United States
Environmental Protection
Agency
 Hazardous Waste Engineering
 Research Laboratory
 Cincinnati OH 45268
Research and Development
 EPA/600/S2-87/086  Dec. 1987
Project  Summary

Field Assessment of Air
Emissions and  Their
Control  at  a  Refinery  Land
Treatment  Facility

B. M. Eklund, T. P.  Nelson, and R. G. Wetherold
  A field assessment was performed to
measure  the emissions of volatile
organics from a petroleum refinery land
treatment site. As part of this study,
the emissions of total volatile organics
from surface-applied and subsurface-
injected oily sludge were measured
over a five-week period. The effect of
soil tilling on the emissions was also
monitored.
  Volatile  organic emission rates were
measured  using the emission isolation
flux chamber method. Emission rates
of carbon dioxide and methane were
also measured for use in evaluating and
estimating apparent biodegradation
rates. Soil samples were collected
during the test periods to determine soil
properties,  oil levels  and  microbe
count. Soil surface and  ambient
temperatures inside the flux chambers
were also measured throughout the
test periods  to determine their influ-
ence on emission rates.
  From the measurements, the emis-
sion rates  of total volatile organics, as
well as the emission rates of selected
individual volatile  organic species,
were estimated over the  five-week
period. The amount of oil biodegrada-
tion was estimated from carbon dioxide
evolution  rates and oil disappearance
from soil samples over the test period.
The measured volatile organics emis-
sion rates and  the  emission rates of
selected species were compared to the
rates predicted with The Thibodeaux-
Hwang land treatment model.
  This Project  Summary was devel-
oped by EPA's Hazardous Waste Engi-
  This Project Summary was devel-
oped by EPA's Hazardous Waste
Engineering Research Laboratory,
Cincinnati. OH.  to  announce key
findings of the research project that is
fully documented in  two separate
volumes of the same title (see Project
Report ordering information at back).
Introduction
  The Office of Air Quality Planning and
Standards (OAQPS) of the U.S. Environ-
mental Protection Agency (EPA) is devel-
oping standards for controlling emissions
from hazardous waste treatment, storage
and disposal facilities (TSDFs). The
purpose of these regulations is to protect
human health and the environment from
impairment by emission of volatile
organic compounds (VOCs*) and partic-
ulate matter. The  Hazardous Waste
Engineering  Research  Laboratory
(HWERL)  has the responsibility of pro-
viding technical support to OAQPS in the
area of atmospheric  emissions determi-
nation from hazardous waste manage-
ment. Part of the research in the HWERL
program involves the study and assess-
ment of  emission  control techniques
which are applicable to TSDFs.
  In the  current study, the emission
characteristics of the land treatment of
oily sludges by surface application and
subsurface  injection techniques  were
evaluated. The results of this assessment
have increased the understanding of
volatile organics emissions from  land
treatment disposal facilities. The subsur-
"VOCs are defined in this study as those purgeable
 volatile compounds determined by a purge-and-trap
 technique.

-------
face injection technique was selected for
this study because  the technique is
believed to represent a potentially major
type of volatile organics emission control
at TSDFs.
  To assess the effectiveness of subsur-
face waste  injection as an  emission
control  technique,  a field study was
performed at the land treatment facility
located at the Chevron, USA refinery in
El Segundo,  California. The major objec-
tives of the study at this site include the
following:

• To determine the percentage of vola-
  tilized organics  as a function  of the
  applied purgeabale organics and of
  the applied oil;

• To estimate the emissions of applied
  volatile organics from the test plots
  for the five-week  testing period and
  annually for the entire land treatment
  facility;

• To determine  the effectiveness of
  subsurface  injection  in  reducing
  volatile organic emissions from land
  treatments  by  comparing the  mea-
  sured emission rates  from  the two
  application methods;

• To determine the extent of oil degra-
  dation and/or  measurable biological
   activity;

• To determine the effects of various
   environmental   and  operational
   parameters  on  emission rates and
   emission rate measurements, includ-
   ing those due to the emission meas-
   urement procedure; and

• To compare the measured emission
   rates to those  calculated using the
  Thibodeaux-Hwang  predictive emis-
   sion model.

Approach
  The land treatment area at the Chev-
ron, El Segundo refinery covers approx-
imately 42,000 square meters (10 acres).
Three test plots, each approximately 420
square  meters (0.1 acre) in area, were
located side-by-side at one corner of the
facility. The  area containing the test plot
has been in land treatment service for
several years.
  Experimental Design—The test design
was a synthesis of two sampling strate-
gies for obtaining emission  measure-
ments: totally randomized sampling over
the test  plots and semi-continuous
sampling at a single location in each plot.
Three test plots were used in this study.
Sludge was surface-applied to  one  plot
(Plot  A)  and  subsurface-injected  on
another plot (Plot C). In between these
two plots was a  third plot (Plot B),  and
no sludge was applied to this plot during
the test period.  Plot  B served as a
baseline  or control plot. Each plot was
divided into 21  equal segments to provide
for randomized sampling.
  Sampling  was performed  during the
first, third and fifth weeks of a five-week
period. Samples were  collected in the
morning  and  afternoon.  On two days,
samples were collected immediately
before dawn so that nighttime emissions
could be  estimated.
  Samp/ing Procedures—The main sam-
pling  technique  employed at this  site
involved the  direct measurement of
emissions using the emission  isolation
flux  chamber. A diagram  of  the  flux
chamber is shown in Figure  1. The flux
chamber is placed on the emitting
surface. Clean, dry air is passed through
the chamber at a controlled  and meas-
ured  rate. The  concentration of  the
specie(s) of  interest is  measured at the
outlet of the chamber. The emission rate
of the measured  species(s) from  the
enclosed surface can then be calculated
from the flow rate and concentration in
the gas.
  Liquid  grab samples of the sludge and
the service water used to irrigate the soil
were collected. Samples of soil were also
collected periodically during the testing.
                        Temperature
                         Readout
Type  K thermocouples  were  used to
monitor the air and soil temperatures
both  inside and  outside of each  flux
chamber.
  Analytical Procedures—-The  on-site
analyses  were limited to gas-phase
analyses of the air samples collected in
gas-tight syringes from the outlet of the
flux chambers. A  Byron  Instruments
Model 401 total  Hydrocarbon (THC)
analyzer was  used to  determine the
concentrations of  total hydrocarbons
(THC), C02 and methane in  the effluent
air samples from the flux chamber.
  The off-site  analyses included the
chemical speciation of the flux chamber
air  samples collected in stainless steel
canisters and  of the liquid sludge and
water samples collected  at  the  land
treatment site. The oil, moisture,  and
microbe  levels in  the soil were  also
determined, as were the physical prop-
erties of the soil samples.
  The oil, water and solids content of the
sludge samples were determined by a
method developed for this  purpose by
Chevron Research  Corporation. This
method is called  the Modified Oven
Drying Technique (MOOT). The oil  and
grease content of the soil samples  was
determined by  EPA Method  413.1.
Standard methods were used to deter-
mine  the bulk density, particle density,
total  porosity,  moisture content  and
particle size distribution of the  soil
samples.
  Sludge Application—The  sludge  was
applied  as evenly  as possible to the
                                          Thermocouple
                                                              Syringe/Canister
                                                                Sampling Port
                                                             Stainless Steel
                                                               or Plexiglas
 Figure 1.    Cutaway side view of emission isolation flux chamber and sampaling apparatus.

-------
  urface-applied  Plot  A  and  to  the
.ubsurface-injected Plot C. The sludge
was injected subsurface using an 8000
liter (50-barrel) capacity injector vehicle
equipped  with four separate  injector
tines. The waste was injected 15-28 cm
(6-11  inches)  below the  surface.  The
average waste loading was 1.4 x  10*
kg/plot,  assuming equal application
rates for both plots. All three plots were
tilled with a disk tiller pulled behind the
tractor during the sludge application.
  Emission sampling with the  flux
chambers was started immediately after
application and continued for four days.
Emission sampling was performed  dur-
ing two other four-day periods in the third
and  fifth  weeks following application.
Each plot was tilled 2-3 times per week
during the five-week test  period.  The
plots were also irrigated with water once
between the first  and second sampling
periods and three times  between  the
second and third sampling periods.
  Emission Estimates—Based  on  the
emission  measurements  performed in
this  study, total-VOC emissions were
estimated for the individual test plots for
a five-week period. To do this, nighttime
emissions were estimated by interpolat-
 ng between  late  afternoon and early
morning measurements. Daily emissions
for days  on which no  measurements
were taken were determined by interpo-
lating between days  when emissions
were measured.
  Annual emissions were also estimated
for the whole  facility so that emissions
reductions potential for emissions con-
trols could be determined. To estimate
annual emissions, five week test-plot
emissions were  extrapolated,  to  the
whole facility  by accounting for facility
waste application rates,  applicaticn
methods and surface area.
  Emission Modeling—Instantaneous
emission rates were calculated using the
Thibodeaux-Hwang model.  Total-VOC
emissions were calculated using average
physical  and  chemical properties for
various classes of volatile compounds
(e.g., volatile aromatics), properties of the
test plot soil  and  site-specific environ-
mental parameters. Similarly, the emis-
sion rates of the 12 compounds studied
individually were  calculated based on
their individual physical and chemical
properties.
Results and Discussion
  The following site-specific findings
were obtained  from this study.
Total Volatile Organic Emission
Rates
  The  average emission rates for each
plot over the five-week test period were
47.1, 6.16, and 53.9 //g/m2-s of volatile
organics for the surface applications
(Plot A), background (Plot B), and sub-
surface application (Plot C) plots, respec-
tively. The instantaneous emissions from
each of the three plots were as  high as
370.7,  38.5,  and  324.9 //g/m2-sec,
respectively.
  The  emission rate decreased  approx-
imately exponentially with  time after
application for each plot. Comparing the
averaged measured emission for  the first
week of sampling to the third and fifth
(final) weeks of sampling,  the emission
rates decreased 93%, 86%, and 91% for
the surface application, background, and
the subsurface application plots,  respec-
tively.  The weekly estimated emission
rates are shown graphically in Figure 2.
The  five-week estimated cumulative
emissions for  Plots A, B, and  C were
33.3, 5.2, and 39.0 kg, respectively.
  It  was  estimated that  the ratio of
volatile organics emitted over five weeks
to purgeable organics  in  the  waste was
0.30 for Plot A and 0.36 for  Plot C. The
ratio of volatile organics emitted over five
weeks to  the  mass of applied  oil was
estimated to be  0.012 for Plot  A and
0.014 for PlotC.
  The  measured  emission  rates were
found to be related to the  ambient air
temperatures above the soil surface. This
resulted in a significant diurnal effect in
emissions. For the two occasions when
sampling was  conducted before dawn,
the average measured emission rate for
the "half-day" (i.e., four-hour) period was
lower than  the average of the half-day
averages for that week for each plot. For
Week 1, the decrease  was 75%, 56%,
and 87% for Plots A, B and C, respec-
tively. For the second week of sampling,
the decrease was 75%, 54% and 59%
for Plots A, B and C, respectively.

Subsurface Injection as  an
Emissions Control Technique
  The application method as practiced at
this refinery did not appear to have a
large  effect on the  emissions. Imme-
diately after  sludge application  and
before the  first tilling, the cumulative
measured emissions from  the surface
application plot  were  slightly greater
than those from the subsurface applica-
tion plot. After the first tilling episode
(two days after the initial application), the
cumulative measured emissions seemed
to be slightly greater  for the subsurface
application   plot   throughout   the
remainder of the  test period. The total
cumulative measured  emissions were
                               Total Est. Emissions/Week
                             A=Surf.  B-Bkg.  C=Subsurf.
  to
  §
Figure 2.   Total weekly estimated emisisons for each plot.

                                        3

-------
  14% greater from Plot C than  Plot A.
  Similarly, the estimated total emissions
  from Plot C (39.0 kg) were 17% greater
  than the total for Plot A (33.3 kg) for the
  five-week test period.
    Because the  difference  between
  cumulative emission rates  for  Plots A
  and C were small and were comparable
  to the uncertainties in emissions rates
  themselves,  it  is not  possible to draw
  conclusions regarding the relative emis-
  sion characteristics of the two applica-
  tion plots. However, any difference  did
  not  appear to be significant.

  Emissions of Individual
  Compounds
    The emissions of 12 selected individual
  compounds consisting  of  alkanes,
  alkenes and aromatics were also studied
  in detail.  The dozen  individual  com-
  pounds examined  behaved  in the same
  manner as the total volatile organics. The
  emissions rates  decreased with time,
  increased  after  tilling, and showed
  diurnal fluctuations. However, the appli-
  cation method  had a  greater effect on
  the  individual compounds than for total
  volatile organics. Also, a greater percen-
  tage of the applied individual compounds
  was emitted than for total volatile
  organics.  The  difference  between  the
  average emissions values and those of
  the  12 selected compounds is  thought
  to be due to the differences in  volatilities
  of the 12 selected compounds  and  the
  average volatility of the oily sludge.
       Biodegradation and Oil
       Disappearance
         Radian measurements indicated that
       little or no biodegradation of the applied
       oil  was observed.  No methane  was
       detected,  implying that no anaerobic
       degradation occurred. Significant quan-
       tities  of  possible products of partial
       degradation were not detected. Because
       of the scatter in the data, neither a
       change in microbial population levels nor
       a decrease in oil content in the soil could
       be  discerned. The  measured  carbon
       dioxide (COz) values were at near back-
       ground levels. Assuming all C02  mea-
       sured was due  to  complete aerobic
       degradation, less than 3% of the applied
       oil was  completely degraded over the
       five-week test period.
         Chevron analyzed  more soil  samples
       using  a recently developed analytical
       technique. One or both of these factors
contribute to more precise measuremen
of the variation of oil content with time
The Chevron results appear to indicat
a decrease  in  oil  levels  in  the  soi
However, even these data show signif
icant  scatter; only  one of the  fou
monitored sampling points exhibited •
change in oil content which had greate
than a 95% certainty of being nonzero.
Emissions Model
  The Thibodeaux-Hwang model for Ian
treatment facilities was found to predic
mean emission rates that were general!
higher than those actually observed, bu
which agreed to within an  order c
magnitude. The model predicts an expo
nential decay  in  emissions  over  tim
which approximately  agrees witl
observed changes in emissions at th<
site.
         B. M. Eklund, T. P. Nelson, and R. G. Wetherold are with Radian Corporation,
           Austin, TX 78766-0948.
         Benjamin L. Blaney is the EPA Project Officer (see below).
         The complete report consists of two volumes entitled "Field Assessment of
           Air Emissions and Their Control at a Refinery Land Treatment Facility:"
           Volume I (Order No. PB 88-1'24540'/AS; Cost: $32.95, subject to change).
           Volume II (Order No. PB 88-1245577AS; Cost 32.95, subject to change).
         The above reports will be available only from:
                 National Technical Information Service
                 5285 Port Royal Road
                 Springfield, VA 22161
                 Telephone: 703-487-4650
         The EPA Project Officer can be contacted at:
                 Hazardous Waste Engineering Research Laboratory
                 U.S. Environmental Protection Agency
                 Cincinnati, OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
                        BULK RATE
                    POSTAGE & FEES I
                           EPA
                      PERMIT No  G-3
Official Business
Penalty for Private Use $300

EPA/600/S2-87/086
             0001961
             LIBRARY  REGION
                                    ST
             CHICAGO
                                       IL   60604

-------