EPA/600/A-92/263

American Wafer Works Association

-------
                           FIELD EXPERIENCES
               VOC AND PESTICIDE REMOVAL USING GAG
                      SUFFOLK COUNTY, NEW YORK
                          David Harris, M.D., M.P.H.
                               Commissioner
                  Suffolk County Department of Health Services
                               Aldo AndreoE
                                  Director
                   Division of Environmental Health Services
                               Joseph H. Baier
                           Office of Water Resources
 INTRODUCTION

 Suffolk County is a community on the eastern portion of Long Hand, approximately
 85 miles long, 10 miles wide, with a land mass of over 850 square miles.
 Groundwater is the sole source of water supply to the L5 million population.

 Community water supply provides water to 125 million people with the remaining
 250,000 using private wells. In addition, over 800 non-community suppliers are active
 throughout the eastern portions of the county.
                                     jf /•
 The Department of Health Services has been screening grouudvrater quality for
 agricultural and organic contaminants since 1977. Thus far, with more than 50,000
 samples analyzed. 19 pesticides and 53 organic compounds have been identified. The
, compounds of greatest concern are: the pesticides, aldicarb and 1,2 dichloropropane
 (DCP); and the organics, tetrachloroethene (PCE — forraaEy tetrachloethylene), and
 1,1,1 tricWoroethanetTCA).

 Acceptable levels of contamination are provided by the USEPA (MCLs) and, in New
 York, by the state health department through standards (generally those of EPA) and
 guidelines. Perhaps the greatest impact to public and private water supply is the
 recently enacted 5 ppb VOC guideline (actually defined as principal organic      :
 contaminants [POC]).  This has caused 15 percent of existing community supply wells
 and 25 percent of private wells to be classiGed unpotable. At this point, the
 appearance of any VOC or blip on a GC analysis is cause to remove a well from
 operation. This same problem now arises with GAC operation since breakthrough of
 1 ppb, slightly over the detection level, leaves little operating margin before S ppb is
 reached.
                                        383-  .

-------
OVERVIEW

Over the past eight years, GAC systems have been installed for treatment of VOC
and pesticides. Several have seen, multiple carbon replacement; changes in water
quality; and experienced a variety of operational difficulties. The following GAC
systems wjE be dismissed:                                    -

     *    Village of Greenport - one 20,000 Ib. GAC operating from 1980 to present
          for removal of aldicarb and carbofuran.

     *    Southold-useof GAC in series for aldicarb, carbofuran
          andDCP         '                                      "

     *    Dix Hills Water District - three 20,000 Ib. units in parallel for removal of
          DCP.

     »    Suffolk Count)' Water Authority - 4 separate well fields treating aldicarb,
          DCP, TCA, TCE, PCE and others, with a variety of operating experiences.

Operational problems encountered with each unit will abo be discussed.

GREENPORT WELL NO. 6-1

This is a 450 gpm well screened at 170 feet.  It was the first community water
supply well to be found contaminated with aldicarb. This pesticide was used
extensively by farmers for 3-1/2 growing seasons before being discovered in
groundwater. Although,  at first, the contamination was considered to be
temporary (1979), It is stHl present in the aquifer and in this community well.

Initial aldicarb concentration was 12 ppb. Since then, the values have gone as high
as 26 ppb and back to the present levels between 7 and 10 ppb. In addition, 6-10
ppb of carbofuran was found; although present levels are 1-3 ppb.

A 20,000 Ib. single  GAC filter was installed in August, 1980. The unit provides a 12-
minute EBCT and has been rebedded six times, with a seventh scheduled for the fall
of 1989.  Table 1 shows the gallons treated (throughput) for each rebed, broken
down to breakthrough and exhaustion.  (New York State established a 7 ppb guideline
for total aldicarb — both metabolites combined). The range of aldicarb values is
also shown.
                                        384

-------
 This filter is sampled weekly when the well is operated so the values presented could
 be plus or minus 3 or 4 MG. The meter OR the filter malfunctioned and was
 replaced three times during the operating period, Pumpage values were often
 esi imaied based on hours of operation and well capacity.

 The important information from this data is that the breakthrough does occur within
 a range of 80-100 MG which the operator can use as a guide for sampling and to
 plan for carbon replacement at non-critical (off-peak) times. The exhaustion value is
 not always at 7 ppb because rebedding is requested after breakthrough occurs and
 scheduling doesn't always coincide with total exhaustion. The data does demonstrate
 a consistency of operation with a fairly consistent contaminant loading.

 The average bed volumes (BV) for breakthrough and exhaustion is 17,000 and 21,645,
 respectively. This data will be compared to that from the Southold removal and
 Long Spring Road (SCWA) later in the report             '    .

 This GAC filter was the first carbon treatment experience for water suppliers, Le.,
 its operating properties, use, and aggressiveness. The initial carbon transfers were
 difficult due to lack of proper equipment and often resulted in carbon losses
 (overfilling vessel, burst fittings, and hoses). Other operating problems include: the
 underdrain system failed and a new set of coEectors was installed; the column
 partially froze [bursting valves] and a building was erected around the unit; meters
 failed, and the main pump was struck by lightning.

 Carbon is abrasive and corrosive; this requires that the tank .interior be inspected
fjf
 after carbon offloading and before rebedding. The interior of this filter is
 sandblasted, cleaned and resurfaced (epoxy) every third carbon change and is due for
 repainting after the present carbon is removed.

 EPA/SCDHS PILOT PROJECT - SOUTHOLD

 A cooperative agreement was initiated with USEPA Drinking Water Research Division
 (DWRD), Cincinnati. Ohio, to examine the effectiveness of certain water treatment
 systems to  remove agricultural chemicals. A 5-gpm pilot plant was constructed to
 examine GAC in series with ion exchange, and reverse osmosis in parallel  (to GAC
 and IX)  to remove: aldtcarb sulfone; aldicarb sulfoxide; DCP; carbofuran; and
 nitrates. Table 2 presents the raw water characteristics observed during the one
 year of pilot plant operation.
                                          385 .

-------
Three GAG columns were used in series.  Each unit provided an EBCT of 5 minutes
with a total contact time of 15 minutes. The process used 10 cu. ft. of Filtersorb-
300. Sampling was performed on each column so that breakthrough and exhaustion
could be observed. A total of five columns were tested with the lead column
changing as exhaustion occurred (example; when column #1 exhausted, it was
rebediied and became column #4, the last in the series).

Tne resells of this work were presented in a previous report (GAC and RQ
Treatment for Removal of Agricultural Chemicals from Groundwater; Baier, Lykins;
EPA Coop. Agreement #CR-811109-02; April 1987). This report presents additional
information regarding series flow for GAC

Multiple adsorbers can be piped for series or parallel flow. Initially, most of the
larger units on Long Island have been installed for parallel flow but concern has
developed as a direct result of a NYSDOH requirement that lowered the VOC
concentration in potable water to 5 ppb. The reason  for concern involves the
margin of error that can be tolerated in operation of a GAC adsorber. The previous
guidelines had been 50 ppb which allowed additional operation of an adsorber for
some time after the initial breakthrough.  This also allowed for s more complete use
of carbon before replacement.  With 5 ppb as the guideline, any breakthrough of a
VOC (current detection is 0.5 ppb) warrants carbon replacement
                                       s*s         .                 ' .
The head loss across on adsorber results in a loss of well capacity and makes
parallel flow more desirable for treatment. This is due to the fact that GAC
treatment is being added  to existing wells not designed to handle the additional head
loss and results in capacity reduction. Recent tests by one supplier showed a 28%
re'duction in well capacity during parallel flow into 2 - 20,000# adsorbers. The same
well capacity reduced 40 percent in series flow.

The type of testing performed at the Southold site allowed the carbon units to be
compared for breakthrough (5 ppb) and exhaustion (influent concentrations). This
results in an increased carbon use or an increased capacity. These results are
presented for DCP, aldicarb sulfone, and aldicarb solfoxide in Tables 3,4 and 5.
With the exception of unit #1 each carbon adsorber  initially starts operating as a
polishing unit and its loading increases as the wave front moves through the previous
unit. By allowing the unit to operate until the entire wave front has passed
increases the capacity of each unit. This increased capacity (more complete use of
the carbon), for each compound, is apparent from the tables.  The capacity increase
can be compared to the lost flow rate (12 percent) that takes place between parallel
                                        386                 '          '

-------
 and series flow.

 The following conclusions have been drawn from the data:

 1.   Unit # I, which received full influent concentration from the start, has the
     poorest performance.  However, k fairness, the other units did process water
     that had no contamination for some period with the loading gradually
     increasing.

 2.   The B V treated to 5 ppb and influent exhaustion did settle into a manageable
     range for each contaminant (similar to Greenport).

 3.   Series flow permitted each column to treat a greater volume of water (40 to
      110 percent added life).

 4.   Although DCP was the least adsorbable contaminant, each column continued to
     remove other compounds without affecting the capacity of columns.

 5.   Series flow allows the operator a comforting degree of operation — not to
     worry when 1-2 ppb in lead column's effluent.

 DIXHILLS

 The Wolf Hill Road well field is located in a residential area and has a 1,000 gpm
 and a 1,400 gprn well that are 326 feet and 562 feet deep, respectively. The deep
's
 well first showed DCP in 1984 and was shut down. The shallow well had been under
 repair and when it returned to service, DCP appeared.  The state VOC guideline was
 50 ppb and, although the concentrations in each well never exceeded the limit, the
 water supplier decided to install GAC treatment.

 The facility design called for 3 - 20,000 Ib. units operating in parallel and providing
 11.5 min. EBCT for the larger well. When both wells operate through the filters, the
 EBCT lowers to 6.5 min. The design  used isotherm results to anticipate 274 MG
 treated at 36 ppb DCP and 5 ppb TCP.

 The filters were loaded in May 1987 (Filtersorb-300) and were used to handle peak
 loads during 1987 and 1988, During this time, no change of influent concentration
 was found. At the end of the 1988 season, 107 MG was treated by the three filters
 and a sampling  tap located at the 80 percent level in the filter showed 2 ppb in the
                                        387

                                         •r

-------
water (all three filters). The district decided to rebed in order to insure the
availability of the site for the 1989 pumping season.

Table 6 presents the actual and design performance of the Dix Hills well.  As shown,
the performances at Southold and Woodchuck Hollow  (similar contaminants)agree
with the design values of Dix Hills, but not the actual value. What caused this
treatment reduction?

Several operational difficulties were encountered during the two seasons of use that
may have contributed to the reduced capacity. Chronologically, the units first
experienced a rupture disc failure (filter #3) when the filtered water was fed into
the system for the first time. The blow-off valve was closing and the system valve
was opening when the disc failed. A faulty disc was thought to be the
cause as the pressure wave should have caused the disc on filter #1 to fail first.

The next problem that occurred was the failure of the underdrains in filter #3
resulting in the loss of about 6,500 Ibs. of carbon. The repair was made and each
filter was inspected to verify the condition of all the underdrains. In order to do
this, the carbon in each filter had to be removed and then returned, again disturbing
the equilibrium.                    /,

The final problem occurred on Thanksgiving Eve, 1987. A rapid drop in temperature
caused the effluent lines and parts of the GAC units to freeze. The system was
turned off and allowed to thaw. The units were drained but the effect that freezing
could have on carbon is  unknown. It should only affect the pore structure and not
damage the adsorptive properties.

Although only 6,666 B V have been processed, another 20 percent could be
anticipated. This would bring the BV to about half of design capacity. It is felt
that future carbon bedding will result in increased capacity.

A postscript can be added.  The filters were emptied in preparation for new carbon.
Inspection revealed that parts of the 10-ml epoxy coating had corroded and the steel
plate was pitted. Improper surface preparations may have been a cause.

WOODCHUCK HOLLOW

There are two wells at this site; one 531 foot deep, 2,400 gpm; and one 508 foot
deep, 1,500 gpm. Both wells are showing contamination  as follows:
                                         388
                                         L

-------
         DCP less than .5 to 7 ppb

         TCE less than 5 to 2 ppb

         TCA less than .5 to 5 ppb

         TCP less than J to 2 ppb

The contamination was first identified in January 1988, and with an impending 5 ppb
regulation, 2 - 20,000 ib. units were installed in November 1988.  Ether well can
operate into the filter with EBCTs of 4.5 min. and 9.2 min. for the 2,400 gpm and
1,500 gpm wells respectively, using paraUei flow. A second similar GAC system is
now being installed.

The contamination has not broken through as of March 1989 and a total of 210 MG
(see Table 6) has been treated. The lower concentration of VOCs has provided
additional capacity for this filter. Because of the difference in EBCTs, it has been
recommended to the owner (Suffolk County Water Authority JSCWA]) that after
breakthrough, the lower capacity well be used to provide longer contact time which
should extend the life of the filter.  A second 2 - 20,000 Ib. unit is presently being
                                       •
-------
A single 20,000 Ib. unit has been operating since July 19S4 and the wells are only
used seasonally. Breakthrough did not occur until early in the 19S6 season and the
column was taken off line (exhausted) in October of 19S6. A tola] of 110 MG was
treated. The single column was replaced by 2 - 20,000 Ib. units that operate in
parallel. A total of 41 MG has been processed through 19S8.

Table 6 shows Greenport (single column) and Long Spring Road to have similar
operating capacities with similar contaminant concentrations and EBCT (12 min).

BRIDGEHAMPTON ROAD

This site contains several wells with contamination but only one well will be
discussed. The well is  148 feet deep with a 700 gpm capacity. In January 1984, TO\
was found in the raw water. The water quality history is summarized below:
            Date                             Contamination
                                        TCA     DCE     PCE
            I/S4                       __ 20 ppb
            7/84                     '"X58ppb     3 ppb
            10/84                         86 ppb     5 ppb
            2/S5                         32 ppb
            2/86                         20 ppb
            2/87                          3 ppb
            2/88                          Ippb
            2/89                                                3.2 ppb

 As shown, the use of the well in conjunction with GAC has led to the cleanup of
 the contamination, although new contamination (PCE) may be appearing.

 One 20,000 Ib. unit was installed in July 1984 and by September 1984, breakthrough
 occurred. The 50 ppb guideline was in effect in 1984 and the unit continued to
 operate until exhaustion in January 1985. A total of 62 MG (11,588 BV) was treated
 (see Table 6). The carbon was replaced and because the water quality was below 50
 ppb, the filter was by-passed until January of 1989 \vhen  the 5 ppb guideline required
 that treatment again be provided. A total of 38 MG has been treated thus far in
 1989-  At this point, each well on the site is now being connected to a separate 2 -
 20,000 Ib. unit.
                                        390

-------
SAMMUEL STREET

The well site has four wells. The two shallowest wells have been removed from
service for over ten years due to VOC contamination. The two remaining wells are
1,200 gpm (349 feet) and 1,300 gpm (373 feet) capacity. All four wells are screened
in the same aquifer.

The contamination took almost ten years to migrate to the deeper wells with trace
of TCA appearing in October 19S7. By July 1988, the TCA was 23 ppb and two
additional VOC's (PCE and dichloretbane-DCA) were present. GAC treatment was
ordered and 2 - 20,000 Ib. units were installed in August 198S to treat 1300 gpm
from well #4. At the time, the TCA had increased to 51 ppb and four additional
VOC's present (PCE, DCA, TCE snd dichlorethylene- DCE).

The  first breakthrough occurred on November 22,1988, after treating 64,7 MG (6,046
BV). TCA was now 113 ppb and the four other compounds had shown some increases
(Table?).      .         '  '              .       .     '

The  second carbon breakthrough occurred on January 22,1989; however, in early
December 1988, benzene was found (initial 11 ppb)  in the raw water. At    *
breakthrough, trace amounts of toluene appeared. New carbon was installed on
February 16, I9S9, after treating only 45.4" MG (4,243 BV):  the 5 ppb guideline
accounts for the reduced capacity. Concentrations of TCA had dropped to 70 ppb;
the four additional VOC's were stiH present in small quantities; benzene was up to
47 ppb; toluene was 3 ppb; and xylene (7 ppb) was found.

DCA has shown to be the least adsorbabte compound even though concentrations of
DCA have not exceeded 10 ppb. Adsorbability is a more limiting factor to carbon
performance than contaminant concentration, especially when 5 ppb guideline must be
followed.

For the third carbon rebed, the columns were operated in series.  This would allow
for the more efficient rebedding of carbon while still allowing continued use of the
facility. Another change included using the other operating well (#3) alternately with
well #4 because it has less contamination (no benzene, lower TCA). As of March 31,
 1989, all eight compounds were still present with TCA backup to  101 ppb and a
brand new contaminant (not previously found on Long Island) identified. Methy!-
 tertiaiy-butyl-ether (MTBE) has been found at levels of 25 ppb. This gasoline
additive is not totally removed by the carbon (even with series flow - a 9 m:n.
                                       391

                                     •V

-------
EBCT). Although the state has initially set a 50 ppb guideline for MTBE, this final
problem could have serious operating consequences for well #4. All the water quality
is summarized on Table ?.                                     .-.''..

Finally, a second set of filters (2 - 20,000 Ib.) has been installed at the site for well
The chronology for well #4 has been a nightmare for the purveyor in terms of water
quality changes and trying to properly monitor the system operation. Because the
well field is in a critical area and the water is needed, daily sampling of raw and
treated water has been instituted,

Table 6 shows the throughput for wel #4 for the first two beds of carbon along
with Bridgehampton Road results.  The effect of the 5 ppb can be seen in the
reduction of treated water at Sammuel Street (30 percent). Bridgehampton Road does
have considerably more throughput but TCA was the only contaminant, 50 ppb was
the guideline, and no DCA was present to limit operation.

SYSTEM COMPARISONS

Two types of designs are in operation on Long" Hand, those provided by Calgon
Carbon Corp. and the others.  The Calgon system contains more valving (series or
parallel flow) and separate carbon loading and unloading piping. Its first cost is
naturally greater but the additional mechanical features can save time and money
with filter operation. As an example, when purchasing carbon, should Calgon be the
chosen vendor, and on Long Island they seem to be the only vendor, the EUing and
empty ing of the units is a simple set of hose connections. For the homemade
(economy design) special fittings for water and air (to transfer carbon) need to be
assembled.

Another problem has occurred with the discharge collection systems. Each of the
four economy models has experienced a collector Failure and subsequent loss of
carbon. Thus far the Calgon units  have not had any problems.

OPERATING CONCERNS

When a GAC is designed, aside from the removal capabilities, the piping, valves,
metering, etc., are important considerations. The following discussion will highlight
operating concerns that should be  reviewed and instituted for GAC treatment.
                                        392                        .

-------
Virgin carbon requires wetting, not only to transfer the material properly but also to
remove the air that is trapped in all the pores. In fact, it is recommended that the
carbon sit in a filled column for a least a day to allow complete saturation.

Before use, new carbon needs backwashing to remove the fines that result from the
abrasiveness of the material. BacfcwasMng should use non-coctaminated system water
and should be completed before the system begins operation. The non-contaminated
water allows the carbon to remain virgin before actual on-line treatment.

Maintaining equilibrium of the adsorption isotherm as it progresses through the media
is extremely important. Disturbing this equilibrium could result in dramatic change
in effluent and total throughput. Even exposing the carbon, once in an operation, to
backwashing with clean water, could disturb the equilibrium. Backwashing, therefore,
should be kept to a minimum.

Before entering the distribution system, GAC effluent should fee sampled for bacteria.
This precaution is two-fold; during construction and loading of carbon, contamination
could enter the system," and the carbon itself is an excellent media for bacteria
growth. Bacteria sampling should also take place if the unit h left out of operation
for several days or longer. Disinfection of the columns should be performed before
the carbon is loaded. Chlorine will dramatically reduce the capacity of carbon.

Finally, sampling of the raw (for fluctuations) and treated (for breakthrough,
exhaustion, etc.) water must be scheduled regularly. The frequency of sampling will
depend on the operating records that accumulate with time, but minimally during the
first run, weekly sampling should occur when you reach 50 percent of the design life
(based on isotherm or published data).

CONCLUSION

GAC technology will be with us for many years. It works but it can't be taken for
granted.  Each treatment system will require Its own set of operating guidelines and
maintenance requirements. For Suffolk County, where 30 - 40 new GAC units are
anticipated in 1989, the field experiences are only beginning. Within 2-3 years, a
wealth of operating data on a variety of contaminants, both singly and in
combination, will develop. This data will be invaluable to the water companies and
should be helpful to suppliers and designers throughout the country.
                                         393

-------
ACKNOWLEDGEMENT

Cooperation and assistance were greatly appreciated from the following:

     Ben Bletsch, superintendent, Da Hills Water District
     Dennis Kelleher, engineer, H2M Corporation
     Paul Kuzuman, hydrologist, Suffolk County Water Authority
     Tom Martin, Carol-Lee Goff and Pat DeLaRosa, Suffolk County Department of
     Health Services
                                       394
                                       ia

-------
                                COMMENTS
                 Field Experiences; VOC and Pesticide Removal
                    Using GAC - Suffolk County, New York
                                Joseph Baier
Dave Paris
     Have you examined the pressure contactors on Long Island for radiation or your
     water supply for radon contamination?

Baier

     Radon is one of the few contaminants we don't have so it isn't a problem we
     have to address.

Jack Dice

     Have any epideraiological studies been done on your customers?

Baier

     New York State has its own epidemiologist and lexicologist and they believe
     that they can set MCLs as well as, if not better, than the U.S. Environmental
     Protection Agency. For example, New York's 7-ppb guideline for aldicarb was
     considered outrageous by the water industry as well as the chemical company
     producing it. The MCL now being considered by the USEPA is 9 ppb, so New
     York feels vindicated about its original assessment. As more research is done
     on these contaminants, the lower the allowable numbers get. Right now, our
     analytical capabilities are forcingjjs to keep the allowable levels of some
     contaminants higher than we'd like. I'm sure the capabilities will come and,   ,
     with these advancements, the allowable levels will decrease.

John Dyksen

     With all the carbon units you mentioned that will be going on line in Suffolk
     County, has regeneration been considered as a cost-reduction measure?

Baier                               -               •

     I don't think it's been considered yet because the primary goal at this point is
     to get everything operational. I do think it's a real possibility in the future.
     But first we have to have the operational data before that kind of decision can
     be made.

Donald Dobey

     I was surprised at the resistance to treatment and the distance traveled by the
     contaminants you're dealing with. Have you found them in any deep wells?
 Baier
      We've found contaminants as deep as 500 feet. I would not be surprised if we
      were to find them even deeper than that.
                                       -395

-------
Table I. Grccnport - Gallons Treated (MG)
Carbon
Loading
1
2
3
4
5
6
Average


Date
6/85
10/85
1/86
5/86
•'V
* Results in ppb

Carbon Unit
1
2
3
4
5

Breakthrough Exhaustion (7ppb)
84 129
92 . 124
104 119
81 101
96 106
76 MG through 3/89
91 116
Table 2. Raw Water Quality Data*
Aldicarb Aldicarb
SulTone SulToxide DCP
48 35 ffr 22
22 17 21
17 14 20
14 12 18

Table 3. 1,2 Dichloropropane - Throughput
Bed Volumes

Aldicarb (ppb)
. 12-20
19-26
14-20
11-14
6-14




Carbofarsn
13
8
6
5


5 ppb Influent Cone. Increased Capacity
12,650 21,219
19,001 53,605
21,656 57,211
22,257 49,837
22,760
1.85
Z82
2.64
2.24
—
                     396

-------
               Table 4. AldicaA SuVbne -Throughput
                          Bed Volumes
Carbon Unit
1
2
3
4
5
Sppb
12,650
28,621
32,262
37,608
34,483
Influent Cone.
21,219
60,075
67,898
. 61,201
_
Increased Capacity
1.67
2.09
2.10
1.62
— .
               Table 5. Aldicarb Sulfoxide - Throughput
                          Bed Volumes
Carbon Unit
Influent Cone.    Increased Capacity
1
2
3
4
. 6
23,766
30,512
48,710
42,635
39,228
22,322
60,075
67,898
61,201
_
1.62
1.96
L39
1.43
_
                                 397 '

                                 IS"

-------
            Table 6. Summary of GAC Removal
Location
Dix Hills (Design)
(Actual)
Southold
Woodchuck Hollow
Greenport
Lone Springs
Samuel St. #lrebed
#2 rebed
Bridgehampton Rd.
BVto
Exhaustion
17,000
6,666
16,000 (avg.)
19,600 +
21,500 (avg.)
20,560
6,046
4,243
11,588
EBCT
Min.
12
15
7
12
12
16

15
Contaminant
36 (DCP)
20 (DCP -f aldicarb)
7(DCP'-s-TCA,TCE)
15 (aldicarfa 4- carbofuran)
10 (aldicarb +• carbofuran}
80 fTCA * 4 others)
120 (TCA + 4 others + BTX)
86 (TCA)
Exhaustion of bed occurs when effluent equals inSuent from the least
adsorbable compound.
            Table 7. Quality Grange - 18 Months
                      Samuel Street
             First Carton
Second Carbon
Third Carbon
TCA
•^PGE
DCA
DCE
TCE
Benzene
Toluene
Xylene
MTBE
51-113 147-75 69-101
^4—5 -7--12 5-9
5-10 6-8-3 7-
-------