NCD980602163_19971001_Warren County PCB Landfill_SERB C_Comparative Evaluation of the Two Phase I Reports on Detoxification Technology Testing of Materials from the Landfill-OCR

P'C B l.dlJRK I t'-JG GFJJUP Fax :[phone] JJct 1 '97 15 :02 P·. 01
JOINT WARREN COUNTY/STATE PCB LANDFILL
WORKING GROUP
CO-CHAIRS:
Dul.I.IE B. llllfl.ln!U.
HEN/IY l,,4/t/CASJ'IIJI
7 20 RIOOEWA.Y STREl'rr
WA.R~ENTON, H .C , 2758Q
PHOM!;: 9 I 9-257-1 Q,48 -FAX [phone]
FAX COVER SHEET
TO:
FROM:
DATE:
Mike Kelly
Doris Fleetwood, Secretary
PCB Working Group
October 1, 1997
Number of pages (including cover sheet): 13
919-7)5-3605
PCB ~JRKING GROUP F,:i x :'319-25 7-1000 Oct 1 ''37 15:02 P.02
COMPARATIVE EVALUATION OF THE TWO PHASE I
REPORTS ON DETOXIFICATION TECHNOLOGY TESTING
OF MATERIALS FROM THE WARREN COUNTY, NORTH
CAROLINA PCB LANDFILL
Prepared by:
Joel S. Hirschhom, Ph.D., Hirschhorn & Associates
in associati.on V\.1.th
Patrick A. Bames, P .G., BFA fnvironmrntal, Inc.
October 1, 1997
1. 0 1 ntroduction
In response to a Request for Proposals to evaluate treatment technologies OJl soils
extracted &om the PCB Landfill. Warren County_, North Carolin.a, issued by the Division of Waste
Ma.b.agcmeot of DENR on January 31, 1997, mo propows were ,elected for funding. The Rf P
had d early identifii!!d two detoxification technologies as meeting the requireme:ots established by
the Joint State.iWarrcn Cowrty Worbng Group ; these were Base Catalyzed Decluorinati()n (BCD)
and Gas Phase Chemical Reductio.n. The llFP notifii!d potential proposers that. the project w as
divided into two phases and that multiple companies might be chosen for Phase I, but that only
oue cornpmy would be awarded a contract for Phase II.
Section 2. l.4 of the RFP specified the following selection c:1iteria for choosing one
(.;O!Ilp&n}' for the Phase n part of the RFP:
a. The ability to meet Phase I performance criteria as demonstrated through Phase I
test data. The performance criteria were pre~nt~d in Sect.ion. 2. ~ of the RFP. For post-
tttatme.nt solids the :Principlll Preliminary Remediation Goals w~e 20 ppb for total PCBs
and l ppt for Dioxin TEQ. The Air Enriwon Performance Goals were 8xl 0-1 micrograms
per cubic meter (ug/cm.) for PCBs and ~xio•R ug/cm for DioX'in TEO.
b. The quality of th.e Phase I te~ repon .
c. The ability of the vendor tt) provide full-icale equipment at the Warren Couuty
PCB Landfill site.
-.. __ _
PCB WJRKI~3 GROUP Fax:[phone] Oct 1 '97 15:03 F·. 03
d. The ability of the vendo.r to provide a safe, reliable and co~•effective full ~ak
application oftbe selected technology at the Warren Comity Landfill. Section 2.2 of the
RFP further clarified the goals and objectives of the testing, particularly in lenN of
determining fea.sibility for full-scale deto,dtication and th.at. in addition to the meeting the
perfonnaoce criteria, feasibility would 1.lso be determmed by cousiderwg the §afety of the
technology, the rate of detoxification, the cost per unit of deto·xification> the reduction in
long term potential for =nvirorunental rcleueg from rem.duals of the treatment proce~s
e. The quality and cost of the proposal for providu:lg Phase 11 services.
This report by the two Scimee Advi~rs for the WoTkm.g Group is the 1;0.mpantive
evaluation of the two Phase I reports mbmitted in this project and is provided to the Working
Group UJd the Di-vi.don of Waste Managemeot for the pwpo§e of assisting with the selection of
the Pluse ll coutractoi: All repons and responses to questions mbmitted by the two complUlies,
ETG Environmental, In.c. (ETG) and ELI Eco Logic International Inc. (EL)~ have been
considered in preparing this report, The format is to present a discussion of the relative pros and
cons. or ,dvantages and disadvantag~,, of each company's proposed technology for each of the
above selection crit.eria. Finally, a swnroary comparuon and recommendation is presented in the
last section..
2.0 l\leetine tht performance criteria
2. I Post•treatment. solids. tota.l PCBs
ETG conducted four test rwu and in all four cases the total PCBs were reduced w levels
Jess than 20 ppb. Ihe average of the four runs was 0. 8 ppb. A consideration is the variation in
PCB levels in tbe input (raw feedstock) mate.rials, which in thi~ Ca$t averaged .508 pplll. Bec.tusc-
nondetects (NDs) arc: often repo:rted, detection limits are also a factor in evaluatmg results In
cowparins two the companies the iMUe th.tt arises is ifNDs are reported but the detection limits
(DLs) arc different, t.hl.'lO the NDs are not exactly the ~me. With higher DLs the NDs a.re less
imptesiive, because the. poteiitial level ofun.detected PCBs is higher. A40T.her way of looking at
this issue is to realize that with lower DL& it is possil>lc to b.ave positive hits or findings of PCB
isomer s. while if higher OLs were used, the.n those fin.din.as would not be pre~t
EL condu~ted three test ruu and reported NDs for all three result&; however, the DU
were significantly h.igher than in the ETG testing, primarily because smaller size samples wer0
used in the EL testing (i.e., 0.010 kg versus 0.030 kg for ETG). The only scienti1ic way to bot:1ter
compare the two sets oftest data, therefore, :i& to recalculate the total PCB levels by \Hing the
worst p ossible case in which it is assumed that the NDs are actually equal tc the DLs. This has
been done IUld the recalculated levels for both companies' data are given in the follo,,,jng t•ble by
using the data given in the re...qpcctive reports. It can be seen that by following thl.~ protedure. that
the r~alculated average fot £TO increases to l ,84 ppb and for EL the average changes from
2
PCB WORK ING GROUP Fax:[phone] Oct 1 '97 15:03 F'. 04
what might be interpreted as zero to J.63 ppb. Thu;, 'Mlile it is com~ct to say that both
companies were able to meet the performance criterion, it is also c<Xrect to conclude th~t ETG
performed better than E~ particularly because ETG conducted four runs while EL presented data.
for three nms. Additionally, the average PCB level in the raw materials tested was 23 '7 ppm for
EL but 385 ppm for ETG, qch makes the ETG results even more significant, because a higher
fraction ofPCBs were removed.
Company/sample Original total PCBs (ppb) Recalculated total PCBs (pph)
ETG WCl-3 _74 1.33 ·-
ETGWC1•4 2 .. SS 3.06
ETGWC2-3 O(ND) l.80
ETGWC2-4 0 (ND) l.18
ETG average 0.8 1.84
EL 1 O(ND) 4.10
EL2 O(ND) 3.40
EL3 O(ND) 3.40
EL average 0(ND) 3.63
2.2 Post-treatment solids-dioxin TEQ
ETG reported data for four t\UlS, with an average dioxin TEQ of 0. 9 1 ppt, however one of
the runs had a value of 2. 96 ppt, but this was for th.e sample with the highest level of PCBs. fo
fact, ETG had optimized its process based on a much lower level of chemical treatment (BCD)
additive, which would explain why in this one run both the PCB and dioxin TEQ l~vels were the
highest in the residual treated solids. In fact, the TEQs for the other three runs were exceptionally
low, with an average of about .23 ppt, which is msignifieant. It $bould a.lso be nokd thar even the
2,96 ppt level is very low and that EPA and most states have not approached this k'\'d fo1· djo:-in
cleanups. For ex.atnplt, for the Koppers Supemmd site clem11p in Morrisville, Nort.h c~robnii th~
dioxill cleanup level was 7,000 ppt.
EL reported data for three runs. However, the dioxin an.aly~~ were redoue for nms l and
3, but not run 2~ because of overly high d~ct:ion li-orits in tbe original testing. Also, some dioxins
were originally found in run 2 material, dei.pite high detecti.on levels, because ofbjgb. dioxin levels
that resuhed from run 2 represa,.~ 2, proces.s failure due to too low a teDlJ)erature .in tbe TR.1\1
desorption unit that did not allow removal of dioxins for chemical reduction in the second high
3
F'CE i_tJIJFK it·.1(3 f3F;IJUF· Fax :·919-257-1000 Clct 1 '97 15 :04 F·. C6
temperature stage. This is discussed in length later in this report. The original TEQ values for
runs 1 and 3, based on using dM:ection. limit values was reported by EL as 1.5 ppt and. 32 ppt,
respectively. If the value for nm 2 is inclu.ded, then the avCTage for al] three runs in the original
data is very high at 142 ppt, and with the data from the retesting it is 127 ppt. The data are
summarized in the follo,ving table.
It is shown bow comparable data can be used from both compani~ in terms of either data
from all rung or ouly th~ best runs. and for EL for original and revised data. For ex:an:iple, t11e
average for the two best EL nuu: cu be compared to the average of the three runs of ETG that
represented their besi process perfonnan.ce. In other words, if an test data are considered for
both companies, tha>. EL clearly fails to meet the penormance criterion, and if only the best run 3,
are con$idered for both companies, then EL meets the criterion but ETG has superior
performance .. From a community perspective, it is valid to judge the companies on the basis of ali
their data, because they are re-spo:nS1'ble for ~boptimal nms.
Average ~oxin TEO Leveli (-ppt)
TEST RUNS ETG ECOLOGIC ECOLOGIC
origiD.al d.tta re-vised data
---
Allnms _91 142 127
Best runs .23 24 .50
It should be noted that the average dioxin TEQ level in the raw materials for E.L runs 1
and .3 (the best runs) was 186 ppt, and 17~ ppt for the three best runs ofETG. Proportionally,
ETG .had slightly better perfomwlce (ie.r, 99.89% ver~ 99_73% removal).
2.3 Air releases• total PCBs
ETG reported data for th.tee 11UlS in terms of both stack discharges and modeling r.:sults
for a property line assumed to be 200 feet from the equip.men.t, which is a proper -procedure for
&ddressing an exposure eoncentration ( and that had been deemed appropriate at the pre~bid
conference for the RFP). ht fa.ct, the performance criteria bad been estab.lidied on the basis of
exposure concentrations for a very low risk level The average of the ETG data was .87x 10-•
ug/cm, or about one-tenth the performance criterion of 8xl0 ... ug/cro.
EL reported data for three runs in terms of stack gas concc..otrations only~ for 'Nbich the
average was l.26xlo·• ug/cm, which is below the perfonnance criterion at the ~ack, and \.\hlcb
autoinatically makes it below the crite.rion at any distance to an exposure point
The ETG data for the stac.k coneentutions wtte significantly higher th.tn for the EL data.
However, air sampling methods and·equipment were not identical and, therefore, a direct
4
PCB WORK ING G~JUP Fax:[phone] Oct 1 '97 15:04 P.06
co:mpariso.n of stack data alone is not necessarily complete. Th.e main problem .is that an air
sampling procedW"e that i£ mo.re efficient and eff'ect{'.'~ :in removing material and obtaining low
detection kvels will have t higher probability of detecting contamina11ts. for exan1ple, sampl:ing a
larger v olume of gas \\-ill increase the p,-obtbility of detecting contaminants. f 01 example, the
stack gas Bow rate in the ETG tesu were about thre.e times larger than in the EL testin~
suggesting a hjgher scn$itivity in the ETG air testing. It must be emphasimd that the p~rfonnance
level set for this teQtmg was e:,,,.-tremely low and stringent and th.at the data from both cowpan1es
indicates that no health haurd would be caused by .PCB air emissions fr\lm the daoxi.fica1ion
process. Indeed, in all probability there may be greater concern about potential PCB air releases
from site excavation and material handling prior to trutment .in equipment, but this issue should
be addressed in the Phase Il work.
Baged on available infonnation it can be Mid that both companies are comparable 011 this
sub•criterion.
2.4 Air releases -dioxin TEQ
ETG reported stack 111d model data for thlee run.$. The model data had an average of
.52xl04' ug/cm. about a tenth of the perfomw,ce criterion.
EL reported that no dioxiru.lfurans were found above detection limits, but that the highest
possible level was 1. 7x:to·~ ug/cm, which i, higher tlan the petformmce criterion. Therefore, El
said that "Due to these sampling and analytical constraints, it is not possibJe to demonstrate the
ability of the ECO LOGIC Proceas to meet the dioxin TBQ petfonunce goal ... '' The highest
possible EL level is some t 000 tim:s greater than the perfonnance goal. However if EL would
have modeled its data to address levels at some rea&0.n.able exposure point. then it would have
beon able to show compliance, because its &tack level was mniw to that found by ETG.
The Science Advisors conclude that ETG has the advanuse on this sub-criterion.
3.0 Quality of the Phase I report
Making a professional judgment about the quality of a technical report can involve man y
coosidcntions Cenainly, tbc reportg abould preaent the information required in the RFP in a
user-fii.endly format . Additionllly, however. the quality of the Tq}Qrt can be ~ocsiclered to be
higher when there is considetab.Je attention to particuLuly important js...\"U.e, and when additio.nal
in!ormatioo is pxovided to usiu the UJ1dersta.nding and. ev.tlu&tion of the testing. It muSt also be
noted that EL submitted a draft report "I.Vb.ich w•s then modified in response tC"> vmons questions
illld comments to produce I final report, while ETG submitted only one initial final repon
Therefore~ to some extent the oo.e submission by ETG is comparable to th.e second submission bv
EL that corroctod deficiencies in. the origins! draft report. · -
5
·---··
F'CB l.1.ICIF:f<ING 13FJJUF· Fax :[phone] Clct 1 '97 1:,: 0:, F'.07
As an important exaq,le of• difference in presentation quality, consider the Tables 4-5
and 4-6 in the ETG report, w.bioh presented in easily understood tabular form the primaty data ou
treatment effectiveness for PCBs aad dio:c:ins/.turans for all the test runs, and for all the individual
isomers or congener, in each category. A reader could immediately s.ee how the post-treatment
levels ooq,ared to tho origmal raw material levels, and \"W.en. nondects wc:n:: r'1)otted tbe
detection limits were also prcac:.oted. In contrast. the EL n;port pre$Cnted the analogous data in
two separate tables (Tables 8 a.u.d 9), and the contractor had to be: instructed to provide detection
limils in the final report, and neither table preseuted the data tor all the indivi.dual isomeTs or
congeners in the U8eful way employed by ETG.
Another factor was that ETG pre&aDted more information than was strictly requited. For
exa.rnple, ETG presented the: results of testing to diow that their process bad actually dts1roy~d
PCBs and mo obtained data on particulate levels in emissions and water con.tent in raw feed
materials.
It tn11st also be noted that the quality of the report is merely a surrogate for the quality of
the bench•scale testing. In this sense, it it important to note th.at. EL conducted three nms versus
four for ETG, and that one of the EL runs (nm 2) was essentially a failw-e because of an operato,.
error t.h:tt caused the initial TitM desorption pan of the proces..c; to :function poorly due to a low
temperature that did not separate dioxins/furans for chemical reduction in. the secund part of the
process. A major aspect of any type of tedtn.ology testing is to see how well a company conduct s
itself and operates its ovvn equip:inmt to demonstrate a very high level of competence that
provides confiden~~ to potetJtial clients that the very best performance will be obtained in foll-
scale usage. The main QllF@QP thl1.11imts attention for this upect oftbe EL..1.~.stiA1Js .. ~b~..1M
senior perwns in £haue of the test did notldcnti& the low lmmm:atutkJW>blcm durini t~ .. J~st
period and repeat the test run 'J'hig wu especially important because only three nws were used
Also, in discussing this event EL has referred to a minimum temperature in the TRM tJDit
of 5 501)C. but the unit is supposed to operate at 600°C according 1.0 tht, information in th~ EL
report. EL did not provide detailed d11ta for run 2 on exactly whtt t~ennure was used or for
how long a suboptimal temperature; existed during the: run. except that the temperature dropped
to 50O"C. However, it reported average mill tmiperatures of 595°C, 572°C, and 628"C for test
nms 1, 2, and 3 respectively_ It is interesting that two of the runs did not have average
ten,,peratmes of600"C. The &et that the average tempetatllre in run 2 w~ above the 55U''C
minimum that BL has mentioned also raises a. question of uncertainty a.bout wh.at letually
ocCUlTed in nm 2 to so thoroughly cause no tJ"eatment of the dioxin contaminatioIJ. The
infonnation suggests that out of the total of 15 minw:es in the TRM unit eve:n a miall amount of
time at suboptimal temperature is sufficien.t to c•use a complete &ilure to retnove ruo>l'iu
contaminants ftom'the feed material, wh.ich is a serious sensitivity to low temperature e.xcuTsions.
EL also had a problem with an important part of its dioxin testing, which it blamed on the
t.estin.g laboratory. How~er. much of what has b"n said. by EL simply does not hold up uude!
close SQTUtmy. Part of the competence of a co~any in the rcmediatio.n technology business is
6
F'CB l.1.ICIF:f< I 1·¥3 13RCIUF' Fax:[phone] Clct 1 '97 1::, :06 F'.08
----expertise for obtaining the best analytical services. If a company provides incomplete or
misleading information to a chemical analysis laboratory, then the laboratory may perform poorty
In the EL testing, Qome of the most critical dioxin testing was inferior because of high detection
limits that resulted from unusually small sample sizes for the testing_ EL clai.--ns that this resuhed
from its telling the laboratory that there were high PCB levels, and that this caused the laboratory
to use lower than normal sample quantiti~-There are two problems, however_ First. rhe overall
PCB levels in the raw nuterws (information giv~n m the RFP) were not exceptionally h:igb for
PCBs_ Second, certainly there would be no rational reason why EL would want the laboratory to
believe that PCB levels were liit,h in their t.t:~ materws. EL also said something that any
experien.ced professional who cleats with dioxin testing of solids knows is incorrect, namely that a
target v.alue of I ppt dioxin TEQ is impossible to obtain or verify for mdhod 8290. lbis is pbin
wrong ; as 1ny ex.a_mmation oftest results for this method clearly mows. mclu~ the dioxin test
data obtained for the project's site investigation work. as well u the diom testing for ETG ftom
the saroe laboratory ui;ed by EL. J:n the opinion of one of th.e Science Advisors, who does
considerable work with dioxin contamination ind cleanup llJld who discussed this issue with both
companies, EL did not exhi.1>.it a high degree of knowledge and experience in tbjs area. ETG
verified tha.t at no time did it change its instructions to the analytical laboratory that both
CQwpanies used, nor is their any logical basis for believing that actio1Js talc.en for EL samples
would have vfeci:ed ETG work in the veey large laboratory. The oo c key &ct fut seems to
explain the cause of the dioxin testing problems faced by EL was the infonnatiou it provid~d t(l
the laboratory ~mch caused them to use irregular, low sample volumes for testing.
Aaothet issue is that EL did not clearly represent what mat ~rials it tested n;lalP-e to the
materials it had received. EL said that it received three S•gallon buckets but tha.t only two were
used_ But no information was provided to indicate enctly what materi.tls were test.ed in EL'"
three runs. In contrast, ETG clearly indictted 'Wh.lt materials had been received and teSted in ea ch
of their four run,_ ·
The professional judgment of the Science Advisors is that the ETG repott followed the
requ.est.ed format of the RFP more closely than the EL report, and that the ETG report was m
iener-al easier to fullow and presented key information in a more concise manner than the EL
report As the above discussion. indicetes, the co11.clusion is that the quality of the ETG report and
its testing '"'ls better titan the EL re.port a.n.d iu te~g.
4.0 Ability to provide full-scale equipment
Both compani06 uo iu the position of roost technology vendors, n.l1lle]y that e.ither au
exu:ting piece of equipment could be tramported to th.e site and usod, or o.cw equipment wou.ld be
co:nstructed for the particular job. Both c-0mpanies have indicated that they would need up to six
months to provide neces...4WY equipment, 'Mtich is consistent with industry practice for large
cleanup projects. But there is more to consider for this evaluation criterion. The tbility to
provide full-.9C.ale equipment is legitimately related to the stability and viability of the co0Jpa11y.
7
F'C B l.1JIJFK I t·.1(3 13F:DUF' Fax :'91'9-257-1000 Oct 1 ''97 15:06
because some &ignificmt investment is necessary for a large project, which this one would be
potentially. ETG is actually part of a business that is weD established commercially and qu.ite
substantial financially, and it has performed in similar remediation applications for lJ. S clients
over some years.
Interestingly, in the BCD treatment technology arena, which .is based on companies
obtainin.g licenses from EPA, ETO has outlasted other companies. In contrast, EL is more of a
start~up technology developer that has had few clients (and none in. the U. S for actual full scale
work) and is not the same level of a stable~ financially successful company with a long commercial
track record as ETG. There may also be some signmcmoe to the &ct that ETG is a US.
company located much closer to North CarolinA., while EL is a more distant Cau.adian company
It is also of some relevance that a very recent study for a major federal Superfund sit~ at which
dioxin is a major contaminant (the Escambia Treating Company site in Pensacola, Horida )
examined potential treatment technologies, and the EPA contractor doing the Remedial
lnvestigation/FeaStoility Study screened out EL and its technology at the earliest stage of
technology evaluation. but included BCD treatment and cited ETG as the source of the BCD
technology. While EL has performed treatability and demonstration tests in the U.S. it has not ye1
secured any actual remediation project. In contrast, ETG is the vendor cleaning up the South~m
Maryland Wood Treatment Supetfund &ite, where it will treat some 145,000 tons of ma1eriaJ
This job .is cl.oscr to the WaJTm County project than any work that EL has perfonned
Tue Science Advisor& conclude that oveull the ability to pro"ide .necessary :fiill-scale
eqwpment for the Warren County project is better for ETG than for EL.
S.O Safety, reliability, cost, and long term release~
5.1 Safety
There are two major ways to look at the safety issue. First, tht intrinsic nature of the.
tec,hnology can be assessed in tem1$ of inherent hazards. For example~ 111y process that utilizes
haz.ardous material$ is fundamentally .more hazardous than one that does not use any hazardou~
materialr, and one that operates at either higher pre11.mre or temperature than anothe1 '.)ne po~es
more potential for unsafe situations.
The second approach is to consider how vulnerable a process is to human error, because
moil .im1aie conditions and accideuts relate to human error. SC'me technologies are inherently
safer bocau,e they are much simpler than other technologies., for example. As for treatability 01
demonstration testing_. it is presumed that a company will logically use its best trained and
qualified personnel If human errori occur in testing, it does not portend well for full-~,;ale
company operations.
ETG technology operate§ at a lower temperature tlun the EL process, roughly about
8
F'CB l.1JIJRf(Il•,j13 13F:ciUF· Fax:[phone] Oct 1 '97 15:07 F·. 10
onehalflower temperature. The ETG process also operates under vacu\llll conditions, while the
EL process operates at tmbimt pressure, and this reduces the potential for ~stem leaks and
discharges of contaminmts into the environment. The EL technology uses hydrogen gas. which is
intrinsically hu:ardouS, while the ETG technoJogy uses no hazardous materials. ¼1th the exc•~tion
of sodium hydroxide which is ~o:rrosivc but not combustibJ.o or toxic. The EL process also
utilize$ a bath of molten cln. whil;ih raises a safety concern.
Although EL has made a good argument th.at industrial processes have use.<J hyd.rogl!o
very succe.ufully, th.ere is still an mhercmt hazard that cannot he entirely dismiiSed, especially
when one rec.oguizes that safety issues generally are centered on unuiUaL unintended and
unfoneen incidents or accidents. While process control instrumentation and monitoring
equipment may be used, tb.e problem. in the reaJ..world are low probability events that consist of
several things going Yn"Ong to create an actual problem. For example, fOT everyone of the nine
factors that EL has cited to assure the saf~ty of using hydr()gcn it is posmbJe to idc:utif)' a
condition that nullifies the &ctor. For example, EL says that no open flames or stuoking are
permitted on site, but any experienced professional has pTobably seen just such human beb.avior 1>0
a,'tUal industrial sties where the prohl'bitions exist. Similarly, leak proof process \.essels in the
field can be rendered unsafe by, for example, a bullet piercing a piece of onsite equipment. The
use ofnitrogea purging to avoid mixing v.itb oxygen can &il when someone us.es a '-VJ"O.o.g gas
tank or makes a wrong connection.
On the issue of human error, it is also relevant that in the bench-scale testing for th.is
project, th.ere was a clear, admitted case of human error in the EL test, when a temperature was
not properly controlled in run 2 in the TRM unit. EL also acknowledged a second "operator
e1Tor" when the e,ccess gas burner that was supposed to be operated at 800"C was st,::tu.afu.·
operated between 300 to 400C'C, apparently for 111 the run~. The company also bad problems with
equipment, including a micro--GC instrument that was $Upposed to be us.ed but was not
operation.al, and a broken mill shaft which happened twice and caused delays.
Another safety issue is the potential for the air pollution control system to function
effectively to prevent unacceptable discharges of hazardous substmces. In this regard. the ETG
report paid considerable attention to this issue and provided e>..1ensive discu~on of how the
comp1:11.y would employ the most sophisticated technology to collect and treat dust.
The Science Advisors conclude that, aSSUmi.ng that either company would employ the b~:.'l
industrial safety practice& in design and operation of their cquipml!mt, the ETG technblogy bas an
advantase over the EL process.
5. 2 Reliability
Reliability can mean many different things., but one key engineering concept is whether a
process technology will offer the best perf'orDWlce under varying field and raw material
9
PCB WORKING GROUP Fax:[phone] Oct 1 '97 1::, :08 P. 11
condition...c;._ In other words, some technologies ar~ more sensitive or vulnerable to variations in
key conditions or parameters and become Jess reliable because there ar~ upsets cre.ated. Some
technologies are very complex with multiple steps or stages and the need for very oomplex
process control and monitoring equipment. All oftb.ese issues can reduce reliability The use of
computers and sophisticated equipment .has not eliminated major problems in indu~rial processes.
An isgje of concern in the current project is the degree to which the detoxificatio.n technology
may become unreliable because of high water content in feed nuterials, for example. To some
degree it seems as if the EL process would he less •sitive to water contmt, but in fact a dost
examination of discussions of this top;c by the contractors rr:veals essentwly the same problem,
That is; both processei can handle higher moisture levels, but both would he negatively impacted
in terms of throughput, efficiency and proce~;ng time. Moreover. ETG actually meaSW"ed water
content and concluded that no dewatering of ma.teritls will be required_
An important feature of the ETG process is that there is solid phase BCD treatroc:.ot of
PCBs and dioxins in the thermal desolJ)tion part of the process and then more BCD liquid pba~
treatment of oily condensate, if necessary. Tum test results mowed that the i<>lid phas,!
deto.icifieation was successful. In contrast, in the EL proceM there is u.o detoxification of PCBs
and dioxins in the first stage thennal desorption unit, and detoxification is dependent on first
separating the contaminants into a vapor phase that then undergoes high temperature reduction.
As the testing showed~ the EL process was vulnerable to poor performance "'he11 th~ thetmal
desorption stage was not operated at optimal conditions and, therefore, dioxins were not
detoxified.
It is the opinio:o of the Science A<Msors that the ETG process is intrinsically s.impler illd
le~ prone to problems created during full~scale use that would redu.ce reliability
53 Cost
EL has said that it foresees a cost of $300 to $350 per ton, based on processing 1 oo to
200 tons per day with a 70 to 80% availability, leading to a project duration of90 to 123 weeks.
The availability range used by EL is not impressive and may indicate, its experience in actual
projects where there has been significant down time. This cost does NOT include cmain
activrties, such as excavation, porfonnallce testing or disposal of processed solids, but it DOES
include system mobilization and commissioning, all Wl.~e preparation and processing, lab co~s
for proces~ outputs analysis and system demobilization_
ETG has said that its estimate of $390 per ton pertains to a processing rate of 300 to 400
t~ns per day with l 00% availability ETG' s cost is all inclusive and includes exc.avatiou. and
b3ekfillmg of soil, which the EL estimate did not. ETG has also indicated that if less stringent
cleanup performance standards were used, then the cost could be decreased by some 35% ifmor~
typical criteria were used for PCBs and dioxins_ In fuct, it is highly probable that the state would
use its nounal c:le.anup critma and that the ETG cost would likely be less than $300 per to1.1 . Jl
F·CB l.1.\IJF:f< I l'-lG 13FDUF· Fax : [phone] \Jct 1 '97 1::, :08 F·. 12
should be n.oted that there is less intrinsic capability to reduce operating costs in the El. µroces»,
as comp ared to the ETG prooes~ and that EL did not offer a similar observation. Note that the
protessing rate for ETG is about twice as bjgh as the EL figure, meaning that total project
du.r1tion might we.ll be reduced by 50% and require about one year rather than two ye.a.rs.
The Science Advisors conclude that for cost and cost ~effecttveness the .ETG process is
superior t o that of EL
S .4 Long term releases from re&i.dll2k
Any detoxification technology will create certain t)"Pes of residuals.. byproduct$! M wast e
effluents that mi.pt pose longer term risks because of releases. Both. companie& havi~ _p.aid
sufficient attention to this issue and provided similar discussions about the safe and effective
band.ling of a.11 pr()(;ess re~-iduals and wastcstreams. and neither has an advantage. ·The more
confidence there is in the company's a.bility to reduce residual levels ofPCBs and dioxins to th!!
lowest possa"ble levels, then there is mirti:nul concern about long term releases from post-tre.atment
residuals replaced on the landfill site.
6.0 Quality and cost of Phase :U proposal
The initial proposals that responded to the RFP included a proposal for the Phase II ,,·ork
For the most part, both compwes submitt6d corttpauble proposals that addressed the
requirements of the RFP. Both compame5 teamed with large, experienced environmental
eugineering firms for the Phase n work. However, the cost of the ETG Phase TT work \\a~
siguificandy less at $89,000, while EL proposed a cost of $115,000.
The Science Advisors conclude that ETG off'er!i a significant cost advantsge for the Phll~~ n work, esp~ in light of oun-etitly limited fimding for all a~tcts of the curreat pro_ioct . Thin
is, the saving of $2~,000 offers the potential ofbeiu.g able to accomplish other fiu1ction!. :m d 11~eds
prior to the state legislature authommg funds for the full.detoxification of the landfill
7.0 Summary comparison and recommendation
Iha following table presents a simplified summary of the conclu$ions reached for th.a
individu.al evaluation criteri.a . Ra.th.er tltan using an artificial and subjective numerical rating
system, the Science· Advisors have indicated a net advantag~ in terms of a + sig.1J for the company
~ith the best capability. When both companies were comparable, both received a .....
11
PCB l.i.lCIF;k I NG 13F;[IUF' Fax :[phone] Clct 1 '97 15 :09 F'. 13
SELECTION CRITEIUON ETG ECOLOGIC
Pcrfonoancc
solids· PCB& +
solids • dioxin TEQ +
air-PCBs + +
air • dioxin TEQ +
Quality of Pbue I report +
Ability to prowie full-scale equipment +
Safety, reliability, cost. long term effects ++++ +
Phase Il proposal, quality and cost +
AB can be seen :from the table, in most categories ETG was judged to have an adva.ntage_
This is not to say or imply that :Eco Logic and its technology was inferior1 poor or completely
unacceptable, but only that, in this particular testing ud for this specific application➔ ETG has a
number of advantages, as presented in the previous discussions. On the basis ofviRiting the te.it
sites, having di9CUS8ions with company personnel and examining the reports. the Science
Advisors agreed that ETG seems a more engineering oriented company, while EL ~ms more
R&D or science oriented. with less of a practical, engineering constructio.n focu.s.
,.\lso, a close examination of all documents provided by EL indicated the followmg ( l ) the
company probably has not met the stringent dioxin clean.up criteria used in this project in previous
projecti; (2) the September 1994 EPA report on the thermal desorption unit found that it ''did not
perform to de$igu sp~cations" and that th.e company 0ex:perienced material handling problems;"
and (3) tho company does not seem to .have been able to perform commercial projects at a profit
tnd may be experiencing unetttain financial conditions_ On the latter point, the last two annual
reports from EL indicate th.at the company has COlltinu~ to be unprofitable, despite rising
revenues, whlJ1, the information supplied by ETG indicates a medium si7--e euvironmental services
company that is profitable. For example. the highest revenue stream for EL in 1996 wu Jess than
SS million annually (for which the annual Joss was nearly $12 million), while the company th at
ETG is a division of grosses about $SO million and is profitable.
In conclu1ion, the Science Advi1ors acne that ETG JS the best qualified compan}'
for receivin& the Pb.1e D cl)n1ract and adviae the Workinz Group to accept thjs !lelectioo
and to formally communicate it.I rttomme1ubtion to the Divi.!lion of Waste Ma.,_agement.
12