABSTRACT

     During FY'88, the EPA Combustion Research Facility  (CRF)  in Jefferson,
Arkansas conducted 58 incinerator tests over 26 weeks. Completed test programs
include: demonstration of the American Combustion,  Inc.  Pyretron Thermal
Destruction System under the Superfund Innovative Technology Evaluation  (SITE)
program; tests to evaluate the use of SFg as a surrogate for destruction and
removal efficiency in liquid injection incineration; evaluation of the fate of
hazardous trace metals fed to the rotary kiln incineration system; and tests
to evaluate the valence state of chromium discharges from the  rotary kiln
incinerator. The presentation focuses on the SFg and trace metal test results.

     During the surrogate-tests, SFg DREs ranged from 99.9 to 99.99999 percent
and POHC DREs ranged from 99.993 to 99.999997 percent. SFg ORE were uniformly
lower than the corresponding POHC DREs and increased with increased combustor
temperatures. SFg and POHC DREs showed no clear dependence on flue-gas 02.

     The trace metals tests studied the fate of arsenic, lead, cadmium,
chromium and barium (appendix VIII) and copper, magnesium, bismuth, and
strontium when these are subjected to incineration. Except for arsenic, metals
discharge distributions were consistent with their volatilities. Cadmium,
lead, and bismuth appeared to be volatile and were found mainly in the flue-
gas.  Barium, copper, strontium, arsenic, chromium, and magnesium appeared to
be refractory and were found mainly in the kiln-ash. Presence of chlorine in
the waste shifted cadmium, lead, copper and bismuth from the kiln-ash to the
flue-gas but did not affect arsenic,  lead,  chromium, barium, magnesium and
strontium.  Neither kiln nor afterburner temperature within the range tested
affected metal discharge distributions significantly. The venturi/packed
column scrubber-system metals-removal efficiency ranged from 31 to 88 percent
and averaged at. 57 percent and were lower for the volatile metals.

     In addition,  reports for several FY'87 test programs were issued.  Several
facility upgrade efforts were completed.  A major facility reconstruction was
initiated to expand the CRF's physical plant and capabilities.
                                     563
-------
            EVALUATING THE COST EFFECTIVENESS OF SITE TECHNOLOGIES

             By:  Gordon M. Evans
                  USEPA, RREL
                  Cincinnati, Ohio
      The goal of the Agency's Superfund Innovative Technology Evaluation
(SITE) program is to develop reliable performance and cost data of unique and
commercially available hazardous waste destruction and treatment technologies.
One major challenge which faced the SITE program was how best to insure that
the cost evaluation process produced estimates which were useful to the
Superfund decision making process.  Toward this end, broad objectives were
established to guide project managers in the construction of cost evaluations
for innovative technologies.  As they have evolved, the goals of SITE
program's cost analysis are:

      1)  To provide the Agency decision makers with the developer's cost
      analysis as well as an independent cost analysis, prepared by the
      Agency, and based on well defined and appropriate operating assumptions.

      2}  To identify and highlight all critical operating and economic
      variables which the engineering analysis suggests are likely to have
      significant impacts on costs.

      3)  To insure that Agency decision makers are provided with the
      opportunity to reconstruct any cost analysis, substituting assumptions
      which are more consistent with their own situation. This in turn means
      that all assumptions must be clearly stated and all equations clearly
      presented.

      This poster will first review the specific legislative and program goals
guiding the SITE program cost evaluations.  Then it will examine the twin
problems imposed on these cost evaluations by: 1) the need to balance the
goals of the public and private sector in the marketplace;  2) the R&D aspects
of demonstrating innovative technology.  The poster will describe how the SITE
program has organized itself to facilitate the attainment of the goals,
mentioned above.  An economic framework will be presented which segregates
estimated costs into one of 12 categories.  Lastly, the poster will examine
the results of completed SITE cost analyses.
                                    564
-------
                         SOLIDITECH  SITE  DEMONSTRATION

                by:  Walter E. Grube, Jr.
                     U.S. Environmental  Protection Agency
                     Cincinnati, Ohio  45268

                     Kenneth 6. Partymiller
                     PRC Environmental Management, Inc.
                     The Woodlands,  Texas  77380

                     Danny R. Jackson
                     Radian Corporation
                     Austin, Texas  78720
                                   ABSTRACT

      The Soliditech SITE technology demonstration was carried out to  obtain
reliable process performance and cost information.  The demonstration  took
place at the Imperial  Oil Company/Champion Chemicals Superfund Site in
Monmouth County, New Jersey.  Contamination
various metals, and petroleum hydrocarbons.
waste material with Urrichem, a proprietary
materials (portland cement was used in this
batch-type concrete mixer.
at this site includes PCB's,
 The Soliditech process mixes the
reagent; additives; pozzolanic
demonstration); and water; in a
      Technical criteria applied to evaluate the effectiveness of the
Soliditech process include contaminant mobility, based upon solute extraction
and permeability tests; and the retention of structural  integrity of the
solidified material, based upon physical  and morphological  tests.

      Extensive sampling and analyses of three different waste types from the
New Jersey site showed (1) a reduction of heavy metals in the TCLP extract of
solidified waste samples, (2) no volatile organic compounds in the TCLP
extract of solidified wastes, (3)  detectable levels of phenols and cresols in
                                   565
-------
the TCLP extract of treated waste, (4) an increase in oil  and grease in
extracts of solidified waste, and (5) good structural stability of solidified
waste under freeze/thaw, wet/dry, and static testing.

      Long-term testing for extractable pollutant solutes, Teachable solutes
mobile through diffusion, morphological  changes in the concrete-like matrix,
and structural properties are planned to continue for an additional  4-1/2
years.
                                    566
-------
                 BIOTROL SOIL WASHING SYSTEM

                    by:  Steven B. Valine
                         BioTrol, Inc.
                         Chaska, Minnesota 55318
                           ABSTRACT

     BioTrol, Inc.  (Chaska, Minnesota) has developed an on-site
treatment system for the remediation of contaminated soils from
wood preserving  sites.   The  system  is based on  a series of
intensive scrubbing and physical separation steps using water
as  a carrier  for  the  soil  and  contaminants.    Contaminant
removal efficiencies of 90  to 95  percent have been achieved.
Process water from the system is treated in a fixed film bio-
reactor using an amended bacterial consortium prior to recycle
back to the  soil scrubbing system.    This  paper presents the
results of two phases of  pilot testing recently completed at
a wood preserving site in Minnesota using a mobile, 500 Ib/hr
pilot-scale unit.
                             567
-------
                   SEPARATION OF HAZARDOUS ORGANICS BY
                LOW PRESSURE REVERSE OSMOSIS MEMBRANES

                          M.E. Williams and D. Bhattacharyya *
                          Department of Chemical Engineering
                                University of Kentucky
                                 Lexington, KY 40506

                                         and

                            Richard P. Lauch, Project Officer
                          U.S. Environmental Protection Agency
                         Risk Reduction Engineering Laboratory
                                 Cincinnati, OH 45268
                                     ABSTRACT

      Membrane processes are gaining considerable attention for the separation of organics and
inorganics from hazardous wastes. Membrane processes can be divided into four categories:
high pressure RO, low pressure RO, nanofiltration, and ultrafiltration. Low pressure processes
with thin-film composite membranes have definite advantages in terms of energy savings and
capital cost. Extensive experiments were conducted with two types of low pressure aromatic
polyamide membranes.  Membrane experiments were carried out with various priority organic
pollutants with and without partial ozonation. Partial ozonation converts many chlorinated
organics to intermediate organic acids. These ionizable intermediates were effectively removed
by charged and uncharged membranes without significant drop in water flux.  Charged reverse
osmosis membranes provide additional advantages because high water flux (11 x 10"4 cm/s) can
be obtained even at low pressures (1.4 MPa).


 INTRODUCTION

      Chemical manufacturers generate millions of tons of wastes containing various hazardous
priority pollutants each year. The development of alternative technologies for the treatment of
various hazardous wastes is becoming increasingly important as concerns grow over its disposal.
These and other wastes, such as leachate from unsecured disposal sites, contain a wide variety of
priority pollutants such as pesticides, herbicides, PCBs, chlorinated hydrocarbons, and heavy
metals.  Much of this waste is relatively dilute and so must be concentrated before further
treatment (1,2).

      Several methods have been used for the treatment of dilute wastewater.  These include
biological treatment, stripping, and carbon adsorption  (3, 4).  Ozonation has also been found to
be effective in oxidizing some hazardous organics (5,6) to less toxic compounds. Membrane
processes can be used to purify wastewater and produce a 20 to 50 fold decrease in waste
volumes that must be treated with other processes such as incineration or wet air oxidation,
greatly reducing energy cost.
* corresponding author
                                        568
-------
 MEMBRANE CONCEPTS AND APPLICATIONS

      Membrane processes are gaining considerable attention for the purification and volume
reduction of dilute hazardous wastes.  The development of low pressure reverse osmosis (RO)
membranes such  as aromatic polyamide and sulfonated polysulfone has resulted in membrane
processes which give high water flux at low pressures (< 2 MPa).  These thin-film composite
membranes can provide high solute separations and have definite advantages in terms of energy
savings, capital cost, and broad pH (2 to 12) operating range. The low pressure membranes can
provide simultaneous separation of hazardous organics and inorganics.  Selective separations of
ionizable compounds are also possible with membranes containing charged groups (7, 8).

      Membrane performance is measured in terms of membrane rejection (R), permeate water
flux (Jw), and extent of water recovery (r).  The rejection is a measure of solute separation by the
membrane and is defined as
                                             C
where C and Cs are the solute concentrations in the permeate and feed streams.

      Reviews of the development of membrane technology can be found in Sirkar and Lloyd
(9), Sourirajan and Matsuura (10), Lee (11), and Belfort (12). Although many of the applications
referred to in these involve the use of cellulose acetate membranes, works with thin-film
composite membranes were more favorable for the separation of hazardous wastes. Chian (13)
used reverse osmosis systems to remove more than 99 % of fifteen major pesticides.
Carcinogenic substances were removed 92.5 % with both spiral wound and hollow fiber
polyamide membranes in studies by Light (14). Shrem and Lawson (15) used reverse osmosis
units to treat wastewater from an organic chemical manufacturing plant and found organic matter
rejection over 90 %. Bhattacharyya et al. (16, 17) have done extensive work with various
priority organic pollutants using thin-film aromatic polyamide (FT30) spiral wound modules.
Experiments with priority pollutants showed that rejections of >98 % for PAH compounds
(napthalene, anthracene, phenanthrene) were possible with little drop in permeate water flux.
For ionizable organics such as phenol, chlorophenols, and nitrophenols, they found that
rejections and flux drops were highly dependent on operating pH values. Membrane rejections
(atpH 11) were 99.5-99.8 % for phenol, 2-chlorophenol, 2,4-dichlorophenol,
2,4,6-trichlorophenol, etc.

      Studies have shown that high separations of hazardous organics such as chlorinated
phenolics can be achieved by low pressure polyamide membranes. However, these studies have
indicated substantial water flux drop for non-ionized chlpro and nitrophenols due to
membrane-solute interactions (17). It is known that partial ozonation will convert many
chlorinated organics to intermediate organics acids (oxalic, formic, etc.) (5, 6, 18). These
organic acids do not interact as strongly with the membrane and so cause less flux drop. Also,
since these intermediates are ionizable, charged membranes can be utilized to achieve high water
flux and separation of partially ozonated hazardous organics at low pressures.


 OBJECTIVES

      This work deals with the use of low pressure, thin-film, composite membranes for the
concentration and separation of selected chlorophenols and chloroethanes with and without feed
pre-ozonation. Membrane feed pre-ozonation will result in the formation of organic acid
intermediates which should result in less flux drop, and, since these compounds are ionizable,
high separations by charged as well as uncharged membranes. Separation and flux
characteristics of both charged and uncharged membranes were studied with both ozonated and
                                          569
-------
non-ozonated solutions to evaluate the effectiveness of the ozonation - membrane process for the
enhanced removal of priority organic pollutants.


EXPERIMENTAL

      Membrane experiments were conducted with a system (Figure 1) containing both a batch
and a continuous flow cell so that the performance of two different membranes could be studied
at the same time. The thin-film, composite membranes used in the experiments were the Film
Tec NF40 and the FT30-BW; the NF40 was placed in the batch cell and the FT30 in the flow
cell. The pump shown in the system served to provide flow of solution through the continuous
cell and mixing in the batch cell; compressed nitrogen was used to supply the pressure driving
force for the system. The operating conditions were: system pressure (P>VB) of 0.3 -1.4 MPa,
feed pH*s of 3.0 - 9.4, and system temperature of 24° C.
avg •
     The procedure used for experiments involving ozonation - membrane process is outlined'in
Figure 2. Membrane feed solutions were ozonated in a 1.8 liter stirred reactor with a flow of
0.20 standard liters per minute (SLPM) O2 containing 2 % ozone. Pre-ozonation times ranged
from 0 to 60 minutes. After ozonation, solutions were mixed for several hours to allow
decomposition of residual ozone. Membrane experiments were then carried out with the
ozonated solutions.

     Membrane performance was measured in terms of flux drop from that of distilled water
flux (DWF) and solute rejection. Membrane feed, concentrate, and permeate samples were
analyzed by Total Organic Carbon and HPLC direct injection.  Total Organic Carbon (TOC) was
measured using a Beckman Model 915-B Carbon Analyzer.  HPLC analysis (phenolics) were
performed with a Varian 5000 liquid chromatograph using a MCH-5 column (reverse phase
octyldecylsilane on silica) and a UV-50 variable detector at 220 and 280 nm.


RESULTS AND DISCUSSION

     Membrane separation of selected hazardous organics with and without feed pre-ozonation
were investigated with the NF40 and FT30-BW membranes. Single component studies were
conducted with trichlorophenol while multicomponent systems examined consisted of
trichlorophenol/humic acid mixtures and mixtures of chlorophenol, dichlorophenol,
trichlorophenol, trichloroethane, and tetrachloroethane.  A wide range of ozonation times and
pressures were studied, and, since many of the compounds studied were ionizable, several
membrane feed pH values were investigated.  Membrane performance results are presented in
terms of solute rejection and flux drop. The % flux drop (at a particular pressure) is defined as:
„   ,      Distilled Water Flux- Flux with Wastewater   inn
Flux drop =	^.  .„  ,„,	—	x 100.
                      Distilled Water Flux

     Membrane stability was checked by standard sodium chloride and sodium sulfate
rejections and distilled water fluxes. For the NF40, a negatively charged membrane made of
carboxylated aromatic polyamide, the DWF at 1.38 MPa was found to be 11 x 10"4 cmVcm2 s
(23.3 gal/ft2 day).  Sodium chloride and sodium sulfate rejections were 30 % and 97 %,
respectively. The DWF of the FT30, an aromatic polyamide membrane, was also 11 x 10"4 cm/s
at 1.38 MPa; sodium chloride and sodium sulfate rejections for this membrane were 96 % and 97
%.  The low sodium chloride rejection of the NF40 illustrates the principal advantage of the
charged membrane over reverse osmosis membranes such  as the FT30. The FT30 membrane
gives high rejections of most solutes, whereas the NF40 membrane can be used to selectively
                                         570
-------
separate compounds with different charges.
 Studies with Trichlorophenol

     Membrane feed solutions of 50 mg/L trichlorophenol (TCP) were ozonated from 0 to 60
minutes. The pH of the solutions decreased after ozonation; this drop in pH indicated the
degradation of the TCP to intermediate organic acid compounds.  HPLC analysis indicated that
after 5 minutes of ozonation the TCP concentration was reduced to 24.7 mg/L, and after 15
minutes, was <0.2 mg/L.  Formation of carbon dioxide during the ozonation was indicated by
reduction in TOC and the presence of carbon dioxide in the high pH trap placed after the
ozonation reactor.

     Figure 3 shows flux drop as a function of ozonation time for the NF40 membrane for low
(3.3-3.6) and high (8.8-9.4) pH membrane feeds. At low feed pH, TCP is not ionized and so
interacts strongly with the membrane; the adsorbed TCP displaces water in the membrane pores,
causing significant water flux drop. Ozonation of TCP results in the formation of organic acids
which do not interact as strongly with the membrane and so cause less flux drop. At the high
feed pH the TCP is ionized; the negatively charged solute does not interact as strongly with the
charged membrane and the flux drop is less than that at the lower feed pH.  Since the ozonation
products are also ionizable, the flux drop for ozonated solutions are also smaller under the high
pH conditions.

     TOC rejections for the NF40 membrane are shown in Figure 4.  It can be seen that both an
increase in feed pH and ozonation time increased rejections. An increase in feed pH results in
the ionization of the TCP or the organic acid intermediates formed during ozonation.  These
negatively charged species are more highly rejected by the negatively charged NF40 membrane.
Ozonation improves rejection since the organic acids that are formed ionize at lower pH's than
does TCP and so these are rejected better than TCP at lower feed pH's.

     Figures 5 and 6 show the flux drop and TOC rejection for the FT30 membrane. Under
non-ionized conditions (feed pH 3.3-3.6), the TCP interacts strongly with the membrane, causing
a large flux drop. The drop in water flux is greater than that for the NF40; the charge on the
NF40 weakens the interactions between the TCP and the membrane. As with the NF40,
ozonation reduces flux drop for the FT30 membrane by reducing solute interactions with the
membrane. Under ionized conditions (feed pH 8.8-9.4), the TCP and organic acids formed
during ozonation do not interact with the FT30 membrane as strongly as under non-ionized
conditions and so the flux drop is smaller. Figure 6 shows that feed pH and ozonation did not
greatly affect TOC rejection for the FT30 membrane; this membrane does not depend on charge
for separation as does the NF40 membrane.


 Studies with Trichlorophenol/Humic Acids

     Experiments to determine the effect of water recovery on water flux and solute rejection
were conducted with mixtures of 50 mg/L trichlorophenol and 10 mg/L humic acids (TCP/HA).
Humic acids are high molecular weight compounds that are present in  soils and so can be found
in ground water containing hazardous organic leachates.  These compounds are highly rejected
by the membrane but can cause large water flux drops due to adsorption.

     The fluxes for distilled water, TCP/HA, and TCP/HA ozonated for 30 minutes are given in
Figure 7 as a function of water recovery for the NF40 membrane. For both non-ozonated and
ozonated TCP/HA, no drop in flux was observed even at high recoveries; the TCP/HA and the
                                         571
-------
 organic acids formed during ozonation were ionized at the high feed pH (9.1-9.3) and so little
 change in water flux occurred. The TOC rejections are shown in Figure 8. Rejection of the
 non-ozonated TCP/HA decreased from 80 % at 10 % recovery to 64 % at a recovery of 80 %;
 however, for the ozonated solution 84 % TOC rejection was maintained even at 75 % water
 recovery. The decrease in TOC rejection for the non-ozonated TCP/HA could have been the
 result of enhanced concentration of the HA at the membrane surface. Since degradation of the
 HA is expected during ozonation, HA concentration on the membrane surface did not increase
 for the ozonated solution, and TOC rejection remained high.


  Studies with Mixtures

      Studies were performed with mixtures of 50 mg/L chlorophenol (CP), dichlorophenol
 (DCP), and trichlorophenol (TCP) with 100 mg/L trichloroethane (TCE) and tetrachloroethane
 (TTCE) to determine the effect of multicomponent systems on flux and TOC rejections for
 ozonated and non-ozonated solutions. The flux behavior of the mixture with the NF40 was
 found to be linear over the pressure range 0.34 -1.38 MPa (50 -200 psi), indicating the absence
 of surface polarization phenomena. Flux drops as a function of pH for the NF40 membrane at
 1.38 MPa are shown in Figure 9. Flux improves substantially for increase in feed pH for both
 the ozonated and non-ozonated mixtures; ozonation did not greatly improve flux drops found for
 the mixture at a fixed feed pH. However, Figure 10 shows that TOC rejection is enhanced
 significantly for the mixture after ozonation for 60 minutes. While the non-ozonated mixture
 rejection is almost constant for the different feed pH's, it is increased to as high  as 80 % for the
 ozonated mixture. The increase is due to the formation of organic acids which ionize and are
 rejected by the membrane. Although the phenolics in the mixture are ionizable, the pKa of these
 compounds are much higher than those of the intermediates, and also, the chloroethanes present
 are not ionizable.  As a result, the rejection did not increase over the pH range studied. However,
 the organic acids formed after ioiu'zation have much lower pKa's and so are rejected by the
 charged membrane.

      The flux behavior of the mixture with the FT30 membrane was also found to be linear over
 the pressure range studied. FT30 membrane flux drops and rejections are shown in Figures 11
 and 12 for an operating pressure of 1.43 MPa. Flux drop is also a strong function of feed pH for
 this membrane, and feed pre-ozonation does improve flux drop over the range of feed pH's
 studied.  As  with the single component TCP system, ozonation produces the intermediates that
 do not interact with the membrane to same extent as the mixture and so flux is enhanced. TOC
 rejection slightly increases with ozonation and feed pH.


  Overall Removal of Trichlorophenol

      Figure 13 illustrates an example calculation of an ozonation - membrane process for the
 separation of hazardous organics; trichlorophenol is used as the model compound. After
 ozonation for 30 minutes the TCP concentration would be reduced to <0.2 mg/L; the TOC of the
 feed solution would be 66 mg/L due to the formation of carbon dioxide during the ozonation.
 For a membrane that rejects 90 % (either the NF40 or FT30), the permeate contains only 6.6
 mg/L TOC which would be due primarily to organic acids; the TCP concentration in the
permeate would be less than 0.2 mg/L.  For 90 % water recovery, the concentrate TOC,  also
mostly due to organic acids formed during ozonation, would be 1194 mg/L. This greatly
reduced volume could be disposed of by incineration. The overall ozonation - membrane
process would produce permeate water of high quality and greatly reduce the volume and risk of
waste that must be further treated.  Also, since the process has been shown to be effective when
utilizing charged membranes, selective separation of hazardous organics from feeds containing
                                         572
-------
high salt concentrations would be possible, allowing high fluxes at low operating pressures even
for feeds with high osmotic pressures.


CONCLUSIONS

      This study has shown that thin film, composite membranes can be used effectively for the
separation of selected hazardous organic compounds. This waste treatment technique offers
definite advantages in terms of high solute separations at low pressures (<1.4 MPa) and broad
pH operating range, and the use of charged membrane would allow the selective separation of
some organics from feeds containing high salt concentrations. Li addition, feed pre-ozonation of
selected organics has been shown to improve flux and rejection characteristics for both charged
and uncharged membranes due to the formation of ionizable organic acid intermediates during
the ozonation that do not interact as strongly with the membrane. It has been shown that the
ovejall ozonation - membrane process could be greatly effective in producing permeate water of
high quality while minimizing the volume of waste that must be further treated.


REFERENCES

 1.  "Industrial Hazardous Waste Management", Industry and Environment Special Issue - No.
      4(1983).

 2.  "Land Disposal of Hazardous Wastes", EPA Report, EPA-600/9-84-007 (1984).

 3.  Shuckrow, A.J., Pajak, A.P., and Osheka, J.W., "Concentration Technologies for
      Hazardous Aqueous Waste Treatment", EPA Report, EPA-600/2-81-019 (1981).

 4.  S.mith, J.K., "Laboratory Studies of Priority Pollutant Treatability", EPA Report,
      EPA-600/2-81-129 (1981).

 5.  Rice, R., and Browning, M., "Ozone for Industrial Water and Wastewater Treatment", EPA
      Report, EPA-600/2-80-060 (1980).

 6.  Baillod, C.R., Faith, B.M., and Masi, O., "Fate of Specific Pollutants During Wet
      Oxidation and Ozonation", Environmental Progress, 3, No. 1, pp. 217-227 (1982).

 7.  Bhattacharyya, D., Adams, R., and Williams, M., "Separation of Selected Organic and
      Inorganic Solutes by Low Pressure Reverse Osmosis Membranes" in "Synthetic and
      Biological Membranes", Alan R. Liss (1989).

 8.  Eriksson, Peter, "Nanofiltration Extends the Range of Membrane Filtration",
      Environmental Progress, 7, No. 1, pp. 58-62 (1988).

 9.  Sirkar, K.K., and Lloyd, D.R., "Membrane Materials and Processes for Separation", AIChE
      Symposium Series, 84. No. 261 (1988).

10.  Sourirajan, S., and Matsuura, T., "Reverse Osmosis/Ultrafiltration Process Principles",
      National Research  Council Canada, Ottawa (1985).

11.  Lee, E.K., "Membranes, Synthetic, Applications", Encyclopedia of Physical Science and
      Technology. 8, pp. 20-55 (1987).
                                          573
-------
12.  Belfort, G., "Synthetic Membrane Processes: Fundamentals and Water Application",
     Academic Press Inc., New York, New York (1984).

13.  Chian, E.S.K., Bruce, W.N., and Fang, H.H.P., "Removal of Pesticides by Reverse
     Osmosis", Environmental Science and Technology. 9 (1975).

14.  Light, W.G., "Chemistry and Water Reuse", 1, Ed. WJ. Cooper, Ann Arbor Science, Ann
     Arbor, Michigan (1981).

15.  Shrem, M., and Lawson, T.,  Industrial Water Engineering, 14 (1977).

16.  Bhattacharyya, D., and Madadi, R., "Separation of Phenolic Compounds by Low Pressure
     Composite Membranes: Mathematical Model and Experimental Results" in AIChE
     Symposium Series, 84> No.  261 (1988)

17.  Bhattacharyya, D., Barranger, T., Jevtitch, M., and Greenleaf, S., "Separation of Dilute
     Hazardous Organics by Low Pressure Composite Membranes", EPA Report.
     EPA/600/S2-87/053 (1987)

18.  Wang, Y., Pai, P., and Latchaw, J., "Effects of Preozonation on the Methanogenic Toxicity
     of 2.5-Dichlorophenol". Journal WPCF, 61, No. 3, pp. 320-326 (1989).
                                         574
-------
                                                        8
                                                CD

                                                E
                                               •c
                                                CD
                                                CL
                                                X
                                               UJ

                                                CD


                                                I
                                               JD

                                                CD
                                               CD
                                               CO

                                               "o
                                               CB
                                               E
                                               CD

                                               O
                                               CO
                                               CD

                                               3
                                               O)
575
-------
UL
O
X
o
H-
O
CO

u.
o

                                                                                      c
                                                                                      o
                                                                                      E
                                                                                      CL
                                                                                      X
                                                                                      LJJ

                                                                                      0>


                                                                                      I
                                                                                      J3

                                                                                      e
                                                                                      o

                                                                                      s
                                                                                       I
                                                                                      c
                                                                                      _o

                                                                                      "cO
                                                                                      c.
                                                                                      o
                                                                                      N
                                                                                      O
                                                                                      01

                                                                                      O

                                                                                      O

                                                                                      O
                                                                                      c
                                                                                      g


                                                                                      'c
                                                                                      o
                                                                                      N
                                                                                      O
                                                                        _o
                                                                        V-*
                                                                         CO

                                                                         E
                                                                         
-------
  20

  IS

  IS

  14

a.12
2

I10
5  81
                                                                     10
                         System: NF40 Membrane
                                 50 me/L TCP
                                 DW = 10.8 x 10~4cm/!
            Membrane
             FeedpH
            °3.3 - 3.6
            0 8.8-9.4
              M""l""l""l"
 Figure
     0  5  10  15  20 25  30 35  40 45  50 55  60 65
                   Ozonation Time (min)

     3.   Effect of Ozonation on Flux Drop for the NF40
         Membrane with Trichlorophenol.
   100
                            System: NF40 Membrane
                                    50 mg/L TCP
    20 i

       0   5  10 15  20 25  30  35 40  45 50  55 80  65
                    Ozonation Time (min)

Figure 4.   Effect of Ozonation on TOC Rejection for the NF40
           Membrane with Trichlorophenol.
                           577
-------
                                                                   11
  30-
  25-
  20-
o.
2
x
  15-
  10-
System:  FT30 Membrane
        50 mg/L TCP
        DWF =10.5 x 10~4 cm/a
        PaTg=I.43MP.
        Temp. = 24°C

        Membrane
         Feed pH
        n 3.3-3.6
        0 a.a - 9.4
     inn [muni HI in|i mi" in ii "i"" i" "i ""i"" i1 "M"
     0  5  10  15 20  25 30  35 40  45  50  55  60 65

                   Ozonation Time (min)

Figure 5.  Effect of Ozonation on Flux Drop for the FT30
           Membrane with TrichlorophenoL






s
75
u
«
1
t?




100-
95'

on •
ou
85"

80-

75'

70"
65
60
55
50

3
a °
3 n n 	
O
a
3



Membrane
Feed pH
°3.3 - 3.8
o 8.8 - 9.4
System: FT30 Membrane
SO mg/L TCP
«vg~ _ *
Temp. = 24°C
111111

      0  5  10  15  20 25  30 35  40 45  50 55  60 65

                    Ozonation Time (min)

Figure 6.  Effect of Ozonation on TOG Rejection for the FT30
           Membrane with Trichlorophenol.
                            573
-------
                                                                   12
    13

    12-

    11-
 •yio-l
     8-

     7"

     8-

     5-
                                           o  a
          ° TCP/HA
          °0zon. TCP/HA (30 min.)
          System: NF40 Membrane
                  50 mg/L TCP - 10 mg/L HA
                  Feed pH = a. 1-9.3
                  Pavg = 1-3a MPa
                  Temp. = 24° C
         —i—'—i—'—i—-—i—•—i—•—i—•—i—.—i—•—r
      0   10   20   30   40   50   80   70   80   90
                        % Recovery

Figure 7.  Flux as Function of Water Recovery for the NF40
          Membrane with Trichlorophenol/Humic Acid
          Mixtures.
  100
  90-
   80-

s
=  70-j
o
V
as
o  80-1
  50-
  40"
  30 i
         °No Ozon.
         A30 Min. Ozon
          System: NF40 Membrane
                  SO mg/L TCP - 10 mg/L HA
                  Feed pH 9.1-9.3
                  P   = 1.38 MPa
                  Temp. = 24°C
          10
               —r—
               20
—r—
 30
—r
 40
                             50
                         Recovery
—1—
 so
—I—
 70
—r-
 80
—r
 90
Figure 8.  TOG Rejection as Function of Water Recovery for the
           NF40 Membrane with Trichlorophenol/Humic Acid
           Mixtures.
                         579
-------
                                                                   13
40-


35"


30-


25"
2*20-1
a
x
!15i
   ID-
    o-
  -5
                       System: NF40 Membrane
                               50 m«/L CP, DCP, TCP
                               100 mg/L TCE. TTCE
                               DTTF = 11.2 x 10 ~4 cm/a
                               P   =1.38MPa
                               Temp. = 24° C
          n No Ozon.
          A60 Min. Ozon.
      234567

                         Feed pH
                                     S     9     10
Figure 9.  Effect of Feed pH and Ozonation on Flux Drop for the
           NF40 Membrane with Hazardous Organic Mixtures.
ioo-|




_o
"a
0^
IT
a:
1
M







.
90-
80-
70'
80-

50-

40-
30-

20-
10-
n-

a No Ozon.
A60 Min. Ozon..
*
S^
/S
.-^
/
./ System: NF40 Membrane
/ SO m«A CP. DCP. TCP
/ 100 mg/L TCE, TTCE
jf P^ = 1.38 MPa
A Temp. = 24° C
Q— 	 . 	 	 	
	 a 	 a

23456789 1C
Feed pH
 Figure 10.  Effect of Feed pH and Ozonation on TOC Rejection
           for the NF40 Membrane with Hazardous Organic
           Mixtures.
                           580
-------
                                                                     14
   50

   45

   40

   35
 2
 X
 3
 b
   20

   IB

   ID

    S

    o
                         System: FT30 Membrane
                                50 mg/L CP. DCP, TCP
                                100 mg/L TCE, TTCE
                                DWF = 10.9 x 10 "*4 cm/a
                                      1-43 MPa
                                      = 84°C
           °No Ozon.
           A60 Min. Ozon.
                       587

                          Feed pH
                                                   10
Figure 11.  Effect of Feed pH and Ozonation on Flux Drop for the
           FT30 Membrane with Hazardous Organic Mixtures.
   100


   95


   90-


c  85-
_o
2
o
•S-  80-1
a:
u
°f  75"


   70-


   65-
  60 i
          ° No Ozon.
          A60 Min. Ozon.
          T"
           3
                nr
                 4
                        System: FT30 Membrane
                               50 mg/L CP. DCP, TCP
                               100 mg/L TCE, TTCE
                               Pavg=sl-43MPa
                               Temp. = 24° C
                      587

                         Peed pH
T—i—r
 9     10
Figure 12. Effect of Feed pH and Ozonation on TOC Rejection
           for the FT30 Membrane with Hazardous Organic
           Mixtures.
                           581
-------
                                                                                    15
                                      CO
                                      ;g
                                      "o
                                      co

                                  •2  o  O
                                  co  7=  u-

                                  $  §  J
                                  cr  oj =1
                                  CD  ^r   OJ
                                  o  O   c
                                  cr       *-
                                  o  >» •*
                                  O  ~  en
                                      co  ^1
                      oo

                       it
                                                      o-S

                                                      O 'O        .
                                                      CD  CO      —j

                                                      "oT „,  o ^  °^
                                                      S .2  X OJ  CO
                                                       „  C  P d   II
                                                                 co


                                                                 Q.
                                                                 CD
-3
cr
i
c
E
C

C

A














c
> £
p-
jj "?
• t.
> cc
) C£
•

CM
)
5

k



c
0
OS
cr
C3
N
O

A




) "J
) ^
| /1^
* s
•S f^
it f~
• ^"»
: 3
IS







^^^
E

o
CO-

k




_J
OJ -^ CO
E ^"^ 3
v '•a'
v^ Q_ CO
f- S i.
^- ^-





















a
c
0
i.
j:
£
0
•s
^



















3
Z
3
^
2
"•
j
>
«

















>.
CD
O
0
CD
DC
rtl
"cO
^o
°
in
C33













O OJ _j 11 OJ CO
Q 0 "OJ CD = ^
E E O E
^fc^ •• jjj* ^C 3«*
S ~ CO — P ^
tS co" "2 ^~ ^^.
<5 Q. "E1
D_ — o
*^ CO
^> _ >—
_ CO o5 ^
5j "*S CO o^
*- CJJ x-T4 Is-;
**" S OT o ^
o ^ A ^ C3J
04 *- II II "
CO 	
II > >
-• 1 1

^"
^ B
•^ 1-
co =
*- CO
CD >-
> CD
0 5
                                                                                    CO
                                                                                    CO
                                                                                    CD
                                                                                    o
                                                                                    o

                                                                                   al

                                                                                    CD
                                                                                    CD
                                                                                    i
                                                                                    cr
                                                                                    o
                                                                                    N
                                                                                   O
                                                                                    a
                                                                                    o
                                                                                    co
                                                                                   O

                                                                                   _CD
                                                                                    a.

                                                                                    co
                                                                                    x
                                                                                   LU
                                                                                    CO


                                                                                    CD

                                                                                    3
                                                                                    O)

                                                                                    ii.
T3
 CD
 CD
LL.
0
o
 0)
 £
•r-   a  °


 CD    "   O)
 E   £   E
•i   -i  O
o   >   co
                                    582
-------
                 Testing of a Leachate  Treatment System
                    Based on a Wood Degrading Fungus

                by:  John A. Glaser, Henry Tabak, and Edward Opatken
                     United States Environmental Protection Agency
                     Risk Reduction Engineering Laboratory
                     26 Martin Luther King Drive
                     Cincinnati, Ohio 45268

                     Susan Strohofer, Carol Hummel, Marge Kupferle,
                     P.V. Scarpino, and M. Wilson Tabor
                     University of Cincinnati
                     Cincinnati, Ohio 45265
                                Abstract

     The fungus, Phanerochaete chrysosporium, has been employed in the
design of a rotating disk reactor for liquid treatment.  Early studies
decolorizing Kraft waste liquor were very promising, and subsequent
applications to individual hazardous waste constituents such as
pentachlorophenol and trinitrotoluene were equally impressive.  The latter
application to hazardous waste has become the subject of an intensive
study to scale-up the early bench-scale reactors.  The research program
and status will be presented with details of projected field activities.
                                    583
-------
                   ASSESSMENT OF KPEG TREATMENT FOR PCB
                            CONTAMINATED SOILS
                by:  Alfred Kernel
                     U.S. Environmental Protection Agency
                     Risk Reduction Engineering Laboratory
                     Cincinnati, OH
                                 ABSTRACT

      The novel reagent generically named APEG (alkaline polyethylene gly-
colate) can be very effective for dehalogenation of a variety of halo aro-
matic pollutants.  The PCBs, PCDDs, PCDFs even PCTs (polychlorinated
terphenyls) can be effectively dehalogenated by APEG reagents to yield non-
toxic products.  The reagent has been shown to be effective on these
classes of pollutants in a variety of matrices from sediment to soils and
waste oils.

      Early work in 1986 with KPEG performed on PCDD and PCDF contaminated
oil (9000 gal.) in Butte, Montana demonstrated the effectiveness of the
KPEG system on these types of hazardous wastes.  Early tests of the in situ
application of KPEG on dioxin contaminated soil in Missouri demonstrated
some limitations of the reagents applicability.  The most severe drawback
for in situ KPEG application is the reagents extreme hygroscopisity.  These
early experiences led to the design of the pilot-scale chemical reactor
system for the effective use of the APEG reagent.  This reactor system was
first demonstrated on a heavily PCB-contaminated site in Guam, USA during
April-May 1988.  This pilot-scale unit was capable of treating from one to
two tons of contaminated soil per batch.  Further refinements were made to
the reactor system after careful examination and analysis of the first
eight reactor batches.  During September-October of 1988 these refinements
were employed in a second series of KPEG treatments in Guam.  Contaminated
soil containing from 500 to 2,600 ppm of PCB closely resembling Aroclor
1260 was treated with the KPEG reagent.  The only residual PCB peak
detected in the soil treatments was a tetrachlorobiphenyl, ranging from
approximately 1 ppm to nondetectable levels.  The KPEG reagent was also
applied to treating all of the contaminted tyvek clothing, gloves and boots
from the combined runs during the 1988 year.  Sampling fo the reactor con-
tents revealed the presence of tetra-, penta-, and hexa- chlorobiphenyls
but at below the one part per million range.  These results demonstrate the
capability of the KPEG reagent to perfrom chemical dehalogenation on
haloaromatics in a variety of matrices, ranging from oils to soils and
contaminated clothing.
                                       584
-------
                    STATE-OF-THE-ART FIELD HYDRAULIC
                 CONDUCTIVITY TESTING OF COMPACTED  SOILS

              by:  Joseph 0. Sal and David C. Anderson
                   K. W. Brown & Associates, Inc.
                   College Station, TX  77840
                                ABSTRACT
    The Congressionally mandated performance standard for soil liners of
hazardous waste management facilities is a hydraulic conductivity of 1 x
10.  m/s or less.  In response to this statutory requirement, the U.S.
Environmental Protection Agency has issued guidance .requiring that facil-
ities demonstrate this hydraulic conductivity in field tests.

    Hydraulic conductivity test methods currently used on soil liners
were evaluated for their ability to meet the minimum requirements for
field tests; i.e., that the test be capable of measuring hydraulic conduc-
tivities as low as 1 x 10   m/s and that the values obtained be represen-
tative of the overall soil liner.  Of the few methods capable of meeting
the minimum requirements, even fewer are both practical to use and rarely
give false low values*  Considering all the advantages of the methods
evaluated, the best, most practical, and currently available technologies
for -evaluating hydraulic conductivity are represented by large single-
ring infiltrometers and sealed double-ring infiltrometers.  If correction
factors are ngeded to bring the values obtained with single-ring devices
below 1 x 10   m/s, confirmatory tests should be conducted with sealed
double-ring infiltrometers.

    A long-term study is needed to allow a comparative evaluation of
candidate hydraulic conductivity testing devices.  A large underdrain
should be incorporated into the study to give the true overall hydraulic
conductivity value with which other values should be compared.
                                   585
-------
   REVIEW OF SOIL HASHING TECHNOLOGIES FOR SOILS  CONTAMINATED  WITH  HEAVY  METALS
by:  Carl Gutterman
     Ramjee Raghavan
     Foster Wheeler Enviresponse, Inc.
     Edison, NJ  08837
                                                 Darlene Williams
                                                 Releases Control Branch, USEPA
                                                 Edison, NJ  08837
                                    ABSTRACT

    Heavy metals are estimated to contaminate forty percent of the Superfund
sites on the National Priorities List (NPL).  Studies have shown that eight
metals (lead, chromium, arsenic, cadmium, copper, zinc, mercury, and nickel)
predominate at these sites.  These facts have led to a review of the current
status of extraction technologies employed on excavated soils.

    The most advanced extraction technologies are identified and their basic
processing steps discussed.  A review is presented of the established
industrial practices incorporated within the technologies for cleaning soils,
sludges, and sediments.  The results reported by the developers and operators
of facilities employing these technologies are discussed, and questions to be
explored with respect to application are addressed.  Recommendations are
given as to the further development of extraction technologies for excavated
soils.
                                     586
-------
                          ROTARY KILN INCINERATION

                    by:    V.A. Cundy, T.W. Lester, S. Acharya, C. Leger,
                                 A.Montestruc,G. Miller
                          Department of Mechanical Engineering
                          Louisiana State University
                          Baton Rouge, Louisiana 70803

                          A.M. Sterling
                          Department of Chemical Engineering
                          Louisiana State University
                          Baton Rouge, Louisiana 70803

                          J.S. Lighty, D.W. Pershing, G.D. Silcox
                          Department of Chemical Engineering
                          University of Utah
                          Salt Lake City, Utah 84112

                          W.D. Owens
                          Department of Mechanical Engineering
                          University of Utah
                          Salt Lake City, Utah 84112
                                       ABSTRACT


       A multifaceted experimental and theoretical program aimed at understanding rotary kiln
incinerator performance is underway.  The program involves university, industry, and
government participation. This poster session summarizes work accomplished to date in the
areas of full-scale insitu sampling from an operating rotary kiln facility, kiln-simulator
experimentation, and numerical modeling.


       Full-scale insitu measurements are obtained from the Louisiana Division rotary kiln
facility of Dow Chemical USA, located in Plaquemine, Louisiana. Summary results obtained
from experiments that were performed during continuous processing of carbon tetrachloride
and preliminary results  obtained during the batch mode processing of toluene-laden sorbent
packs are presented.  In situ measurements from a rotary kiln incinerator during steady and
transient loading are helping to clarify the local behavior inside an operating kiln. The infor-
mation is needed to develop reasonable mathematical models of the process for use in inciner-
                                      587
-------
ator evaluation and incinerator design. Ultimately, the model may be developed to the degree
that knowledge of the flame-mode chemistry is combined with knowledge of the transport rates
throughout the kiln to tailor kiln design and operation achieving optimum destruction per-
formance.  Results to date demonstrate the existence of significant spatial and temporal
variations in temperature and species concentrations in the kiln even under quasi-steady oper-
ating conditions.  The addition of turbulence air (to promote mixing) reduces, but does not
eliminate, these variations. A consequence of the enhanced mixing, however, appears to be a
reduction in processing rates in the upper regions of the kiln and an increase in the demand on
the afterburner. Data obtained from the kiln/afterburner/stack train are presented to demonstrate
these trends.

       Kiln simulator work conducted during this study has aided in providing an under-
standing of the results observed at the full-scale. An objective of the laboratory kiln simulation
work has been to determine bed evolution rates. Results clearly demonstrate the utility of the
kiln-simulator in helping to understand processes occurring at the full-scale. The importance of
kiln rotation rate (hence, bed motion) on the evolution characteristics of the studied sorbent-
contaminant  matrix has been demonstrated.  The intermittence in contaminant evolution
observed at the full-scale has been, at least in part, explained through careful interpretation at
the kiln-simulator scale. The results also demonstrate that careful experimental design, ex-
ecution of experiments and interpretation of results must be employed to yield meaningful
information from the kiln-simulator. Various species and temperature data obtained from the
kiln-simulator are presented to demonstrate these trends as well.

       The poster session also involves a presentation of the modeling efforts that have been
undertaken to predict rotary kiln performance.  Work initiated with two-dimensional modeling
efforts. Tliree-dimensional models that, although considerably more complicated, are required
to obtain even qualitatively correct predictions. Realistic bed evolution models and
fundamental chemical kinetic parameters that will be used in the three-dimensional models are
also under development  Various results from these components of the modeling effort are
shown.
                                       588
-------
BIOLOGICAL TREATMENT OF CHLOROPHENOL-CONTAMINATED 6ROUNDWATER

                   by:  Thomas J. Chresand
                        BioTrol, Inc.
                        Chaska, Minnesota  55318
                           ABSTRACT

     BioTrol, Inc. has developed a system for the treatment of
chlorinated phenols  in wastewater streams.  The  phenols are
degraded  by  microorganisms  which   are  immobilized  in  a
submerged, fixed-film bioreactor.  An  indigenous consortium of
microorganisms  is augmented  by  inoculation  of  a  specific
bacterium with capability to degrade penta chlorophenol  (PCP)
as  well  as  tetra-  and  tri-chlorophenol.     BioTrol  has
demonstrated treatment  of wastewater  containing up  to 90 ppm
PCP.  Removals from 95 to 99 percent were achieved with a one
hour  hydraulic  residence  time.    The  system is  primarily
applicable to treatment of groundwater; however,  treatment of
process and lagoon waters has also been demonstrated.
                             589
-------
         COMPUTERIZED MANAGEMENT AND DISSEMINATION
        OF INFORMATION FOR RESEARCH AND DEVELOPMENT
          OPERATIONS AT THE TECHNICAL INFORMATION
                 EXCHANGE (TIX) EDISON. NJ
        May Smith,  Francine Everson,  and Pacita Tibay
                     Enviresponse, Inc.
             6SA Raritan Depot, Woodbridge Ave.
                      Edison,  NJ  08837
                        Hugh Masters
                   Releases  Control  Branch
            Risk Reduction Engineering Laboratory
            U.S. Environmental Protection Agency
                  Edison, New Jersey 08837
                          ABSTRACT
    The Technical Information Exchange (TIX) at EPA's Edison,
New Jersey Facility was established in 1980 to  support the
research personnel at this location.  It  is operated  by a
Research and Development  (R&D) on-site contractor, currently
Enviresponse, Inc., for EPA's Releases Control Branch (RCB) of
the Risk Reduction Engineering Laboratory  (RREL).

    TIX provides a wide range of support services for RREL
research activities being conducted at the Edison Facility,
including the preparation of  slides; selection  of slides,
movies and videotapes from the extensive  library developed by

                                1
                             590
-------
RGB over the years for presentations; and dissemination of
descriptive literature.  TIX provides immediate  access to its
collection and assists  in locating or procuring materials from
other sources (e.g., EPA's Center for Environmental Research
Information, other Agency  information sources,  the National
Technical Information Service, etc.).  A library  of commercial
software is maintained  so users  may  evaluate various packages
before purchasing additional copies.

    TIX also pursues research studies of its own into the
computerized retrieval  and dissemination of this  information.
It developed and maintains the computerized On-Line
Information System (COLIS) designed to meet the needs of
several specific R&D efforts. COLIS searches  by key words to
retrieve files for display.  While COLIS is expandable to
other needs as they arise, the intent is not to compete with
broad-based, commercial types of information databases.
Currently, while not all tasks are equally advanced in their
development, COLIS is designed to service program efforts.

    These include a Case History File of hazardous materials
spills, remedial  and removal  actions for Superfund sites, and
corrective actions for underground storage tank problems; a
SITE Applications Analysis Reports program containing
performance and cost data from Superfund  Innovative Technology
Evaluation  (SITE) Program  projects;  a Library Search System
that provides descriptive records and abstracts  for documents
on RREL research  topics in  the TIX files, including a special
collection of Stormwater Research publications containing EPA
reports, technical papers, etc., developed by the Storm and
Combined Sewer Overflow Program during 1967-1982.

    In addition to a modest operating budget,  TIX is funded
from other R&D projects to support their special needs.  To
meet growing demands for technology  transfer of  RREL research
at Edison, TIX recently expanded its workshop/conference
capabilities with the addition of three modular units.
                            591
-------
                    The EPA Treatability Database
              by:  Stepahanie A. Hansen,
                   Radian Corporation
                   Milwaukee, ¥1   53214

                   Kenneth A. Dostal, and
                   Glenn Shaul
                   Environmental Protection Agency
                   Cincinnati, OH   45268
                               ABSTRACT

     The purpose of this paper will be to introduce to the public the
EPA Treatability Database.  Since the mid-1970's, EPA has accumulated
a wealth of data on the treatability of compounds found in industrial
and domestic wastewaters, groundwater, leachates, ponds, lagoons and
other surface waters.  To date, various attempts have been made to
organize selected segments of information, but a comprehensive
evaluation of peer reviewed information encompassing compounds
regulated by all pertinent environmental laws has never previously
been undertaken.  This major research activity was initiated with the
overall objective of providing a database on the treatability of
priority pollutants and other hazardous compounds in water and/or
wastewater.  The database summarizes years of studies on the treat-
ability of compounds regulated under the Clean Water Act, Safe
Drinking Water Act, Resource Conservation and Recovery Act, Toxic
Substances Control Act, Superfund Amendments and Reauthorization Act
and the National Priorities List.

     The database contains such compound specific information as
physical-chemical properties (formula, melting point, boiling point,
molecular weight, etc.), reference to environmental data (health
advisories, water quality criteria, etc.), carbon isotherm
(adsorption) data, as well as actual treatability data.  The treat-
ability data summary tables present the percent contaminant removal
achieved, effluent concentration for a compound, treatment technology
(9 biological and 8 physical-chemical technologies were evaluated),
treatment scale (bench, pilot or full-scale) and water matrix (12
matrices were studied).

     The database has been computerized and can be accessed using a
personal computer.  The program for the database was developed using
dBase III PLUS programming and compiled with Clipper.  Thus,  no
special language or software is needed to run the program.   The
                                     592
-------
.program is menu driven so  that persons with minimal computer experience would
be able to access any information provided in  the database.

     This paper will highlight important aspects of the database and provide
an overview of program availability and use by the public.
                                     593
-------
CHEMICAL TREATMENT OF METALS IN WASTEWATERS AND
SLUDGES AT THE T&E FACILITY

Douglas W. Grosse
U.S. EPA, RREL
Cincinnati, OH

Sardar Q. Hassan and Mark J. Briggs
University of Cincinnati
Cincinnati, OH
Hazardous waste treatment  technology  studies  are currently being
conducted at  the  U.S.  EPA'S Test and  Evaluation (T&E)  facility.
As  a part of this program,  alternative technologies  are being
evaluated for the  treatment of hazardous wastes containing heavy
metals and cyanide.  Both  bench-and pilot-scale research studies
are being performed to evaluate the removal of cyanide from spent
electroplating solutions and residues.  Findings to date have shown
that it  is  important to destroy all  forms  of cyanide (amenable,
non-amenable, and  complexed)  in cyanide/metal bearing wastewater
prior to metals precipitation.  Otherwise, significant quantities
of cyanide (complexed)  can be precipitated with heavy metals into
a hydroxide sludge.  Although conventional alkaline chlorination
is  a widely  employed technology used  to  destroy amenable (free)
cyanide  in  solution,   often  strongly complexed  (non-amenable)
cyanides  remain in  solution  impervious  to  chemical  oxidation.
Several processes have shown potential for breaking these complexes
(e.g.,ferricyanide)     and    destroying     the    cyanide.
Ozonation/photolysis (UV light)  is one of them.  Bench studies will
be  conducted  evaluating this technology.  In the area  of metals
treatment, work will be performed which will evaluate the treatment
of  an  EDTA-lead complexed  wash generated from  the treatment  of
lead-contaminated  soils.   Additional  studies will  focus  on the
treatment of chromate wastes resulting from tanning, wood-treating,
and  ink  manufacturing  operations.   Results generated  from these
studies will be presented.
                               594
-------
         BIOLOGICAL DEGRADATION OF CHLORINATED PHENOXY ACIDS

by:  R. A. Haugland, D. Schlemm, R. Lyons and A. M. Chakrabarty
     Department of Microbiology & Immunology
     University of Illinois at Chicago
     Chicago, Illinois

     P. R. Sferra
     Risk Reduction Engineering Laboratory
     U.S. Environmental Protection Agency
     Cincinnati, OH 45268
                               ABSTRACT

     Previous studies have shown that Pseudomonas cepacia strain
AC1100 and Alcaligenes eutrophus strain JMP134 metabolize the chloro-
phenoxyacetate herbicides 2,4,5-T and 2,4-D. respectively.  Our
results indicate, however, that both of these activities are inhibited
when co-cultures of the two organisms are confronted with mixtures of
2,4,5-T and 2,4-D.  The basis for this inhibition may be related to
the physical separation of the 2,4,5-T degradative activity in AC1100
cells from the 2,4-D degradative activity in JMP134 cells since we
have shown that the effect can be largely eliminated by transfer of
the 2,4-D degradative genes via the plasmid PJP4 from JMP134 to AC1100.
Results of studies on the biochemical and physiological factors respon-
sible for this inhibition will be presented.
                                  595
-------
           USE  OF  FOAM TECHNOLOGY  FOR  CONTROL OF TOXIC  FUMES DURING
                        EXCAVATION AT SUPERFUND SITES
     by: Ramjee Raghavan
         Patricia Brown
         Foster Wheeler Enviresponse, Inc.
         Edison, NJ  08837
Mark Torpey
Foster Wheeler Development Corp.
Livingston, NJ  07039
                            John E. Brugger
                            Releases Control Branch, USEPA
                            Edison, NJ  08837

                                   ABSTRACT
    This poster depicts the development of "foam scrubbing" as a technique
for minimizing hazardous gas and/or airborne particulate releases during
cleanup operations at Superfund sites.

    In this application, foam is used to 1) incorporate and neutralize
vapor phase emissions and 2) capture particulates.  A foam concentrate that
contains neutralizing agents is converted to a foam by using the vapor and
particulate-contaminated air to actually generate the foam; the foam is not
preformed for use as a blanket.  Once the foam bubbles are formed, a large
interior liquid surface area is available for gas, vapor, or particulate
interaction.  A neutralization agent present in bubble walls reacts with
the absorbed gas or vapor to render it innocuous.  The foam can then be
collapsed and recycled with makeup reagents.
                                    596
-------
SOIL WASHING-REMOVAL OF SEMIVOLATILE ORGANICS USING AQUEOUS SURFACTANT
SOLUTIONS
    by:    Edward Coles
           Ramjee Raghavan
           Foster Wheeler Enviresponse, Inc.
           Edison, NJ  08837
Darlene Williams
Releases Control Branch, USEPA
Edison, NJ  08837
                                    ABSTRACT

    Semivolatile organic chemicals from soil (size range -20+200 mesh) may be
removed using aqueous surfactant solutions as the extracting medium and high
mass transfer rate ultrasonic cavitational excitation for liquid-solid
contacting.  Tests were performed to confirm performance of the critical
processing steps: 1) soil deagglomeration and clay separation; 2) ultrasonic
cavitational extraction; 3) philic to phobic conversion of spent wash
solution; and 4) hydrophobic organic removal.
                                     597
-------
                  PREPARATION FOR SITE DEMONSTRATION OF A
              POWDERED ACTIVATED CARBON TREATMENT (PACT)  UNIT
             by:  John F. Martin
                  Risk Reduction Engineering Laboratory
                  U.S. Environmental Protection Agency
                  Cincinnati, OH  45268
                                 ABSTRACT

     As part of the Superfund Innovative Technology Evaluation (SITE)'
program, Zimpro/Passavant Inc. will work with EPA to conduct a field
demonstration of its mobile Powdered Activated Carbon Treatment (PACT)
unit.  The project is designed to produce data on the feasibility and
cost-effectiveness of using the process for Superfund site remediation.

    The PACT process is a biological treatment system which incorporates
physical adsorption using powdered carbon.  It is particularly well suited
to the treatment of combined industrial/municipal wastes.  In the PACT
process,  powdered activated carbon is added to the aeration basin of a
standard activated sludge system at a dose rate that varies considerably
depending on the biodegradability and adsorptivity of the waste
components.  Carbon dosages may range from a low of 10 mg/1 to more than
1000 mg/1.

    During operation of the PACT unit, excess solids (biomass and carbon)
are removed from the system by wasting a portion of the clarifier
underflow.  The SITE demonstration of the PACT system will employ wet air
oxidation (WAO) to  regenerate carbon and destroy organic contaminants
carried with the  solids.

  .  The Zimpro demonstration will be performed at a Superfund site near
Newark, New Jersey.  The 15-acre site had been used for production of
alkyl resin carriers for paint and varnish products as well as for
reprocessing off-specification resins from other suppliers.  Contamination
at the site includes volatile and semi volatile organics, pesticides,
polychlorinated biphenyls, and metals.  The demonstration will concentrate
on treatment of the shallow aquifer (2-6 feet below the surface) which
contains mostly organic solvents.

    Field work is scheduled for start-up in the summer of 1989.  During
the spring of 1989, personnel from Zimpro/Passavant will be conducting
treatability studies on wastewater collected from the demonstration site.
These data will enable design of the treatment protocol for the full-scale
field demonstration.
                                    598
ftU.S.GOVERNMENT PRINTING OFF I CE t t 990-748-1 5 9/00404
-------
-------
-------
-------
-------
-------
-------
-------
03 O
£ S

S1™
^|

=•3

18
00
0>
CO
o
o
rn
PA/600/9-
<£>
0

o
o

















If you do not wis
detach, or copy 1
upper left-hand
o 3- 3"
5 w o
m ° ~*
r 1 i

QJ CD
D —


21 O
3 5
|E

3- w
CD O
Q, X
r Q. m
a. o
3 ^
w -f-
w m
3 33
— m

Please make all
detach or copy,
left-hand corner
3
Q_ <0
-i CD
as
C Q)
3 •<
o 3"
<-* Q)
IT 3
CD (Q
Q. M
CL O
3 3
« s:
^ffl



T3 S"
m ^*
"~
                > m
                
-------