SEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-96/145
December 1996
Enhanced
Bioremediation of
Using Immobilized
Nutrients
Field
Demonstration and
Monitoring
-------
-------
EPA/600/R-96/145
December 1996
Enhanced Bioremediation of BTEX
Using Immobilized Nutrients
Field Demonstration and Monitoring
by
Robert C. Borden, Russell Todd Goin and Chin-Ming Kao
Department of Civil Engineering
North Carolina State University
Raleigh, NC 27695
Charlita G. Rosal
Characterization and Monitoring Branch
Environmental Sciences Division
Las Vegas, NV 89193-3478
Cooperative Agreement Number
CR820468
Project Officer
Charlita G. Rosal
Characterization and Monitoring Branch
Environmental Sciences Division
Las Vegas, NV 89193-3478
This study was conducted in cooperation with
Department of Civil Engineering
North Carolina State University
Raleigh, NC 27695
NATIONAL EXPOSURE RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
US ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
Printed on Recycled Paper
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development
(ORD), partially funded and collaborated in the research described here. It has been peer reviewed by the
Agency and approved as an EPA publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
Abstract
A permeable barrier system was developed for controlling the migration of dissolved contaminant plumes
in ground water. The barrier system consisted of a line of closely spaced wells installed perpendicular to the
contaminant plume. Each well contained concrete briquets that released oxygen and nitrate at a controlled rate,
enhancing the aerobic biodegradation of dissolved hydrocarbons in the downgradient aquifer.
Laboratory batch reactor experiments were conducted to identify concrete mixtures that slowly released
oxygen over an extended time period. Concretes prepared with urea hydrogen peroxide were unacceptable
while concretes prepared with calcium peroxide and a proprietary formulation of magnesium peroxide
gradually released oxygen at a steadily declining rate over a three- to six-month period.
A full-scale permeable barrier system was constructed at a gasoline-spill site near Leland, NC. Initially,
increased dissolved oxygen and decreased benzene, toluene, ethylbenzene, and xylene isomer (BTEX)
concentrations in the downgradient aquifer indicated that oxygen released from the remediation wells was
enhancing biodegradation. Over time, treatment efficiencies declined, suggesting that the barrier system was
becoming less effective in releasing oxygen and nutrients to the aquifer. Field tracer tests and soil analyses
performed at the conclusion of the project indicated that the aquifer in the vicinity of the remediation wells was
being clogged by precipitation of iron minerals.
in
-------
Contents
Abstract iii
Figures vi
Tables vii
Acknowledgment viii
Sections
1. Introduction 1-1
Background 1-1
Previous Studies of Permeable Barrier Systems 1-2
2. Conclusions 2-1
3. Recommendations 3-1
4. Methodology ' 4-1
Laboratory Methods: Pre-Barrier Construction 4-1
Oxygen Retention in Solid Peroxide Concretes 4-1
Oxygen Release Over Time from Solid Peroxide Concretes 4-1
Effect of Nitrate Addition on Bioremediation 4-2
Field Monitoring of Permeable Barrier System 4-2
Site Description 4-2
Barrier Design 4-3
Well Placement 4-4
Well Construction 4-4
Ground-Water Sampling 4-5
Iron Content of Soil Adjoining Remediation Wells 4-5
Specific Discharge Measurements 4-6
5, Results and Discussion 5-1
Laboratory Results: Pre-Barrier Construction 5-1
Oxygen Retention in Solid Peroxide Concretes 5-1
Oxygen Release Over Time from Solid Peroxide Concretes 5-1
Effect of Nitrate Addition on Bioremediation 5-2
Field Monitoring of Permeable Barrier System 5-3
Background Ground-Water Quality 5-3
Ground-Water Monitoring 5-3
Variability in BTEX and Indicator Parameters Upgradient of the Barrier 5-4
Evaluation of Permeable Barrier: Test Period 1 - Day 0 to Day 242 5-4
Evaluation of Permeable Barrier: Test Period 2 - Day 242 to Day 361 5-6
Evaluation of Permeable Barrier: Test Period 3 - Day 361 to Day 498 5-7
Remediation Well Clogging 5-9
Specific Discharge Measurements 5-9
IV
-------
Contents (Continued)
Iron Content of Soil Adjoining Remediatiop Wells 5-10
Overall Evaluation of Permeable Barrier Sysleraft .'* 5-11
Reference's' ' SS'JEI ooj>»-«s>i-tjja^Av guciueiwiti is-'i u^ca iu,!jr>3i lu ^umiicuimfs i
Abbendices aliaw snhoiinom brie lamed sldssnrreq gnrworfa q.em sJi8 .£
i*' u & S****yj!/ fSfifffiJT*1"i^f*"! ^"^ l^"ci>*m:*O_'(:%M"lf**iOT*f'cif"i f'^A^i^P ^\ft£* £ fiilFT> f*tf"\ "~Ff\ rW>?1!>«3*''i.rt T 5-'
Appendix A'- ' Laboratdry-Smaies-'dxygenMease dvef Ti₯e rronf1 L3L'&£> bfls Ssm§no 10 no.te.xU ,t
, , Solid Peroxide Concretes , , . A-l
Appendix B - Field Monitoring Data .. .^.'.T . .*..!."'?.'"?.. ..'. . ."*. V..... .*..'.'.".". V.. .". i'.'.i. .^.".' B-l
Appendix C - Summary of Tracer Test Results .. ^ C-l
i.-c , (latemBifc mo-\.i) atsapnd aisionoo fO§Mmoil aassiai n&^vxO ,?.
lot sraitj riiiw gajs's seBslsi-naavxo ni nortemv gnivorie Ksnil bsSsmilas JR 3833 .d
£-c asxim sisisnoo soixoiaq mublfio bns abixoiaq moig;
-9«iiO2sg raoit islsw bnuoig ni noitebingsboid X3T9 no nonlbba sisilin lo :
t-S , DM .bnfilsJ TBSH sJis
£-£ bohaq Josjp-sq sri) isvo insiteig bnB noiJoaiib wofi isifiw-bnoo-jg ni noltshsV ,8
5VJ13S sril 10 trssibfiragi? allsw gnnolinora nt noiteunsonoo XHT3 kloJ ni noitebeV .Q
£~c , isrrrad sldesfmaq arij 10 (8U8) eabia fo'iJno-j fans (VUg)
snoiiBiinaonoa ns§>(xo bsvloaaib (d) bns anohsiinoonoo X3TS IsJot (s) ni noitehfiV .01
(VU8) i'iisibingqii elFsw §nhoJinoin ni f.S^S vsb ol 0 vf?fc; f faohsq J£ vsb o 0 \j£b) I bohsq J33i ^nrnfb
'?o in3Jb6ignwob alfsw gnhosinom ni gnohsihreonoo XHTS IBJDJ ni nci?.''iij«V .SI
d< -lomBd aldBsrrmq Ms la a^bia (QU2) [oiinoo bne (OiLT^) aviiofc odi
nrsg\fxo bsvfoasib (d) bne anobsilnsoaoo XHT9 \&) ni noilshsV .£ F
«1S9W gnhotinom ni HSP vsb ol £!»£ vsb) £ boh^q iggj gnhob
nsg^xo bs'/ioaaib (d) farm ?nortei?03Daoo XHTH ifiloJ SB) ni noiteiDiV .M
VIU8 .I^IUo ,TU3 gfi&w gnholinom ni !'8Q^ reb ot Idt veb) £ bohaq J89i gohub
3-~; dFJ?, bns .ef'Jg ,FJU2 to 59873vs srii bms
>a n&yyxo bsvfo8?:ib (d> bns !5noi}fiitnsoaoo X5T3 ijjjoJ Cs) ni noiteiiuV .c!
tfifj3 ,£ITJ2 sllsv/gnhoJrnom nt (SQ^Ysbc) Idf. \fcb) £ bonsq iasl gniiofa
~. ' " SIU8 bos
so nagvxo bavJoReib Cd) DOB ^noitSTUnsDnoo XHT3 IrJol (B) f-ssM .dl
fcfss abohsq JfisrnJB^'i.1 isubivibni 10* B'!S>;!' gfihotioofft ni''
.. . . , bonoc {jinoJtii'iiiqo 'tarnsd sibna
v
-------
Figures
1. Schematic of reactor used for measuring oxygen-release rates 4-1
2. Site map showing permeable barrier and monitoring wells 4-3
3. Location of original and added permeable barrier remediation wells 4-3
4. Schematic of remediation well containing oxygen-releasing concrete 4-3
5. Oxygen release from MgO2 concrete briquets (1.7-cm diameter) 5-2
6. Best fit estimated lines showing variation in oxygen-release rates with time for
magnesium peroxide and calcium peroxide concrete mixes 5-2
7. Effect of nitrate addition on BTEX biodegradation in ground water from gasoline-
contaminated site near Leland, NC 5-3
8. Variation in ground-water flow direction and gradient over the project period 5-3
9. Variation in total BTEX concentration in monitoring wells upgradient of the active
(SU7) and control sides (SU8) of the permeable barrier 5-4
10. Variation in (a) total BTEX concentrations and (b) dissolved oxygen concentrations
during test period 1 (day 0 to day 242) in monitoring wells upgradient (SU7)
and downgradient (SU13) of the barrier 5-5
11. Variation in (a) total BTEX concentrations and (b) dissolved oxygen concentrations
during test period 1 (day 0 to day 242) in monitoring wells SU14 and SU5 5-5
12. Variation in total BTEX concentrations in monitoring wells downgradient of
the active (SU10) and control (SU9) sides of the permeable barrier 5-6
13. Variation in (a) BTEX concentrations and (b) dissolved oxygen concentrations
during test period 2 (day 242 to day 361) in monitoring wells
SU7, SU13, and SU14 5-7
14. Variation in (a) total BTEX concentrations and (b) dissolved oxygen concentrations
during test period 3 (day 361 to day 498) in monitoring wells SU7, SU14, SU17
and the average of SU13, SU15, and SU16 5-8
15. Variation in (a) total BTEX concentrations and (b) dissolved oxygen concentrations
during test period 3 (day 361 to day 498) in monitoring wells SU13, SU15,
and SU16 5-9
16. Mean (a) total BTEX concentrations and (b) dissolved oxygen concentrations
in monitoring wells for individual treatment periods and
entire barrier operational period 5-11
vi
-------
Tables
1. Mass Ratios of Components in Concrete by Treatment Period 4-4
2. Sample Collection and Preparation Protocol 4-5
3. Average Oxygen Contents of CaO2, MgO2, and CO(NH2)2»H2O2 in Original
Form and in a Concrete Matrix 5-1
4. Model Oxygen-Release Rate Equations for Magnesium Peroxide and Calcium
Peroxide Concrete Mixes 5-2
5. Specific Discharges for Remediation Well Groups Estimated from Tracer Tests 5-9
6. Extractable Iron Content in Soils Adjoining the Remediation Wells and
Upgradient of the Barrier 5-10
7. Average Concentrations of BTEX in Monitoring Wells over the Entire
Treatment Period 5-11
8. Mass of Oxygen Released from Original Remediation Wells on Day 459 5-12
9. Mass of Oxygen Released from New Remediation Wells on Day 459 5-12
vn
-------
Acknowledgment
Regenesis, Inc. and the North CarohnaPi.vtao^^ .
for this project. We would like to thank Ms. Linda Mintz for providing access to her property and Rebecca
Stager for her assistance in the laboratory. ............. locc-toil noiteisqsiq bus nohosiioD slqrms?, .£
0 ni SO<.H»£{SHI/1)OD bns ,sOgM .rOeD 1o ainaJnoD nsg^xO agsisvA .£
xhlsM aisionoD .e ni bnB raio';I
bf,R abfxoia*! mi/ia^ngeM io1 zooiJBupH. sisS aassls^-nsg^xO SsboM >
, asxiM atsiDfioD afaixoTta'-J
isosiT rnofl batoniM-I RquoiO HsW noiJcibsmsM lo't gsgisrioaiQ oi'fiosqg ,?.
bne albW noiJBibsmsH sdt gninro[faA slio?, ai JnsJnoD noil sldBtoB-rtxH .5
. , , ismsS sri) to:
3ii) T3vo aflsW gnhoJinoM ni XHTSlo anoite'j^nsonoD sgBTrsvA .V
, bobs*! JasmJBSiT
no ?lfsW noilBibsms^ IsnighO mo'fl bsgssfaM nsg^xO ^o 83fiM .8
no giisW noiteibsma^ v/sM mcnft bsgBsJsM nsgy;xO ^o ggsM .6
-------
Section 1
Introduction
Background
The U.S. Environmental Protection Agency
(U.S. EPA) is studying the performance of enhanced
bioremediation systems to evaluate the effectiveness
of the technology. The goal of this study was to
design and monitor the field performance of a
permeable barrier treatment system for controlling
the downgradient migration of dissolved gasoline
components. The system operates by enhancing the
biodegradation of contaminated ground water that
passes through the barrier and could be a less
expensive method for treating contaminated ground
water than the techniques currently employed. The
potential advantages of a permeable barrier treatment
system include low maintenance requirements, no
above-ground facilities, and in-situ biodegradation of
contaminants with no requirement for disposal of
contaminated treatment media or ground water.
Contamination of ground-water supplies by
gasoline and other petroleum-derived hydrocarbons
released from underground storage tanks (USTs) is
a serious and widespread environmental problem.
Corrosion, ground movement, and poor sealing can
cause leaks in the tanks and associated piping. As of
1990, there were about 2 million underground tanks
storing gasoline in the United States with 90,000
confirmed releases reported between 1989 and 1990
(OUST, 1990).
In large spills, gasoline may penetrate the soil
and reach the saturated zone. Once gasoline comes
in contact with ground water, the more water-soluble
components, including benzene, toluene,
ethylbenzene, and the xylene isomers (BTEX), will
dissolve. Benzene has been identified as a carcin-
ogen, and the compounds TEX have been identified
as neurotoxins (NIOSH, 1990). Although these
aromatic hydrocarbons are relatively water-soluble,
they are contained in the immiscible bulk fuel phase
that serves as a slow-release mechanism for sustained
ground-water contamination.
Biodegradation and irreversible sorption are the
two main natural mechanisms that remove organic
materials in aquifers. Of these two mechanisms,
biodegradation is the major removal mechanism
(Major et al., 1988). Biodegradation of organic
contaminants within the subsurface results from the
activity of microorganisms as they obtain energy and
carbon to generate new cells. Microbial degradation
of a contaminant can result in mineralization
(complete degradation of the parent molecule to
inorganic end products) or biotransformation that
may yield other organic compounds as end products.
Biodegradation rates can vary two to three orders of
magnitude between aquifers or over a vertical
separation of only 1 or 2 m in the same aquifer
(Wilson et al., 1986). These rates are controlled by
environmental parameters such as temperature
(Thorton-Manning et al., 1987), community inter-
actions (Lewis et al., 1986), pH, electron acceptors
(Nakajima et al., 1984), salinity, mineral nutrient
availability (Lewis et al., 1986), competing organ-
isms, concentration of primary and secondary
compounds (Wilson et al., 1986; Schmidt et al.,
1987), and adaptation of microorganisms to the
pollutant (Spain and Van Veld, 1983; Lewis et al.,
1986).
Under favorable conditions, soil micro-
organisms will degrade most fuel hydrocarbons. In-
situ aerobic bioremediation has been shown to be
effective for many fuel spills. Controlled laboratory
and field studies have demonstrated that a variety of
1-1
-------
indigenous microbes can aerobically degrade
mixtures of aliphatic and aromatic compounds found
in gasoline and distillate fuels. All BTEX compon-
ents have been found to be biodegradable under
aerobic conditions (Wilson et al., 1983; Swindell et
al., 1988; Chiang et al., 1989; Song et al., 1990).
Early work on enhanced in-situ bioremediation at
contaminated aquifers involved sparging air in a
well, but the low solubility (9.2 mg/L at 20°C) of
oxygen increased the difficulty and expense of
maintaining aerobic conditions in ground water.
Using hydrogen peroxide (B^O^ to provide oxygen
to contaminated ground water can increase the
effective solubility of oxygen. However, disadvan-
tages of using H2O2 at elevated levels include its
toxicity to microorganisms, reactivity with inorganic
species, and rapid oxygen-release rate.
One approach for remediation of contaminated
aquifers that is attracting increased attention is the
installation of permeable reactive zones within the
aquifers. As contaminated ground water moves
under natural or induced hydraulic gradients through
a permeable reactive zone, the contaminants are
scavenged or degraded, and uncontaminated ground
water emerges downgradient of the permeable zone
(Gillham and Burris, 1992).
The full-scale permeable barrier system
examined in this study employs concrete prepared
with a proprietary formulation of magnesium
peroxide (MgO^. The concrete is loaded into
permeable filter socks and placed in fully-screened
polyvinyl chloride (PVC) wells (remediation wells)
installed perpendicular to the ground-water flow
direction. When ground water passes through a line
of remediation wells, the MgO2in the concrete reacts
with water, producing oxygen. Indigenous
microorganisms then use the released oxygen to
aerobically biodegrade the petroleum hydrocarbons
present in the ground water. Sodium nitrate
(NaNO3) may also be added to the concrete to
provide nitrogen, further enhancing biodegradation.
Laboratory batch experiments were conducted to
determine the oxygen-release characteristics of
several solid peroxide-concrete mixtures. A full-
scale barrier system was then installed at a UST
gasoline-spill site near Leland, North .Carolina.
Monitoring wells were installed upgradient and
downgradient of the barrier in the contaminated
portion of the aquifer. Ground-water samples were
monitored and analyzed for dissolved oxygen (DO),
individual BTEX components, and other relevant
parameters to assess the effectiveness of the barrier
system. According to the system design, high DO
and low BTEX concentration should be observed in
the remediation wells and downgradient monitoring
wells. At some distance downgradient of the barrier,
the BTEX concentration should be degraded below
regulatory levels.
Previous Studies of Permeable Barrier
Systems
Burris and Antworth (1990) and Hatfield et al.
(1992) performed bench-scale experiments modeling
subsurface sorption systems (SSSs) which are zones
of treated soil within an aquifer positioned
downgradient of a contamination source. These
zones retard the flow of contaminants through the
aquifer. Burris and Antworth (1990) performed
experiments using cationic organic surfactants to
form SSSs. Sorption coefficients for common
ground-water contaminants were shown to increase
by two to three orders of magnitude through
surfactant modification of aquifer sediment. Hatfield
et al. (1992) proposed SSSs consist of existing soils
or fill soils that contain a residual saturation of a non-
toxic sorbing organic phase (SOP) into which
hydrophobic ground-water contaminants partition.
These researchers performed experiments with
aquifer material containing decane at residual
saturation and observed increases in retardation
factors for common hydrophobic ground-water
contaminants of at least two orders of magnitude.
These hydrophobic contaminants partition
preferentially to organic material and are scavenged
from ground water by the SOP.
Starr and Cherry (1994) developed the Funnel-
and-Gate concept in which contaminated ground
water is forced to pass through a small permeable
reactive zone by the installation of low hydraulic
conductivity cutoff walls. The advantage of this
system, over a system in which the contaminated
ground-water plume is not funneled, is a smaller
permeable treatment zone may be used. The
1-2
-------
researchers presented various system configurations
including single-gate systems, multiple-gate sys-
tems, fully penetrating gates, and hanging gates.
They also presented five classes of in-situ reactors
that could be employed: 1) an in-situ reactor with
material that alters pH or redox potential; 2) a reactor
containing a material that dissolves and causes
precipitation of a mineral phase that immobilizes the
contaminant; 3) a reactor which removes contam-
inants via sorption; 4) a reactor which supplies
nutrients whose normal in-situ availability limits the
rate of biodegradation; and 5) a reactor in which a
physical removal or transformation of the contam-
inant occurs. As a result of ground-water modeling,
the critical factors in the performance of Funnel-and-
Gate systems were determined to be funnel width,
gate width, gate hydraulic conductivity, and retention
time.
Bianchi-Mosquera et al. (1994) performed a
short-term field study of the effectiveness of oxygen-
releasing concrete and slurry in the reduction of
injected benzene and toluene concentrations. These
researchers installed 20% MgO2 concrete briquets
and MgO2 "pencils" (MgO2 water slurry) in separate
treatment lines and injected benzene and toluene in-
to the aquifer at the Canadian Forces Base Borden to
achieve a concentration of 4 mg/L for each contam-
inant. Contaminant levels in ground water passing
through the MgO2 concrete line were below detection
limits in downgradient wells about 18 days after
installation. DO levels increased in downgradient
monitoring wells after installation of concrete
briquets, with peak values of 15 mg/L approximately
0.5 m from the concrete line. The installation of the
MgO2 "pencils" yielded reductions in benzene and
toluene levels and increases in DO concentrations
but not to the extent observed in the concrete briquet
zones. Field testing of the treatment zones was
performed over a 39-day period with no major
system inefficiencies encountered.
Cohen et al. (1996) proposed the use of peat and
nutrient briquets as media in permeable treatment
zones. They performed a series of experiments to
identify types of peat that had good potential for use
in a permeable barrier. High sorption capacity and
reasonably high hydraulic conductivity were
identified as important characteristics for peats to be
used as permeable barrier media. In addition, the
researchers developed nutrient briquets to supply
nitrate as an electron acceptor for the microbial
denitrification in a simulated contaminated ground-
water system. In bench scale studies, a combined
nutrient (nitrate) briquet and peat barrier removed up
to 85% of toluene and 71% of ethylbenzene from the
system. Cohen et al. (1996) and Thomson et al.
(1990) suggested possible field construction of
permeable barriers by trenching and backfilling with
treatment media in shallow, contaminated aquifer
systems.
Davis-Hoover et al. (1991) reported the use of
hydraulic fracturing to create permeable channels
that could be filled with granules of slow-dissolving
nutrients or oxygen-releasing chemicals. Hydraulic
fractures filled with sand act as permeable channels
to increase the rate of delivery and the area affected
by the injection of nutrient- or oxygen-bearing fluid.
Encapsulated sodium percarbonate was suggested as
a possible solid oxygen-releasing compound, but this
compound has a very short oxygen-release life. The
authors suggest that a longer lasting and less toxic
oxygen-releasing compound should be developed.
1-3
-------
-------
Section 2
Conclusions
1. Portland cement concretes incorporating solid
peroxides, which release oxygen at a controlled
rate, can be easily prepared. Concretes con-
taining either calcium peroxide or a proprietary
formulation of magnesium peroxide (ORC)
have desirable oxygen-release characteristics,
including high retention of the original oxygen
content and slowly declining oxygen-release
rates. Both concretes have useful oxygen-
release lives of 100 days or more. Concrete
prepared with urea hydrogen peroxide was
unacceptable for two reasons: 1) chemical
assays revealed that most of the original oxygen
was lost during the preparation of the urea
hydrogen peroxide concrete; and 2) oxygen-
release testing revealed that the oxygen that had
been retained by the concrete during
preparation was released in less than 10 days.
2. BTEX concentrations decreased and DO
concentrations increased during passage
through the active side of the permeable barrier
system. Reductions in BTEX concentrations
were statistically significant but were not
sufficient to contain the plume. BTEX
reductions on the control side of the barrier
were much greater than on the active side of the
barrier. The cause of this reduction is
unknown. Consequently, it is not possible to
determine whether the decline in BTEX was
due to the barrier system or due to natural
variations in BTEX concentration throughout
the site. The modifications made to the barrier
during the course of the project did not
dramatically improve BTEX removal
efficiency.
Batch reactor experiments indicated that nitrate
addition enhanced the aerobic biodegradation of
BTEX in ground water from the site.
Incorporating sodium nitrate (NaNO3) into the
concrete briquets at 0.5 to 0.7% by weight
during the second and third treatment periods,
respectively, did not cause regulatory levels for
nitrate to be exceeded. The highest nitrate
concentration observed downgradient of the
barrier was 2.9 mg/L NO3-N. Nitrate
concentration declined to near background level
further downgradient.
Remediation well clogging had a major impact
on oxygen delivery to the aquifer. Tracer tests
conducted at the end of the project indicated
that the average specific discharge through the
control remediation wells (no concrete) was
over 4 times higher than in the original
remediation wells that received concrete.
These results indicated that the active
remediation wells clogged over time. Also,
soil iron concentrations were significantly
higher around the active remediation wells
than in upgradient site soils, indicating that the
clogging was at least partially due to the
precipitation of insoluble iron oxides.
The average specific discharge in the original
active remediation wells was significantly
higher than in the new active remediation
wells. This difference is believed to be due to
well construction techniques. The new
remediation wells were vibrated into place
with no filter pack while the original
remediation wells were installed with a large
diameter hollow stem auger and filter pack.
2-1
-------
The vibration may have caused local
densification of the soil surrounding the new
wells with an accompanying decrease in aquifer
permeability.
6. The oxygen-releasing permeable barrier
installed near the Leland, NC site was not
effective in fully containing the dissolved
BTEX plume. Only toluene was reduced to
below regulatory standards in monitoring well
25 m downgradient. Benzene, ethylbenzene,
and the total xylene concentrations in
downgradient monitoring wells were
consistently above regulatory levels.
The failure of the oxygen-releasing permeable
barrier system to meet remediation objectives
was primarily due to two factors: 1) high total
BTEX concentration entering the barrier, and
2) high dissolved iron concentration entering
the barrier. The high total BTEX concen-
tration entering the barrier resulted in a high
demand for oxygen, which was difficult to
meet with a reasonable number of remediation
wells. The high iron concentration entering
the barrier caused clogging of the remediation
wells and reduced oxygen delivery to the
aquifer.
2-2
-------
Section 3
Recommendations
1. Future work on oxygen-releasing permeable
barriers should focus on sites with lower
concentrations of biodegradable organics and
dissolved iron. At sites where oxygen demand
or dissolved iron concentrations are high,
delivery of sufficient oxygen to the aquifer will
be very difficult.
2. The tracer test measurements of specific
discharge conducted at the completion of this
project were very useful in evaluating the
performance of the remediation wells. In future
work, before construction of the full-scale
barrier, field measurements of specific discharge
should be combined with laboratory
measurements of oxygen and nitrate release to
more precisely predict the amount of oxygen and
nitrate that will be introduced into the aquifer.
These measurements will allow a more rational
design of the permeable barrier and significantly
improve the probability of success.
3. The nitrate content of the concrete should be
increased for further enhancement of aerobic
biodegradation and for use as an electron
acceptor after the available oxygen is depleted.
A small increase in the nitrate content of the
concrete generally should not result in any
violations of water quality standards since the
maximum nitrate concentration observed in the
monitoring wells downgradient of the barrier
was 2.9 mg/L NO3-N, a value well below the
current ground-water standard of 10 mg/L
NO3-N.
4. Near the end of this project, significant con-
centrations of DO were reaching wells immed-
iately downgradient of the permeable barrier, yet
BTEX was not being biodegraded. The lack of
biodegradation could be due to stratification
within the aquifer, which reduces mixing of
oxygenated- and BTEX-contaminated, ground
water. In future work, variations in oxygen and
contaminant concentration with depth should be
examined to evaluate the importance of
stratification on mixing and subsequent bio-
degradation.
3-1
-------
-------
Section 4
Methodology
Laboratory Methods: Pre-Barrier
Construction
Prior to building the permeable barrier, labor-
atory studies were conducted to determine the
oxygen-releasing characteristics of solid peroxide
compounds when incorporated into concrete and to
determine the effect of nitrate addition on the aerobic
biodegradation of BTEX. Methods used in these
analyses follow.
Oxygen Retention in Solid Peroxide Concretes
Three solid peroxide compounds, magnesium
peroxide (MgO^), calcium peroxide (CaO2), and urea
hydrogen peroxide [CO(NH2)2»H2O2], were
analyzed for their oxygen-releasing characteristics
when incorporated into concrete. The first part of
this analysis involved performing chemical assays on
each compound and corresponding concrete mix to
determine the amount of available oxygen retained in
the concrete.
The oxygen content was determined by adding
75 to 100 mL of 1 M sulfuric acid to a flask
containing 0.5 to 0.6 grams of compound or
powdered concrete mix and titrating the solution
with standardized potassium permanganate (approx-
imately 0.1 N) to a light pink-purple end point
(Applied Power Concepts, Inc., 1992). This assay
was performed in triplicate on each compound and
on each corresponding concrete mix. The amount of
oxygen in the compound or concrete mix was
calculated as:
V = volume of potassium permanganate
solution added (mL)
W = weight of sample assayed (g).
(The factor 7.9997 in Equation 1 is the
grams per equivalent of oxygen.)
Oxygen Release Over Time from Solid Peroxide
Concretes
A technique based on soil respiration measure-
ment (Page et al., 1982) was used to measure the
oxygen-release characteristics of the three concrete
mixes. The procedure used an enclosed reactor to
measure changes in pressure, which were subseq-
uently related to the amount of oxygen released by
the concrete. Figure 1 presents a schematic of a
reactor used to measure oxygen-release rates. The
reactor consisted of a 500-mL jar and an attached
manometer for measuring pressure changes. The
reactor was initially filled with a known volume of
water, which was subsequently saturated with
oxygen. A known mass of concrete was then
completely submerged in the oxygenated water, and
the reactor was sealed.
Tubing
g sample
_(AQ(VQ(7.9997)
W
Clamp
Rubber Stopper
Water Level
Concrete
Briquettes
(1)
Manometer
Water
/
where: N = normality of potassium perman-
ganate solution
Figure 1. Schematic of reactor used for measuring
oxygen-release rates.
4-1
-------
One reactor was used for each concrete mix.
Measurements of change in the water-column height
in the manometers were taken periodically, then the
jars were vented. Pressure changes due to oxygen
release in the enclosed reactors were measured as
changes in water levels of the attached manometer.
The number of milligrams of oxygen released was
calculated as:
released = (&)(A/r
'manometer
(2)
where: k
= is the reactor constant
= change in water-column
height in the manometer.
The reactor constant (k) was determined for each
reactor based on the physical properties of the reactor
and environmental conditions as presented in
Equation 3.
k=-
(3)
where: k = reactor constant
Vg = volume of total headspace in
reactor
Vf = volume of water added to reactor
T = temperature in degrees Kelvin
a = solubility of oxygen at ambient
temperature
P0 = standard pressure
= 10,336 mm for distilled water.
Water level changes in the manometers were
corrected for barometric pressure fluctuations by
subtracting the water-level changes in a controlled
reactor constructed without concrete.
The reactors were loaded with different types
and sizes of solid peroxide concrete (10-cm-diameter
concrete cylinders or 4-cm-diameter briquets).
Oxygen-release rates were measured for the
following concrete mixes: 21% CO(NH2)2»H2O2
concrete briquets, 14% CaO2 briquets, 21% MgO2
cylinders and briquets, and 37% MgO2 cylinders and
briquets. A lower percentage CaO2 was used
because of the higher percentage available oxygen in
CaO2.
Effect of Nitrate Addition on Bioremediation
A batch biodegradation experiment was
performed to assess the influence of nitrate addition
on the aerobic biodegradation of BTEX. Gasoline-
contaminated ground water from a site in Leland,
NC, was collected from a monitoring well upgradient
of the concrete permeable barrier (SU7) and trans-
ferred to 125-mL serum bottles (100 mL in each).
Three bottles were amended with NaNO3 to produce
a final concentration of 100 mg/L NO3-N. To
another three bottles, only hydrochloric acid (HC1)
was added to reduce the pH to less than 2; these
served as controls. The remaining three bottles
received neither nitrate nor acid (ambient). All
bottles were sealed with an aerobic headspace and
incubated at 16°C in the dark. Aqueous samples
were periodically taken from each bottle and
analyzed for BTEX to determine the effect of the
nitrate addition.
Field Monitoring of Permeable Barrier
System
Site Description
Soil and ground-water contamination are present
at the Jennifer Mobile Home Park near Leland, NC,
due to the release of gasoline from a former UST
present on an adjoining property. The spill was
detected when dissolved hydrocarbons were found in
nearby domestic water supply wells. The water table
was less than 3 m (10 ft) below grade in the area
adjoining the former UST and is shallower in the
downgradient aquifer. Ground-water flow has
transported the gasoline components at least 150 m
(500 ft) downgradient from the UST. The former
UST and some petroleum-contaminated soil were
removed six months prior to installation of the
barrier. Excavation of the contaminated soil was
limited by the shallow water table and the foundation
of the nearby store.
The geology underlying the site consists mainly
of a medium gray-brown silty sand to a depth of 0.6
to 1.2 m (2 to 4 ft). This material changes to an
4-2
-------
orange-brown clayey silty sand for approximately 0.6
m (2 ft), becoming a medium to very coarse light
brown to blond sand at greater depth. This upper
sand unit extends 15 m (50 ft) or more below the
surface. The medium to very coarse sand layer
results in a single, unconfined aquifer within the
relevant depth of contamination throughout the site.
The average hydraulic conductivity of the aquifer
was estimated to be 23 m/d from drawdown and
recovery tests (LaTowsky, 1993). The vertical extent
of contamination is limited to within approximately
7.6 m (25 ft) of surface grade based on monitoring
from clustered wells.
Barrier Design
The permeable barrier intersected the BTEX
plume approximately 27 m downgradient from the
former UST location and consisted of a series of 15-
cm (6-in)-diameter PVC wells installed
approximately 1.5 m (5 ft) on center (Figure 2).
Each well was designed to release a plume of DO,
enhancing biodegradation in the downgradient
aquifer. Preliminary modeling indicated that the
plumes from each well would mix over a 6- to 15-m
distance resulting in complete biodegradation of the
BTEX plume. Field delineation of the BTEX plume
indicated that the barrier would need to be 40-m
wide and extend approximately 3 m below the
ground-water table.
MW2
Store
House
Ground-water
Flow Direction
Remediation Well
O Control Well
Monitoring Well
t.
MW1
Scale: 1 cm : 12 m
Figure 2. Site map showing permeable barrier and
monitoring wells.
Twenty remediation wells were initially installed
in the remediation line perpendicular to the plume at
a distance of 1.5 m on center (Figure 3). (Ten new
remediation wellsNR1 to NR10 were installed
later in the project. See Section 5 for further
discussion.) The nine original wells on the eastern
half of the plume did not receive concrete and were
operated as a control to evaluate the barrier
effectiveness. One of the remediation wells (R6) had
to be installed downgradient of the other wells due to
an overhead power line. A schematic of a remedia-
tion well is presented in Figure 4. Originally, 3-m-
long concrete columns with 10-cm diameters were
encased in filter fabric socks and hung inside
remediation wells. This design was modified over
the course of the project in an attempt to further
enhance the barrier system effectiveness.
Ground-
Water
Flow
NR9 NR7 NR5 NR3 NFU
'O
( C8 C7 C6 C5 C4 C3 C2 C1 R11R10R9 R8 R7 R5 R4 R3 R2 R1
Approximate limits of plume
Figure 3. Location of original and added permea-
ble barrier remediation wells (not to
scale).
Manhole Cover
ORC Concrete
In Filter Sock -
BTEX-
Bacterla -
BTEX '
^ Locking Cap
Ground Surface
Rope
\7 Ground-water Level
- 6-in-diameter
PVC Well
Biologically
Active Zone
BTEX + 02"C
-Pa
Figure 4. Schematic of remediation well
containing oxygen-releasing concrete.
4-3
-------
Three different mixes of concrete were used in
the operation of the barrier system. The concrete
was prepared by blending Portland cement, sand,
water, a proprietary formulation of MgO2 (Plant
Research Laboratories, Corona Del Mar, CA), and
NaNO3. The compositions (by weight) of the three
mixes are presented in Table 1.
Table 1. Mass Ratios of Components in Concrete
by Treatment Period
Treatment
Period
1
(Days 0
to 242)
2
(Days 242
to 361)
3
(Days 361
to 498)
Mg02
1
1
1
Portland
Cement
0.694
0.260
0.330
Sand
0.388
0.470
0.600
Water
0.643
0.670
0.770
NaNO3
0
0.013
0.020
Well Placement
Twenty monitoring wells (Fl, MW1, MW2, and
SU1 to SU17) were installed to define the plume,
monitor ground-water flow, and aid the permeable
barrier system design (Figure 2). North Carolina
State University (NCSU) installed the SU wells.
Monitoring wells SU1 to SU6 were used to
define the width and depth of the subsurface
hydrocarbon plume. They are situated approximately
50 m (175 ft) downgradient (Figure 2) from the
former UST along a transect whose center point is
roughly in line with the contaminated wells in the
study area. Monitoring data (Goin, 1995) indicate
that SU2 and SU6 are located at the edges of the
plume, and Fl and MW1 are located outside of the
plume area. The BTEX plume is reasonably well
defined and extends in a northeast direction away
from the former UST location and through the
central portion of the monitoring wells.
Monitoring wells SU7 and SU8 are 10 m (33 ft)
upgradient of the remediation line with SU7 located
on the active side 5 m from the plume centerline and
SU8 located 5 m from the centerline on the control
side. Monitoring wells SU9 and SU10 are similarly
located llm (36 ft) downgradient of the remediation
line. Monitoring well SU10 was damaged during
grading of the site and could no longer be sampled
after 86 days of barrier operation. Monitoring well
SU17 was installed to replace SU10.
Monitoring wells SU13 and SU14 were installed
3 m (10 ft) and 8 m (25 ft), respectively, down-
gradient from the remediation line along the same
stream line as SU7 and SU10. These wells allow for
ground-water sampling at positions immediately
downgradient of the remediation line. Monitoring
wells SU15 and SU16 were installed 0.75 m west
and east of SU13, respectively, to evaluate transverse
dispersion of oxygen and nutrients away from the
remediation wells.
Well Construction
Monitoring and remediation wells were
constructed in accordance with the applicable
standards of the NC Division of Environmental
Management. Monitoring wells were installed with
a hollow stem auger and consisted of 5.1-cm (2.0-in)
diameter PVC well casing with a 1.5-m (5-ft) long,
0.025-cm slot, PVC screen and end plug. A natural
sand pack was placed around the screened interval of
the well casing and a bentonite pellet layer was
installed above the sand pack to prevent infiltration
of surface water into the well. The well was
completed with the installation of a locking well cap,
metal identification tag, and steel cover set in
concrete. A dedicated Waterra model D-25
inertial pump attached to a section of high density
polyethylene tubing was installed in each monitoring
well. Ground-water samples were obtained by
vertically oscillating the tubing, thereby advancing a
column of ground water to the surface. During
sampling, a short section of new vinyl tubing was
attached to the surface end of the polyethylene tubing
to allow for easier sample collection.
Seventeen of the original 20 remediation wells
were installed by a private contractor. These
remediation wells consisted of 3 m of Schedule 40
PVC well screen with a 0.050-cm slot size attached
to 1.5 m of Schedule 40 PVC casing. The
installation procedure was similar to that previously
described for the monitoring wells but with a coarse
4-4
-------
filter pack installed along the entire length of the
well screen. A bentonite seal and locking well cover
were installed to prevent infiltration of surface water.
NCSU installed three of the original remediation
wells (R7, R8 and R9) and the ten new remediation
wells by vibrating 4.6 m of 15.2-cm-diameter well
screen (0.05-cm slot size) into a pre-augured pilot
boring. Due to the nature of the installation, no filter
material could be placed around the well screen. All
remediation wells were developed by repeated
surging with a high capacity pump.
Ground-Water Sampling
Ground-water samples were collected and
handled according to the protocol described in
Barcelona et al. (1988) with the following sequence
of events: 1) well purging; 2) sample collection; 3)
field blanks; 4) field determination; 5) preservation/
storage; and 6) transportation.
Prior to sampling, the monitoring well headspace
was purged with pre-purified argon gas to prevent
the introduction of atmospheric oxygen to the ground
water during purging and sampling. A minimum of
five well volumes were pumped from the well prior
to sample collection. Samples were collected,
filtered, labeled, and preserved according to the
information shown in Table 2. Field and equipment
blanks were collected and treated in the same manner
as all other samples. Field samples were stored on
ice in insulated ice chests and transported to the
NCSU Environmental Engineering Laboratories.
Upon arrival at the laboratory, samples were stored
in an ignition-safe refrigerator at 4°C and analyzed
within 48 hours.
Field analysis of ground-water samples included
measurement of DO, temperature, and pH. Ground-
water temperature and DO were measured using an
Orion Model 840 dissolved oxygen meter. The DO
meter probe was introduced into the well and placed
approximately mid-height in the existing water
column. The probe was slowly oscillated vertically
and readings recorded after equilibration. Sample
pH was measured using an Orion Model 920 ISE
meter with an Orion pH triode.
Table 2. Sample Collection and Preparation
Protocol
Analysis
Volatile
Organics
Back-up
Volatile
Organics
Metals
Nutrients
Alkalinity
Field
Analysis
PH
Container
40-mL
VOA Vial
40-mL
VOA Vial
40-mL
VOA Vial
40-mL
VOA Vial
225-mL
Polyethyl-
ene Jug
500-mL
Beaker
Label
ID
MW-X,
GC-1
MW-X,
GC-4
MW-X,
SS-6
MW-X,
SS-7
MW-X
None
Filter
No
No
Yes-
45 Mm
Yes-
45 //m
Yes-
45 ^m
Yes-
45 fj.m
Preserved
2.0 N HCI
added to pH 2
2.0 HCI
added to pH 2
2.0 HCI
added to pH 2
No
No
No
Volatile organic compounds (BTEX) were
analyzed using a Perkin-Elmer Model 9000 Auto
System Gas Chromatograph fitted with a flame
ionization detector, Tekmar Purge-and-Trap Model
LSC 2000, and a 75-m DB-624 megabore capillary
column. Sample analysis for Cl", Br", and SO4" was
conducted on a Dionex Ion Chromatograph. A
Perkin-Elmer Plasma II Inductively Coupled Plasma
Atomic Emission Spectrometer (ICP-AES) was used
for determining soluble concentrations of Na, K, Ca,
Mg, Fe, Al, Cu, and Mn. Nitrogen compound
analysis was performed using a LACHAT auto-
analyzer. Alkalinity was determined by titration to
pH 4.5 with 0.1 N HCI, and phosphate was deter-
mined using the ascorbic acid method (APHA,
1989).
Iron Content of Soil Adjoining Remediation Wells
In addition to collecting ground-water samples,
the field team collected soil samples to determine if
iron was precipitating next to the remediation wells.
Soil samples were collected for iron analysis
immediately adjoining three remediation wells (RIO,
Rll, NR10) and three locations 7 m upgradient of
the barrier. Soil samples adjoining the wells were
taken by inserting a modified 60-mL-syringe barrel
horizontally into the soil mass surrounding each well
until the screen was encountered, removing the
syringe, and placing the soil sample in an air-tight
4-5
-------
container for transport. Six soil samples (three
points spaced 15 cm apart at two different depths)
were collected at each well for iron content
determination. For comparison purposes, soil sam-
ples were collected from two depths at three
locations approximately 7 m upgradient of the active
remediation wells in the contaminated portion of the
aquifer. Upgradient soil samples were obtained by
drilling to a depth of approximately 10 cm above the
ground-water table with a stainless steel hand auger
and placing the soil samples in an air-tight container
for transport. After transport to the laboratory, the
readily extractable iron was determined by extracting
1-g soil samples with 1.0 N HC1 for 3 hours followed
by ICP analysis of the filtered extract (Lovely and
Phillips, 1986; Goin, 1995). Analyses were per-
formed in triplicate for each soil sample.
Specific Discharge Measurements
Tracer tests were conducted on several of the
active and control remediation wells to evaluate the
effect of the oxygen-releasing concrete on specific
discharge through the well. The tracer tests and
specific discharge calculations were performed
following the procedure outlined by Hall (1993).
The concrete briquets were removed from the wells
and the water level allowed to equilibrate before
tracer addition. Background specific conductivities
and ground-water temperatures were measured at
several depths using a YSI Model 33 meter and a
YSI33000 Series probe. The tracer, consisting of a
solution of 100 or 250 g of sodium chloride (NaCl)
in 1 L of distilled water, was then vigorously mixed
into each well. Specific conductivity and temper-
ature readings were taken at the same depths in each
of the wells over a two-day period. Specific
conductivity was converted to NaCl concentration
using a standard curve developed by measuring
conductivities of solutions with known NaCl concen-
trations at the ambient ground-water temperature.
The slope of the natural log of the tracer
concentration versus time was determined by linear
regression. Equation 4 (Hall, 1993) was used to
calculate the specific discharge (q) of ground water
through the well.
(4)
q = \ v.
where: V
A
C
t
dt
volume of the water-filled test
interval
cross-sectional area of the test
interval normal to flow
tracer concentration
time.
4-6
-------
Section 5
Results and Discussion
Laboratory Results: Pre-Barrier Construction
Concretes prepared contained CaO2, MgO2, and
CO(NH2)2»H2O2 as potential sources of oxygen to
enhance BTEX biodegradation. Urea hydrogen per-
oxide releases oxygen due to the decomposition of
hydrogen peroxide (H2O2). Reaction of CaO2 and
MgO2 with water releases oxygen as shown below:
CaO2 + H2O - Ca(OH)2 + 0.5O2
MgO2 + H2O - Mg(OH)2 + 0.5O2
Oxygen Retention in Solid Peroxide Concretes
In the first part of this analysis, different concrete
mixes were prepared and analyzed in triplicate to
determine the fraction of the original oxygen retained
in the final concrete mix. Average oxygen contents
for each compound and mix are shown in Table 3.
Table 3. Average Oxygen Contents of CaO2,
MgO2, and CO(NH2)2>H2O2 in Original
Form and in a Concrete Matrix
Compound or
Concrete Mix
CaO2
14%a CaO2 Concrete
MgO2 (ORC)b
21% MgO2 Concrete
37% MgO2 Concrete
CO(NH2)2.H2O2
21%CO(NH2)2.
H2O2 Concrete
Oxygen
Content
(mg O2/g of
material)
49.75 ±1.21
7.36 ± 0.64
66.82 ±1.59
14.55 ±1.08
24.25 ±1.80
64.69 ± 3.54
1.67 ±0.36
Oxygen
Recovery0
(%)
106
104
98
12
Percentage of compound by weight in concrete.
A proprietary formulation of magnesium peroxide (Plant Research
Laboratories, Corona Del Mar, CA) was used in this study.
Oxygen Recovery = (oxygen content of concrete/percentage of
compound in concrete)/(oxygen content of compound); i.e., 14%
CaO2 average oxygen recovery = ((7.36 mg CVg mix) /
(0.14))/(49.75 mg O^g of compound) = 106%.
The oxygen recoveries for MgO2 and CaO2 con-
crete mixes were close to 100% based on the oxygen
content of the original compound. In contrast, a
large portion of the available oxygen in the original
CO(NH2)2»H2O2 was lost during preparation of this
concrete.
Oxygen Release Over Time from Solid Peroxide
Concretes
The CO(NH2)2«H2O2 concrete was highly re-
active and released oxygen at rates as large as 2.5 mg
per gram of CO(NH2)2»H2O2 per day. This high
release rate, along with the low-oxygen retention for
the CO(NH2)2»H2O2 concrete, resulted in rapid
depletion of the available oxygen (Appendix A).
After 10 days of operation, no measurable oxygen
was being released from the CO(NH2)2«H2O2
concrete. The rapid decline in oxygen-release rate
indicates that CO(NH2)2»H2O2 concrete would not be
acceptable for use in long-term bioremediation
activities.
Experimental results for MgO2 and CaO2 con-
crete were modeled assuming the oxygen-release rate
declined linearly with time. The modeled linear
regression equations for each material are provided
in Table 4. Figure 5 shows oxygen-release rates over
time for 21% and 37% MgO2 concrete briquets along
with the best fit line. Oxygen-release rates from the
37% MgO2 briquets and cylinders closely matched
the modeled linear regression. However, oxygen-
release rates from the 21% MgO2 and 14% CaO2
were much more variable, resulting in a much poorer
model fit (r2 < 0.7).
5-1
-------
Table 4. Model Oxygen-Release Rate Equations
for Magnesium Peroxide and Calcium
Peroxide Concrete Mixes
needed, the 37% MgO2 or 14% CaO2 concretes may
be more useful.
Oxygen-
releasing
Media
37% MgO2
Briquets
37% MgOj,
Cylinder
21%Mg02
Briquets
21%MgO2
Cylinder
14% CaOa
Briquets
Oxygen-Release Equation
Rate*= 1 1 .30 - 0.06268 Time**
Rate = 7.88 - 0.0461 9 Time
Rate = 3.55 - 0.01 065 Time
Rate = 2.41 - 0.00777 Time
Rate = 27.24 - 0.2853 « Time
r2
0.94
0.94
0.65
0.69
0.57
* Rate = units of mg CVday/g available O2,
** Time = days.
A 37% briquets
37% model
D 21% briquets
21% model
100 150 200 250 300 350
Days
Figures. Oxygen release from MgO2 concrete
briquets (1.7-cm diameter).
Figure 6 is a plot of oxygen-release rates over
time, estimated using the linear regression equations
for different concrete mixtures and sizes. Use of the
large cylinders (10-cm diameter) slowed the oxygen-
release rate somewhat, presumably due to the slower
rate of water entry into the cylinders. The 21%
MgO2 concrete cylinders and briquets released
oxygen at measurable rates for up to 300 days, while
the 14% CaO2 briquets were exhausted by 100 days.
Where a slow constant release of oxygen is required,
the 21 % MgO2 concrete briquets and cylinders will
be most useful. When a higher O2 release rate is
-©- 37% MgO2 bnquets
-- 37% MgO2 cylinder
-0- 21% MgO2 briquets
21% MgO2 cylinder
14% Ca02 briquets
150 200 250 300 350
Days
Figure 6. Best fit estimated lines showing vari-
ation in oxygen-release rates with time
for magnesium peroxide and calcium
peroxide concrete mixes.
Effect of Nitrate Addition on Bioremediation
Figure 7 shows the results of the nitrate addition
experiments. In the bottles amended with 100 mg/L
NO3-N, significant BTEX degradation was observed
after 10 days. Total BTEX concentration dropped
from 22.2 mg/L to below 0.035 mg/L after one
month of incubation. In the bottles with no added
nitrate or acid (ambient), no significant change in
BTEX concentrations was observed after 10 days.
Significant BTEX degradation was observed only in
two of three bottles after 30 days of incubation.
Over this period, the average total BTEX
concentration dropped from 22.2 to 7.4 mg/L in this
comparison group. In the abiotic control bottles,
there was a small initial abiotic loss, then BTEX
concentrations remained constant. These results
indicate that nitrate addition enhanced the rate of
aerobic biodegradation. Since an aerobic headspace
was maintained, the added nitrate is believed to have
enhanced biodegradation by increasing nitrogen to
non-limiting levels, thereby increasing biomass
synthesis; the increased biomass caused the higher
BTEX degradation rate.
5-2
-------
10 15 20 25
Time (Days)
30
35
-Q- Abiotic Control -jAr Ambient -^- Nitrate Addition
Figure?. Effect of nitrate addition on BTEX bio-
degradation in ground water from gaso-
line-contaminated site near Leland, NC.
(Symbols are the mean of three replicates.
Error bars are ± one standard deviation.)
Field Monitoring of Permeable Barrier
System
Ground water upgradient and downgradient of
the full-scale permeable barrier was monitored over
an 18-month period to determine the barrier system's
effectiveness and to identify areas where the design
could be improved. Field monitoring data are on
Appendix B. The barrier first built at the site
consisted of 20 six-inch-diameter PVC wells
screened from 1.5 to 5 m below grade (Figure 2).
Oxygen-releasing concrete was installed in the
western line of wells (R1-R11). The eastern line of
wells (C1-C9) served as a control and did not receive
concrete.
Background Ground-Water Quality
Monitoring well MW1 is located outside of the
plume and is representative of background ground-
water quality. BTEX was consistently below
detection in MW1 (<5 ^g/L of each component).
The average DO and pH were 3.5 mg/L and 5.1,
respectively. Average concentrations of nutrients
and major ions were 9 mg/L bicarbonate alkalinity,
1.4 mg/L NO3-N, <0.1 mg/L NH4 -N, <0.1 mg/L
PO4-P, 9 mg/L SO4, 9 mg/L Cl, <0.1 mg/L Fe, <0.1
mg/L Mn, 2 mg/L Ca, 1 mg/L Mg, 9 mg/L Na, and
1 mg/L K. These results indicate that the back-
ground water quality was somewhat acidic, was very
low in dissolved solids, and contained low levels of
inorganic nutrients.
Ground-Water Monitoring
Monitoring-well data indicate an average
hydraulic gradient of 0.0043 m/m over the project
period. Figure 8 presents hydraulic gradients and
ground-water flow directions for selected days within
the project period. The length of the arrows are
proportional to the hydraulic gradient on that day.
The hydraulic gradient data indicate that, while there
were small fluctuations, the average ground-water
flow direction is closely aligned with the row of
monitoring wells: SUV, SU13, SU14, SU17, and
SU5 (Figure 2).
t
© Remediation Well
O Control Well
Monitoring Well
Remediation
Wells
Arrow Scale:, 1 cm: 0.0026 m/m
Figure 8. Variation in ground-water flow direction
and gradient over the project period.
(Monitoring day is shown next to each
arrow.)
Ground water upgradient and downgradient of
the permeable barrier was monitored from 33 days
before startup of the barriers to 498 days after
startup. The row of monitoring wells SU7, SU13,
SU14, SU17, and SU5 were monitored
approximately twice per month. These wells were
installed on a single stream line to determine the
variation in BTEX, oxygen, and indicator parameters
as ground water was transported through the barrier
and the downgradient aquifer. Other wells at the site
were monitored less frequently.
Over the course of this project, two major
modifications were made to the barrier system in an
attempt to improve treatment efficiency: 1) use of
smaller concrete briquets containing MgO2 and
NaNO3, and 2) installation of additional remediation
wells. The smaller concrete briquets were used to
increase the oxygen-release rate from the existing
wells. Addition of NaNO3 was to enhance bacterial
5-3
-------
growth and resulting biodegradation rates in the
downgradient aquifer. Ten new remediation wells
(NR1 to NR10) were added 1.5 m upgradient of the
existing barrier to further increase the oxygen supply
to the aquifer. Figure 3 shows the placement of the
new remediation wells. To evaluate the effect of the
original and modified barrier on the contaminant
distribution, the monitoring data has been separated
into three periods: 1) day 0 to 242 (original barrier
cylinders or briquets without NaNO3); 2) day 242 to
361 (original barrier - briquets with NaNO3); and 3)
day 361 to 498 (barrier with additional wells -
briquets with NaNO3).
Variability in BTEX and Indicator Parameters
Upgradient of the Barrier
Monitoring results for wells SU7 and SU8 are
shown in Figure 9. These wells are located 10 m (33
ft) upgradient of the active and control sides of the
barrier, respectively, and showed an almost identical
trend in total BTEX concentration over time. The
similarity in concentrations between these two wells
indicates that the distribution of total BTEX
concentration across the width of the plume is
relatively symmetric with respect to the longitudinal
axis upgradient of the barrier. BTEX concentration
started out very low and increased steadily over the
first 100 days of barrier operation. Prior to startup of
the barrier, the site experienced a period of very
heavy precipitation. The high precipitation is
believed to have diluted the contaminants, resulting
in lower BTEX concentrations in the aquifer
immediately before startup. Over time, the effects of
the high recharge diminished, and the BTEX
concentrations in both wells returned to the 15 to 40
mg/L range. The low BTEX concentrations
observed around days 180 and 390 were also
associated with periods of high ground-water re-
charge and high water table elevation. Dissolved iron
concentrations in SU7 and SU8 averaged 19 and 22
mg/L, respectively. The pH values in both wells
were approximately 6. The higher pH in the con-
taminated wells is believed to be due to Fe(OH)3
reduction in the upgradient aquifer. The reduction of
the Fe(OH)3 releases OH" ion and this release
increases the ground-water pH.
-100
100
300
400
500
200
Days
Figure 9. Variation in total BTEX concentration in
monitoring wells upgradient of the
active (SU7) and control sides (SU8) of
the permeable barrier.
Evaluation of Permeable Barrier: Test Period 1 -
Day 0 to Day 242
The original barrier system was installed on two
different days. The first three remediation wells (R7,
R8, R9), which are directly upgradient of SU13,
were loaded with 37% MgO2 concrete on day 0(1-
28-93). The remaining eight active remediation
wells were completed and loaded with MgO2
concrete on day 9. The first treatment period
extended from this initial loading of the remediation
wells to day 242 when concrete containing NaNO3
was installed. Discussion of monitoring results
primarily focuses on data from the row of monitoring
wells SU7, SUB, SU14, SU10, and SU5, since
information from these wells illustrates the effects of
the remediation system. Complete monitoring results
for all parameters and wells are reported by Goin
(1995).
Monitoring well SU13 is located 3 m (10 ft)
downgradient of the active side of the barrier on
approximately the same stream line as upgradient
well SU7. The total BTEX and DO concentration
data for wells SU7 and SU13 are compared in
Figures lOa and lOb for the first treatment period.
These plots show that total BTEX concentration in
SU7 was low initially, then climbed to the 15 to 30
mg/L range after completion of the barrier. DO in
SU7 was typically low (<0.5 mg/L) although high
oxygen measurements were observed on days 0 and
63 and were associated with periods of high
recharge. The average total BTEX and DO concen-
trations in SU7 from day 0 to day 242 were 17 and
0.4 mg/L, respectively.
5-4
-------
On day 0, the total BTEX concentration in wells
immediately upgradient (SU7) and downgradient
(SU13) of the barrier were similar (~7 mg/L).
However, by day 9, the total BTEX concentration
had started to decline in SU13 while total BTEX
concentration in SU7 continued to increase. After
some initial fluctuations, the BTEX concentration in
SU13 appeared to have stabilized below 2 mg/L until
day 139. During this period, DO in SU13 followed
an opposite pattern to BTEX. On day 0, DO
concentration in SU13 was low (0.7 mg/L) and then
increased to between 1.5 and 3.0 mg/L. On day 139,
DO concentration dropped to 0.8 mg/L in SU13 and
on the subsequent sampling BTEX increased to 11
mg/L. The drop in oxygen and increase in BTEX
concentration in SU13 was probably due to reduced
oxygen release from the concrete. At that point, the
oxygen-releasing concrete had reached the end of its
operating life and was probably not releasing
sufficient oxygen. On day 170, the old concrete
cylinders were replaced with fresh concrete briquets.
Briquets were used in place of cylinders to increase
the oxygen-release rate. Replacement of the concrete
250
appeared to improve barrier performance, and on the
next sampling event, DO concentration increased and
BTEX concentration decreased in SU13.
Figures lla and lib show total BTEX and DO
concentrations for wells SU14 and SU5 located 8 m
(25 ft) and 23 m (75 ft) downgradient of the barrier.
Immediately after construction of the barrier, total
BTEX concentration in SU14 and SU5 continued to
increase following the same pattern as the upgradient
well SU7. After day 40, DO concentrations in SU14
began to increase which corresponded to a BTEX
decline. The lag in oxygen arrival and BTEX
removal in well SU14 is believed to be due to the
travel time from the barrier to these wells. Using the
non-reactive transport velocity of 0.3 m/day, oxygen
released from the barrier would not be expected to
arrive at well SU14 until day 25. A small increase in
DO and decrease in BTEX was observed in SU5 on
day 50. These changes are not believed to be due to
the barrier since oxygen released from the barrier
would not be expected to arrive at SU5 until day 75.
250
50 100 150 200
Days
250
50 100 150 200
Days
250
Figure 10. Variation in (a) total BTEX concen-
trations and (b) dissolved oxygen
concentrations during test period 1 (day
0 to day 242) in monitoring wells up-
gradient (SU7) and downgradient
(SU13) of the barrier.
Figure 11. Variation in (a) total BTEX concen-
trations and (b) dissolved oxygen
concentrations during test period 1
(day 0 to day 242) in monitoring wells
SU14andSU5.
5-5
-------
On day 115, DO concentrations were 1.0 mg/L in
SU14 and only 0.3 mg/L in SU5. These concen-
trations indicate that whatever oxygen was being
released to the aquifer by the permeable barrier, it
was essentially depleted before it reached SU5.
After 150 days, DO decreased in both SU14 and
SU5. After the concrete in the remediation wells was
replaced on day 170, the DO increased and BTEX
decreased in SU14 as seen by results on day 220.
As part of the original experimental design,
monitoring wells were installed 11 m (36 ft)
downgradient of both the active (SU10) and control
(SU9) sides of the barrier. Figure 12 shows total
BTEX concentrations in these two wells for the first
86 days of barrier operation. The total BTEX
concentration on the control side (SU9) is much
lower than on the active remediation side (SU10).
The reason for this difference is not known. The
barrier could not be the cause because water released
from the barrier should not reach these wells until at
least day 50. Regardless, BTEX concentrations are
much lower on the untreated (control) side, than on
the side which was treated with oxygen-releasing
concrete.
80 100
Figure 12. Variation in total BTEX concentrations in
monitoring wells downgradient of the
active (SU10) and control (SU9) sides of the
permeable barrier.
In general, concentrations of inorganic nutrients
(N03-N, NH,-N, PO4 -P) and ions, as well as pH,
upgradient and downgradient of the barrier were very
similar. The average pH in SU7, upgradient of the
barrier, and in SU13, immediately downgradient of
the barrier, was 5.9 and 6.1, respectively.
Concentrations of PO4-P and NO 3-N were below
detection (<0.5 mg/L) both upgradient and
downgradient of the barrier. Ammonia (NH4-N)
decreased slightly from 1.8 mg/L to less than 0.5
mg/L during passage from SU7 to SU13. Two
notable exceptions to this trend were Ca and Mg. On
the first sampling date after the concrete installation,
there was an abrupt increase in Ca and Mg
immediately downgradient of the barrier (SU13); but
by the second sampling date, both Ca and Mg had
returned to background levels. The temporary
increase in Ca and Mg is believed to be due to the
dissolution of fine powder produced during handling
of the concrete. Once this powder was depleted, the
rate of Ca and Mg release returned to background
levels. The only inorganic parameter which was
consistently affected by the barrier was dissolved
iron. From day 0 to 242, the average concentrations
of dissolved iron upgradient (SU7) and downgradient
(SU13) of the barrier were 19 and 7 mg/L,
respectively.
Evaluation of Permeable Barrier: Test Period 2 -
Day 242 to Day 361
After the completion of the first treatment
period, it was apparent that while the oxygen-
releasing barrier was having some beneficial effects,
the existing barrier was not fully effective in
containing the contaminant plume. Field monitoring
had shown that dissolved nitrogen and phosphorus
concentrations in the ground water were very low
(PO4-P < 0.5 mg/L, NH,-N < 0.5 mg/L). Based on
this initial work, laboratory batch experiments were
conducted to evaluate the effect of nitrate addition on
the BTEX biodegradation rate (see page 19). Results
of these experiments indicated that nitrate addition
could potentially increase the BTEX biodegradation
rate. In test period 2, concrete briquets were
prepared from a mixture of 41% MgO2 and 0.5%
NaNO3 and installed in the existing remediation
wells. This formulation was selected to provide
sufficient nitrogen for bacterial growth but to be low
enough in nitrogen to ensure that the ground-water
quality standard of 10 mg/L NO3-N was not violated.
Total BTEX and oxygen concentrations in
monitoring wells SU7, SU13, and SU14 are shown
in Figures 13a and 13b for the second treatment
period. Total BTEX concentration in the upgradient
well (SU7) was relatively consistent during this
period, ranging from 18 to 32 mg/L with an average
of 26 mg/L. BTEX levels in SU13 and SU14 were
5-6
-------
lower than in SU7, but were much higher than in the
previous treatment period. At this point, there was
no clear explanation for the drop in BTEX removal
efficiency. Oxygen levels in SU13 were somewhat
lower than in the previous treatment period,
indicating that oxygen was not penetrating the
aquifer in sufficient quantities to biodegrade the
plume. In contrast to the disappointing results with
BTEX and oxygen, nitrate concentrations behaved as
expected. Average nitrate concentrations in SU13
and SU14 were 0.88 and 0.91 mg/L, respectively. In
comparison, the average nitrate concentration in SU7
upgradient of the barrier was less than 0.5 mg/L.
The maximum nitrate concentration detected in any
well was 2.9 mg/L NO3-N, which is well below the
ground-water quality standard of 10 mg/L NO3-N.
Sodium (Na) release from NaNO3 present in the
concrete was negligible. During test period 1, the
average Na concentration in SU13 was 5 mg/L,
while in test period 2 the average concentration was
7 mg/L.
240 260
b)
280 300 320
Days
340 360 380
.i
2.5
2.0
^
c
§>1-5
I1-0
I
8 0.5
in
5
0
.A.
240 260 280
300 320
Days
340 360
380
As previously observed, there were no
significant changes in pH or most dissolved ions
during transport through the barrier. Dissolved iron
concentrations were consistently lower in wells
downgradient of the barrier than in wells upgradient
of the barrier.
Evaluation of Permeable Barrier: Test Period 3 -
Day 361 to Day 498
During the first two test periods, BTEX
concentrations were reduced by approximately 12 to
16 mg/L over an 18-m distance. While this result
was promising, significant concentrations of BTEX
persisted downgradient of the barrier. Since the DO
levels in downgradient wells were low, availability of
oxygen was believed to be limiting further
biodegradation of BTEX. Ten new remediation
wells were installed at the beginning of the third
treatment period to provide additional oxygen and
further enhance BTEX biodegradation.
The new wells were installed between the
existing wells just upgradient (~1.5 m) of the active
side of the barrier. The net effect of this installation
was that oxygen-releasing wells were spaced
approximately 0.75 m on center over the western half
of the BTEX plume. Both the new remediation wells
and the original active remediation wells were loaded
with fresh concrete briquets prepared with 37%
MgO2and 0.7% NaNO3 on day 361. Two additional
monitoring wells (SU15 and SU16) were installed
directly adjoining SU13 to evaluate oxygen and
nutrient transport transverse to the direction of flow.
Total BTEX and DO concentrations in wells
SU7 to SU17 are shown in Figures 14a and 14b for
the third treatment period. SU17 replaces SU10,
which was damaged during site grading. Due to the
proximity of monitoring wells SU13, SU15, and
SU16, the BTEX and oxygen concentrations in these
three wells are averaged to represent contaminant
levels immediately downgradient of the barrier.
Figure 13. Variation in (a) total BTEX concentrations
and (b) dissolved oxygen concentrations
during test period 2 (day 242 to day 361) in
monitoring wells SU7, SU13, and SU14.
5-7
-------
360 380
400 420 440
Days
460 480 500
360 380 400 420 440
Days
460 480 500
Figure 14. Variation in (a) total BTEX concentrations
and (b) dissolved oxygen concentrations
during test period 3 (day 361 to day 498) in
monitoring wells SU7, SU14, SU17 and the
average of SU13, SU15, and SU16.
Total BTEX concentration upgradient of the
barrier (SU7) was 19 mg/L at the beginning of the
third treatment period and then decreased to 6.1
mg/L on day 410 before rebounding to over 20 mg/L
for the remainder of the treatment period. DO
followed an inverse trend in SU7, increasing from
1.6 mg/L on day 361 to 3.3 mg/L on day 410 and
then gradually declining. The high DO and low
BTEX concentrations on day 410 were associated
with a period of high water-table elevation.
DO concentrations in all wells downgradient of
the barrier increased after the installation of MgO2-
nitrate briquets in the new remediation wells. The
average of the oxygen concentrations in wells SU13,
SU15, and SU16 increased from 1.0 mg/L on day
361 to 5.2 mg/L on day 410, indicating that oxygen
released from the barrier was being transported
through the aquifer. Farther downgradient in wells
SU14 and SU17, oxygen also increased after a lag
period. DO levels in the active remediation wells
were high, ranging from 7 to 30 mg/L more than 100
days after the beginning of the third treatment period.
One surprising observation was that BTEX
levels remained fairly constant in the wells
downgradient of the barrier even though DO
concentrations increased. While there was a
measurable decline in total BTEX concentrations in
wells SU13, SU15, and SU16 when oxygen
increased on day 410, substantial concentrations of
BTEX persisted. This trend continued when the data
from individual wells were examined. Figures 15a
and 15b show total BTEX and oxygen
concentrations in the individual wells SU13, SU15,
and SU16. All of these wells were located 3 m
downgradient of the barrier and were spaced 0.75 m
apart perpendicular to the ground-water flow
direction. Total BTEX and DO concentrations were
similar in these three wells with no consistent
differences. This similarity indicates that oxygen
was being distributed laterally through the ground
water. While high concentrations of oxygen were
reaching these wells, high concentrations of BTEX
continued to persist. For the period from day 382 to
438, the average DO and total BTEX concentrations
in well SU16 were 6.7 and 2.3 mg/L, respectively.
We hypothesize that the continued presence of
high concentrations of both DO and BTEX in several
of the monitoring wells may be due to inadequate
mixing between layers in the aquifer. If high oxygen
concentrations were present in one layer and high
BTEX concentrations were present in an adjoining
layer, there would be little opportunity for
biodegradation, yet monitoring wells screened over
the two layers would show high concentrations of
both oxygen and BTEX.
Nutrient and indicator parameter concentrations
followed the same general trends as observed in the
second treatment period. Nitrate concentrations in
downgradient wells ranged from 0.7 to 2.9 mg/L
NO3-N, indicating that small amounts of nitrate were
being transported into the downgradient aquifer.
Dissolved iron concentrations continued to be
somewhat lower in wells downgradient of the barrier
than upgradient. The 'prf^and other dissolved ion
concentrations were similar to values observed in the
second treatment period.
5-8
-------
360 380 400 420 440 460 480 500
Days
500
Figure 15. Variation in (a) total BTEX concentrations
and (b) dissolved oxygen concentrations
during test period 3 (day 361 to day 498) in
monitoring wells SU13, SU15, and SU16.
Remediation Well Clogging
A preliminary review of the field monitoring
results was performed to determine if dissolved iron
concentrations were being reduced during passage
through the barrier. Dissolved iron concentrations
in SU7 were used to represent upgradient conditions.
Before day 382, iron concentrations in SU13 were
used to represent downgradient conditions. After
day 382, the average of SUB, SU15, and SU16 was
used. Results of a Student's t-test indicated that the
iron concentration in SU7 was statistically greater
than the iron concentrations in the downgradient
location at the 99% confidence level. The mean
difference in concentrations was 12.1 mg/L with a
standard error of 3.0 mg/L.
The observed decline in iron concentration
during passage through the barrier could be due to
oxidation of soluble iron (Fe*2) by oxygen released
from the remediation wells and precipitation as
insoluble iron oxides (Fe+3). If the iron oxides
precipitated near the active remediation wells, the
aquifer could become clogged resulting in large re-
ductions in aquifer permeability and reduced barrier
efficiency. Brown and Norris (1994) observed that
elevated dissolved iron concentrations are often
found in conjunction with hydrocarbon plumes.
These authors indicate that oxygen from all sources
causes iron precipitation and subsequent plugging
around oxygen injection points. Heersche et al.
(1994) reported that iron hydroxide precipitation
reduced the withdrawal rate in a pumping system at
a gasoline-contaminated site when using hydrogen
peroxide as the oxygen source in a bioremediation
system. Brown and Norris (1994) suggested the use
of tripolyphosphate as both a phosphorous source
and to reduce the impact of iron precipitation on
bioremediation system operations.
Specific Discharge Measurements
The potential for iron clogging of the remedia-
tion wells in the permeable barrier system was
investigated by conducting tracer tests and specific
discharge measurements on the remediation wells
after the end of test period 3. Several wells were
then excavated to visually examine the aquifer
material and collect samples for iron analysis.
The mean and standard deviation for the specific
discharges (q) of the original active remediation
wells, the new active remediation wells, and the
control wells are presented in Table 5. Tracer test
data for individual wells are reported by Goin
(1995). Regression coefficients (r2) for all wells
were greater than 0.85 indicating a strong correlation
between changes in tracer concentration and time.
Table 5. Specific Discharges for Remediation Well
Groups Estimated from Tracer Tests
Well Group
Control Wells
Original
Active Weils
New
Active Weils
Number
of Wells
Tested
5
9
6
Mean q
(rn/day)
0.2233
0.0476
0.0174
Std.
Deviation
(m/day)
± 0.0362
±0.0164
± 0.0032
5-9
-------
The mean specific discharge of the control well
group was significantly greater than the specific
discharges of both the original active well group and
the new active well group at the 99% level using a
one-tailed Student's t-test. The mean specific
discharge of the original active remediation well
group was also significantly larger than the new
active well group at the 99% confidence level.
The difference in specific discharge between the
control remediation well group and the two active
remediation well groups was hypothesized to be the
result of iron precipitation around the active
remediation wells. The difference in the specific
discharges of the original active remediation wells
and the new active remediation wells is believed to
be partially due to differences in well construction
techniques. With the exception of remediation wells
R7, R8, and R9, the original active remediation wells
were installed in an oversized augered borehole with
a coarse sand pack placed along the full length of the
well screen. The new remediation wells were
installed by vibrating the well screen into place
without an artificial sand pack. The vibratory
method of installing the new remediation wells may
have caused localized densification of the aquifer
material with an accompanying reduction in
permeability.
Iron Content of Soil Adjoining Remediation Wells
Ferrous iron (Fe+2) will rapidly precipitate in the
presence of DO forming an insoluble ferric
hydroxide Fe(OH)3 according to the equation:
Fe+2 + 0.25O2 + 0.5H2O + 2OH' >
Fe(OH)3 (solid)
We hypothesized that this reaction occurred in the
imme'diate vicinity of the barrier, resulting in a loss
of permeability and lower specific discharge through
the remediation wells. To evaluate this hypothesis,
the aquifer adjoining two of the original active
remediation wells (RIO and Rl 1) and one new active
remediation well (NR10) was excavated to a depth of
approximately 10 cm above the ground-water table
and visually examined for evidence of clogging.
There was no apparent change in color between
the soils upgradient and directly adjoining the
barrier. The lack of distinct color change is not
surprising. Soils in the saturated zone are a very
coarse light brown to blond sand. Minor red or
brown staining due to iron hydroxide cement would
probably be obscured by the existing soil color.
While there was no apparent change in color,
there was a distinct change in the cohesive strength
of the soil. Soils upgradient of the barrier flowed
easily with essentially no cohesive strength. Soils
immediately adjoining and downgradient of the
barrier had considerable cohesive strength and
formed a vertical face immediately adjoining the
active remediation wells. Upon drying, soils
collected adjoining the remediation wells retained
this cohesive strength although they could be easily
crumbled by hand. This indicates that some type of
cement had formed binding the soil particles.
The moisture contents of the soil samples were
calculated in conjunction with the iron content
analyses so that the iron contents could be expressed
in terms of soil dry weight. Extractable iron
concentrations for each location sampled are
presented in Table 6.
Table 6. Extractable Iron Content in Soils Adjoining
the Remediation Wells and Upgradient of
the Barrier
Sample
Location3
R10
R11
NR10
Upgradientb
Iron Concentration
(mg Fe/g of dry soil)
0.415
0.272
0.254
0.159
Standard
Deviation
(mg Fe/g of dry
soil)
0.091
0.080
0.071
0.110
Soil samples were taken in triplicate at two depths for each
sample location.
Upgradient refers to soil samples obtained approximately 7 m
upgradient of the permeable barrier.
The soil iron contents around the three remed-
iation wells were compared with the upgradient soil
iron content using a one-tailed Student's t-test. The
iron contents adjoining all three remediation wells
were significantly greater than the upgradient soil
iron content (p < 0.01). The mean differences
between soil iron contents around RIO, Rll, and
5-10
-------
NR10 and upgradient soil iron content were 0.256,
0.113, and 0.095 mg/g, respectively, with
corresponding standard errors of 0.034, 0.032, and
0.031 mg/g.
As a rough check on the horizontal extent of the
iron precipitation, the mass of iron precipitated by
the barrier was estimated using the mean difference
in upgradient and downgradient dissolved iron
concentrations (12 mg/L), the specific discharge, and
the time the barrier has been in place. This mass was
then converted to a volume of clogged soil using the
maximum difference between the upgradient and
remediation well soil iron contents (0.256 mg/g).
Following this procedure, the iron-clogged zone was
estimated to be 1.5 m wide. This distance seems
reasonable, suggesting that iron precipitation is the
probable cause for low specific discharges in active
remediation wells.
Overall Evaluation of Permeable Barrier
System
The permeable barrier system examined in this
project was designed to control the migration of
dissolved gasoline components by enhancing the
aerobic biodegradation of these compounds in the
aquifer immediately downgradient of the barrier.
Ideally, all contaminants would be degraded to below
regulatory limits before reaching the furthest
downgradient monitoring wells. The permeable
barrier examined in this project did not achieve this
objective. Table 7 lists average concentrations of
Table 7. Average Concentrations of BTEX in Monitoring
Wells Over the Entire Treatment Period
Well
SU7
SU13
SU14
SU5
NC
Standards
Distance
from
Barrier8
-10m
+3m
+8m
+25 m
Benzene
(mg/L)
2.419
0.757
1.123
0.877
0.001
Toluene
(mg/L)
8.326
2.406
3.469
0.853
1.000
Ethyl-
benzene
(mg/L)
1.391
0.383
0.595
0.272
0.029
Total
Xylenes
(mg/L)
6.060
1.627
2.366
0.745
0.400
Negative distances are upgradient of
distances are downgradient.
the barrier; positive
BTEX in monitoring wells immediately down-
gradient of the barrier over the entire treatment
period and current North Carolina ground-water
quality standards. While the average concentration
of all BTEX components decreased substantially
with distance downgradient, none of the BTEX
components met ground-water quality standards 8 m
downgradient of the barrier. Farther downgradient at
well SU5 (25 m), only toluene met water quality
standards.
Figures 16a and 16b show average concen-
trations of total BTEX and DO concentrations in
monitoring wells SU7, SU13, SU14, and SU5 for the
three treatment periods and for the total project.
a)
30
I*
£ 20
UJ 15
I 1°
£ 5
0
Period 1
Period 2
Period 3 Average
Period 1 Period 2 Period 3 Average
Figure 16. Mean (a) total BTEX concentrations and
(b) dissolved oxygen concentrations in
monitoring wells for individual
treatment periods and entire barrier
operational period. (Note: SU5 not
included in period 2 graph because only
one measurement was taken.)
Total BTEX concentrations in wells downgradient of
the barrier are significantly lower than upgradient of
the barrier for each treatment period at the 95%
confidence level, indicating that some biodegradation
is occurring. The barrier also appears to be effective
5-11
-------
at increasing the DO concentration in the wells
immediately downgradient of the barrier. The
increase in DO was most notable during the third
treatment period when the average oxygen
concentration in SU16 was over 4 mg/L.
One possible cause of the poor barrier
performance is inadequate delivery of DO to the
aquifer. The total mass of BTEX passing through
the active side of the barrier was estimated to be
approximately 80 grams per day, using the
background specific discharge and the average
BTEX concentration in SU7 during the third
treatment period (17.5 mg/L total BTEX
concentration). Samples collected directly from the
remediation wells indicated that essentially all of the
BTEX compounds that actually entered the
remediation wells were biodegraded. This would
result in a net loss of 5.6 g of BTEX/day. The total
oxygen delivered to the aquifer on day 459 from Rl
to Rll and NR1 to NR10 was estimated to be 6.24
grams of oxygen per day. This release rate was
estimated, using the measured specific discharges
and the oxygen concentration in the remediation
wells on day 459 (Tables 8 and 9). Assuming
complete mineralization of BTEX (3 to 1 mass ratio
of oxygen delivered to BTEX biodegraded), the
delivered oxygen should be sufficient to biodegrade
an additional 2.1 g of BTEX/day in the downgradient
aquifer. This would result in a total BTEX bio-
degradation of 7.7 g/day or 10% of the total BTEX
load. While there is considerable uncertainty in
these calculations, they do clearly illustrate the
problem in delivering sufficient oxygen to the
aquifer using oxygen-releasing wells. This problem
was only partially due to clogging of the remediation
wells. Assuming an average DO concentration in the
remediation wells of 20 mg/L and no reduction in
specific discharge, the total maximum BTEX
concentration that this barrier could effectively treat
is 6 mg/L.
Tables. Mass of Oxygen Released from Original
Remediation Wells on Day 459
Well
R1
R2
R3
R4
R5
R6
R7
R8
R10
Mean
C. V.
Specific
Discharge
(m/day)
0.045
0.038
0.052
0.079
0.034
0.055
0.059
0.045
0.022
0.048
34%
Oxygen
Concentration
(mg/L)
26.5
26.4
20.2
. 7.4
18.0
29.5
22.6
26.3
26.0
22.5
30%
Oxygen
Release8
(mg/day)
542
458
476
269
276
739
609
542
264
464
36%
Oxygen Release = (q)(DO concentration)(well cross-sectional
area) with normal well area = (diameter of well)(3 m saturated
thickness).
Table 9. Mass of Oxygen Released from New
Remediation Wells on Day 459
Well
NR1
NR2
NR3
NR4
NR7
NR10
Mean
C.V.
Specific
Discharge
(m/day)
0.023
0.015
0.017
0.020
0.016
0.014
0.0175
19%
Oxygen
Concentration
(mg/L)
12.7
18.3
12.5
11.3
12.3
21.5
14.8
28%
Oxygen
Release"
(mg/day)
132
125
97
101
91
139
114
18%
Oxygen Release = (q)(DO concentration)(well cross-sectional
area) with normal well area = (diameter of well)(3 m saturated
thickness).
In summary, the permeable barrier constructed in
this project was not fully effective in containing the
hydrocarbon plume. This was due to two factors: 1)
the high concentrations of BTEX entering the
barrier, and 2) the clogging of the barrier wells by
oxidized iron precipitates.
5-12
-------
References
Applied Power Concepts, Inc. 1992. Operation
Procedure, 92-MGO2-1. Allied Power
Concepts, Inc., Tustin, CA.
APHA (American Public Health Association). 1989.
Standard Methods for the Examination of Water
and Wastewater. 17th ed.; APJA-AWWA-
WPCF: Washington, DC.
Barcelona, M.J., T.R. Holm, M.R. Schock, and O.K.
George. 1988. Spatial and Temporal Gradients
in Aquifer Oxidation-Reduction Conditions.
Water Resour. Res. v. 25, no. 5, pp. 991-1003.
Bianchi-Mosquera, G.C., R.M. Allen-King, and
D.M. Mackay. 1994. Enhanced Degradation of
Dissolved Benzene and Toluene Using a Solid
Oxygen-Releasing Compound. Ground Water
Monitoring and Remediation, v. 14, no. 1, pp.
120-128.
Brown, R.A., and R.D. Norris. 1994. The Evolution
of a Technology: Hydrogen Peroxide in In Situ
Bioremediation. In: Proceedings of
Hydrocarbon Bioremediation. R.E. Hinchee,
B.C. Alleman, R.E. Hoeppel, and R.N. Miller,
eds. Lewis Publishers, Boca Raton, FL, pp. 148-
162.
Burris, D., and C. Antworth. 1990. Potential for
Subsurface In Situ Sorbent Systems. Ground
Water Mgmt. no. 4, pp. 527-538.
Chiang, C.Y., J.P. Salanitro, E.Y. Chai, J.D.
Colthart, and C.L. Klein. 1989. Aerobic
Biodegradation of Benzene, Toluene and Xylene
in Sandy Aquifer - Data Analysis and Computer
Modeling. Ground Water, v. 27, pp. 823-834.
Cohen, A.D., R.C. Borden, E.M. Stack, and J.R.
Durig. 1996. Evaluation of Peat-Based and
Nutrient Media for Remediation of Hydrocarbon
Contaminated Ground Water in Permeable
Barrier Applications. EPA/600/X-96/001.
Davis-Hoover, W.J., L.C. Murdoch, S.J. Vesper,
H.R. Pahren, O.L. Sprockel, C.L. Chang, A.
Hussain, and W.A-Ritschel. 1991. Hydraulic
Fracturing to Improve Nutrient and Oxygen
Delivery for In-Situ Bioreclamation. In:
Proceedings of In-Situ Bioreclamation. R.E.
Hinchee and R.F. Olfenbuttel, eds. Butterworth-
Heinemann, Stoneham, MA, pp. 67-81.
Gillham, R.W., and D.R. Burris. 1992. In-Situ
Treatment Walls - Chemical Dehalogenation,
Denitrification and Bioaugmentation. The Third
International Subsurface Restoration
Conference, Dallas, TX, June 21-24, 1992. pp.
66-68.
Goin, R.T. 1995. Enhanced Bioremediation of
BTEX Using Immobilized Nutrients. Masters
Thesis. North Carolina State University,
Raleigh, NC, 122 pp.
Hall,S.H. Spring 1993. Single Well Tracer Tests in
Aquifer Characterization. Ground Water
Monitoring and Remediation, pp. 118 -124.
Hatfield, K., D. Burris, T.B. Stauffer, and J. Ziegler.
1992. Theory and Experiments on Subsurface
Contaminant Sorption Systems. Journal of
Environmental Engineering, v. 118, no. 3, pp.
322-337.
Heersche, J., J. Verheul, and H. Schwarzer. 1994.
Evaluation of an In Situ Bioremediation Using
Hydrogen Peroxide. In: Hydrocarbon
Bioremediation. R.E. Hinchee, B.C. Alleman,
R.E. Hoeppel, and R.N. Miller, eds. Lewis
Publishers, Boca Raton, FL, pp. 388-397.
LaTowsky, H.D. 1993. A Field Study of an
Enhanced-Bioremediation System at a Ground
Water Contamination Site. Master of Civil
Engineering Project Report, North Carolina
State University, Raleigh, NC.
R-l
-------
Lewis, D.L., H.P. Kollig, and R.E. Hodson. 1986.
Nutrient Limitation and Adaptation of Microbial
Populations to Chemical Transformation. Appl.
Environ. Microbiol. v. 51, pp. 598-603.
Lovely, D.R., and E.J.P. Phillips. 1986.
Availability of Ferric Iron for Microbial
Reduction in Bottom Sediments of the
Freshwater Tidal Potomac River. Appl.
Environ. Microbiol. v. 52, no. 4, pp. 683-689.
Major, D.W., C.I. Mayfield, and J.F. Barker. 1988.
Biotransformation of Benzene by Denitrification
in Aquifer Sand. Ground Water, v. 26, pp. 8-
14.
Nakajima, M., T. Hayamizu, and H. Nishimura.
1984. Inhibitory Effect of Oxygen on
Denitrification and Denitrification in Sludge
from an Oxidation Ditch. Water Research, v.
18, pp. 339-343.
NIOSH. 1990. Pocket Guide to Chemical Hazards.
U.S. Department of Health and Human Services,
National Institute for Occupational Safety and
Health.
OUST. 1990. LUST Trust Fund Monthly Progress
Report. Office of Underground Storage Tanks,
U.S. EPA, Washington, D. C.
Page, A.L., R.H. Miller, and D.R. Keevey. 1982.
Methods of Soil Analysis. Am. Soc. of Agro.,
Inc., Madison, Wisconsin. Part 2, 2nd ed.,
Chap. 41, pp. 831-872.
Schmidt, S.K., K.M. Scow, and M. Alexander.
1987. Kinetics of p-Nitrophenol Mineralization
by a Pseudomonas sp: Effects of Second
Substrates. Appl. Environ. Microbiol. v. 53, pp.
2617-2623.
Song, H.G., X. Wang, and R. Bartha. 1990.
Bioremediation Potential of Terrestrial Fuel
Spills. Appl. Environ. Microbiol. v. 56, no. 3.
pp. 652-656.
Spain, J.C., and P.A. Van Veld. 1983. Adaptation
of Natural Microbial Communities to
Degradation of Xenobiotic Compounds. Appl.
Environ. Microbiol. v. 45, pp. 428-435.
Starr, R.C., and J.A. Cherry. 1994. In-Situ
Remediation of Contaminated Ground Water:
The Funnel-and-Gate System. Ground Water.
v. 32, no. 3, pp. 465-476.
Swindoll, C.M., C.M. Aelion, and F.K. Pfaender.
1988. Influence of Inorganic and Organic
Nutrients on Aerobic Biodegradation and on the
Adaptation Response of Subsurface Microbial
Communities. Appl. Environ. Microbiol. v. 54,
pp. 212-217.
Thomson, B.M., S.P. Shelton, and E. Smith. 1990.
Permeable Barriers: A New Alternative for
Treatment of Contaminated Ground Water. 45th
Annual Purdue Industrial Waste Conference.
Thorton-Manning, J.R., D.D. Jones, and T.W.
Federle. 1987. Effects of Experimental
Manipulation of Environmental Factors on
Phenol Mineralization in Soil. Environ. Toxi.
Chem. v. 6, pp. 615-621.
Wilson, J.T., J.F. McNabb, B.H. Wilson, and M.J.
Noonan. 1983. Biotransformation of Selected
Organic Pollutants in Ground Water. Dev. Ind.
Microbiol. v. 24, pp. 225-235.
Wilson, J.T., L.E. Leach, M. Henson, and J.N.
Jones. 1986. In Situ Biorestoration as a Ground
Water Remediation Technique. Ground Water.
v. 6, no. 4, pp. 56-64.
R-2
-------
APPENDIX A
Laboratory Studies: Oxygen Release Over Time from Solid Peroxide Concretes
Day
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
130
135
140
145
150
155
160
165
170
175
180
185
190
195
200
205
210
215
220
225
230
235
240
245
250
37% MgO2
briquets
0.2740
0.2664
0.2588
0.2512
0.2436
0.2360
0.2284
0.2208
0.2132
0.2056
0.1980
0.1904
0.1828
0.1752
0.1676
0.1600
0.1524
0.1448
0.1372
0.1296
0.1220
0.1144
0.1068
0.0992
0.0916
0.0840
0.0764
0.0688
0.0612
0.0536
0.0460
0.0384
0.0308
0.0232
0.0156
0.0080
0.0004
37% MgO2
cylinder
0.1910
0.1854
0.1798
0.1742
0.1686
0.1630
0.1574
0.1518
0.1462
0.1406
0.1350
0.1294
0.1238
0.1182
0.1126
0.1070
0.1014
0.0958
0.0902
0.0846
0.0790
0.0734
0.0678
0.0622
0.0566
0.0510
0.0454
0.0398
0.0342
0.0286
0.0230
0.0174
0.0118
0.0062
0.0006
21% MgO2
briquets
0.05170
0.05093
0.05015
0.04938
0.04860
0.04783
0.04705
0.04628
0.04550
0.04473
0.04395
0.04318
0.04240
0.04163
0.04085
0.04008
0.03930
0.03853
0.03775
0.03698
0.03620
0.03543
0.03465
0.03388
0.03310
0.03233
0.03155
0.03078
0.03000
0.02923
0.02845
0.02768
0.02690
0.02613
0.02535
0.02458
0.02380
0.02303
0.02225
0.02148
0.02070
0.01993
0.01915
0.01838
0.01760
0.01683
0.01605
0.01528
0.01450
0.01373
0.01295
21% MgO2
cylinder
0.03510
0.03454
0.03397
0.03341
0.03284
0.03228
0.03171
0.03115
0.03058
0.03002
0.02945
0.02889
0.02832
0.02776
0.02719
0.02663
0.02606
0.02550
0.02493
0.02437
0.02380
0.02324
0.02267
0.02211
0.02154
0.02098
0.02041
0.01985
0.01928
0.01872
0.01815
0.01759
0.01702
0.01646
0.01589
0.01533
0.01476
0.01420
0.01363
0.01307
0.01250
0.01194
0.01137
0.01081
0.01024
0.00968
0.00911
0.00855
0.00798
0.00742
0.00685
14% CaO2
briquets
0.20052
0.19002
0.17952
0.16902
0.15852
0.14802
0.13752
0.12702
0.11652
0.10602
0.09552
0.08502
0.07452
0.06402
0.05352
0.04302
0.03252
0.02202
0.01152
0.00102
A-l
-------
-------
i
PQ o
OX)
PH
d
Ii
w'l g
Is I
S I
S
*^
o|
^4
5! f
&
S U
H
w
G,
E§ 1
/-v bJ
p^
s s
^2I
Q ^
lit
o C oo
H S 3
(D
d j
i^'oT)
£3
i
pq
1
TEST
PERIOD
W")
?
VO
en
oo
oo
^t-
VO
oo
oo
0
£
oi
r---
t~~
T <
O4
VD
04
O
^)
en
en
w-j
?
VO
*
w>
*
o
o
o
1 1
* I
1 1
04
VO
>r>
i
en
&
vo
0
o
o
o
o
o
ON
V)
O
V
*
04
VO
g
v>
o
2J
in
en
r
r-
>o
o
i i
*-H
I 1
O!
en
04
ON
>o
VO
T I
>0
r 1
VO
r-
04
WO
oi
r-
ON
*
O
rf
0
S
*
en
oo
ON
en
>o
?
"*
VO
>o
O4
10
VO
r-H
ON
en
r
in
f-
E^
VO
en
T-H
en
O
04
>n
en
r-H
in
r *
>O
?
01
t-
r~-
m
ON
*
OO
oo
r~-
r *
in
f~;
r <
ON
ON
O
-
O
VO
en
» i
r (
r-H
01
0)
in
S
in
O
vb
ON
c~-
en
m
in
» i
in
_)
<"!
j
z »
Z S
f^
2 S
-, fcj
^ ^v
U |>
S |
a|
.j
dB|
j
*^
U |»
S'f
^ tP
g^
£H
i"t
i
TEST
PERIOD
cs
04
o
en
VO
r t
00
o
o\
ON
T i
04
O
o
8
?
04
Tf
in
?
in
t>
ON
od
in
?
en
en
IT:
?
04
O
04
r-
cs
en
r-H
1 1
1 I
cn
TI-
O
o
8
O
ON
r-H
in
?
TI-
od
""i
od
in
?
?
T-H
1
O\
t
O
rt
o
es
in
in
VO
TT
O
en
O4
f-
V)
04
S
?
en
0
m
o
V
in
\d
in
C-^
in
?
?
en
04
ON
en
-
04
O
00
en
£
o
S
04
S
o
s
o
V
*f
in
^
in
00
r~
in
V
V
in
i i
00
O
04
O
VO
in
r-
0
04
8
o
8
$
en
in
?
in
VO
VO
ON
V)
?
?
S
ON
N
OI
3
ON
O
04
0
VO
VO
ON
00
O
T^H
ON
T-H
T-H
o
o
8
H
in
V
i i
od
ON
vd
t
?
?
VO
en
04
in
CD
04
O
5)-
VO
00
VO
O
vo
t~-
T-H
O)
o
o
8
o
V
T-H
m
o
V
VO
ON
vq
t-^
i i
?
V
04
OO
en
*o
0
04
O
04
VO
S
O
s
es
T-H
o
0
8
o
V
TI-
CS
?
04
vd
t-~
T-H
o
V
d
V
o
T-H
Ti-
en
oo
d
1 (
?
VO
VO
s
o
VO
in
04
T-H
o
o
8
d
V
cs
o
O4
?
C--
04
CS
00
(
o
V
d
V
oo
en
Tt-
i/^
d
cs
o
ON
in
T-H
r-
d
T t
OO
T-H
§
o
8
3
VO
in
cs
?
i i
T-H
Tf
VO
T-H
o
V
?
04
ON
Th
1
.£
'3
H
3rH
&s
1 &
0 "0
° ,-,
w ^
ft
2 "S
s ^
R &
ilbarrie
s were r
.s s
i
O ffl
ON
.t
O (-!
cells on day
tion wells 01
a H3
o jj
a ^
S §
a) S
C c
CO GO
S S
g?
i i
iers were m
iers were in
Original ban
Original ban
B-l
-------
O
Q
i
It
M
CL,
E U
£ °
w
Q,
5§ 6
§1
S o
Q ^
11*
4*
>^s ^
if *
* i ^
U3 _g
H
"o s
E-
1
1 M
M
^
c3
Q
H O
CO HH
m 2
H W
P-,
-i
T 1
» 4
.0
oo
o
in
r--
5
vb
cn
ON
«
o
o
o
,_,
0
cn
cn
i
iri
O
VO
VO
s
u~>
0
T-H
VI
cn
T 1
r-
ON
T-H
-
o
VO
CN
cn
cn
vo
00
*
O
o
o
^
0
i-H
\f}
^,
00
VO
oo
VO
o
J~>
14
3
<*
O
o
o
o
o
o
o
r-4
VO
cn
CN
oo
vq
cn
r-
o
vi
o
VO
CN
cn
CN
C5
ON
O
O
O
O
O
O
CN
oo
cn
cn
I
«
U
d
1
CO
tJ
2
fe§
VI
CN
VO
VO
o
V
V
cn
CN
oo
CN
O
C5
S
t-^
CN
cn
CN
2
ON
cn
cS
-------
CN
o
o
o
.
^ t^
< t
1* u
H
'ltL
SI
st
U 0 ^
^ CM* ^
U
x-k W
-& i ^
^ a 1
m al ~
f>
(O
1 3
C3
CJ toJ)
Q
£ S
W C^
H W
i
CN
t
CO
vd
r-
m
vd
o
T I
CN
CN
T-H
Z
m
CN
vd
CO
i i
n
CN
in
r-
o
CO
CO
CO
1
i
ON
in
vd
VD
CN
i
T-H
T 1
Co
CN
00
VO
oo
T 1
OO
CO
T-H
T t
1 1
*
T-H
T-H
1 1
T-H
T-H
CO
>o
vd
-
d
Co
C-;
CO
vl
S
T-H
§
t 1
CO
0,
CO
0
in
vd
o
CN
d
Co
CO
o
r-
ON
oo
T-H
So
co
CO
ON
CO
W~i
d
V
oo
^
ON
CO
vd
ON
CN
OO
d
Co
CN
T-H
T 1
CN
i i
oo
CO
CN
T-H
r-
VD
T 1
in
-
d
V
co
ON
vd
O
T-H
1
d
Co
o
co
VO
-
T-H
CO
ON
T-H
§
in
T t
T t
T t
W-l
d
V
vq
OO
CO
in
vd
ON
i
o
oo
C--
00
ON
-
m
VO
VD
r-
T 1
CN
d
£
CN
T-H
vd
T <
CO
T-H
Co
P
in
T t
T-H
CN
O
' T-H
CN
T-H
1 1
T-H
oo
CO
T-H
VO
co
CN
T 1
d
CO
vo
T-H
vd
CN
m
T-H
h
co
00
rt-
T-H
oo
T <
CN
CO
ON
CO
00
o
T-H
CN
CN
CN
CN
OO
CO
CO
d
m
CO
oo
CO
3
00
oo
ON
T-H
1 1
CN
T-H
5
oo
w P1
,1
« <
* t
s t
Cd *^v
O a1
« <
s t
d 1
._]
^ ^
cs
PQ j|
u U
cT^
°* S
II
^-
S1
Q
TEST
PERIOD
O\
T (
?
00
uS
CO
oo
00
ON
vo
O
C)
S
O
V
f~~
CO
>n
?
«n
VD
ON
l>
>n
?
>o
?
CO
CO
oo
T-H
0
VO
ON
vo
oo
CN
OO
co
CO
O
S
CJ
V
CN
*
>0
?
0
T 1
^
VD
>o
?
>T)
^
T-H
T-H
t
ON
ON
CO
CO
CN
O
ON
CO
VO
CN
O
i i
oo
T 1
O
O
8
CD
V
v~>
10
>o
^
o
T-H
2J
10
?
v~>
$
T-H
»n
T-H
"^
T-H
?
t-;
in
T 1
VD
co
o
o
s
o
V
T 1
CN
T-H
V~l
$
*
0
*
t>
in
?
>n
^
T-H
T-H
T-H
?
»0
T t
S
in
ON
S
O
8
d
V
T-H
T t
T t
*o
?
VO
Tf
T-H
T-H
1 t
?
in
?
n
^
CO
c~.
r~
\d
T-H
^
2;
C5
T-H
VO
CO
CN
CO
T-H
d
CO
VO
T-H
T 1
CN
CN
CO
O
d
8
?
in
in
in
?
ON
f-^
CO
VO
T 1
?
T t
d
V
CN
OO
CO
CO
1
c~-
CN
?
CN
vo
CN
oq
T 1
S
d
8
?
>n
CO
CN
CN
?
OO
oo
s
$
T-H
3
oo
ON
*
B-3
-------
o
o
i
£ J
||
« ^
5j|
cu
fa
E^ '
w
ex
& 1
S |
tfc-4
O
O ^5
ii »2
?
lit
^
°$»
ft*
S^ 3
j. 8 ,
^ 8 a
[13 u 3
N
e s
u
§ 1 ]
g ^)
§ »
pa
&
Q
Q
^ §
S3 S
G m
S
>n
?
oi
04
«n
a
\d
vo
en
en
en
C5
T <
t (
04
5
J5
r-M
3
S
f-
0
r-l
t
04
» i
r-*
04
* (
r (
n
» <
O
oo
VO
n
m
r-H
oo
en
en
04
ON
«T)
?
VO
*
VO
10
s
vd
i i
VO
o
W)
C5
04
oo
s
ss
en
>n
oo
VO
>n
VO
en
o
W)
* i
vo
O4
T 1
en
04
O
>n
£
vd
VD
OO
i
O
r
T <
04
Tf
>n
04
T t
S
T-H
,4
S
>0
r^
a
g
>o
s
T (
»o
^
04
en
>o
^t-
oo
T (
vd
04
t-p
in
C5
vb
04
OO
oo
oo
VO
*
t I
t I
Si
a
* t
3-
en
en
en
t t
in
>n
^
oo
04
f~
VO
VO
04
vd
04
VD
T-H
o
_
en
O
-*
14
t~-
en
s
VO
ON
t i
*
es
04
ON
m
i i
» t
* t
V 1
v>
?
0
en
°
vd
oo
o
Tl-
>n
oo
VO
ID
S
T t
£
« 4
04
0
o
*
1-H
T-H
vd
oo
^1-
en
CJ
o
r-4
04
8
0^
00
04
*
§
oo
ON
9
en
en
en
t i
oo
04
00
en
Si
0
t>
>n
ON
O
^
>n
04"
04
vo
14
9
O
*
en
vo
04
«n
en
en
ON
04
o
5
» (
*
>n
VO
14
^
>0
ss
*
T 1
oo
ON
04
l-H
OO
ON
*
04
04°
CNl
I
vq
04
oo
oi
en
o
1
OX
'
ON
OS
0
VO
s
04
n M
ON
CM
04
>o
o
vq
ON
ON
en
vd
04
8
8
IT)
vd
Ox
04°
cs
04
04
oi
en
vd
vq
VO
O
04
ON
en
S
en
O
o
o\
en
o-
o
""!
o
04
ON
ON
C5
ON
vd
VO
en
04
a
oo
-------
on
O
i
Sf J
i»
.M t:
< 1
1-U
H
,«
ex
E g
8|
f si
Q £
-a X ,
l$i
H pq 3
ID
6 |i
£ *
II*
ill
w J5 3
g
ii
S^
S .
8 'a
c ££>
0 3
«
1
^ Q
S S
w 3
H W
OH
vq
VO
en
f-
*
""1
vi
cs
es
en
T t
s
T i
T-H
cs
S
T-H
0
o
Tf
>n
en
CO
i
/">
csi
T)-
o
r
>n
0
T 1
VO
"*
en
l>
ON
en
T-H
r-
V)
&
cs
C3N
P>;
O
CS
O
VO
ri
>n
>n
,
0
rq
0)
S
oo
en
S
T-H
o
VO
Tt-
S
oo
VO
m
i i
en
S
S
>o
v~,
en
oc
>o
*r>
cs
0
C
i (
cs
5
en
CS
o\
t-H
oo
en
S
ON
en
i i
**
CS
cs
*
T^
CN
>o
z
«M
o
^
^1-
^
T-H
<-H
>O
oo
>n
vd
VD
vq
D
o\
t (
c^
cs
en
*
S
T-H
£
^H
m
t~
p
>o
f-;
IT
§
rs
O
00
en
S
ts
*
T 1
en
>o
>n
cs
r-
T I
1 1
VO
en
cs
oo
oo
Tj-
^1-
oo
oc
i/S
ON
er
fS
O
b
T i
cs
o
8
cs
1 <
VO
en
§
cs
cs
oo
en
ON
cs
v->
en
CS
t~-
r-
en
^ (
^-H
Tl-
00
VO
oo
en
*
en
*-*
V
O
>n
vo
ON
i i
T-H
VO
CS
VO
00
oo
ON
*
00
o
1
U
P
3
W
PH
ON
cS
cs
v
ON
vd
cs
10
VO
VO
S
wo
n
VO
ON
cs
00 ON
VO
cs
W-)
VO
S
o
V
B-5
-------
s
2*
J
j^
IN
H S
. B *^
JL Jsl fefi
X
til
"5» 1 t
1^
^ P
H
3 . 1
H &b
t> =i
2?
Q
Q
feo
H U]
0.
"1
?
O
Tl-
X
a
VO
R
m
"?
VO
CO
n
ON
1
ON
5
NO
T;
1
c
o
CO
*
0
cs
vo
g
o
o
in
t
5
VO
OO
g
oo
o
CO
en
ex
vo
O>
m
?"
CN
en
00
vo
CN
vo
>n
fr
CO
d
en
§
CO
CN
o
V)
o
CN
T
r~
V
0\
ON
oo
g
VO
t~
n-
VO
O
E
in
CO
in
ff
CN
t-
t~-
ON
CO
c*"
CN
oo
oo
ON
in
vn
p*
-,
o
VO
VO
in
V
i
o
vb
m
CO
S
*
oo
oo
CO
en
?
CN
o
en
o>
VC
in
?
O
CN
cn
vo
vo
in
en
CO
o
ON
oo
Cf
c-
t~
CN
oo
01
oo
0
VO
CO
c-
o>
on
n
?
m
cs
o
VO
3
o
VC
ON
o
T
0\
oc
c
r-
co
CN
\r
CN
V
o
CN
0
VO
s
c-
o
^
VO
VD
S
X
§
VI
S
VO
cs
ON
OC
n
VO
»n
c<
»n
t^
»n
;c
CN
CO
o
oo
)
*^
E
5! I
7 M
z S
00 fcJ
S S
,1
tf|
|1l,
s s
a|
^ S
^ t^
m If
^J
0|
^
co g
^ t
O"c
Cu «
IT"
I1 t^
gt
Js"
Q
Q
U3 &
p c [i
PH
(N
*-*
* t
?
vq
m
Tf
CN
r-H
Tt
in
O
p
VO
in
<
in
G5
CO
r
CO
c<
in
o
V
m
o
V
CO
CO
eo
rt
°
f-
VO
ON
O
r-
S
o
c
oo
,n
C5
VO
in
in
o
in
rt
in
o
,-H
^
o>
^
CO
o
ON
VO
c--
0
CN
r-
S
o
in
C
OC
c~
V
T:
in
CO
m
Q
in
n
C*"
cs
c
CN
1-1
^
?
ON
Tt
vo
o
00
Tt
s
o
in
c
°
V
?3
in
oo
CO
in
o
V
in
o
V
ir
CO
VO
-
VO
OC
11
eo
O
in
in
VD
O
CN
OO
CO
s
o
v>
c
CO
c
v>
^
*n
-------
3
CO
Z _j
M fcTx
O I?
j
L> oil
^ S
ex
£ °
'o.
4= >
2 !a>
43
8-2 |
Q £
3^
O f i Ofl
E""* rn *-^
0
i ID ti
*5L. ^^
X 3
tli
££ 3
*8^
-G M 63)
pq <3 3
3 S)
r ,
,
Q
g§
AH
in
c
V
rs
^f
^.
in
T-H
VD
o
l-H
CO
o
=
T <
o
OS
CO
VO
^"
TJ-
OO
VO
VO
^o
CO
0
x>
CO
1-1
1
1 1
CO
CO
1
^_
in
O
VO
f!
O
"3-
0
OO
in
cs
co
cs
OS
1 1
»^-
T I
co
CS
rs
in
OS
T 1
1
CS
CO
OO
OS
in
^
T H
o
op
o
os
m
VO
t\ js
VO
^
* H
OO
ON
W)
so
T H
CS
S;
1-1
OS
in
O
V
"^*
cs
in
OS
o\
^
o
r
O
ro
Os
wo
rs
CS
CS
CO
00
O
vo
[--
CO
cs
m
in
i i
o
CO
* 1
0
WO
~
T 1
Tt
vo
vo
i r
CS
CS
OO
CO
^J-
co
OS
CO
in
o
v
^-
1 1
in
cs
* i
vo
n
0
WO
r>
i <
T 1
CS
cs
o
in
cs
52
co
C^
(N
r-
O
CO
in
i i
wo
fi
V
OO
CO
wo
VO
S
VO
0
cs
0
ON
CO
wo
OS
OO
wo
wo
T I
VO
rj-
Os
ro
DO
VO
*H-
cs
£;
1 <
^
wo
f~l
v
o
CS
O
OS
wo
vo
o
r-
ih
no
OS
i i
wo
0
CO
cs
OO
co
vo
1
o
OS
CO
T 1
00
wo
0
O
in
r-
cs
ro
wo
OO
cs
0
CO
CO
^.
^
ro
or,
CS
VO
DO
0
T 4
o
vn
CS
CO
"I
o
r_l
VO
cs
o
o
cs
in
CO
00
cs
0
M-
co
M-
Os
co
Xt"
CS
CO
^5
* '
S
00
CO
CO
o
CO
r-
wo
^^
00
rrs
on
wo
os
cs
on
CS
r-
vn
in
wo
on
OS
^ M
G
o
V
VO
wo
wo
v
oo
Os
~
^ {
v
CO
o
vo
CO
cs
CS
rl
o
CO
CO
oo
p
S
vo
8
o
8
o
V
OS
» I
wo
°
VO
CO
oo
^
y
ro
0
CS
oo
CO
ON
. i
CO
o
^
Tf
wo
» I
VO
vo
oo
0
p
V
Os
wo
T 1
CS
°
ON
-
1 1
°
^
0
oo
co
CS
^H
CS
o
ON
CO
00
^
VO
O
P
OO
CS
cs
"
ON
06
wo
oo
wo
00
o
00
\D
0
oo
ON
CO
B-7
-------
B
CO
O
5" .
Sf
2 s
J4«J
5 f
to
£°
%
el
S|
t<5
o c3
*"* p£
£ g =
6 -ft a
8°
n*
s#=
e
IT I "a
JU
&°
gd
S a
o 3
1
Q
c~* O
Cu
>n
3
o
co
m
CM
CN
vo
en
in
°?
0
s
>n
ON
S
CN
ivo
CN
!co
co
Iin
CN
S
0\
CO
n
9
vo
PI
CO
CO
8
vo
CO
i
r~
0
ih
*
CO
4
o\
s
in
CO
8
CO
ON
S
n
"?
o
in
CO
» i
vo
ON
CO
5
CO
CO
in
?!
CO
o
21
o\
§
ON
CO
in
n
ON
r-H
in
CN
t-
n
o
CO
*
o
rt
i I
8
ON
c
s
o
T (
cs
1 1
^~
1 (
8
CO
CN
CO
O\
0>
0
5-
CO
m
o
CO
>n
CO
o
0
n
vo
CN
in
8
CO
OO
5
w «
,1
*i
S 6
6 |
e ti
S 6
d "I
.&i
*"* a
of
O *&i
co g
OH
o*f
0L, B
17-
Elf
z s
Q
i|
fXl
o
rt
1?
-
CO
1 1
1 1
i H
T-H
S
O
CM
in
CO
CO
CO
Tf
in
*
m
^
CO
CO
o
. I
?
m
t~
CO
§
CO
o
*
o
CM
r
in
vo
t-
m
*
in
^
1
ON
CO
i i
t
°
CM
t>
3
9
ON
O
8
ON
in
in
^.
CM
in
°
in
^
ON
CO
i i
CM
CO
o
t
vd
CO
p
vo
CM
CO
s
o
in
0
in
in
c-
o
in
"\?
in
%
in
-
CM
CO
o
in
r-
-
^
CM
8
o
in
o
^
'-'
in
CO
ON
in
^
>n
*3
i
ON
CO
r
^ j
?
CO
CO
in
ON
O
CM
o
CO
s
o
8
ON
f-
in
vd
CO
*
in
^
i i
^
^ [
o
od
CO
CM
ON
CO
g
O
8
CO
CM
in
m
ON
in
n
i i
°
ON
i i
O
i i
«
c-l
r-
{
o
ON
ON
CO
CO
r-
co
3
o
s
m
ON
t i
in
O
in
vo
i i
?
r-
i i
o
CM
OO
CO
OO
t (
^
CM
c-i
r-;
ON
CO
vo
CO
o
o
s
^
ON
CO
CM
OO
{
CM
1 1
?
T_H
CO
O
o
1 1
CO
VO
CM
CM
O
~
CO
VD
O
m
g
o
S
c-~
ON
CM
CM
2
CO
CO
1 I
?
,
CO
o
OO
CO
OO
CM
o
m
CO
d
ON
CM
CM
S
O
s
°
CM
CO
CM
t^
CM
CO
s
o
CO
1 I
o
CO
3
B-8
-------
t)
CO
O
O
^
£t
**
ex
a u
H
a
&,
fl 1
^
S S3
§"* -S ^
Q £
o S 6b
H pq 3
, § d
o -^ ab
II*
u
i £*
W ^ ^
,
Q
CO §
H CJ
?
CN
T I
OO
in
ON
Tj-
^
CN
i i
in
E-
o
o
o
o
o
o
en
en
i
"
-
~.
ON
Tt-
» I
i 1
CN
r-
CN
=
T-H
o
o
o
o
o
o
T 1
1 1
ON
en
oo
T 1
in
oo
» t
oo
( 1
in
r-
cn
o
o
o
o
o
o
VO
?
-
ON
en
ON
-3-
CN
ON
T I
=
ip
CN
en
«
CN
CN
i [
M-
m
CN
en
CN
oo
Tj-
CN
00
ON
» I
in
5P
CN
CN
CN
O
CN
ON
O
ON
en
i t
^
°
CN
en
en
o
in
R
^
CN
in
ON
in
o
o
o
ON
in
o
T <
in
t
0
CN
CN
in
Tt-
in
in
i i
in
^t"
VO
o
o
o
CN
Tfr
^
^
i i
CN
r-
<^-
ON
en
^.
O
z
en
en
O
O
O
en
O
^
n
in
CN
T (
CN
=
en
O
O
O
O
O
O
(N
OO
en
\o
i i
in
E-
-
l-H
O
o
en
0
on
en
en
CN
*
S
CO
1 1
O
O
O
i-H
0
00
J
^ t
3f
« <
* 5
00 tJ
"5 M
2 e
._j
6 |
j
s f
d |
j
a1!
s |
a |
0* ^
CO g
eu ^
o' ^
PL, B
!& ">
1 s
&
Q
TEST
PERIOD
CO
i
\o
in
en
in
m
T I
2
o
V
i 1
^
Iin
?
i>
oo
en
in
o
V
in
3
en
en
oo
S
o
en
O
o
Ti
?
m
?
VO
t
in
o
V
in
?
i i
^ i
ON
CN
CN
0
VO
en
*
*
\o
VO
en
in
O
2
?
T 1
O
V
n
^
CN
^r
\D
V~l
O
V
m
?
en
CN
ON
en
T 1
OO
d
CN
*'
00
\o
*
T t
0
O
S
?
T 1
O
in
?
CN
iri
in
in
o
V
in
3
i t
in
CO
CN
0
in
Tf
t-;
5
*
g
o
S
^
14
O
V
in
?
vo
en
in
O
V
m
3
i*
T i
i t
r-
CN
C5
O
oo
CN
m
en
en
S
0
2
?
T-H
o
V
in
3
i t
OO
* 1
i 1
?
n
?
in
t
i i
r^
CN
0
OO
*
VD
r-
en
Tf
O
0
2
?
CN
O
in
o
V
P-;
OO
T 1
?
1 I
?
r-H
^0
CO
CM
ro
T-H
0
r>
t>
oo
vq
*
in
TJ-
^H
O
0
2
?
l
o
V
>n
?
en
oo
en
T 4
?
C5
CN
oo
en
o>
CN
C)
*
r-^
in
vq
r-
xt-
xt-
o
o
2
3
i i
o
V
CN
O
V
en
r t
V- 4
1 H
^
*-H
?
oo
en
*
en
0>
CN
i-H
C3
00
r~
^H
vq
in
c-~
Tf
T-H
O
o
2
?
* 1
?
CN
CD
V
ON
OO
"*
ON
S
^
?
oo
o\
*
B-9
-------
5
vt
O
If
M
H*
a
si
Si
2 s
?*
IN
4 IS
X 3
t-JN
-i-?3
*Jl
!§
t2 a
1*
1
Q
g°
?
8
r-
s
0
9
en
0
fc
V
2
51
3
£
8
s
CD
5?
0
?
*
en
*
ON
v>
a
"1
«i
o)
g
5
3
r-
DO
§
I
8
r*
i
o\
v»
g
in
SI
00
o
SO
n
&
n
oo
c~
in
i
00
in
t>
i
«
X
OO
3
s
s
t~
m
K
m
s
m
o
eV
y>
*
eS
1
OO
5?
CO
^
VJ
*-4
1
1
S
VI
?
s
*
t-
g
in
oo
*
O
s
<0
in
\o
oo
fl
1
OO
CO
\o
r^-
01
$
oo
->t
1
5
in
o\
oo
oo
in
S3
o
in
n
oo
in
\o
oo
g
1
CM
oo
0
Ov
CO
VI
?
oo
n
VO
T-<
e^i
1 1
o
ON
oo
*
en
«
VO
7
§
p
R
r-
r-
&
s
5?
in
p»
^
NO
*?
«
O
in
S
1
t-
r-
v>
oo
oo
en
»n
CM
1
s
r--
O\
CS
en
en
n-
oo
m
"3
?
O
VO
i i
> c
n
S
?1
1
o
oo
O
1
S
S
§
"J
?
CO
S
NO
«s
s
o
so
r-
en
S
v»
§
i
m
§
CM
\D
VO
i-H
Tf
i i
CO
0
^H
^*
VI
a
CO
i
s
3
NC
NO
s
s
VI
CO
\o
1
in
oo
04
es
NO
r-
1
p-
*
i
ON
2
O\
OO
o
CO
oo
a
en
o
rt
NO
O
NO
NO
0?
VO
1-H
oo
§
1
CO
v>
T-<
CM
t4
o\
r-
s
t~
in
CN
s
en
en
c>
in
-*
c^<
NO
S
eN
eN
y
n
en
c- 1
en
s
S3
CM
CM
£
r-
r-
in
o
oo
en
«
oo
en
*
o
r~
*
C-J
NO
?
CO
CO
i
§
s
?
oo
CO
§
o
s
C-4
en
NO
o
f-H
**
en
n
o
oo
m
in
M
£>
in
en
in
oo
°
NO
in
rl
i
1
S
&
g
?!
OO
5
CO
oo
f-
vo
oo
in
-
o\
o
*
oo
NO
in
0
s
en
R
SS
OO
o\
r-
co
CM
r-
s
CO
5
en
c-
o
ON
oo
oo
in
?3
in
NO
n
1
I
S
2
OO
CO
CO
CM
OO
5
-I
'S
,!
2|
61
|d
s s
k_J
aa
18 1
1
»j
0 a
o'1^
eo E
If
ffi'l
z E
&
Q
TEST
PERIOD
?
en
in
en
CT\
o
r-
NO
S
o
S
n
in
?
in
-
s
in
Ox
C3
NO
OO
OO
c-4
o
11
oo
NO
es
NO
NO
T-H
OX
o
o
8
^
P-;
CO
m
3
en
t-;
oo
in
?
2
oo
en
Ox
Ti
r~
OX
0
en
*
o
r--
NO
in
en
0
S
-------
00
OT
^
O
i
*J
g"l
si
P-n
£0
a.
s I
Hi
S S
CL. o +3
tT 2 ,«
Q £
glf
H pg 3
(D
, s d
o -a a
V 3
i>s
!!*
o
fi|
ri-1 S P
w .0
H
e "
(U
§ 1-4
If
ffl
I
TEST
PERIOD
V~i
3
oo
CO
r-
S
VO
CO
1 1
CO
10
ON
S
CO
0)
T 1
*
VO
1 1
s
1 1
CM
10
CO
CO
«
?
10
CM
Tj-
vo
vo
10
0
=
10
f-
Tt
CO
ON
CM
VO
CM
vo
9
VO
r--
a
ON
ON
5
* 4
i 1
1
wS
s
vo
oo
n
10
0
So
CM'
Tf
s
VO
£
9
CO
r 1
ON
CO
1O
ON
10
» t
CO
§
T 1
o
vo
>o
S
vo
VO
^r
-*
0
z
10
r-
o
10
CO
10
ON
CO
VO
o
CO
CM
01
10
CM
ON
ON
i
>n
CM
ON
CO
ON
w-1
3
o
-*
vo
-*
s
vo
CO
10
CO
0
^t-
10
oo
§
2
CM
CO
s
£
oo
10
vo
r-M
>o
>n
m
o
CO
CO
c->
-*
2;
vo
o
VO
i
cs
0
) 1
1 1
10
oo
*
8
2
o
oo
§
Si
c~-
1
s
oo
Cn
o
CO
T H
s
i
>n
->t
S
t-H
O
§
oo
CM
OO
CN
VO
>o
10
8
vo
f
Tt
0
o
^J-
oo
1
i
CO
§
C-)
IO
oo
CO
C-)
CO
vo
t-~
CM
*
CO
CO
vo
* i
oo
10
oo
oo
1O
1O
CM
o
F~-
*
T 1
o
%
ON
-*
f-^
CO
ON
CM
IO
f5
s
vo
oo
»J->
^
00
CO
CO
00
10
T 1
VO
CO
CO
1
cs
o
^t-
r-
oo
1
s
CO
oo
ON
1O
VO
1
o
10
ON
vo
ON
oo
* 1
1 t
CO
ON
S5
vo
>o
tj-
o
>h
oo
ON
§
CO
CM
CM
s
00
Tf
5*
8
vo
1 1
oo
CO
CO
oo
s
oo
1O
ON
CO
v>
3
VO
fl
3
vo
CO
.1
1
o
CO
ON
CO
ON
s
CM
CO
r~
s
CO
CO
ON
s
vo
ON
s
§
CM
1O
t-
?l
CM
vo
ON
*
o
io
t-^
r~-
VO
oo
5
CO
o
o
ON
s
CM
1O
i
VO
10
p
1 I
ON
CO
*-H
t
ON
1 1
cs
CO
s
VO
CO
1 I
1
_
o
r-
i
"*
r
ON
CO
SN
s
10
s
CO
oo
s
CM
o\
i
CO
04
0
vo
10
s
1O
*
o
ON
CO
o
;_
vo
s
CO
CO
9
r~
§
CO
CO
CO
CO
ON
VO
§
1
» <
vo
CO
CM
2;
0
CM
r^
(
VO
3
CO
o
5o
VO
"O
o
"O
>o
s
vo
vo
CO
1O
r t
CO
VO
00
CO
CM
CM
Tt-
%
CM
oo
CO
CM
0
CM
vd
*
o
r-
10
g
s
CM
VO
ON
ON
ON
5?
oo
*
>o
ON
5
ON
oo
>o
9
00
CO
*
CO
1 I
r
0
Tt
t
CM
ON
vo
p
IO
CM
T 1
Tt-
» 1
CO
ON
9
s
to
1 1
1 I
»n
a
1 1
vo
oo
>o
oo
s
v .
* s
5! f
&i
si
6i
if
a |
*f
*f
" it
s
c? d.
5j g
^
0 |
o, K
11
1
fcS
OH
0
^
°
OO
VO
o
O
i i
CM
O
O
1
r-
r-
1O
^
oo
VO
CM
10
1O
3
CO
CO
0
CO
o
ON
CM
CM
CM
o
o
r 1
O
O
^
CO
CM
1O
ON
CO
> 1
10
1O
* H
1 1
1
0
ON
0
rt
?
VO
f-
10
r-
o
CM
CO
o
O
1
CM
1O
^
VO
1O
CM
IO
1O
CO
CM
CO
ON
CO
2
CM
o
r-
ON
o
CO
CO
CM
o
o
2
?i
IO
r-
VO
CO
10
IO
5J
CO
vo
vo
oo
1 1
11
*
p-
ON
s
o
vq
CM
o
o
£
0
CO
1O
CM
r-
00
CM
~
o
VO
CO
CM
L
d
rt
o
c~-
ON
10
oo
o
CO
8
0
o
ON
ON
10_
CO
oo
00
ON
o
1 1
o
CM
oo
CO
1 1
1 1
°
CM
00
10
ON
o
ON
CO
8
0
1
CO
8
CM_
VO
00
10
1 1
-i
o
00
H
0)
o
00
-*
ON
IO
CM
CM
ON
oo
o
0
1
r-
s
CM
0
10
CM
CM
c>
CM_
T t
oo
CO
B-ll
-------
O
Z j
*l
tcj
o
f-
0.
si
§1
f£>
2 ra
^
3m «4
^P
i 1^
^s
?g eJ
i&§
til
w §^
o
Ji
^ s
o ^
c> ^
CO
I
Q
£ S
tlj CC
H S
S
o
?
CO
1
*
o
VO
o
r>
vd
ts
§
8
5
CO
&
§
en
«?
>n
^
5?
o
*
3
VO
r^
c
en
O
t ^
o\
i
o\
n
f
O
o
r-
OJ
00
S
§
vS
5
e>»
en
g
S
00
en
S
vo
oo
^T
eN
O
z
in
*-H
en
I
S
>n
p~
VO
^r
en
m
OS
en
cs
en
en
vo
in
in
VO
O
r-
IN
8
CN
T «
en
t-»
c~
8
VO
in
S
OS
en
-
eM
vo
cs
o
p~
in
ON
Os
* 4
C4
S
i-H
*
-
en
in
en
p~
in
0
rj-
n-
oo
in
t
oo
o
So
T-H
CM
s
8
§
5f
VO
VO
CN
S
~
0
i
VO
1
CN
in
m
oo
5
VO
in
en
n
^
T-H
o
oo
S5
in
i
*-H
u-i
v>
T-H
S
vo
S
OS
m
CN
o\
S
3
o
00
en
eS
O
VO
R
vo
OJ
*
CO
in
Tt
i t
CM
Os
T-H
in
5
S
CN
^1-
* i
O
CO
VO
VO
CO
CN
s
o
t
CM
oo
o
VO
oo
oo
CO
T-H
^d-
CO
s
5
CO
oo
5
oo
s
CM
oo
CO
vo
o
00
in
*
OS
CM
m
VO
CN
OS
s
g
^H
S
s
00
9
en
oo
vo
o
CM
in
in
vo
VD
OO
VO
T-H
m
CM
in
5?
o
s
s
8
f-
00
9
k/l *>
* e
*|
*t
$1
^
6 g
e ^
^ e
5|
J
fct?
M ^
6
a |
0^
w 6
1 fr-3
gf f
^tj
1
^o ^~
rjj pg
CL,
Tt
CM
CM
O
CO
r-
CO
"
CO
Os
1 1
m
o
0
1
*^J*
?
*
VO
in
°
m
?
CO
CO
rt
T-H
rt
?
m
m
S
0
o
1 1
CM
in
vo
CO
in
°
in
~
o
Os
VO
CM
o
m
^.
CO
Os
CM
O
8
Os
VO
in
t
00
in
?
in
CO
CM
T-H
CO
CO
CM
Os
CO
o
oo
1 1
T-H
o
o
1
C*4
3
CN
CM
CM
in
V
in
in
CO
VO
VO
oo
CM
0
CO
o
s
VO
1 1
CM
o
0
in
p
f-
, i
in
in
CJ
in
*
*
T-H
OS
CO
m
r-
i i
OS
T-H
CM
,
^
VO
CO
o
CO
00
85
0
Os
in
0
0
8
VO
r~-
in
S
r-
rM
T-H
O
T-H
O
VO
CO
CM
m
CO
0
vq
in
gj
0
CS
r-;
T-H
o
o
1
CO
oo
in
OS
in
T-H
T 1
1 1
O
?
S3
CO
oo
00
CO
0
s
OS
CO
1 1
T-H
oo
t~
o
o
1
oo
in
i
OS
OS
S
1 1
^
5
oo
t
00
cs
0
in
m
i i
Os
CO
o
o
1
^
in
CM
m
CO
CO
£
o
o
3
oo
CO
E-12
-------
o
3
o
o
i
1
^ ^J
*""{ £h
^ s
1" u
%
E§ 1
8|
S j§
Q *" ^
H § 3
(D
t Q}
i _ Q
pq a> 3
S ,
u t-J
1*
D 3
PQ
Q
H|
M Cfl
S
i
CM
en
CM
vo
T 1
in
cs
o
in
cs'
CM
o
1 1
1 1
cs
1 1
-
CN
CO
1
i
s
en
OS
in
oo
o\
CS
f-
oo
CM
en
O
vo
cs
1 1
cs
r <
5
* 1
f-;
VO
VO
en
i
en
O
r^
» i
CM
CS
O
en
» t
oo
CM
* i
S
O
VO
en
1
O
2
8
vd
T 1
VO
o
in
i t
o
CM
OO
en
in
CM
t--
oo
CM
CM
OO
VO
o
in
vo
OS
i
^
8
vd
r-
in
s
»h
n
en
T 1
en
i i
r-
en
T I
en
S
vd
o
i t
o
cs
cs
o
VO
VO
§
oo
T-H
CM
T-H
VO
1 1
o
cs
en
en
CM
O
OS
en
?
CM
CM
in
S
vd
cs
cs
o
» I
T 1
o
f-
en
CS
g
en
m
CM
in
T 1
T-H
T <
in
5
s
vd
in
o
OS
VO
en
vo
cs
en
r-
o
OS
en
oo
Os
en
CM
en
vo
>n'
in
en
CM
O
in
OS
r <
CS
00
in
in
HJ
o
OS
en
CM
CM
O
VO
vo
oo
M a
s
-
*r
rf
>n
T 1
S
o
8
?
^
Os
n
>n
>n
^
>n
C3
V
en
en
» i
T <
?
>n
>n
T t
*
S
o
s
o
t 1
cs
in
?
VO
en
in
3
in
o
V
1-H
T 1
1
O
ON
IS)
1-H
1 H
0
oq
in
en
i i
0
T-H
*0
r-H
S
o
vo
00
in
^
en
in
oq
in
>n
?
in
O
V
en
CM
T 1
en
OS
en
oe>
T-H
O
VO
-
vq
in
S
C3
8
?
en
T 1
in
?
in
in
en
in
in
?
in
?
i t
in
en
vo
vo
00
^-1
B-13
-------
D
oo
O
O
§
!5 j
i"t
3^
< 6
Cu
s u
H
%
fl 1
.j
8f
5 b
% 0 CB
Q £
^ B «d
o C w>
H m P
oJU
£ 3
f JU
££ 3
1 £
IP
J«
e =
di k_J
1 <2>
5 3
PQ
&
Q
Q
s i
CD U
H 53
Pk
in
?
CM
» 4
OS
*
s
in
o
CO
n
?
\o
CM
CO
8s
in
s
CO
^-1
in
T-H
o
o
0
o
<
t t
1
t~~
CO
55
IT)
CO
r-l
1
cs
o
co
s
cs
» (
CN
VO
CO
§
£
in
CO
o\
n
O
£
VO
o\
o\
CO
§
CM
>o
CO
VO
CO
t t
T <
VO
cs
» (
CO
cs
0
CO
0
* (
>o
S
i t
0")
VO
CO
t (
1 1
cs
*
>0
CM
o\
CO
-
m
^
O
* (
*
cs
§
>o
VO
00
l>
T *
CM
£
CO
in
in
-*
CO
t t
§
VO
1
^ Qa
* e
d |
J
* t
J
« 1
0 |
n ^
co g1
PH j
'- ^
2 s
£ J
S s
5-
Q
TEST
PERIOD
in
t-H
1 1
3
in
r-;
o
o\
OS
t t
8
?
OS
Iin
?
in
CO
r--
in
?
in
?
CO
CO
vo
i i
t t
?
CO
CM
C-;
m
CO
CM
CO
o
s
o
m
*
in
?
*
OS
in
?
in
^
i 1
f-H
1
ON
00
9
in
CM
VO
CO
Tl-
oo
t-H
m
VO
CM
S
?
CM
CO
in
?
CO
CO
in
?
m
?
S
OS
CO
-
t~-
t-H
t-H
?
CO
CM
s
VO
t-H
O
o
2
?
OS
t-H
in
?
CM
CO
OS
in
?
m
?
t-H
in
in
t-H
CM
O
CO
t-;
in
in
S
O
S
O
t
t-H
IT)
?
VD,
CO
CO
in
?
in
?
T-H
t I
^J.
CN
T-H
^
OS
CO
»
t-H
CO
*'
1 1
o
o
s
?
in
t-H
m
^
CM
iri
in
^^
?
in
?
in
f-~
t-H
Tt
t-H
C5
CO
in
?S
VO
*
t-H
O
o
cS
?
OS
m
^
OS
iri
in
T-H
?
T-H
?
VO
CO
CM
OS
t-H
t-H
O
vq
in
CO
*
CM
oo
*
»-H
O
0
s
?
oo
o
in
?
r-;
in
c~
t-H
?
T-H
?
CM
CO
CO
CO
in
t i
t-H
3
in
n
co
CO
CM
iri
t i
o
0
s
^
CM
0
CM
?
co
*
-------
CN
O
o
i
gml
SI
0,
S U
H
a
a.
pq g
O Ijjjj
f s I
Q t£
l j?U
H pq 3
6 1 *&>
A, u
p| 1 1
jO
o
l£
-0 3
P3
O i-J
i 1
m
Q
PH
3
VO
CN
OO
VO
Tfr
ON
IO
oo
CN
O
io
O
1 1
P-
1O
T-H
T-H
oo
T-H
o
ON
ON
en
en
1
O
CN
T-H
10
oq
IO
oo
T-H
1
ON
O
00
VO
T 1
o
CN
CN
T-H
S
» I
V 1
VO
*t
vq
0
C5
oo
CN
T-H
CN
oo
o
oo
CN
OO
8
T I
$
ON
i
CN
__,
sj-
S
en
t
O
CN
T-H
T-H
r-H
O
OO
en
en
en
10
S
is
ON
OO
CN
en
oo
en
en
O
1
01
o
E
T 1
T t
1
en
ON
CN
CN
en
ON
T-H
CN
ON
en
-
i
2
10
en
S
t
ON
O
jo
T 1
CN
VO
en
T 1
en
oo
IO
oo
5?
w-i
a
en
vo
o
t
i-H
O
O
T 1
CO
i-H
-
-
T t
CO
-
i-H
1 1
1)
V
ON
2
CN
d
ON
en
-
-
0
-
O
en
g
d
1O
VO
en
vq
0
en
O
o
o
0
o
o
VO
en
CN
s
oo
en
oo
CN
0
en
T-H
>o
c>
r
1O
CN
cs
o
o
0
o
CN
CN
OO
en
en
C)
en
t
oo
io
*
CN
O
en
T-H
1 !
00
ON
w *a
Q
si
tq p^
f-4 fTl
OH
^
T <
t I
?
CN
<0
en
CN
0
>o
Ien
O
8
^
-ri-
r-H
,_,
0
o
CN
>0
?
C3
i-H
to
?
o
en
CN
>o
?
>o
?
T-H
T-H
1
ON
en
CN
en
0
>o
vd
1O
CN
en
en
o
O
8
3
o
T-H
>o
?
en
r~
O
CN
>O
?
>o
$
a
ON
en
T-H
si;
?
*
vd
oo
VO
10
o
o
o
o
?
00
>o
?
ON
C<3
>o
?
10
?
T-H
IO
ON
?
t>
0
Tf
o
o
o
?
IO
o
3
oo
l>
>o
?
VO
en
CN
10
^
>0
?
. t
OO
?
*'
VO
en
O
00
' *
O
0
8
?
oo
t>
IO
?
t-~
^
ON
T-H
T-H
?
<0
?
IO
f-
T-H
wr
o
VO
vd
1O
IO
g;
*
o
^
8
?
ON
vd
IO
?
en
06
oo
T-H
T-H
?
en
O
T-H
VO
en
CN
\
CN
O
vq
ON
o
oo
a
10
0
0
8
3
T-H
Tt
10
?
en
CN
-------
m
t
3
p
i
s
R-
-A
&i
Q
o
15
ft
£S
cs1
8
s
SI
64 M
,f
a *d
Z |
60 d
s 1
*f
If
J
5 S
£6
«|
o i
p
0^
VI g
*V tJ
of
si
z E
&
£3 5
HS
cs
?
oq
CO
v>
^
CO
CO
3
o
s
o
en
m
?
oo
en
f-
m
?
vq
en
es
ON
m
?
m
?
en
es
?
as
en
*
?
en
in
n
?
VI
~
IE
vo
oo
s
n
o
V
2
o\
er
vo
in
?
*
oo
*
>o
» 1
in
S
S
o
in
o
?
in
cs
in
?
c
oo
vq
r~
?
in
o
V
in
f-
C3\
<-5
§
en
*
?
00
VC5
S
T 1
CO
\o
oo
s
o
8
?
oo
>n
?
*
>o
t~
n
V
in
?
in
in
(N
in
00
e-4
es
oo
o\
cs
in
o
00
>o
s
CN
r~
o
vo
o
o
8
?
CS
t-H
i 1
m
?
s
m
V
??
o
Jc
en
^
O
o\
ON
oo
CS
r-
'-^
S
o
S
?
*
en
in
?
r-;
00
r-H
rt
?
s?
o
So
en
es
en
o
t~
\o
vo
en
CS
S
ey\
vo
o
o
cS
?
*
OS
<1
?
t
rt
>o
^
^
in
O
o
5-
en
CS
-------
D
co
£
O
i
|l
3^
< B
B U
H
%
SI
§1
f<3
2 3
p£
£« 3
6JU
Jv 3
><
I? J "=i
± ^, M
Sx, a
u
IB
(D
.N
£*
84
81
«
1
TEST
PERIOD
V,
3
'fr
OS
in
o
oo
>n
£
r-
o
in
rS
~
OO
CO
1 1
co
en
in
8
s
9
CO
B
so
o
(
r-
m
o
so
3
in
o
so
cs
oo
oo
a
R
Ol
1
g
1
1
os
~3-
*t
oo
o\
m
oo
*?
»-H
VI
t-
SO
01
oo
1
0
m
%
s
S5
m
o\
CO
oo
o\
T 1
v->
?
a
vo
*
in
O
so
f
r-
o
o\
ol
o
%
m
SO
s
§
1 1
01
m
^t
Si
s
so
2
CO
01
o
in
s
so
o
"?
0)
o
*
CO
01
CO
oo
f-
os
S3
oo
t
CO
t 1
0\
in
in
§
VO
en
1
CO
01
*
ss
in
ol
os
01
>n
in
CO
o
a
s
01
01
£
os
o
VI
in
os
£
0\
en
*o
?
o
0
>n
OS
OS
>T)
O
CO
1
OO
o
01
CS
t--
l-H
s
v>
oo
>n
VO
§
>n
so
CO
I
ol
p-
t-H
m
-
*
*
>n
Os
>n
oo
i
01
h
CO
0)
Os
%
s
oo
cs
V-l
t
§
»-H
01
*-H
1 1
SO
00
in
9
v>
oo
so
l !
T-*
v>
01
t^
o
o
Ol
i
r-
i-H
CO
3
1 1
>n
< i
SO
oo
v.
?
01
CO
00
s;
>n
CO
OS
*
00
§
in
>n
oo
a
2
oo
00
>n
o
Os
* t
-
01
o
CO
OS
>n
in
in
*
so
O
Tf
r-
11
T-H
T-H
01
s
CO
5
oo
CO
oo
s
Os
CO
>n
^
o
yH
g
so
01
1
*
o
SO
SO
1
CO
8
CO
oo
oo
CO
T 1
9
so
so
£
V <
01
>n
r-
i i
o
ft
8
SO
CO
YH
in
i i
in
Os
CO
s;
1
§
Ol
i
oo
§
1
Os
01
oo
n
s
so
*?
m
o
1
?
m
i
01
3
T 1
o\
3
so
01
! 1
Os
i
^i;
CO
n
CO
SO
o
CO
01
=
1 1
in
SO
o
B
t <
ol
01
-*
s
00
s
9
s
2;
m
m
Ol
i i
t
3
in
t-H
SO
OS
Ol
CO
OS
c
1
so
§
cs
m
oo
Ol
ol
M-
in
so
in
Os
CO
i
00
CO
oo
Ol
m
2
>n
9
so
SO
o
f 1
01
oo
Os
01
g
o
SO
m
S
m
oo
oo
oo
o
in
*
£
2
Os
OS
01
CO
t~
1 1
c-
1 1
m
S
SO
SO
CO
u->
w-i
t-
TJ-
-
SO
OS
CO
m
o
t-~
CO
oo
in
%
oo
Os
OS
1 1
s
c
in
so
o
§
s
so
01
oo
CO
ol
V~l
o
T 4
*
s
so
?
t-
co
0
01
I
OS
in
» <
S
CO
in
f-
m
ss
s
o
5
OS
C*1
O
01
SO
01
CO
0
CO
s
T t
CO
in
Os
CO
5
so
M-
in
01
CO
*
OS
§
oo
9
CO
t
oo
OS
in
oo
» i
i
SO
1 1
s
OS
oo
so
so
1 1
1
in
£
8
-a-
§
5
^
o
in
oo
f-
r-
g
CO
so
oo
s
g
ol
CO
§
1
W a
^
*i
git
2 e
«3|
ll
3 ^
3 i
£1
«l
o |
o"^
V) Q
%*&
O cj
OH H
lz ,
^^
z B
&
Q
&§
CO M
PQ 2
C-H frl
CM
e-J
CO
0
t~-
CO
p~
01
CO
oo
r-
IS
o
in
m
?
ro
*
SO
r^
in
?
in
?
o
OS
OS
*
^^
0
in
Os
O
01
so
S
o
S
?
01
r~
in
?
so
*
f-
SO
in
?
m
?
CO
ol
CO
OS
CO
V)
01
T-H
o
f-
SO
CO
in
8
o
g
?
o
m
?
oq
in
CO
so
in
?
m
?
>n
-
so
SO
oo
oo
^H
?
t-~
in
Os
O
in
ol
CO
S
o
2
?
01
Tf
m
?
SO
*
t
in
?
n
o
IS
ol
in
oo
N
Ol
OO
OS
ol
<*-
CO
^ !
?
r~
00
oo
oo
*
cs
'd-
1 1
oo
o
o
in
o
?
in
t i
m
?
CO
Tf
rt
?
-
o
c!
CO
l
0
CO
CO
01
s
*
oo
00
m
o
8
?
OS
oi
m
?
s
o
CO
^
?
>n
00
O
CS
oo
CO
*o
in
(
^
Os
P-
C^
CO
oo
1 1
1 I
s
o
S
?
CO
OS
01
?
Os
00
Tf
n
rt
?
in
o
0
5-
CO
00
CO
rt
o
'S-
oo
SO
oo
01
s
c-
s
o
8
?
*
SO
oo
CO
t~
Os
r 1
rt
?
so
n
o
oo
CO
*
OQ
CO
ol
o
01
SO
*
oo
01
n
Os
r-
S
o
S
?
m
so
ol
?
OS
Os
in
S
?
in
*
O
oo
9
B-17
-------
co
d
o
o
§
p^ 1
i"t
^ B
1*0
a) o
H
ffi
CL,
w S
sf
.B
D« O |j
2 W d
0 P g)
H pq a
u->
1 I
in
m
CM
3;
»n
^
(**i
"i
i*
"*
co
CO
CM
f^
f^
VO
CO
ON
CM
n
vn
T 4
VO
ON
-
T^-
"*
in
CM
00
CO
in
oo
o
CO
vo
in
VO
ON
ON
CO
in
CM
11
rM
oo
CO
ON
r^
O
vq
^
ON
in
in
t
CO
ON
CM
VO
OO
in
CM
,^-
CM
1 (
in
oo
m
CM
0
00
o
i t
^
co
in
O
oq
in
-
ON
CO
VO
CO
ON
in
0
t"
CM
vn
00
*o
^
o
oo
ON
^
ro
VO
oo
co
*
CO
Ji
CO
CM
o
VO
oo
VO
in
CM
00
VO
1 (
f^
CM
CO
^
s?
oo
1-1
in
T^-
f
^
14
VO
t>
C3
T^-
^
OO
t-~
co
00
in
m
CM
14
CM
ro
^
VO
vn
CO
f^_
CM
no
^
(J
W ob
<^ bb
^ ^i
Z S
60 tj
rt "ti.
u S
1 f
el
*t
W B
U M
B
Q*1 'g.
oo g
^ H>
gl
^"t
Z E
Q
C/3 i t
w &
H W
PM
oo
ON
CM
i i
VO
co
CM
i i
CM
O
CM
^H
VD
in
o
co
O
o
q
VO
CM
in
°
vq
00
oo
i i
"
CO
CM
O
CM
OO
CO
C-;
14
1 1
CM
VO
CO
ON
i 1
5
CO
p
q
^.
in
CM
°
vd
ON
TJ-
i t
^
O
O
1-H
vq
CO
CM
O
ON
VO
CO
vq
T 4
VO
r-
in
o
0
q
ON
ON
0
OO
vd
ON
oo
^
^
in
d
oo
CO
CM
j
O
(
VO
14
ON
i i
ON
ON
in
o
1
1 4
ON
CM
O
V
oo
VD
VO
q
v
ON
oo
ON
co
B-18
-------
D
00
d
o
o
8
Q
o
H
s
r. 53
00 HH
w S
H W
/ 1
-V.D
3s
r-
01
_.4.
3-f
cs
3-
CS
ZXt*
3
csc
rS3
"
cs
ZBO
-"Q
^cs
r
Vi
-4
15
CS
"0°
I
1
^ >
LI10
"I
j!
D
l"i
l
rv
G 1
oo
-ON
01
H
Q O
.
2
-TO
or,"
-vo
! ^.
CS
VO
v
V*
ft
>
ft!
_J
y O !
I 5 !
-i
CV3
P
Q i
*tr
j
J ^
B-19
-------
D
co
d
o
o
o
S
ft
"3 M
< 6
0,
S U
ffl
Q.
e 1
8|
*£j 4>
S 2 $
£ 3
U
i U *J.
Q ^l Ql
1 o
i& s
(D
,-i, C3 -J
^ KJ ^i
"£3 H ^*
w JS 3
S
1 1*
H
o
C 1
M ^i
o S
w
Q
H o
oo a
m p/
P PC
PH
o
"*
n
vo
VO
* 1
t-~
vn
VO
CM
CO
^-f.
VO
CM
S
f-
ro
fs
CM
no
ON
CM
^
^
0
ON
»n
?
d>
ON
CO
oo
ON
VO
OO
r>
CM
O
CO
Tf
M-
o
VO
o
VO
t
VO
o
w>
CO
r <
t-H
S
CM
CM
CM
*o
ON
C"
^ 1
»o
Tt
CO
r-H
0
^
ON
CO
C5
^
VO
vq
» t
^j~
CM
OO
£J
O
OO
n\
CM
CO
VO
CM
ro
ON
oo
ro
^
CO
T-H
oo
CM
oo
>o
CM
i i
CO
CO
oo
CM
i
t 1
S
CM
^^
oo
o
CO
co
ON
VO
oo
CO
ON
r-
ON
O
f
t t
VO
oo
CM
VO
CO
t 1
rs
oo
^"
S
CM
*o
oo
oo
ON
^
J tJ
W ^b
,f
^ a
00 t^
6 |
It
a |
*l
a|
0 |>
rr J
00 g
o't
52;
^1
1
^§
H (d
PH
OO
ON
CM
VO
CO
CM
vq
co
l-H
O
CO
oo
CO
CO
VO
a
ON
O
O
S
°
^
ON
?
O
I (
CO
n
i
00
0
CM
00
CO
co
O
ON
co
CO
s
§
o
S
°
CM
1 I
CM
00
ON
CO
OO
J 1
t
o
1 1
10
T-H
-------
I
Q
^ k~J
Vn ^,
i 6
^ j
^ 8
S 0
n
o
v
t~-
1 (
en
i i
T 1
es
VO
VO
en
0
0
o
o
o
o
ON
in
o
v
~
^-'
Y *4
in
o
t-H
VO
en
&J
cn
^H
cn
i^-j
T-H
oo
ON
es
O
r-
VO
^^
en
oo
^
vt
1 1
O
1-H
VO
Tt
ON
i-H
-
o
*
,_,
m
<-<
in
o
^
oo
en
m
r-~
» *
on
oo
r-
00
t-
=
in
0
o
o
-
0
T (
0
o\
en
i-H
en
s
es
oo
T 1
VO
en
ON
oo
en
vn
ON
O
(N
^_|
ON
O
T (
en
vo
i i
VO
oo
en
es
C5
& it
es
oq
en
VO
U
|
eS
O
V
U
ON
ON
cs
es
s'
in
in
o
C5
c5
V
S
>n
Q
CS
ON
en
H
oo
es
cs
ON
o
cs
VO
en
es
es
VO
ts
cs
B-21
-------
: 1 : . i I
i -? ' i ^ i '- ;^~* { r jo L
' - E ' !- ' i '
. . >- '< >. M !
\ j . . 1
> . «* ; » JfO i !
~ '- : : i ;b ! I13 |S 1
:_j.~-j.-..:._J !_ !_ I
' i M i» !
^ "C ; 17° * ' i"^ -"
" " " " ~. ~ T~ ~\ ~<
l-*J !-"1 >
I f i-V ! ! O- ! O '
~ i : . -O iO O <
. . ^ . . I ,
: ; , j
:* i ' \ =~^ ;
? »-? ' \ ~-> < \s°) rv L
. (V '
S :C ': . ' !o !o --) :
;- -'- i ; ic to o
V v ; ; ,V | : '
. ! . ' t i
___; : J ; r
*w ^^ t i ^w J ' '*~, , ,f~*
CT p . , i l. ;.w f :>- O
«-,*». , *-*. . i >*. ^.
*-.«*- , I ^* I i1-.? *w*
M i . iv ! iv v
i * ;
C -"" ; ' i ' !O ; !» !O
'- i '-il, i-L
!" , j ! t ' 1
I" '-> :.,!."« ru ,-V
;. __,..,%.. r0 J-..
'J 1 :V V
: | f j
i i ;
* > ' » : '
** t - , . . ,
-' t ' i ; j
r , i j
,O- ' !f-J
i i . j i
1 ; '- '
~ i . . : trti i irv .-v
:~. " -~ ' ' < j o j ' o o
1 * ' J v 1
1 ^ '
- -" . - - : 1 [
-.--' 1 1
'C ''~ ' :" ' 1 ! : ;O
| !v ! '' ,'v'
'. !
: JT :- '"' ."\- ~ ' ; ':">> .'/O
'" ' ^-- J^ -^.' "f"- "' ' '~ --'"*< '": i
; - ' ' j '' 5- ~" - f *"" t '
i 1 t
1
f
n >^
1
f
i>
^' !
.
r
^ oc ,;
^ ^ '
-^** &5
a
b.
3^ ^ E
a ' r
j.
f &
u ,
& Si
3 _
3
!
r
J? rt
43 CJ
3 j
f"s
r t^5
v? "'
3 o ,
3 tfl
1
< 3. i
Js" V
" ria
- s
i
'£
W-i
- ; -
rv
L-T
1
CV
Jo
.
p'
Q^
Ti
0
-
10
tr
s'
i
'
S
/D
H'
i
OO
i
,
00
CV
-_
p'
tr>
ru
_,
O
t <
t ,
hwf
O/
-n
/o
n
T **
r.,
O
V
oo
ro
<^>
1--^
CX5
-1
OS
-H
-i
oo
^
/o
rv
i r
O
O
o
V-,
o
I
5
", -
-1
__
> T
1 f
o
v
CV
r
ro
'-1
-,
o
17
T
I
ru
0
V
CV
ru
"i^"
rv
0
'
(V
CV
2
rv
13
t r
r^
fr-
CX!
/o
-1
lo
V
<-1
"
1 r
"1
O
1
-v
2i
_
=
,
o
o
(~1
0
o
o
.-o
X? <^_
a O
2s
f ^
_
03
0 B
H
ac
j^
-' B.
3' W
Is
h
t x =?
{&* ^ rt
c m H
~ 1
b 1 ,
35 J*sw O
K X
^ 8 $
iw ^-J >
G ?< &
.
1 1 f
O-,
cp
tu
^3 a
e c7
01
dS" ^
« 8
n
t?
Q
Q
;-
r
' :
.!
s
u,
L- 1
w
s
h^
S.
C
Q
tu
S
P
to
.
I
-------
u ts
I
*«
1
S J
r
-8
^a^
>^N
B
p*3
e
S
0
Q
WellLD.
oo
ON
O
oo
r-
cs
0
cs
cs
CO
ON
O
o
^
ON
H
M
cs
3
ON
O
1
cs
0
st-
, (
S
o
ON
(3
*=?
V
cs
cs
U
ON
O
0
oo
g
0
oo
o
o
o
ON
c
s
o
3
O\
o
s
0
8
o
ON
S
o
s
ON
O
cs
cs
cs
0
^.J.
[*^
s
o
ON
A
S
o
s
VO
ON
o
T 1
i
0
1 1
ON
O
O
ON
3
O
i i
ON
O
CS
T4
0
o
ON
l/"J
O
o
ON
c
3
O
CS
1
oo
o
CO
ON
VD
t-H
0
o
ON
00
^O
O
O
ON
C
O
S
oo
ON
O
ON
T (
0
cs
00
O
ON
^
3
O
^
m
ON
O
§
0
o
VO
in
o
ON
A
3
O
c-
ci
r-
oo
o
0
O
ON
s
O
st-
ON
C
3
O
0
T 1
ON
ON
O
I
O
oo
in
T i
0
o
ON
§'
4
*-H
O
ON
o
CS
ON
r-~
S
o
o
CS
CO
0
o
si-
c
3
4
S
r-
ON
O
r>
CO
S
O
ON
T-H
T-H
0
o
ON
3
S
oo
ON
O
oo
i
o
c^
ON
T t
0
o
st-
ON
C
3
S
in
ON
O
o
ON
OO
8
0
o
m
O
s
o
st-
ON
n
3
i
S
ON
ON
O
ON
O
in
0
o
r-.
in
i t
0
o
ON
A
3
4
op
ON
O
a
CS
s
0
in
^^
o
o
st-
ON
3
4
o
C-l
-------
m
"0
o
o
35
to
C3
sit
" CO
-05"
^- CD
< CO
OJ. CO
CD
O 3D O m C
CD 3 3
^ < ?
o
CD o Q.
Si :T O
(-> E5" m 2.
X o" < §L
^1 Is
W = i 2.
O5 O J CD
00 3 CD o
G>
ro
=*
m °
-3
ro
CO
CD
m
33
Q
^
-o
O
U)
§s
------- |