United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-83-096 Dec. 1983
Project Summary
An Evaluation of Subsurface
Conditions at Refinery Land
Treatment Sites
K.W. Brown, L.E. Deuel, Jr.
Soil cores were collected from five
land treatment facilities being used for
the disposal of various solid wastes
from oil refineries. Cores from similar
but untreated soils adjacent to each
facility were also collected for compari-
son. The samples were analyzed for
chemical constituents to help determine
the movement of wastes in the soil.
The selected sites represented diverse
climatic regions, and the texture of the
soils ranged from clay to sand. The
facilities had been in operation from 1
to 7 years before sampling and had
received a wide range of waste applica-
tions.
Data from this study indicate that
metals from the applied waste typically
remain in the treatment zone, and that
concentrations generally are within
ranges considered normal for soils.
Only at one site (which had acidic soil)
did chromium move to depths below
the zone of incorporation. The potential
exists for possible downward migration
of land-treated hydrocarbons. At most
sites, only very low concentrations of
hydrocarbons were found at limited
depths below the zone of incorporation.
Since these materials remain in the
aerobic zone, they are likely to degrade
with time. At one site with sandy soils,
hydrocarbons were detected at a depth
of 224 cm (88.2 in.), where degradation
would be expected to occur only very
slowly.
This Project Summary was developed
by EPA's Municipal Environmental Re-
search Laboratory. Cincinnati, OH, to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Land treatment has been demonstrated
to be an effective method of disposing of
waste streams that contain biodegradable
organic materials. The various constitu-
ents of the land-treated waste that do not
degrade accumulate in the soil, but some
of these constituents could migrate down
through the profile if the retention
capacity is exceeded.
One waste category for which landf arm-
ing has been extensively used is oily
refinery wastes. Such wastes come from
wastewater treatment facilities, tank
cleaning, or specific process facilities.
The oil in these wastes is not effectively
decomposed in anaerobic landfill envi-
ronments, and the waste's water content
typically makes incineration an expen-
sive, fuel-consuming process. Thus land-
farming has become an increasingly im-
portant mechanism for treating these
materials.
Studies have indicated, however, that
some of the components of oily wastes
applied to soils are not completely
degraded and may be mobile in the soil.
Water quality could be adversely affected
by saturated soil conditions or by overload-
ing with high-BOD materials if contamin-
ant solubilities are significantly enhanced.
Such enhancement may occur as the
result of reducing environments (unoxy-
genated), soluble metal-organic complexes,
or acidic soil conditions. The primary
objective of this study was therefore to
determine the potential for downward
-------
migration of constituents following the
long-term use of specific sites for land
treatment of refinery waste sludges.
Methods and Materials
Site Descriptions
With the assistance of the American
Petroleum Institute (API) task force, five
sites were selected that had been used
for land treatment of oily wastes for more
than 5 years. Selections were based
principally on geographical and climato-
logical diversity. Because of the broad
range of site characteristics (i.e., various
wastes handled, application rates, soil
types, and application methods), this
effort should adequately reflect current
refinery landfarming operations.
Site A comprises two experimental
plots, approximately 1 acre each, within a
larger enclosure. The terrain is rolling
hills, and the elevation in the area ranges
from 823 to 1189 m (2700 to 3900 ft). The
native vegetation consists mainly of
western wheatgrass, green needlegrass,
and sagebrush. No cultivated vegetation
existed on the site. The area has a 4- to 7-
percent slope The soil is a Kyle series
consisting of a silty clay, well-drained,
nearly level, gently sloping, and fine
textured soil. The soil is moderately to
strongly alkaline and occurs to a depth of
152 cm (60 in.), below which is clay.
Site B consists of roughly 2.4 ha (6
acres) and is located some 1.2 km (.75 mi)
from a coast. The elevation is 152 m (500
ft) above sea level. The terrain includes
upland and terrace depressions. Typical
area vegetation consists of alsike and
white clover, Italian ryegrass, Kentucky
bluegrass, Douglas fir, and cedar. No
cultivated vegetation existed on this site,
which has a 1- to 2-percent slope. The
land treatment soil samples revealed an
undulating complex of Norma silty clay
loam and Cagey silt loam. Both soils have
a gravelly sand layer occurring at a depth
of approximately 46 to 51 cm (18 to 20
in.). The Norma silty clay loam occurs to a
depth of 198 cm (78 in.) and is slightly
acid to a depth of 144 cm (58 in.), below
which the soil is neutral to slightly
alkaline. The Cagey silt loam occurs to a
depth greater than 122 cm (48 in.) and is
moderately acid to that depth, below
which it is neutral to mildly alkaline.
Site C covers an area of about 2.4 ha (6
acres) and is surrounded by a 7.6-m (25-
ft) concrete wall. The site is located in a
lagoon area at an elevation of approximate-
ly 24 m (80 ft). Vegetation is typical of
species occurring only on extremely
disturbed areas. The slope is less than 0.5
percent. The soil is aTypicXeropsamment
mixed thermic (Oakley fine sand).
2
Site D lies on a nearly level coastal
terrace with a slope of 0 to 1 percent,
bordered on one side by a narrow strip of
land with moderately and strongly
sloping loamy soils. The site lies on top of
a hill, and the vegetation consists of
native and sprigged coastal Bermuda.
The area is approximately 165 x 55 m
(180 x 60 yd), and it is divided into 71
equally wide strips. A control strip occurs
at each end, with alternating grass and
treated strips in between. The soil is
Miguel fine sandy loam consisting of
deep, loam, and sandy soils containing
moderate clay pan. The loam occurs to a
depth of 76 to 138 cm (30 to 55 in.), and
the pH ranges from 6.0 to 7.0 from the
surface down.
Site E consists of a diked area surround-
ing a tank on the refinery site. The
existing vegetation is 60-percent St.
Augustine grass and 40-percent Bermuda
grass. Some St. Augustine grass was
grown on the site. The typical slopes were
0 to 3 percent, but the treatment area had
up to 5-percent slope in places. The soil is
a complex of Lake Charles and Urban
Land. The soil consists of remnants of
deep, clayey soils that have been altered
by cutting, filling, and grading. The Lake
Charles and Urban Land soil ranges from
slightly acid to mildly alkaline, and it
occurs to a depth of 188 cm (74 in.).
Experimental Design
A sampling program was initiated to
evaluate the presence of waste constitu-
ents in surface and subsurface soils.
Samples were collected both from soils
that had been treated with refinery
wastes and from adjacent untreated
soils. The samples were all collected from
similar depths, and the results of the
analyses were compared to determine
the mobility of waste constituents.
Groundwater and soil pore analyses were
not conducted. A schematic diagram of
the analytical system appears in Figure 1.
Sampling
The sampling scheme was dictated by
field conditions. Multiple soil cores were
taken at each site and composites were
made for various depth intervals. Analyses
were made for the parameters of interest
following transport of the samples to the
laboratory. The samples were transported
and maintained in cold storage until
analyzed.
Analysis
Soil Properties
Appropriate soil analyses were made to
provide a data base for correlative
interpretation of the overall study results.
These included pH, specific conductance,
texture, cation exchange capacity (CEC),
and soluble and exchangeable cationic
distribution. In addition, soils were
analyzed for NOa-N, chloride, andsulfate
by specific anion electrode techniques
(Table 1).
Soil Sample
Taken in Field
Transport to
Laboratory
I
III 1
Soxhlet TOO HN03 Physical and
Extraction and Digest Chemical Properties
Dichloromethane 44 pfj
Sil/Cc
Co/i
Fractio
\
Analysis CEC
for ci
1 Gel Metals Texture
imn 1
nation I I
As Cd
Cr Hg
Pb Ni
1
EC
NO3
S0t
Exchangeable
Cations
\
Saturates Aromatics Polynuclear Aromatics
GLC GLC GLC
Figure 1. Schematic diagram of the analytical system.
-------
Table 1. Mean Cl , SO*', andNO3~ Values
Averaged Over Depth for Both
Treated and Untreated Soils.
Site
A-Untreated
A-Treated
B-Untreated
B -Treated
C-Untreated
C -Treated
D-Untreated
D-Treated
E- Untreated
E -Treated
Cl ~
1.91
3.38
0.67
2.51
40.3
404
60.8
29.5
1.48
6.00
Anion
S04°
leg/liter--
272
258
9.54
31.6
8.58
41.2
13.7
12.8
4.25
6.68
N03~
0.95
0.76
0.54
1 08
2.24
1.71
9.01
16.8
0.25
0.31
Metals
Subsamples of each composite were
digested with nitric acid and hydrogen
peroxide. The latter was added to
facilitate the destruction of organics and
oxidation of the various metallic species.
Following digestion, metals were analyzed
according to EPA protocol. Atomic
absorption spectroscopy was used for
specific metal analyses, except for
arsenic. Colorimetry was used for arsenic
analysis following conversion to its
hydride and complexing with silver
diethyldithio-carbamate in a pyridine
base. An aliquot of the metal digest was
evaporated to a very low volume in the
presence of a sulfuric/hydrochloric acid
matrix to purge traces of nitric acid from
the sample before arsine generation.
Organics
In addition to the above inorganic soil
properties and constituents, each seg-
mented core was analyzed specif ically for
total organic carbon and extractable oil
and grease.
Organics extracted from the first three
depth intervals were subjected to column
fractionation before gas chromatographic
(GC) analysis. No attempt was made to
fractionate organics in samples from lower
depths before GC analyses because of
relatively low extractable levels. Extracts
were dried over anhydrous Na2SO4 and
reduced to a known volume by vacuum
distillation. Methylene chloride was
evaporated from an aliquot and gravime-
trically assayed to determine the extract-
able residue level. Another aliquot was
evaporated with a gentle stream of dry
nitrogen, and constituents were resolubi-
lized in hexane for loading onto a silica gel
column and for subsequent fractionation
into saturates, aromatics, and higher
condensed polynuclear aromatics.
Fractionation was achieved by loading
a 0.2% solution of hydrocarbons onto 10 g
of activated silica gel. The sample vial
was rinsed with approximately 2 ml
petroleum ether, and the rinse was
transferred to the column. Saturates
were then eluted with 25 ml of petroleum
ether. The sample vial was rinsed with 2
ml of 20% methylene chloride in petroleum
ether and loaded onto the column.
Aromatics were then eluted with 50 ml of
20% methylene chloride in petroleum
ether. Afinal rinse of the sample vial was
made with methylene chloride, which
was then loaded onto the column. This
step was followed by elution of carbazoles
and some higher condensed polynuclear
aromatics. Last, the silica gel was rinsed
with methanol to recover some of the
higher-molecular-weight materials retained
by the column. This fraction was analyzed
by high performance liquid chromatog-
raphy (HPLC).
Characteristic GC profiles were devel-
oped using a temperature-programmed
Tracer Model 560GC* equipped with a
flame ionization detector. The GC was
fitted with a 1.8-m by 0.65-cm (6-ft by
1 /4-in.) ID glass column packed with 3%
OV-1 on 80/100 mesh phrochromosorb
W. Column temperature was programmed
between 100° and 240°C at 3° C/min,
with an initial hold setting of 10minanda
final hold of 40 min. Quantification was
accomplished by comparing detector
response (measured electronically) as an
integrated peak area relative to the total
carbon injected for known materials. A
myriad of saturates, aromatics, and
polynuclear aromatics were routinely
used in assessing detector response.
Results and Discussion
Particle Size Distribution
Physical properties of a soil are defined
as those characteristics, processes, or
reactions of a soil that are caused by
physical forces but are for all practical
purposes integrally related to particle
size. Site A reflects a medium-textured
(loam) surface over a clay. Treated and
untreated soils at Site B ranged from
medium to coarse texture throughout the
profiles. A coarse texture (sand) dominated
the profile developed at Site C. Site D was
typically medium textured, becoming
coarser with depth. Site E can best be
described as a heavy clay.
'Mention of trade names or commercial products
does not constitute endorsement or recommenda-
tion for use
Cationic Distribution
The CEC is the total number of exchange-
able cations that a soil can adsorb. Some
of the calcium and magnesium reported
as exchangeable cations exceed CEC
values because of sparingly soluble
sulfate or carbonate salts or both. CEC
levels generally reflect corresponding
clay contents.
Some of the treated sites have higher
CEC values than their control areas, a fact
that is attributed to organics in the wastes
applied. High sodium saturations are
generally reflected in both treated and
untreated profiles. The data suggest that
after these sites have been used in the
land treatment of refinery waste sludges,
only a slight alteration has occurred in the
cationic distribution toward sodium.
Some evidence shows that the sodium
levels were attenuated through the land
treatment of refinery sludge at Site A.
Soil Reaction
Comparative soil reactions as pH
profiles were developed for both untreated
and treated soils at the respective sites.
Site A demonstrates the typical acidifying
effect of organics undergoing degradation
in soil. The influence of land treatment on
soil pH was mostly found to be attenuated
within the upper 0.9 to 1.85 m (3 to 6 ft).
Attenuation is reflected by convergence
of the values for treated and native soils.
The divergent profiles for Sites B and C
are attributed to coarse soil texture and
associated low buffering capacity (CEC).
Soluble Constituents
The distribution of soluble cations
found in the soil samples does not clearly
relate to migration tendencies because of
the complicated anionic interactions
involving precipitory mechanisms of less
soluble species. The presence of soluble
sodium salts within treatment facilities
tended to decrease solution levels of
background calcium and magnesium
salts.
Profiles were developed for the electri-
cal conductivities (EC's). Values generally
reflected near-surface salt accumulation
in waste-treated soils and downward
migration in all but the more acid Site A.
The extreme salinity noted between 0.5
and 1 m (1 and 3 ft) of the control soil at
Site D indicated a subsurface saline seep
(a natural phenomenon).
Chloride, sulfate, and nitrate anions
were measured for each depth interval sam-
pled and correlated to the corresponding
EC value using the following multiple linear
regression model:
EC = b0 + (CD + bz (SO/) + b3 (NO3~)
-------
Computed values were linearly correlated
with the observed values.
The data show that the variability of EC
values can be described by a three-
component anion model when EC is
adjusted to a saturated paste value.
Treated Site E did not conform well to a
linear model (r2 =0.33). The variability in
EC measured for a 1:1 soil-to-water ratio
was somewhat mitigated when the EC
was converted to a saturated paste
moisture level such that a linear model
could not resolve subtle differences
below 46 cm (18 in.).
Though somewhat scattered about the
idealized regression line, the data
demonstrate a strong positive correlation,
which supports the fact that variability in
EC with depth can be discerned by the
changes in the respective anion concen-
trations. Data of this nature are always
difficult to extrapolate, but the impact of
salt loading from land treatment of
refinery wastes does appear to be low.
Heavy Metal Distribution
Trace minerals generally tend to have
an enhanced soluble phase in soils with
high organic contents — principally
through chelation mechanisms. Under
such conditions, mobility is controlled by
soil surface adsorption and precipitory
mechanisms.
Chromium was significantly higher in
all treated surface horizons. Compared
with untreated background soils, lead
was higher in all but Site E. The only
other significant element detected was
the mercury accumulated at treated Site
B. Though chromium, lead, and mercury
accumulated m significant quantities at
one site, only the mercury level could be
considered abnormally high relative to
naturally occurring metal concentrations.
Considering the low CEC of this site, the
mercury concentration is probably at a
maximum safe level.
No evidence showed the downward
migration of metals, including those
normally considered as anionic in charac-
ter (As, Cr, V).
Organic Distribution
An analysis of variance (ANOVA) using
total organic carbon, oil, and grease as
duplicate measures of the same parame-
ter was used to evaluate hydrocarbon
levels in treated and untreated soils and
at various depths within sites. Hydrocar-
bon levels at Site A were significantly
higher in the treated soil, with an F test
indicating the difference to be significant
at better than a 1 -percent level. The least
significant difference (LSD) computed for
Site A was used to compare oil and
grease levels with depth at the treated
site. This test suggested that oil and
grease are retained within the surface 23
cm (9 in.) of soil.
The ANOVA for Site B indicated that
the greater hydrocarbon levels in the
treated soil were significant at a 1-
percent level. Variance with depth was
significant at the 5-percent level. Oil and
grease at this site were attenuated within
the top 60 cm (2 ft) of soil. No attempt was
made to split out the variability because of
the technique of measuring hydrocarbons
from that of the error mean square. This
technique reduces the sensitivity of
assessing real differences in the hydrocar-
bon concentration as a function of depth.
Data evaluated for Site C do not reflect
a statistical difference between hydrocar-
bon levels of treated and untreated soils
but values decreased significantly with
depth. Simple comparisons with the
untreated soil are not possible because
high hydrocarbon levels are present in
the reference soil to appreciable depths.
Hydrocarbons at Site D reflect the
general trends found at Sites A and B in
that the treated soil differed significantly
from the untreated soil. The organics
were attenuated within the top 30-cm (1 -
ft) of the soil.
Site E reflected no statistical differences
as a result of treatment, and correspond-
ingly no differences with respect to depth.
These data suggest that hydrocarbons
loaded onto the soil have degraded
without appreciable migration of degrada-
tion products.
High Performance Liquid
Chromatography (HPLC)
Surface horizons for all treated sites
and subsurface samples showing signifi-
cant hydrocarbons (detected by GLC
flame ionization) were analyzed by HPLC.
To help detect phenolic derivatives,
analyses were made by injection of a
methanolic extract of the sample. Extracts
were generated by high-speed blending
of sufficient sample to provide a detection
limit of 1-ppm phenol, based on the
integrated area of the standard.
Comparative retention time and area
ratio analyses show only two surface
samples and one subsurface sample that
possibly contain phenolic materials.
These are the immediate surface samples
collected at treated Sites A and E and the
76- to 91 -cm (30- to 36-in.)depth interval
sampled at treated Site B. Though several
phenolic derivatives fit the retention time
criteria, only pentachlorophenol passed
the area ratio test for the surf ace samples
at treated Sites A and E. Based on the
detector's response to lindane, aldrin,
dieldrin, heptachlor, and arochlor 1254,
no halogenated hydrocarbons were
detected in quantities exceeding 1 ppm.
This result strongly suggests that material
detected in the phenolic screening as
potentially pentachlorophenol was some-
thing other than a chlorinated hydrocarbon.
Gas Liquid Chromatographic
Characterization
Detector Response
Gas liquid Chromatographic (GLC)
analyses were used in conjunction with
column fractionation on silica gel to
develop characteristic chromatograms of
untreated and treated soil with respect to
the depth interval sampled.
A complex mixture of standards was
injected to evaluate GC column conditions
and detector response over periods of a
day, a week, and a month. The standard
deviation reflects variability over a 4-
month period and innate differences
among compounds, particularly xanthene.
A detector response of 10,000 integration
units requires 100 ng C ± 18 ng C for
Detector 1, and 80 ng C ± 10 ng C for
Detector 2. Excluding xanthene does not
numerically affect the standard deviation,
so 10,000 integration units correspond to
100 ng C ± 8.5 ng C for Detector 1 and 80
ng C + 4.5 ng C for Detector 2.
Column Chromatographic
Fractionation
The standard fractionation on silica gel
reflects the usefulness of this procedure
for potentially reducing the complexity of
chromatograms to evaluate the fates of
the various classes of compounds contained
in the wastes added to soils. Soxhlet
extracts of samples collected at the first
three depth intervals were subjected to
column fractionation on silica gel.
Fractions 2 and 3 were combined in a
manner similar to that used for the
standard before GLC analyses. Methanol
was used as a final rinse in an attempt to
extract higher condensed polynuclear
aromatics (e.g., asphaltenes) not eluted
with the other solvents. No attempt was
made to use column fractionation on
Soxhlet extracts of samples collected at
the lower depth because of the relatively
low hydrocarbon levels and the potential
for lowered concentrations of hydrocar-
bons recovered in multiple fractions.
Thus to improve the sensitivity level to
hydrocarbons extracted from samples
collected at the lower depths, chromato-
grams were developed for total extracts
-------
following the removal of an aliquot for
gravimetric analyses.
Vlolecular Weight and Carbon
Number
A linear regression model was used to
describe the relationship between retention
time (RT) and molecular weight (MW). RT
corresponding to peak sensitivity for
compounds used in the standard mixture
and others (including naphthalene,
biphenyl, methyl heptadecanoate, 1,3,5-
triphenyl benzene, triphenylethylene,
tetraphenylethylene, and 9,9-bifluorene)
increased linearly with increased molecu-
lar weight. RT values averaged for
multiple injections of known compounds
with various molecular weights gave a
regression coefficient (r2) of 0.84. These
data suggest that hydrocarbons with less
than 76 g/mole would be eluted with the
solvent front under the same column and
instrument conditions used for the
standards. A 20-min increase in RT rough-
ly corresponds to 5 carbons added in a
chain configuration for saturates, or a
benzene ring added for aromatics.
Similar results are reflected in the linear
regression retention and carbon number.
Again, the fit of the model is reflected in
the high regression coefficient value (r2 =
0.83).
The full report was submitted in fulfill-
ment of Cooperative Agreement
CR807868 by Texas A&M University
under the sponsorship of the U.S. Envi-
ronmental Protection Agency.
K. W. Brown andL, E. Deuel, Jr., are with Texas A&M University, College Station
TX 77843.
Carlton C. Wiles is the EPA Project Officer (see below).
The complete report, entitled "An Evaluation of Subsurface Conditions at Refinery
Land Treatment Sites," (Order No. PB 84-102 169; Cost: $14.50. subject to
change) 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:
Municipal Environmental 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
U.S. POSTAGE
PAID
Cincinnati, Ohi
Permit No. G35
Official Business
Penalty for Private Use $300
U.S. GOVERNMENT PRINTING OFFICE: 1984-759-102/0814
------- |