United States
Environmental Protection
Agency
Risk Reduction
Engineering Laboratory
Cincinnati, OH 45268
Research and Development
EPA/600/S2-91/025  Aug.  1991
Project  Summary
 Landfill  Leachate  Clogging  of
 Geotextile (and  Soil)  Filters


Robert M. Koemer and George R. Koerner
  The primary ieachate collection sys-
tem of most solid waste landfills con-
tains a filter layer which has historically
been a granular soil. Recently, however,
various types of geotextlle filters (both
woven and nonwoven) have been used
to replace the natural soil filters. A project
using six different landfill leachates and
aimed at Investigating the functioning of
different geotextlle filters was the focus
of this 36-mo long study.
  In the Initial 12-mo, referred to as Phase
I, flow rates In various filters were inves-
tigated under aerobic conditions at six
different landfill sites with the use of the
site-specific leachates. The study Inad-
vertently found that the overlying granu-
lar soil clogged as much as the geotextlle
filter that was located downstream. The
effects of different types and styles of
geotextiles  were generally masked by
the upstream soil clogging. A separate
anaerobic Incubation task under no-flow
conditions showed clogging to be
present but to a  significantly lesser ex-
tent than occurred with the aerobic flow
tests. This clogging was believed to be
completely  biological in nature rather
than a combination of sediment and bio-
logical processes. An Important finding
In this task was  that blodegradatlon of
the geotextiles was  not evidenced and
was concluded to be a non-Issue.
  The subsequent 24-mo study, referred
to as Phase ll(a), led  to the development
of a vastly Improved flow rate monitor-
ing device.  With the use of these new
flow columns, which are made from PVC
fittings, locally  available at hardware
stores and very inexpensive, a wide range
of variables were evaluated: I.e., four
different styles of geotextiles, geotextlle
alone versus sand/geotextile filters,
anaerobic versus aerobic conditions, and
six different landfill leachates. The re-
sulting 96 columns (4x2x2x6) were
evaluated for their flow rate  behavior
over time and found to essentially repli-
cate the first year's aerobic test results.
Varying degrees  of clogging  by sedi-
ment or particulates and microorgan-
isms did occur for the various geotextlle
and natural soil filters that were evalu-
ated. After establishing this point, a se-
ries of backflush  remediation  attempts
were evaluated. In general, flow rates
were partially restored, but only tempo-
rarily.
  In a separate task, referred to as Phase
ll(b) and conducted simultaneously with
Phase ll(a), bloclde-treated  geosyn-
thetics were used at the two sites with
the most aggressive leachates. Although
the biocldes may have been effective In
killing microorganisms, the dead bacte-
ria were as troublesome as the viable
bacteria In  creating  subsequent clog-
ging.
  This Project Summary was developed
by EPA's Risk Reduction  Engineering
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
  The primary Ieachate collection and re-
moval systems at solid waste landfills are
generally overlain by a filter layer consisting
of a geotextile or a  natural soil. Such filters
must serve the dual tasks of allowing the
Ieachate to pass into the underlying drain
while retaining the upstream particulate
matter without excessive clogging. This rep-
resents a severe challenge to the filter since
                                              7v-> Printed on Recycled Paper

-------
many leachates contain large amounts of
suspended solids and/or microorganisms.
  To investigate the behavior of several
geotextile filters and of an Ottawa-sand soil
filter, six landfill leachates were evaluated
under different experimental conditions. The
characteristics of the leachate are shown in
Table 1. Note that all of the sediment and
microorganisms contained  in the six
leachates fall into a relatively tight particle
size distribution entirely within the silt-size
classification, i.e., they range from 0.074
mm down to 0.002  mm.
Phase I - Initial Flow Rate
Evaluations
  This phase of the project, which lasted for
12 mo, used flow boxes for aerobic evalua-
tion and large containers for anaerobic in-
cubation with subsequent testing.
  In the aerobic portion of the study, 12 in.
x 12 in. wooden flow boxes, 24 in. high were
used. The boxes were all constructed using
a base plate, a geonet drain, a geotextile
filter, and 6  in. of free draining sand. The
remaining 18 in. of the boxes were empty so
that falling head permeability tests could be
conducted.  Leachate passed through the
sand and geotextile and then flowed within
the  geonet,  which was open at one end
only. The time of flight forgiven quantities of
leachate to pass through  the system was
measured. Each of the six sites had at least
four boxes, the only difference being the
type of geotextile filter. Both woven and
nonwovengeotextileswere evaluated. They
consisted of various polymer  types and
manufacturing styles.
  The following findings are based on the
flow rate behavior over the 12-mo evalua-
tion period at each site.
(a)  The flow rale measurements from the
    original values all decreased but varied
    considerably.
(b)  The relatively tightly woven geotextile
    filter, with a 4% open  area, performed
    the poorest. For each of the four differ-
    ent sites in which it was used, it clogged
    beyond  our  detection limit. The time
    periods were from 4-1/2 to  12 mo.
(c)  Opening up the void space of the same
    type of woven geotextile to a 10% open
    area helped  considerably.  Flow rates
    still decreased but were more in line
    with the needle-punched  nonwoven
    geotextile types.
(d)  The  needle-punched, nonwoven
    geotextiles performed  equivalently.
    They were similarly constructed but
    were of  different polymer types. The
    results  indicate that  polypropylene,
    polyester, and polyethylene fibers do
    not appear to give significantly different
    values in their flow rate response be-
    havior.
(e) A heat-bonded, nonwoven geotextile
    was used at two sites. Its response was
    somewhat poorer than that of the
    needle-punched nonwovens but better
    than that of the 4% open-area woven
    geotextile.
(f)  The Phase I study indicates that use of
    open-woven  geotextiles and  each of
    the  needle-punched, nonwoven
    geotextiles resulted in equilibrium flow
    conditions lasting between 6 and 12
    mo. The flow rate was reduced from as
    little as 20% of the original values (at
    four sites) to as much as 80% (at two
    sites). These reductions appeared to
    be related to the strength of the leachate
    insofar as their total solids (TS) and
    microorganism content (BOD) were
    concerned. In the  worst cases, flow
    rates were usually greater than 1.0 gal/
    min-sq ft. This is equivalent to 6.2 x 107
    gal/acre-day, which probably far ex-
    ceeds most  design requirements for
    leachate collection system filters.
(g) The cause of the flow reductions cre-
    ated somewhat of a dilemma. By cut-
    ting a cross section of the boxes at the
    end of the 12-mo period it was clearly
    evident that the 6 in. of sand over the
    geotextile was a major source of the
    flow reduction. Clearly, the experiments
    showed that soil clogging is every bit as
    serious as geotextile clogging. Further-
    more, the soil used was a very open-
    graded, rounded  sand (it was Ottawa
    sand) having a permeability coefficient
    of approximately 0.02 cm/sec (0.04 ft/
    min). Thus, it actually meets, and even
    exceeds, most regulatory criteria for a
    drainage soil, let alone for a filter soil.
(h) Microscopic examination of the cross
    sectioned soil/geotextile systems
    showed heavy particulate clogging in
    the upper half of the soil layer. Thereaf-
    ter, the clogging was either fibrous or
    consisted of  very small clusters. Al-
    though not conclusively proven, we
    believed that the upper portion of the
    soil column filtered the suspended sol-
    ids out  of the leachate and thereafter
    biological activity spread throughout
    the remaining portion of the soil column
    and into the underlying geotextile. This
    biological activity took numerous forms
    including the deposition of precipitates
    in the soil and in geotextile voids. Thus,
    different geotextiles (all other things
    being equal)  responded differently to
    the same site's leachate.
(i)  The relative amounts of flow  rate re-
    duction between  leachate sediment,
    biological precipitates,  and biological
     growth could not be distinguished in
     these particular tests.
   The anaerobic portion of the study was
 performed  under completely submerged
 conditions in 55 gal drums. Twelve samples
 of each type of geotextile were suspended
 on stainless steel racks and placed in the
 various sites' leachate. One sample of each
 type was removed for testing each month.
 Four geotextile types were  evaluated for
 each of the six landfill leachates. After the
 samples were removed, they were brought
 to our laboratory and were tested for their
 retained flow capability and possible strength
 reduction. The general findings follow:
 (a)  Relatively minor flow reductions  oc-
     curred in all types of geotextiles evalu-
     ated. The reduction values varied from
     10% to 20%. Note that these amounts
     are distinctly less than those that oc-
     curred in most of the aerobic tests. We
     believed that sediment clogging did not
     form since flow was not occurring dur-
     ing the incubation periods. Furthermore,
     the absence of  a  soil column had a
     dramatic (but quantitatively unknown)
     effect on improving the flow rates.
 (b)  All  of  the exhumed geotextiles  had
     heavy  biological growth that could be
     easily seen and felt.
 (c)  Very informative scanning electron mi-
     crographs taken at various times of
     incubation were compared with the as-
     received geotextiles. After 3 mo of in-
     cubation, complete growth around the
     individual fibers or growth in clusters
     could generally  be seen  (Figure 1).
     Although difficult to quantify, the amount
     of growth was clearly related to the time
     of immersion.
 (d)  The micrographs also revealed that the
     biological growth was easily removed
     from the fiber's surface. There appeared
    to be  no  fixity  or  attachment to  the
    fibers.
 (e)  The above observation was corrobo-
     rated by various strength tests per-
    formed on the geotextiles after immer-
     sion. Within the limits of our testing,
    there was no strength reduction over
    the 12-mo period.  This suggests that
     for  these leachates, bbbgical degra-
     dation  of geotextiles is not a problem.
     Phase II studies did not include the
    polymer degradation issue.

 Phase ll(a) - Improved Flow
 Rate Columns and Remediation
 Attempts
  This phase of the project, which lasted
 24-mo, used vastly improved flow rate mea-
 suring systems. This type of improved sys-
tem has been developed into a test method
 and  procedure adopted by the American

-------
Table 1.  Details of Municipal Landfill Leachates Evaluated in this Study and Approximate
        Leachate Characteristics
Site
Designation
PA-1
NY-2
DE-3
NJ-4
MD-5
PA-6
Approximate Leachate Characteristics at Start-Up
pH
8.0
5.5
5.8
7.4
6.8
6.5
COD*(mg/L)
15,000
20,000
40,000
45,000
1,000
10,000
TS(mg/L)
8,000
8,000
17,000
16,000
100
5,000
BOD, (mg/L)
2,000
5,000
24,000
25,000
150
2,500
'COD = chemical oxygen demand; TS - total solids content; and
BODS = biochemical oxygen demand at five days.
Figure 1. Scanning electron micrograph ofgeotextile fiber after 3 mo immersion in leachate
        from Site NY-4 (400X).
Society for Testing and Materials as a Stan-
dard Test Method (ASTM D1987-91, Test
Method for Biological Clogging of Geotextile
or Soil/Geotextile Filters").  The flow  rate
measuring column appears as shown  in
Figure 2 for evaluations being performed in
a variable or falling head test mode.
  Ninety-six of these devices  were used
under the following set of conditions:
 • four different geotextiles,
 • without soil and with Ottawa sand above,
 • aerobic and anaerobic conditions, and
 • six landfill leachates.
Continuous flow rate testing was monitored
for 6-mo. Once trends were established, a
series of remediation procedures were at-
tempted.
  The following comments apply to the first
6-mo of flow testing, i.e., before the first
remediation was attempted.
(a)  The columns with sand  above the
    geotextiles clogged considerably more
    than those with the geotextile alone,
    i.e., 23% flow was retained for sand/
    geotextiles versus 34% flow retained
    for geotextiles alone. Note that if the
    heat-bonded nonwoven  fabrics are
    eliminated from the geotextile  group,
    the flow rate retained by the geotextile
    group would be 45%. This suggests
    that geotextiles can clog less than natu-
    ral soil filters.
(b)  Of the four geotextiles evaluated, the
    highest retained flow was achieved with
    the  lightweight  needled  nonwoven
    (38%), with the heavyweight needled
    nonwoven (34%),  and  woven  mono-
    filament (32%) slightly behind. The non-
    woven heat-bonded fabric had the low-
    est retained flow of only 10% after 6 mo
    of evaluation.
(c)  Of the various landfill leachate types,
    the lowest retained flow rate resulted
    from use of the NJ-4 (14%) and DE-3
    (17%) leachates. Recall from Table 1
    that these  are the leachates with the
    highest TS and BOD concentrations.
    The  other four landfill leachates and
    their percentages of flow retained after
    6 mo of testing were PA-6 (26%),  MD-
    5 (29%), PA-1 (38%), and NY-2 (41%).
  After the initial 6 mo of flow rate testing
confirmed the results of the Phase I study,
remediation of flow columns was attempted.
The first remediation, a leachate backflush,
improved flow rate but to varying amounts
between the different columns. After 4 mo of
continued flow  testing, the flow rates de-
creased  and  allowed  for a  second
remediation. This remediation used a water
backflush. Again flow rates increased, but
over the next 5 mo they again decreased.

-------
                           Clear
                           Plastic
                           Standpipe
                             Geotextile
                             Test
                             Specimen
                              How
                              Column
Flgun 2. Flow rate columns used in Phase ll(a)
         and (b) studies.
The third remediation utilized a nitrogen gas
backflush, ft improved flow rates, but 3 mo
later they were once again reduced. The
fourth, and last, remediation was vacuum
extraction, which only nominally improved
flow rates when it was performed. Thereaf-
ter, the flow rate again decreased. The
overall average behavior of the 96 columns
is shown in Figured. It visually describes the
decreasing flow rate trends  between
remediations and the rapid  increase in flow
rates immediately following remediation. The
individual curves for each of the 96 columns
are given in the full report.
  To quantitatively assess  the overall per-
formance of the remediation attempts and
their relative performances in contrast to
one another, the data were analyzed with
respect to their percent of flow rate improve-
ment. Within each combination, however,
there  were decided differences. For ex-
ample:
(a) Backflushing  of geotextiles by them-
    selves was more  efficient  than
    backflushing  of geotextile/sand sys-
    tems. The average recovery efficien-
    cies were 29% and 13%, respectively.
(b) With sand overlying a geotextile there
    was no measurable difference from
    one type of geotextile to another.
 (c)  With only a geotextile, remediation was
     most  effective  with  the woven,
     monofilament geotextiles (38% recov-
     ery efficiency), slightly  less effective
     with the nonwoven, needled lightweight
     (31%) and  the heavyweight  (30%)
     geotextiles,  and relatively ineffective
     with  the  nonwoven, heat-bonded
     geotextiles (16%).
 (d)  When sand is placed overthe geotextile,
     there is no difference between anaero-
     bic and aerobic remediation schemes.
 (e)  With only a geotextile, remediation was
     slightly better under anaerobic condi-
     tions than with aerobic conditions.
 (f)   When sand is placed over the geotextile,
     the  remediation recovery  efficiency
     rankings  were: water > nitrogen >
     leachate > vacuum
 (g)  With only a geotextile, the remediation
     recovery efficiency rankings were: wa-
     ter > leachate > nitrogen > vacuum

 Phase ll(b)  - Blocide Treated
 Geosynthetlcs
   Because of the relatively large flow rate
 decreases observed in the course  of this
 study, an attempt at using biocides in the
 flow system was undertaken. This was done
 under the assumption that the  biocide would
 kill the microorganisms that come into con-
 tact with it and that the nonviable (i.e., dead)
 matter would pass through the system in
 much the same way that fine particles or
 sediment  moves through any other filtra-
 tion/drainage system. Because we believed
 the biocide should be introduced on a long-
 term basis rather than as one bulk dose, the
 biocide was added to the polymer com-
 pound during  fabrication of  the  selected
 geonets or geotextiles. The reasoning was
 that the biocide would release over time (via
 molecular diffusion) through the polymer
 structure and migrate to the surface of the
 ribs or fibers over a long period of time.
   From the results of these biocide treated
 geosynthetics, we believe the location  of
 the biocide vis-a-vis the initial formation of a
 biofilm layer was critical.  This was con-
 firmed at the end of the tests after solidifying
 the test columns with epoxy and cutting
 them apart. Clearly, the biofilm layer was
 occurring  at the top of the sand column
 some 2 to 3 in. above the biocide-treated
 geosynthetics.  Although there may have
 been some flow rate improvement due  to
 high concentrations  of biocide, it was very
 subtle (at best) and was masked by the
 inherent scatter in the test data. There was
 essentially no difference between flow rates
 in anaerobic versus  aerobic conditions.
  These findings led to additional tests with-
out  sand  above the biocide-treated
geosynthetic that forced the leachate  to
 interface directly with the biocide. Rather
than use a single type of geotextile, three
different types were utilized. The opening
sizes varied from 0.15 mm (nonwoven,
needle-punched style), to 0.21 mm (a wo-
ven monofilament), to 0.42  mm (another
woven  monofilament). Quite clearly, the
flow rates through the largest opening size
geotextiles, i.e. the 0.42 mm, were the high-
est. This  suggests that microorganisms
(dead or alive) must be able to pass through
the system. Whenever these microorgan-
isms reside on or within the small pores of
the filter, partial, or even complete, clogging
is possible.

Conclusions
  A simulated  field-oriented project con-
cerning biological clogging of landfill drain-
age systems was focused on geotextile
filter clogging. A number of domestic landfill
leachates were employed. The filter was
singled out (versus the geonet drain, drain-
age stone or perforated  pipe) since it has
the smallest openings and is likely to be-
come clogged before other  components.
Geotextiles were emphasized because they
are relatively new materials for this particu-
lar application.
  Phase I results reoriented our initial goals
since the granular soils covering the filters
were clogging before the geotextiles were.
Furthermore, sediment and/or particulates
were a major factor in flow rate reductions,
which appeared to be synergistic with the
biological clogging. Clearly, partial fitter clog-
ging was occurring with a gradual reduction
of flow rate over time. These trends were
common to all six landfill leachates being
used. All of the landfills were domestic (Sub-
title "D") facilities; but their waste stream,
volume of waste deposited, and liquid man-
agement schemes differed. We recognized
early in this Phase I activity that remediation
attempts would  be a necessary part of the
overall study, but the Phase I experimental
setup could not accommodate such activi-
ties. New and different test devices would
be  needed if such attempts were to be
made. We did, however, draw the following
conclusions from Phase I activities.
•  Filter clogging (as indicated by flow rate
  reductions) over the 12-mo test  period
  varied widely, the range  being between
  10% and complete (i.e., to the limit of our
  testing capability).
• A geotextile filter must be relatively open in
  its pore structure if it is to limit the amount
  of clogging, i.e., the geotextile must be
  capable of passing the sediment or par-
  ticulates along with the associated micro-
  organisms into the down-gradient drain-
  age system.
• The polymer type (polypropylene, polyes-
  ter or polyethylene)  comprising the
  geotextile fibers appears to be a nonissue.

-------
              100
   Flow Rate
  (% Retained)
                 0    2    4    6    8    10   12   14   16   18   20   22   24
Figure 3. Average response of 96 How rate columns from Phase II (a) activities.
• Both anaerobic and  aerobic conditions
  promote clogging; the relative amounts,
  however, were not capable of being iden-
  tified because of differing test setups.
• The strength of the geotextiles was not
  adversely affected by the 12-mo expo-
  sure to the various leachates. This finding,
  coupled with numerous micrographs which
  showed no chemical attachment of bacte-
  ria to the fibers, led us to conclude that
  biological degradation of polymeric based
  geotextiles does not occur.
  Phase ll(a) of the study saw the develop-
ment of a new and vastly  improved test
device for flow rate evaluation. The 4-in.-
diameter flow columns developed  during
this project have the following capabilities.
• All types of cross sections can be evalu-
  ated: geotextiles  by  themselves,  soil/
  geotextile systems, soil/geotextile/geonet
  systems, or soil/geotextile/gravel systems.
• Anaerobic or aerobic conditions can be
  maintained.
• Flow rates can be evaluated using falling
  head or constant head measurements.
• The devices are relatively small and quite
  portable. Therefore, they can be  stored
  indoors and taken to a site for evaluation,
  or stored at the site, or even stored within
  the leachate storage tank or sump,
• Various methods for remediation of dogged
  systems can be evaluated.
• The test devices and their measurement
  protocol have  recently been adopted as
  an ASTM Test Method underthe designa-
  tion of D1987-91.
• The test devices and their contents can be
  solidified by epoxy and cut in half to visu-
  ally observe the conditions existing within
  the cross section.
• Since all parts of the device consist of PVC
  plumbing and  swimming pool accesso-
  ries,  they  are readily available, easily
  sealed by chemical wipes, and inexpen-
  sive.
  The following conclusions were reached
from this Phase ll(a) study.
• Flow rate reductions were similar to  the
  results of Phase I, and the conclusions
  drawn earlier have been substantiated.
• If geotextile and/or soil filters are to be
  used in leachate collection systems, they
  should  have sufficiently open voids to
  pass the sediment or particulates along
  with the microorganisms contained in  the
  leachate into the downstream drainage
  system.
• The limiting or equilibrium flow rate  re-
  tained must exceed the site specific  de-
  sign requirement. If flow rates over time
  are not adequate, remediation is neces-
  sary. It was found that the water backf lush
 technique gave the best results (35%  im-
  provement), nitrogen gas backf lush (23%),
  and leachate backf lush (17%) methods
 were next. The vacuum extraction was  the
  least effective; it provided only nominal
  improvement (2%).
 • The periodicity of backflushing to open up
  a clogged or semi-dogged filter system
  appears to be approximately  6 mo.
   Incorporating biocides into the geotextile
 (or geonet) polymer structure to keep the
 flow system open was Phase  ll(b) of the
 study. The concept was to add various
 amounts of a time-released biocide into the
 polymer compound as the product was
 manufactured — biocide that would essen-
 tially  diffuse to the surface of the fibers
 during its service life. On contact, it would
 kill the viable m icroorganisms in the leachate.
 In the tests that were conducted  on 16
 separately built flow columns, some experi-
 mental evidence indicated that 2% and 4%
 biocide was partially effective. The remains
 of the dead bacteria must, however, be
 permitted to pass through the system, and
 this apparently could not happen for our
 particular tests setups. Thus, the idea of a
 very open  filter system was further rein-
 forced.

 Recommendations
   Based on the major findings of th is project,
 namely,
 • under continuous flow of landfill leachate
   a gradually decreasing flow rate will occur
   for  all types of  filters (soil or geotextile)
   and eventually reach an equilibrium value,
 • the equilibrium value of flow rate will vary
   according to the type of filter, the type of
   leachate, and the hydraulic gradient, and
 • the equilibrium flow rate for any given
   filter system must be compared with the
   design required fbw  rate to ultimately
   assess the adequacy of the filter's de-
   sign,
 we feel that the following recommendations
 should be considered regarding geotextile
 and soil filters placed over different types of
 leachate collection drains.
 (a) Design criteria  should be developed
    that considers the amount, size, and
    type of microorganisms and sediment
    present in the leachate along with con-
    ventional issues such as hydraulic gra-
    dient and type of filter.
 (b) Leachate collection systems at  land-
    fills that are  decommissioned or ex-
    humed for other reasons  should be
    investigated in light of the results of this
    study.
 (c) This particular  project should be fol-
    lowed  by another effort aimed at a
    larger variety of geotextile filters along
    with design guidance and field perfor-
    mance of existing systems.
  Thefull report was submitted in fulfillment
of Assistance  ID  No. CR-814965 by the
Geosynthetic Research Institute of Drexel
University  under  the sponsorship of the
U.S. Environmental  Protection Agency.
                                                                           •&U.S. GOVERNMENT PRINTING OFFICE: 1991 - 548-028/40059

-------

-------

-------
 Robert M. Koemer and George R. Koemer are with Drexel University, Philadelphia, PA
    19104.
 Robert E. Landreth is the EPA Project Officer (see  below).
 The complete report, entitled "Landfill Leachate Clogging of Geotextile (and Soil)
    Filters," (Order No. PB91- 213 660/AS; Cost: $23.00, 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:
         Risk Reduction Engineering 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 PAID
EPA PERMIT NO. G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-91/025

-------