Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Hazardous Waste Engineering Research Laboratory
Cincinatti, Ohio 45268
Contract No. 68-03-3242
Work Assignment No. 3
EPA Project Officer EPA Work Assignment Manager
L. H. Garcia Harry M. Freeman
CASE STUDIES OF EXISTING TREATMENT
APPLIED TO HAZARDOUS WASTE
BANNED FROM LANDFILL
PHASE II
SUMMARY OF WASTE MINIMIZATION
GASEiSTUDY RESULTS
Final Report
October 1986
Prepared by
Thomas Nunno
Stephen Palmer
Mark Arienti
Marc Breton
ALLIANCE TECHNOLOGIES CORPORATION
(Formerly GCA Technology Division, Inc.)
213 Burlington Road
Bedford, Massachusetts 01730
-------
DISCLAIMER
This Final Report was furnished to the Environmental Protection Agency by
the Alliance Technologies Corporation, (formerly GCA Technology Division,
Inc.), Bedford, Massachusetts 01730, in partial fulfillment of Contract No.
68-03-3242, Work Assignment Nos. 1 and 3. The opinions, findings, and
conclusions expressed are those of the authors and not necessarily those of
the Environmental Protection Agency or the cooperating agencies. Mention of
company or product names is not to be considered as an endorsement by the
Environmental Protection Agency.
11
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and governmental concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled land
are tragic testimony to the deterioration of our natural environment. The
complexity of the environment and the interplay between its components require
a concentrated and integrated attack on the problems.
Research and development is the first necessary step in problem solution;
it involves defining the problem, measuring its impact, and searching for
solutions. The Hazardous Waste Engineering Research Laboratory develops new
and improved technology and systems to prevent, treat, and manage hazardous
waste pollutant discharges. This publication is one of the products of that
research.
This document presents information on waste minimization practices
currently employed in the printed circuit board (PCB) and semiconductor
manufacturing industries. Case studies conducted at six facilities evaluated
the technical, environmental and cost impacts associated with the
implementation of technologies for reducing the volume and toxicity of PCB
metals-containing sludges and solvent wastes. The analyses of these data are
the basis for demonstrating waste minimization technologies to reduce
hazardous waste.
111
-------
CONTENTS
Figures
Tables
Project Summary 1
Metal Plating Bath Waste Minimization Case Studies 2
Resist Developing Solvent Recovery Case Studies 14
Conclusions 19
1 Introduction 21
Background 21
Waste Minimization Case Study Selection 21
Report Organization 23
2. The Electronics Products Industry 24
Background 24
Waste Generation 24
Waste Management 31
3. Facility A Case Study 35
Facility Characterization 35
Process Testing and Results 40
Economic Evaluation 53
4. Facility B Case Study 55
Facility Characterization 55
Process Testing and Analytical Results 64
Economic and Environmental Evaluation 72
5. Facility C Case Study 76
Facility Characterization 76
Process Testing and Analytical Results 85
Economic and Environmental Evaluation 89
Environmental Evaluation 94
6. Facility D Case Study 96
Facility Characterization 96
Process Testing and Analytical Results 102
Economic and Environmental Evaluation Ill
7. Facility E Case Study 113
Facility Characterization 113
Process Testing and Analytical Results 121
Economic and Environmental Evaluation 130
8. Facility F Case Study 135
Facility Characterization 135
Process Testing and Analytical Results 140
Economic and Environmental Evaluation 148
-------
Contents (continued)
9. QA Summary 153
Introduction 153
Project Organization and Responsibility 153
Precision, Accuracy, Completeness, Representatives and
Comparability 153
Sampling Procedures 160
Corrective Action 161
Quality Assurance Reports 161
10. Conclusions and Recommendations 163
Electronic Industry Waste Management 163
Case Study Findings 164
Recommendations 165
References
-------
PROJECT SUMMARY
The purpose of this project was to investigate the effectiveness of
various waste minimization practices or technologies in the printed circuit
board and semiconductor manufacturing industries. The most significant waste
streams in these industries are waste halogenated solvents from photoresist
stripping and developing operations (RCRA Waste Code F001-F003), and
metal-bearing sludges (RCRA Waste Code F006) from the treatment of metal
plating and etching rinsewaters. This paper summarizes the findings of case
studies conducted at five printed circuit board manufacturing facilities and
one commercial treatment/recovery facility. Each facility investigated
employs some practice that requires offsite disposal. Two of the case studies
focus on the recovery of spent halogenated solvents, and the remaining four
discuss the recovery or reduction of metal plating and etching process
wastes. Table 1 summarizes characteristics of facilities investigated which
range from small job shops to large integrated facilities.
TABLE 1. SUMMARY OF FACILITIES TESTED UNDER WASTE
MINIMIZATION CASE STUDY PROGRAM
Facility nami
Facility A
i Description
Treetment storegc dispoeel
facility handling electro-
plating baths, uaeta
etchants. spills, etc.
Capacity: 1,000 gph
(24,000 gpd).
Technology
- Sodium hydroxide precipitation
- Sodium borohydnda reduction
- Alkaline chlorinatioa
Waates treated/reduced
- Nickle plating bathe )
- Copper placing bathe I Sludge
- Cyanide * '
product
Facility B Contract FCB manufacturing ahop.
Employees: 77
Production: 500,000 af/yr
Salia: STHM/yr
Facility C Computer manufacturer.
Employee!. 10,000
Facility 0 Electronic equipment nfgr.
FC board manufacturing uaing
the aubcractive technique in
the HacOermid proccaa.
Employeei. 260
Ficility E Computer manufacturer.
FC board manufacturing uiing
additive techniquee.
Employeei: 600
Production: 600,000 if/yr
Facility F FC Board manufacturer.
2-sided aingle layer circuit
boarda.
Production: 680,000 af/yr
Sodium borohydnde reduction
Hemtek ultrafiltration ayetem
- Solvent diatillation/
fractionation recovery of
reaiat developera.
- 2-atage aolvenc dietillation
(1) DuFonc RISTON SRS: 120
solvent recovery atlll
(2) Recyclene Froducta, Inc.
- Activated carbon regeneration
of apent plating batha.
Agmet Equipment Corp.
electrolytic recovery uniti.
Cupric chloride exchant
Electroleaa plating rinaea
Electroplating rinaea
Sludge product
Methyl chloroform reiilt developer
Freon renst developer
1,1,1-tnchloroechane reaiat developer
1,1,1-tnchloroethane still bottoms
Acid copper plating bath
Acid copper plating rinaewaters
Tin/lead plating rinaeweters
-------
METAL PLATING BATH WASTE MINIMIZATION CASE STUDIES
Metal plating wastes generated from plating bath dumps, rinses, etching
machines and scrubbing operations generate copper-, nickle-, tin-, and
lead-contaminated wastes. Four of the six case studies investigated under
this research project focus on the minimization of sludges generated primarily
by copper plating and etchant baths and copper and tin/lead rinsewaters.
The common objectives of each of the technologies evaluated are:
(1) minimization of metals sludges generated; (2) compliance with effluent
guidelines or local discharge limitations; and (3) reduction in operating
costs over other conventional alternatives. The following discussion briefly
summarizes each case study, the nature of the minimization technology, the
measurements data collected and the results obtained.
Facility A Case Study
Description—
Facility A is an offsite TSD facility which processes concentrated dumps
from the metal plating and printed circuit board industries, including
alkaline etchants, acid plating baths, nitric acid rack strip baths, and
electroless plating cyanide baths. The average metals concentration in the
incoming waste is reportedly 12 g/L (12,000 ppra). These waste streams are
classified into the following four categories: (1) acidic metals solutions;
(2) alkaline metals etchant solutions; (3) cyanides; and (4) chelated metals
solutions. The case study for this facility focuses on the use of a sludge
minimizing treatment technology for the metals and cyanides wastes.
Initially, the facility was designed to operate using lime and ferrous
sulfate precipitation of metals as the primary means of waste treatment. When
the high cost of land disposal of the lime sludges was considered, and alternate
means of treating and disposing of the waste was selected.
The unit processes employed to detoxify the wastes and recover metals at
Plant A currently include sodium hypochlorite oxidation of cyanides (alkaline
chlorination), sodium hydroxide precipitation, pH adjustment, sodium
borohydride (SBH) reduction (with sodium metabisulfite stabilization),
sedimentation, plate and frame filter press (for sludge dewatering), rapid
sand filtration, and ion exchange columns for effluent polishing.
Results—
The primary purpose of the Facility A case study was to evaluate sodium
borohydride as a viable waste treatment alternative for reducing RCRA
Hazardous Waste Code F006 spent electroplating baths and effluents. The
evaluation criteria were the ability of sodium borohydride (SBH) to
effectively meet local compliance standards and produce a high density,
low-volume sludge. The test program evaluation relys mainly on the trace
metals results to evaluate system performance.
The SBH reactor was sampled for trace metals on the influent, effluent,
and sludge streams. Both filtered and unfiltered samples were collected for
eight selected metals and the results are summarized in Table 2. The
2
-------
unfiltered sample showed little or no reduction as expected. However, the
filtered sample showed individual metals reduction efficiencies which ranged
from 16 Co 99.8 percent. The observed range in efficiency data was attributed
to variations in^concentration and chemical potential of each of the metallic
ions contained in the solution. Overall, SBH was able to reduce 6.91 kg of
the initial influent metals loading of 7.25 kg.. These results represent a
greater than 95 percent reduction in total metals for a complex waste stream.
The remainder of the metals influent loading (0.337 kg) consisted of over
70 percent calcium.
TABLE 2. SODIUM BOROHYDRIDE FINISHING REACTOR TRACE METALS
CONCENTRATIONS AND REMOVAL EFFICIENCIES
Element
AS
All
Cd
Cr
Cu
Hi
Fb
Zn
Reactor
influent (ng/L)
24.0
5.7
0.015
0.031
237.0
0.96
0.32
5.10
Reactor
effluent (mg/U
unf il tared"
6.2
4.76
0.01
0.03
207.0
0.902
0.31
4.76
Reactor
effluent Img/L)
filtered b
0.06
. 0.15
0.01
0.026
0.47
0.422
0.14
0.79
Percent
of
removal
¥9.7
V7.U
c
16.1
yy.8
56. U
56.2
y8.4
•Nonfiltered sample.
bFiltered onsite at Plant A'c Lab. In addition, a blank DI water sample was
filtered at the Plant's Lab onsite as a QC measure. Results for that sample
showed less than detection limits in all eases.
cUnable to obtain adequate precision.
An additional objective of this program was to evaluate the ability of
Facility A to consistently meet local pretreatment requirements. Table 3
presents observed metals effluent concentrations in comparison with sewer
discharge standards. The resultant data for two separate batch runs show
exceedences of effluent limits, apparently due to incomplete polishing caused
by cation exchange column breakthrough. Since the test program was completed,
Facility A has instituted the use of a quality control holding tank and
further waste processing optimization to remedy these problems. Follow-up
discussions with the local sewer authority revealed that Facility A's effluent
quality has improved considerably and is now consistently meeting compliance
guidelines.
-------
TABLE 3. PLANT A FINAL EFFLUENT COMPLIANCE DATA
Element
«g
AS
Cd
Cr
Cu
Ni
PD
Se
Zn
Total metals
Batch 1
0.05b
0.62C
0.01
0.29°
0.146
0.767C
0.14
0.2
0.98e
concentrations
Batch 3
0.05b
0.29C
0.01
0.03
1.826
0.bblC
0.1
0.2
0.05
(nC/L)
Effluent
limit4
0.03
0.01
0.05
0.2
0.4
0.5
0.15
1.0
0.5
'City of Warwick, RI effluent limits.
^ Unable to obtain great enough precision.
c£xceedence of pretreatment effluent limits
In addition to assessing wastewater effluent characteristics, the testing
program was designed to evaluate uncontrolled process air emissions. Table 4
summarizes the results of grab sample and integrated sample analysis of
process reactor exhaust ducts based on Draeger gas stream analysis. The
emissions results given in Table 4 show a continuous presence of hydrochloric
acid and hydrogen gas accompanied by occasional presence of ammonia and sulfur
dioxide. One of the hydrogen emissions grab sample results (6.0 percent) is
significant since this value is greater than the lower flammable limit for
hydrogen (4.0 percent). Note that grab sample concentrations for ammonia and
sulfur dioxide exceeded adopted short-term exposure limits (STEL) for these
substances.
Table 5 presents Facility A sludge characterization data for trace
metals. Analysis of the nickel/cyanide and SBH sludges shows total metals
contents (dry weight) of 35.7 and 6.6 percent, respectively. Neither sludge
result supported Facility A's claim of 60 to 70 percent metals content (dry
basis). While the SEN sludge result was significantly below performance
expectations, the exact cause of these results was not discernable. Possible
explanations include: (1) a possible process upset; (2) sampling error; arid
(3) analytical error. It seems most probable that a process upset was
responsible for these results, since blinding of the sludge press occurred on
the SBH press. Based on other SBH reduction case study results conducted
under this program, it is reasonable to assume that these results are not
representative, since typical sludge metals contents should be greater than
70 percent.
-------
TABLE 4. SUMMARY OF DRAGER TUBE ANALYSIS RESULTS FOR
UNCONTROLLED PROCESS AIR EMISSIONS3
Parameter
Hydrogen Cyanide
hvdrogen
Sulfur Dioxide
Hydrogen Sulfide
Ammonia
Hydrochloric Acid
Gas Concentrations (ppm
Crab Sanple Results
2 ppo
1.7 - 6.0d Z
1-20 ppo
1 ppm
5 - 180 ppm
1 ppm
or 1 as noted)
Integrated
Sanple Kesultt
2 ppm
0.4 Z
1 ppm
1 ppm
5 ppm
2 ppm
Threshold
limit value
IT^V>
exposure
limit"
1U ppme
™~
5 ppm
15 ppm
J5 ppm
5 ppmc
•Drager detector cubes are compound-specific for the parameter indicated.
Accuracy is estimated at ^S-20i of reading. Test conditions were as follows:
Flowrate • 3,600 afpn
Duct diameter 6 inches
Duct area - 22.274 in2 or 0.196 ft2
Volumetric flowrate at actual conditions • 0.196
ft2 X 3,600 afpm - 706.86 acfm.
kSource: ISBN 0.936712 - 61-9, 1985.
cTime weighted average value used in lieu of short term exposure limit.
dFive pump strokes were required (10 scrones standard) to reach saturation
concentration of 3Z, thus excrapolated reading is 3.OS (10) • 6.0Z
T57
TABLE 5. FACILITY A SLUDGE CHARACTERIZATION RESULTS
Element
*g
AS
Au
Ba
Ca
Cd
Cr
Cu
Fe
MB
Hi
Pb
Rh
Se
Sn
n
Zn
Total
Dry weight
concentration
(percent)
Ni/Cn SBH
sluage sludge
0.019 0.017
0.004 0.017
0.134 0.328
0.001 0.001
0.205 0.089
0.008 0.001
0.029 0.003
11.000 5.250
0.720 0.049
0.042 0.003
19.400 0.293
0.130 0.046
0.865 0.050
0.003 0.001
0.305 0.015
2.840 0.361
35.705 6.624
EP
Ni/CN
Toxicity
results
Img/i.)
bBH
sluage sludge
0.03
0.0'.
_-_
0.224
0.589
0.294
_._
0.06
0.05
—
0.163
0.016
0.032
-—
0.0018 0.0022
—
4.6
—
0.04
-—
0.03
___
0.04
—
tP ToxiPiPv
air IWAlCifcy
i>tanaarasa
(m6/L)
5.U
5.0
___
10U.O
—
1.0
5.0
___
0.2
™
5.0
_.
1.0
—
-— -
—
^Source U.S. Environmental Protection Agency
Federal Register V. 45 No. 98 98. 331:2 May 14, 1S60
-------
Table 5 also presents Facility A sludge EP Toxicity leachate results for
both Che nickel/cyanide and SGH reactor sludges. The results of the tests
clearly show that for Facility A influent metals concentrations, the SBH
sludge produced is fairly stable in that its leachate characteristics are
below EP Toxicity limits for all metals. However, note that the waste is
still classified as F006 hazardous waste.
An additional objective of the Facility A case study was to evaluate the
ability of sodium borohydride to economically reduce F006 waste streams. At
the time of testing, Facility A reduction chemistry was very inefficient at
$19.80/lb of copper reduced. However, through process optimization, chemical
costs have reportedly decreased over 63 percent, bringing process economics
within acceptable limits. The case study follow-up for Facility A has
indicated that the cost of copper reduction has been lowered to $7.27/lb of
copper.
Facility B Case Study
Description—
Facility B is a captive printed circuit board manufacturing facility
employing 77 people in Santa Ana, California. Gross sales are approximately
$7 million annually on production of 500,000 sf of board. Production at
Facility B uses a special hybrid process, employing elements of both additive
and semi-additive printed circuit production techniques. Process wastes of
interest to this study include rinsewaters from the electroplating and etchant
baths. The principle components of the acid copper electroplating baths are
copper sulfate and sulfuric acid. Facility B uses a slower acting etchant
(sodium chloride, sodium chlorate, and muriatic acid) which etches copper from
the board, and yields cupric chloride in the waste stream.
Facility B uses a rather unique end-of-pipe treatment system employing
sodium borohydride treatment and ultrafiltration (Meratek) technology for
solids separation. In this process, incoming plating and etching wastes are
adjusted to pH 7-11 by addition of sodium hydroxide or sulfuric acid. Sodium
borohydride is added to obtain an oxidation reduction potential (ORP) of
approximately -250 or less. The reacted waste then feeds from the
concentration tank to a Meratek ultrafiltration unit from which the permeate is
discharged to municipal treatment, and the concentrate is returned to the
concentration tank. A small plate and frame sludge filter press dewaters the
sludge which is drawn from the bottom of the concentration tank.
Salient points of interest in evaluating the Facility B waste treatment
system for this case study were: (1) compliance of the ultrafiltration
permeate (wastewater discharge) with local and Federal discharge standards;
(2) the volume and EP toxicity of the sludge filter cake; and (3) economic '
evaluation against comparable technology (lime and ferrous sulfate treatment).
The objective of the sampling program was to evaluate the effectiveness
of the sodium borohydride technology in use by Facility B. The effectiveness
was measured in terms of metal reduction efficiency and minimization of
hazardous waste streams. The data in Table 6 show influent and effluent
stream concentrations for metals of interest in this study. Data derived from
-------
Che metals concentrations in the effluent stream were used to determine the
effectiveness of the SBH reduction system in both meeting effluent guidelines
and minimizing releases to the environment. Table 7 describes effluent
loading characteristics in terms of reduction efficiencies £nd effluent
compliance.
TABLE 6. SUMMARY OF TEST DATA FOR SODIUM BOROHYDRIDE
TREATMENT AT FACILITY PLANT B
t'aramuter
local organic carbon
Total organic halide
Total trace metals:
Cu
Ni
It
Zn
EP toxic metals
Ar
ba
ul
ir
fb
Ug
Se
Ag
Influent waste
(Stream 3)
(tie/L)
40. 0
1.7b
786.0
(J.055
0.57
J.Bb
_
-
-
-
-
-
-
~
Effluent wascewater
(Stream 5)
-------
Analysis of these characteristics showed chat copper was reduced most
efficiently (99.82 percent), while nickel reduction was the Least efficient at
(45.5 percent).Differences in removal efficiencies were attributed to variations
in concentration (higher removals for higher concentrations), but the chemical
potential (quantity of free energy required for an ionic species to obtain
equilibrium) may also have been a factor. Approximately 144.7 Ibs of combined
metals were reduced to elemental form by the SBH reaction system, representing a
combined reaction efficiency of 99.8 percent. Despite deviations from design
operating conditions, the SBH/ultrafiltration system performed very well. EP
Toxicity leachate test results for Facility B filter press sludge clearly show
that the sodium borohydridge sludge produced is fairly stable in that its
leachate characteristics are below EP Toxicity limits for all metals. However,
note that the waste is still classified as F006 hazardous waste.
Table 8 presents the results of an economic comparison of the use of sodium
borohydride versus lime-ferrous sulfate chemistries. The results demonstrate
that in this application, sodium borohydride would be superior to lime-ferrous
sulfate for the following reasons: (1) sludge disposal costs and volumes would
be reduced by 93.5 percent; (2) overall operating expenses would be 48 percent
lower; and (3) sludge generated by the SBH reduction process was 78 percent
copper and suitable for reclamation.
The use of the sodium borohydride and ultrafiltration treatment at
Facility B is favored by the use of the chloride etch process in lieu of the
more commonly preferred ammonium peroxide etch. The ammonium-based etchants
create borohydride sludge stability problems which require tighter treatment
process control and the use of stabilizers such as sodium metabisulfite.
Additional factors which favor the economics of sodium borohydride treatment at
Facility B include: (1) the use of cupric chloride etchant; (2) high copper
concentrations and low organic loadings seen at this facility; and (3) low
effluent limitations required by the sanitation district.
TABLE 8. PLANT B ANNUAL TREATMENT AND DISPOSAL COSTS FOR SODIUM BOKO-
HYDRIDE AND LIME/FERROUS SULFATE PRECIPITATION TECHNOLOGIES
Sodium oorohyonde Lime/tcrrous
treatment suliate treatment
Basis Unit cost (S) system cose (S) system cost (S)
Cnemieal costs
SBH solution
Sodium hydroxide
Ferrous sulfate
Hydra t«d line
Total Chemical Cost
Disposal costs
Sludge disposal*
Annual costs
Total annual cost
Cost/lb metal reduced
2.7/lb
0.32/gal
0.11/lb
50.0/ton
200/ton
100,298
12,500
-
-
112.798
13,278°
126,076
3.5
-
-
35,ss«
-95
36,343
205,ibOe
241,943
6.7
"35 percent total solids
b78 percent metal in the solids
C5 percent metal in the solids
-------
Based on Che above results, it appears that sodium borohydride reduction is
an effective technology which can be utilized to reduce complex metal
electroplating sludges and render them reclaimable, and possibly less
hazardous. Note that the economics of SBH technology is highly dependent on
site-specific factors and warrants a detailed study prior to implementation.
Facility F
Description—
Facility F is an independent manufacturer of printed circuit boards. The
normal production volume of the facility is 40,000 ft^/month. The major waste
streams of interest to this case study are rinsewaters that follow
electroplating and etching processes. Prior to implementation of the
electrolytic recovery technology being studied, these rinsewaters contained
copper and lead at concentrations of up to 3,000 mg/L. Because of this, the
concentration of these metals in the final effluent exceeded pretreatment
standards (4.5 mg/L for copper and 2.2 mg/L for lead) for discharge to the city
sewer system. To decrease the concentration of metals in the effluent, the
facility converted several rinse tanks into static dragout tanks in order to
recover metals from rinse baths following copper, electroplating, tin/lead
electroplating, electroless copper plating, and a copper microetch process. The
quantity of metal recovered from the electroless copper rinse and the copper
microetch was small. Thus, the reactors were removed from these baths and
installed at the copper and tin/lead rinse baths where there was more potential
for metal recovery.
The electrolytic reactors used at this facility are Agmet Equipment Corp., Model
5200 reactors. They consist of a wastewater sump, a pump, and the anode and
cathode, contained within a rectangular box with dimensions of approximately
22 in. x 10 in. x 22 in. The anode is cylindrical and is encircled by a
stainless steel cathode with a diameter of 8 in. and a height of 6 in. The
anode material used for copper plating solutions is titanium. For tin/lead
plating solutions, however, a columbiura anode is required because the
fluoroboric acid in the tin/lead plating solution is extremely corrosive to
titanium. The columbium anode increases the cost of these electrolytic units to
$4,500, as opposed to $3,500 for the titanium anode units.
At the time of testing, four electrolytic reactors were being used for
recovery of copper, and three were being used for recovery of tin/lead. To
evaluate the performance of these units, samples of the plating bath, dragout,
and rinse bath were analyzed. A summary of the results of these analyses is
presented in Tables 9 and 10. Conclusions that were drawn based on these and
other data include:
• Recovery of copper from the acid copper solution is very effective—
rates of recovery were 4 to 5 grams/hour/unit, representing a current
efficiency of nearly 90 percent.
• Recovery of tin and lead was not effective at the time of testing-
concentrations of these two metals in the dragout were not
significantly less than in the plating bath. However, evaluation of
the data was difficult because the analytical results for some of
these samples were inconclusive due to matrix interference.
-------
Use of in-line eloctrolytic recovery was not able to reduce metal
concentrations to a nign enough degree to enable this tacLlity to meet
pretreatment standards.
Electrolytic recovery would significantly reduce the amount of sludge
generated if a lime precipitation system were utilized to remove
metals from the final plant effluent. For this facility, a reduction
of 32 tons/year would be realized (see Table 11).
At a sludge disposal cost of £200/ton, the annual cost of electrolytic
recovery would exceed the savings. However, if sludge disposal costs
increased to S300/ton, the savings (at least for copper recovery)
would exceed the processing costs.
TABLE 9. COPPER ELECTROPLATING SYSTEM - ANALYTICAL RESULTS
overage concentration Ing/LJ
Location Lop pur 1 ina Lvad 1UC
Plating bach 21,100 17 3.2 oil
uragouc ttol J.b U.45 lb.4
Second rinse ttU.b 4.7 U.2 —b
aAll Tin daci considered invalid, see Seecions It and b.
bruC analysis not done on clus Lath.
TABLE 10. TIN/LEAD PLATING SYSTEM - ANALYTICAL RESULTS3
Average concentration (mg/L)
Location topper Tinb Lead
Plating bach
U ragout
Second rinse
J.U
4.3
U.7
2.55U
l,Jo2
17.8
5,JJOB
J.,1.61
16.8
"No TOC measurements for tin/lead due to corrosivicy of the solution.
"Data considered invalid, set- Sections & and ft.
10
-------
TABLE 11. ECOWMIC AND ENVIRONMENTAL DATA FOR ELECTROLYTIC RECOVtKY
Plating Annual Annual
process cose (JO savings U)
Clipper 7,»oi» 4,o64
liu/le*d V,J7i l,88o
Metal hydroxide sludge
recovered* generation
( los) avoided (tons)
»«
2ito y
a based on average quancicion recovered by facility
Electrolytic recovery methods remove metals from an aqueous solution in a
metallic form which allows for the use of the recovered material as scrap
metal. Conversely, hydroxide precipitation removes the metal from solution and
generates a sludge with a low metal concentration. In most cases, the only
method of handling this sludge is landfilling at a high cost. Therefore,
electrolytic recovery is useful in minimizing the quantities of metal-bearing
sludge that must be landfilled. The cost effectiveness of this type of
technology will increase as sludge disposal costs increase in the future.
Facility E Case Study
Description—
Facility E began operations in January 1982 as a manufacturer of
customized, fine-line multilayer printed circuit boards. Facility E initiated
an ambitious waste minimization program in mid-1984. Since that time,
production has roughly doubled, but liquid discharge to the wastewater treatment
plant has remained constant and wastewater sludge generation has dropped roughly
30 percent. Waste minimization efforts continue to center around in-process
modifications to use nonhazardous or reclaimable solutions, to reduce water
consumption and bath dump frequency, and to optimize wastewater treatment
operations.
At Facility E, boards are pattern plated with eight acid copper and one
aqueous tin/lead plating baths in a 48-tank plating line. The line begins with
a nitric acid HN03 rack strip tank. After the racks are stripped, boards are
loaded and then undergo rinsing, cleaning with phosphate solutions (H^PO^,
Electroclean PC2000), and more rinsing before being plated. Acid copper baths
contain CuS04, organic brighters, and chlorides with copper concentrations of
24 oz/gal. The general processing procedure is to activate the board surface
(HC1), plate, clean/rinse and replate.
In plating operations, addition agent and photoresist breakdown products
will incrementally accumulate and contaminate an electrolytic (charge carrying)
plating bath. In the absence of a bath regeneration system, the manufacturer
would typically be forced to either discharge the spent'plating bath to the
wastewater treatment plant or send it offsite for disposal. In either case,
large quantities of metals containing sludge (RCRA Waste Code F006) would be
generated and subsequently land disposed. At Facility E, these spent plating
baths are regenerated through activated carbon filtration (used to remove
built-up organic bath contaminants) and then returned to the process. Copper
and solder plating baths are treated with activated carbon once every three
months and every month, respectively. The frequency of cleaning is determined
by organic contaminant build-up. Electroplating baths never have to be dumped
with this arrangement under normal processing conditions.
11
-------
Activated carbon treatment is performed in a batch node for acid copper,
solder and nickel raicroplating baths in three separate systems. The bath
reclamation system consists of a holding tank, mixing tank, and MEFIAG
paper-assisted filter. For acid copper treatment, 2,400 gallons of contaminated
solution is pumped into"*^ 3,000 gallon mixing tank. Hydrogen peroxide is added
and the temperature of the bath is maintained at 120 to 130°F for 1 hour.
Powdered activated carbon (80 Ibs) is added and the contents are mixed for 3 to
4 hours to oxidize volatile organic species. The solution is recirculated
through a paper-lined MEFIAG filter several times to remove the activated
carbon. The filter solids and paper are removed as needed when a predetermined
pressure drop across the filter is reached. When the bulk of the activated
carbon has been removed (generally after three passes of the solution through
the filter), the filter is precoated with 5 gallons of diatomaceous earth. The
solution is again recirculated through the filter until a particulate test
indicates sufficient solids removal (no residue detected on visual examination
of laboratory filter paper). Total spent solids from plating bath purification
is 1-1/2 drums every 3 months which is landfilled. I
Results —
The purpose of this case study was to evaluate the extension of
electroplating bath lifetimes (and subsequent waste reduction) by activated
carbon removal of organic brightner breakdown products. The acid copper baths
were selected for study since recovery of this solution results in the most
significant amount of waste minimization.
Sampling and analysis was conducted on three process streams associated
with activated carbon bath reclamation. A summary of the analytical results is
presented in Table 12.
TABLE 12. ACTIVATED CARBON BATH REGENERATION SYSTEM TEST RESULTS
Contaminated Recovered
solution solution spent carbon
Parameter (og/U (ng/U
Total Organic Carbon 257.9 218. A
Volatile organics:
Sulfur dioxiae 2.1
Methyl formate 1.9 2.3
Methyl acetate 0.43 O.t>5
Acetone - 0.08
Unknowns 0.34
Trace metals:
Copper 21,500 21.600 107,000
Lead 1.1 0.66 99b
Tin 6.9 £.3 *20
Cyclic voltaic stripping:8
Organic brightner 6.42 3.6
aThis analysis was conoucted by the bath manufacturer for additional
information purposes only. No QA data were obtained for these results.
b£P Toxicity test results for lead showed 0.17 cg/L wnich is significantly
less tnan the EP Toxicit> standard of 5.0 mg/L for lead.
12
-------
Based on these results shown in Table 12 the following conclusions can be
drawn:
• Forty-seven percent of the organic by-products and brightners were
removed from the contaminated solution;
• Low molecular weight organics such as carboxylic acid derivatives are
not preferentially adsorbed;
• Reduced sulphur (a brightening and leveling agent) is oxidized and
volatilized during treatment; and
• Inorganic contaminants such as tin and lead are also removed
(37.5 percent and 24.5 percent, respectively) as a beneficial
by-product of the treatment process.
In recovering spent electrolytic plating baths, Facility E was able to save
over $50,000 in hazardous waste disposal and raw material purchase costs (see
Table 13). These savings represent a payback period of only 3 months for
purchasing the activated carbon recovery system. This relatively short payback
period, combined with the volume of plating solution regenerated, make activated
carbon treatment a cost-effective and environmentally safe technology for
reducing the quantity of hazardous waste that would otherwise be land disposed.
TABLE 13. ECONOMIC EVALUATION OF FACILITY E
PLATING BATH TREATMENT SYSTEM
Cost item Unit cose ($) toot liJ
Capital costs
(1) Model 3020Y filter 6.J56 6,Jib
treatment systeo*
Miscellaneous 101 of purcnase price fcjb
Total capital: s.i*^
Annual O&M
.lefiag filter papers' 166/250
Electricity6 0.05/kwh
Maintenance 102 of total capital
Labor IS/nr
Powdered accivated carbon1- 0.96/lb
S02 hydrogen peroxide0 0.56/lb-
Total O&H:
Annual costs
Annual ized capital (12!, 10 yrs) 0.177
Annual OUi
Annual spent caroon disposal^ 140/drum
Total costs:
Annual savings
Hazardous waste disposal4 1.15/gal 12,<*20
Recovered plating solution' (copper) 10,000/oatn ill,000
(tin/lean) 15,000/batn li.uuu
Total savings: o7,«.^u
Net Annual Savings: 57,2o7
aBaker Brothers Technical Bulletin.
bDepartment of Energy, Energy Information Administration, National Average,
Decenoer 1986.
cMciCesson Cnemical Technical Brocnure.
dAs quoted by Clean Harbors Inc.
eOMI SEL-REX Tecnnical Brocnure.
-------
RESIST DEVELOPING SOLVENT RECOVEKY CASE STUDIES
Two case studies evaluated under this program focused upon Che minimization
of developer solvent wastes and sludges which might require either land disposal
or incineration. In general, the recovery of resist stripping and developer
solvents is not unique within the PC board manufacturing industry. However, the
recovery systems evaluated at the two facilities discussed below represent
state-of-the-art technology applications. In the case of Facility C, the
technology involves the separation of a two-solvent system with subsequent
recovery and reuse of each solvent. In the case of Facility D, the technology
evaluated further recovers the solvent bottoms product of the initial recovery
unit.
Facility C Case Study
Description—
Facility C manufactures computing equipment including logic, memory and
semiconductor devices, multilayer ceramics, circuit packaging, intermediate
processors and printers. One of the major hazardous waste streams that is
generated is spent halogenated organic solvents (RCRA Code F002). The solvents
and their uses are: (1) methylene chloride used in resist stripping of
electronic panels; (2) methyl chloroform (1,1,1-trichloroethane) used in resist
developing of electric panels and substrate chips; (3) Freon used in surface
cleaning and developing of substrate chips; and (4) perchloroethylene used in
surface cleaning of electronic panels.
The spent solvents from photoresist stripping and developing are
contaminated with photoresist solids at up to 1 percent, and the solvents used
for surface cleaning are contaminated by dust, dirt or grease. Waste solvents
are recovered at Plant C by distillation or evaporation and returned to the
process in which they were used. Several types of equipment are used including
box distillation units to recover methylene chloride and perchloroethylene,
flash evaporators to recover methyl chloroform, and a distillation column to
recover freon.
There are two identical flash evaporators at the facility, each with a
capacity to recover 600 gallons of methyl chloroform (MCF) per hour. The flash
chamber operates at a vacuum of 20 in. Hg, allowing the MCF to vaporize at 100
to HOT. The units are operated one to two shifts/day depending on the
quantity of waste solvent being generated.
A packed distillation column is used to recover pure freon from a waste
solvent stream containing approximately 90 percent freon and 10 percent methyl
chloroform. Waste is continuously fed to a reboiler where it is vaporized and
rises up the packed column. Vaporized freon passes through the column, is
condensed and recovered at a rate of 33 gal/hour. MCF condenses on the packing
and falls back into the reboiler. The distillation bottoms are removed when the
concentration of methyl chloroform reaches 80 percent (approximately 1 to
2 weeks).
-------
There are also two identical box .stills at the facility, each with a
capacity to recover 475 gph of raethylene chloride. These are very simple units
consisting of an 800 gallon still pot with hot water heating coils. The
contaminated raethylene chloride is heated to between 103"F and 108°F, and clean
solvent is condensed overhead.
Results—
Sampling and analysis was conducted on process streams associated with two
of the solvent recovery processes. One of these processes was the flash
evaporator used for recovery of methyl chloroform (1,1,1-trichloroethane), and
the other was the distillation column used to recover Freon TF from a
Freon/methyl chloroform mixture. A summary of the analytical results is
presented in Tables 14 and 15. The conclusions drawn from these results are:
• At least 95 percent of the solids are removed from the solvent waste
influent;
• The recovered product is at least as clean as the virgin material; and
• The still bottoms from recovery of contaminated solvent still contain
a high fraction (90 percent) of solvent.
As shown in Table 16, the recovery of spent solvents at the facility is
motivated primarily by economic benefit. In recovering spent solvent, the
company saves over $10 million annually, compared to offsite recovery. The
savings per pound of solvent recovered is $0.18, $0.18, and $0.61, respectively,
for methyl chloroform, methylene chloride, and freon.
The high cost savings are primarily due to the fact that the solvents
recovered are reused onsite, thus reducing the quantity (by greater than
95 percent) of new or virgin solvent that must be purchased. Offsite recovery
could be conducted, but at much higher cost as indicated in Table 16. Since the
rate of generation of spent solvent is so high, the initial expense of
purchasing recovery equipment is quickly returned.
To landfill or dispose of such a large quantity of spent solvent by any
other method would be economically unacceptable. Incentives other than economic
reasons for onsite recovery include:
• Reduction in the risk of a spill of solvent in transporting the waste
to a TSDF; and
• Reduced liability related to an accident at the TSDF resulting in the
release of spent solvent.
Facility C is trying to further reduce the quantity of waste solvent that
must be sent offsite for recovery. They intend to do this by recovering the
still bottoms generated by distillation of freon/methyl chloroform waste. In
addition, they eventually plan to phase out the use of methyl chloroform and
methylene chloride and replace these materials with aqueous-based photoresist
developers and strippers.
15
-------
TABLE 14. FLASH EVAPORATOR TEST RESULTS
Faramece r
Virgin
Recovered Still Methyl
Influent Product Bottoms Chloroform
Volatile Organics (Z w/w)
- Freon TF <0.1 <0.1
- Methyl Chloroform 99.9 99.9
- Other <0.1 <0.1
Solids (mg/kg) 460 3.0
92
78,000
TABLE 15. DISTILLATION COLUMN TEST RESULTS
99.9
2.2
Faramete r
Recovered Still Virgin
Influent Product Bottoms Freon TF
Volatile Organics (Z w/w)
- Freon TF
- Methyl Chloroform
- Other
Solids (mg/kg)
96
3.9
<0.1
1.2
99
0.9
<0.1
0.06
52
47.9
<0.1
27
99.9
<0.1
<0.1
0.13
TABLE 16. RECOVERED SOLVENT QUANTITIES AND COST SAVINGS
Methyl Methylene
Chloroform Chloride
Freon TF
Quantity Recovered (l984)a
Quantity Still Bottoms Sent Offsite
(1984)a
Cost Savings ($)
aThousands of pounds.
38,500
1,200
27,500
800
2,310
1,355
7,000,000 5,000,000 1,400,000
16
-------
"acility D
Description-
Facility D manufactures mobile communications equipment components in their
Florence, S.C. facility. The operation consists of a small metal-forming shop,
prepaint and painting lines, electroplating, printed circuit board manufacture,
and a 30,000 GPD onsite wastewater treatment plant.
Printed circuit boards are produced using the subtractive technique and
solvent-based photoresists. Methylene chloride resist stripper and
1,1,1-trichloroethane (TCE) developer are continuously recycled in closed-loop
stills. The TCE developer wastes (Waste Code F002) are recovered in a DuPont
Riston SRS-120 solvent recovery still (referred to as the primary still) and
returned to the developer line. Until recently, all still bottoms from the
primary still were drummed and shipped offsite for reclamation at a solvent
recycling facility. Facility D purchased a Recyclene Industries RX-35 solvent
recovery system (referred to as the secondary still) in October 1985, to further
remove TCE from still bottoms onsite.
The Recyclene Industries RX-35 solvent recovery system is a batch
distillation system with a 30 gallon capacity, silicone oil immersion heated
stainless steel boiler, a non-contact, water-cooled condenser, and a 10 gallon
temporary storage tank. The boiler is equipped with a vinyl liner inside a
Teflon bag. The Teflon bag provides temperature resistance and the vinyl bag
collects solid residue, eliminating boiler clean-out and minimizing sludge
generation after distillation. Two thermostats control the temperature of the
boiler and the vapor, automatically shutting down the boiler when all the solvent
has evaporated. The maximum operating temperature of the still is 370°F, so
recovery of solvents with higher boiling points would not be practical. Recovery
of a 20 to 25 gallon batch of still bottoms requires approximately 90 minutes at
Facility D, and four to six batches are completed each day.
Results—
Evaluation of the system consisted of the analysis of the contaminated feed,
overhead product, and distillation bottoms. A mass balance based on the results
of these analyses is presented in Table 17. Based on these data, the following
conclusions can be made:
• Purity of recovered solvent was 99.99 percent;
• Total solvent recovery was 99.78 percent;
• Still bottoms contained 7.5 weight percent 1,1,1-trichloroethane; and
• Reduction in waste generation was 97.5 percent..
TABLE 17. RECYCLENE STILL MASS BALANCE
Loading (Ib/oatch)1
Parameter
Solvent
Solid
Otner
Influent
(stream 4)
1853.7
i7.0
142. 9
Distillate
(stream 5)
1849.45
0.01
142.79
Bottoms
(scream o)
4.25
46.99
0.11
1 7
-------
An additional objective of the study was to evaluate the economics of the
batch solvent recovery unit. Table 18 lists the annual cost savings and waste
reduction calculated for Plant D, based on the first year of RX-35 operation.
In addition, the investment payback period for the RX-35 was calculated
considering credit for reclaimed solvent and reductions in waste transportation
and disposal costs. The estimated payback period was 7.3 months, given the
current level of solvent reclamation. Thus, the low capital cost of the unit
and the relatively high costs of virgin solvent favor the second-stage recovery
of TCE developer still bottoms.
TABLE 18. ANNUAL COST SAVINGS AND PAYBACK FOR RECYCLENE RX-35 AT PLANT D
Number of Coit per Cost Prior Cost alter
Cost Item Units (per yr) Unit ($) to installation (S) Installation (SJ
Contaminated Solvent 10,625 gal .35 3,719
Reeyelene Bottom* 3.2 com 200 - 640
Differential Solvent
Purchase 10,602 4.50 47,709
Differencial Energy
Coniumpcion 20,092 kwh 0.06 - 1,205
Replacement Liners
Teflon 52 bags 45.15 - 2.34b
Nylon 155 bags 6.50 - 1 .UlU
Additional Labor 208 hr* 15.00 - 3,120
TOTAL COST 51,428
ANNUAL COST SAVINGS (1st year) 43,105
RECYCLENE RX-35 PURCHASE AND INSTALLATION COST 26,150
PAYBACK PERIOD 7.3 mo.
There are several potential drawbacks to the implementation of RX-35 batch
still that should be discussed. The first is that since the bottoms product
contains 7.5 weight percent 1,1,1-TCE, it remains classified as RCRA Waste Code
F002 (halogenated organic solvents) and is among those solvent wastes being
considered under the land disposal ban. Thus, while this technology
significantly reduces the volume and toxicity of the solvent still bottoms, it
continues to generate a hazardous waste. A second potential concern is the
accumulation of contaminants and/or breakdown products. For example, 6.7 to
11.0 percent concentrations of carbon tetrachloride were found in process feed
and exit streams, indicating a build-up of this contaminant. Another
significant contaminant found was 2-Butanone, which represents 3.6 percent of
the solvent waste feed stream. It could not be determined whether a build-up of
2-Butanone was occurring or if it is harmful to the system. However, its
presence and effect on the solvent properties of 1,1,1-TCE should be considered.
18
-------
A final consideration in the implementation of any solvent recovery still
is the issue of safety. The unit at Plant D was housed in a separate
structure and provided with adequate ventilation to minimize the risk of
exposure or explosion. The RX-35, according to the manufacturer, is safe for
flammable materials, and is rated for NFPA Class I, division I, Group D
environment (Recyclene, 1985). These safety considerations should help to
minimize the risk of chronic exposure or danger from explosion to personnel.
Nevertheless, explosion risks from solvent recovery operations should be
carefully evaluated in planning the layout and installation of the unit.
CONCLUSIONS
The findings of the waste minimization case studies evaluated under this
program are presented in Table 19, which includes data collected by the
facilities and verified by sampling and laboratory results. These results
indicate that a good variety of technologies exist to minimize
metals-containing and solvent wastes produced by the PCB and semiconductor
industries. The technologies discussed range from simple changes in treatment
system reagents with nominal capital costs to large onsite solvent reclamation
facilities with significantly higher capital costs.
TABLE 19. SUMMARY OF FINDINGS OF WASTE REDUCTION CASE STUDIES
Facility name Technology
Facility A
Facility It
Facility C
Sodium borohydride reduction
Sodium borohydride reduction
Solvent batch distillation
Waste reduction
Metals sludge
Metals sludge
Methylene chloride
Annual waste
Projected
annual coat
reduction Capital costa Havings
achieved (t) indicates negative value.
19
-------
Four of Che case studies investigated under this program focussed on
technologies to reduce metal-plating rinsewater sludges. The use of sodium
borohydride as a substitute for lime/study ferrous sulfate was found to be
viable in one case and appeared to be marginally acceptable in another. The
case study on carbon adsorption recovery of plating bath wastes found that
this technology significantly reduced both disposal costs and waste volume.
The case study of electrolytic recovery indicated that this technology is
highly waste stream specific. An acid copper electroplating rinse is an ideal
waste stream for electrolytic recovery. However, other metal-bearing rinses,
such as those from solder (tin/lead) plating or etching, are not appropriate
for use of electrolytic recovery. Electrolytic recovery units are, however,
generally inexpensive to purchase and can be used in many cases to supplement
an end-of-pipe treatment process.
Two of the case studies presented in this paper involve the recovery of
spent halogenated solvents using batch distillation units. Both of these case
studies indicate that onsite solvent recovery is successful from a technical
and an economic standpoint. In both cases, over 95 percent of the waste
solvent was recovered and reused onsite. Solvent recovery appears to be a
technology that could be applied to a number of printed circuit board
manufacturing facilities.
The results of this project indicate that waste reduction can be achieved
through the use of appropriate technology, and it can be achieved with
significant reductions in cost. The case studies also indicate that the
success of waste reduction is in many cases waste stream specific. The
technologies will not necessarily be successful in all cases. A slight
variation between one waste stream and another may make waste reduction either
technically or economically impractical. Therefore, successful waste
reduction is dependent on a thorough knowledge of waste quantities and
characteristics.
20
-------
SECTION I
INTRODUCTION
BACKGROUND
With the enactment of the Hazardous and Solid Waste Amendments (HSWA) in
November 1984, Congress set forth a schedule for evaluating the restriction of
various classes of hazardous wastes including: (1) solvents; (2) metals and
cyanides; (3) halogenated organics; (4) corrosives; and (5) dioxin wastes. A
key issue identified in the evaluation of the waste bans is the availability
of commercial treatment capacity to handle the wastes proposed for banning.
Therefore, Congress also asked EPA to evaluate the potential for onsite waste
minimization to reduce the quantity or toxicity of wastes being considered
under the ban.
In an effort to identify successful waste minimization technologies,
EPA*s Office of Solid Waste (OSW) and Office of Research and Development IOKI;)
Hazardous Waste Engineering Research Laboratory (HWERL) set forth on research
efforts aimed at assessing the viability of waste minimization as a means of
reducing the quantities of land disposed hazardous waste. OSW's research
focused on an exhaustive literature review identifying a broad spectrum of
waste minimization technologies and their various applications. The primary
emphasis of HWERL1s work was on demonstrating the effectiveness of specific
minimization technologies through case studies and process sampling.
WASTE MINIMIZATION CASE STUDY SELECTION
The case study development work was divided into two phases with Phase I
involving:
• Waste category assessments;
• The identification of the data requirements and organization of the
case studies; and
• The selection of specific sites/streams for use in the case studies.
The waste category assessments were a series of five reports aimed at
identifying key industries that generate wastes which are being considered for
restriction from land disposal. The five waste categories assessed included:
(1) solvent wastes; (2) metals-containing wastes; (3) cyanide and reactive
wastes; (4) halogenated organic nonsolvent wastes; and (5) corrosive wastes.
The findings of these reports were used in conjunction with the findings of
other aspects of the case study selection approach to help direct the final
sections.
21
-------
As part of Che case study identification/selection process, the project
team contacted trade associations and state agency representatives to solicit
ideas and advice. As a result of these meetings, it was determined that case
study selection should focus on a single industry or waste stream. The
electronics industry was initially judged as a good choice because it is a
growth-oriented industry and ranks in the top 20 industries generating solvent
wastes.
The criteria for selecting case studies was further narrowed down to
those facilities generating waste described by RCRA codes F006 or F001 and
F002, which are respectively, waste treatment sludges from electroplating
operations, and spent halogenated solvents or still bottoms from recovery of
those solvents. These waste types were selected because they are two of the
largest volume hazardous waste streams generated by the electronics industry,
particularly by manufacturers of printed circuit boards and semiconductors.
Facilities which met the selection criteria were contacted to determine
whether they practiced some fort of onsite waste minimization or recycling.
Preliminary site visits were scheduled for cooperating facilities after
determining the willingness to participate. The purpose of the preliminary
site visit was to evaluate the practicality of testing the waste minimization
process to determine its performance and to gather information necessary to
conduct the testing.
During the case study selection process over 50 facilities were contacted
by mailings or telephone to explain the case study program and determine their
interest and anticipated level of cooperation. Based on the initial
screening, 15 metals waste case studies and 12 resist strip solvent case
studies were identified. Ten facilities were visited for pretest site visits
to assess the facility's suitability for testing and further explain the
intent and scope of the case study program. In the final section, six
facilities were determined to be suitable to the scope of the program and
willing to cooperate.
Part II of this study was devoted to testing of the waste minimization
process and development of the case study reports. During this phase of work
under this program QA Project Plans (Test Plan) were ^zg^pared for the testing
proposed at the six facilities selected in Phase I. Following approval of the
Test Plan by the facility and EPA, testing was conducted. During the case
study testing, process information was collected by the investigators or
provided by the facility where appropriate. Mass throughput data and samples
for analyses were collected according to the test plans. These data were
occasionally supplemented by plant-supplied data where necessary to obtain a
more representative picture of the long-term operation.
The case study assessments presented in this report discuss the results
of analytical measurements used to discuss the performance of each
technology. In addition, measurements of process residuals and/or other
discharges are presented in the case studies. Finally, an assessment of the
economics of each technology is also presented to assist the cost
effectiveness of each technology.
22
-------
REPORT ORGANIZATION
The remainder of chis report presents the results of this research
effort. Section 2 presents pertinent production and waste management
information on the electronics products manufacturing industry. Sections 3
through 8 present waste minimization case study results for plants A through
F, respectively. Section 9 summarizes Quality Control statistics for each
case study. Section 10 presents conclusion of the project and recommendations
for further research efforts.
23
-------
SECTION 2
THE ELECTRONICS PRODUCTS INDUSTRY
BACKGROUND
The electronic components manufacturing industry (SIC 367), includes
eight specific product areas identified by four-digit SICs. These product
areas include capacitors, transformers, semiconductors, and printed circuit
boards. As determined by the case study selection criteria, the semiconductor
and printed circuit board industry were assessed as the product areas of
greatest interest. Total worldwide production of printed circuit boards was
approximately $4.5 billion in 1984, but has declined by 40 percent in 1985
(Electronic Business 9/1/85). Worldwide production of semiconductors also
experienced a setback in 1985 as evidenced by the 1984 production of
$33 billion down to $29 billion in 1985 (Electronic Business 3/1/86).
However, total U.S. production of semiconductors is forecasted to experience
growth from $8.3 billion dollars in 1985 to $15.9 billion dollars in 1988
(Industry Week 10/14/86), while world production of printed circuit boards
will reach $9 billion in 1989 (Electronic Business 9/1/85).
The industry consists of both small, independent job shops with limited
product lines to large automated facilities with integrated operations
generating large quantities of hazardous waste. In 1980, there were reported
to be 545 companies in the U.S. involved in the manufacture of semiconductors,
and 345 involved in the manufacture of printed circuit boards. Only
12 percent of the companies surveyed employ over 2,400 persons, while
80 percent employ 100 or less (EPA, 1983). Due to the high degree of design
diversity within product areas and the large disparity between generator
volume, wastes are categorized by the primary constituent of the waste, not by
raw material usage or manufacturing process.
WASTE GENERATION
In the manufacture of printed circuit boards (Figure 1) and
semiconductors (Figure 2), major waste streams of concern are spent organic
solvents (RCRA codes F001-F005) or metals containing wastes (characterized in
Tables 20 and 21). Organic solvents are used for wafer/board cleaning and for
the developing and stripping of photoresist materials used in the image
transfer and/or circuit fabrication processes. Photoresists are light
sensitive, organic, thermoplastic polymers available as either liquids or dry
solids. Negative image photoresists polymerize upon exposure to light, after
24
-------
BOARD
PREPARATION
•— wstavater with solid*
BOARD
CLEANING
SURFACE
PREPARATION
vaitc organic tolvcnti
addle or alkaline rlnaevatari
. wait* organic solvents
acidic ot alkaline rinaewateri
CAIALYST
APPLXCAIIOH
—— — • rinaewaiara containing Mtala
ELECT&OLESS
PLATING (FLASH)
IMAGE
TRANSFE&
ELECTROPUIIUe
(coma)
ELECnOPLATIHG
(SOLDER)
•pent platini solution
rlnamrntar containing eoopleztd coppar
vaate organic lelventi
• waitcwattl vlth pbotereslit
-•pent plating solution
•rlaaevatcT eonra1n1ng copper
-•pent plating solution
-riuenter containing tin and lead
ETCHING
ELECTROPLATING
(TABS)
10
••pent etcbants
-rlaaanter containing Mtali
-rlnaevatei* possibly contaialag
cyanldt
Figure 1. Subtractive printed circuit board production flowsheet.
Source: EPA-600/2-83-033.
25
-------
SEMICONDUCTOR
INGOT
WAFER CUTTING,
SMOOTHING, AND
POLISHING
1
1
CHEMICAL
CLEANING AND
POLISHING
2
1
EPITAXIAL
wKUw in
3
|f
CIRCUIT
• FABRICATION
4
DIFFUSION
5
1
__^^J
— — — — ^— wastewater with solids
— — — ^ spent organic solvents
^ acid and alkaline rinsewaters
— ^ spent acids
^ waste organic solvents
— — ^- rinsewaters with solvents and acids
1
"
METAL — — ^ waste organic solvents
INTERCONNECTION — ^— — ^^ rlnsewater wltn metals
6 ^ spent acid baths
1
t
( MOUNTED ^\
V WAFER y
Figure 2. Integrated circuit production flowsheet.
Source: EPA-600/2-83-033.
26
-------
TABLE 20. CHARACTERISTICS OF RAW WASTE STREAMS FROM SEMICONDUCTOR
DEVICE MANUFACTURING [EPA-600/2-83-033]
Parameter
Antimony
Arsenic
Beryl 1 i urn
Cadmium
Chromium
Copper
Cyanide
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
Phenols
Oil and grease
Total suspended solids
Total organic carbon
Biochemical oxygen demand
Fluoride
1,2,4-trichlorobenzene
1 ,1 ,1-trichloroethane
Chloroform
1,2-dichloro benzene
1,3-dichlorobenzene
1,4-dichloro benzene
1,1-dichloroethylene
2,4-dichlorophenol
Ethyl benzene
Methylene chloride
Napnthalene
2-nitrophenol
4-m'trophenol
Phenol
Di-n-octyl phthalate
Tetrachl oroethyl ene
Toluene
Trichloroethylene
Concentration
range, mg/liter
<0. 001-0. 187
<0. 003-0. 067
<0.001-<0.015
<0. 001-0. 008
<0. 001-1. 150
-0.005-2.588
<0. 005-0. 01
<0. 04- 1.459
<0. 001-0. 051
0.005-4.964
<0. 002-0. 045
<0. 001-0. 013
<0. 001 -0.01 2
0.001-0.289
<0.002-6.1
ND-20.8
ND-203
ND-80
9-202
ND-330
<0. 01-27.1
<0.01-7.7
<0. 01-0. 05
<0.01-186.0
<0.01-14.8
<0. 01-14. 8
<0. 01-0.071
<0.01-0.017
<0.01-0.107
<0.01-2.4
<0. 01- 1.504
<0. 01-0. 039
<0. 01-0.18
0.014-3.5
<0.01-0.01
<0.01-0.80
<0.01-0.14
0.007-3.5
Mean
concentration,
mg/liter
0.021
0.018
0.002
0.003
0.129
0.570
0.005
0.145
0.004
0.502
0.021
0.005
0.015
0.093
0.630
5.058
31.61
55.676
52.763
62.0
4.643
1.395
0.015
15.972
1.450
1.341
0.029
0.012
0.021
0.244
0.214
0.024
0.061
0.519
0.01
0.122
0.018
0.322
Industry
wide pollutant
discharce,
kg/day*
13.2
13.2
1.9
1.9
99.9
540.7
3.8
61.5
5.7
655.6
6.9
3.8
11.3
46.5
812.6
2,778.3
30,470.6
17,094.2
38,848.1
35,909.0
257.5
928.2
15.7
499.3
174.0
156.4
9.4
9.4
6.3
276.1
19.5
27.6
15.1
203.5
6.3
363.0
33.9
177.1
Flowrate weighted.
NO - Not detected.
27
-------
TABLE 21. CHARACTERISTICS OF RAW WASTE STREAMS FROM PRINTED CIRCUIT
BOARD MANUFACTURING
Constituent
Total suspended solids
Cyanide , total
Cyanide, amenable to chlorination
Copper
Nickel
Lead
Chromium, hexavalent
Fluorides
Phosphorus
Silver
Palladium
Gold
EDTA
Citrate
Tartrate
NTA
Range,
0.998 -
0.002 -
0.005 -
1.582 -
0.027 -
0.044 -
0.004 -
0.648 -
0.075 -
0.036 -
0.008 -
0.007 -
15.8 -
0.9 -
1.3 -
47.6 -
mg/liter
408.7
5.333
4.645
535.7
8.440
9.701
3.543
680.0
33.80
0.202
0.097
0.190
35.8
1342
1108
810
Source: EPA-600/2-83-033.
28
-------
which unexposed areas are dissolved by developer solvent. Developers and
strippers for this type of resist are generally organic solvents such as
1,1,1-crichloroethane, methylene chloride, xylene, and ethyl benzene.
Positive-image photoresist materials become soluble upon exposure to light,
after which developer solvent is used to remove resist material under the
transparent areas of the photomask. Developers and strippers for this type of
resist are generally aqueous solutions which are either alkaline in nature or
contain organic compounds such as glycol ethers and alcohols.
A recent trend in electronics component manufacturing is the switch-over
from negative to positive photoresist materials. This is particularly evident
in states such as California where the Air Resources Board guidelines will
require a 90 percent decrease in the emission of volatile organic compounds
(VOCs) by 1987. Since positive photoresists utilize aqueous solutions, their
use can aid in the compliance with the new standards (Electronic Business
3/1/86). In the absence of spent organic solvents, the aqueous solution can
be released to the sewer, with solids removal being the only required
treatment.
The electronics component industry ranks high relative to other
industries in the generation of solvent waste (as shown in Table 22).
Semiconductor manufacturers are ranked 12th and electronics component
manufacturers not elsewhere classified (which includes the manufacture of
printed circuit boards) are ranked 19th. These data, however, reflect 1981
practices. If an increasing number of companies continue switching to
photoresist materials with an aqueous or semiaqueous as opposed to an organic
solvent base, then the quantity of hazardous waste should be decreasing.
Metals are essential to all electronic components due to their conductive
and resistive properties toward electricity. Silver, gold, copper, tin, and
their alloys are utilized because their high conductivity is essential to the
operation of components or because their use in leads and connectors keeps
electrical power loss to a minimum. Many metal parts must be protected from
corrosion by plating with nickel, silver, gold, or tin/lead. The most common
forms of application are electroless and electrolytic plating, in which an
adherent metallic coating is deposited on an electrode (the part being plated)
to produce a surface with properties or dimensions different from those of the
basic metal (EPA SV-140c 1977). These metals are introduced into the waste
stream through either the disposal of concentrated plating baths or running
rinses directly following the electroplating process. A second major source
of metallic contaminants is the chemical etch step utilized as part of the
electroplating preclean operations or in the removal excess surface metal.
Etching rinses will also contain relatively high concentrations of metals
along with dilute levels of etching solution. Chemical etch baths typically
contain ammonium chloride, ammonium persulfate, or sulfuric acid/hydrogen.
peroxide as the active ingredient and are applied in either a batch mode or in
a conveyorized spray apparatus. Conventional waste treatment includes
chemical precipitation, clarification, and dewatering, which results in the
landfilling of hazardous sludges (RCRA code F006).
29
-------
TABLE 22. TOP 20 INDUSTRIES GENERATING SOLVENT WASTES
U)
o
Nn
of
estab.a
2145
1160
1529
4287
541
2902
4656
4151
393
337
15490
2237
2563
6506
560
1040
32
861
5392
235
57017
SIC
Code
2851
2869
2821
3471
2833
3479
3662
3714
9711
3721
3079
3674
2899
7391
3411
3711
2067
2879
3679
3951
SIC description
Paints & Allied Products
Industrial Organic Chemicals
Plastics Materials
Plating and Polishing
Medicinal, Botanical Products
Metal Coating & Allied Serv.
Communication Equipment
Motor Vehicle Parts
National Security
Aircraft Equipment
Plastic Products, Misc.
Seraiconduc tors
Chemical Preparations
Research & Devel. Labs
Metal Can Fabrication
Motor Vehicle Bodies
Chewing Gum
Agricultural Chemicals
Electronic Components
Pens & Mechanical Pencils
Weighted0 number
Halogenated
solvents^
105
327
215
471
137
136
186
241
166
107
120
93
85
103
35
57
57
59
96
66
•Z— ? TT ? T-T1S-~'~11S1TT S f " T- ~ '— T ~T-T — •
of solvent waste streams
Nonhalogena ted
solvents^
1436
654
536
176
323
279
225
161
178
230
201
194
189
163
154
127
87
85
40
59
aNumber of establishments based on Dun's Marketing Services, a company of Dun and Bradstreet Corp.,
1983 Standard Industrial Classification Statistics.
Information on generators taken from 1981 data (National Survey of Generators).
CFor weighting procedure refer to Westat, Inc., 1984.
Source! Engineering Science, 1984.
-------
WASTE MANAGEMENT
As effluent discharge limits for the electronics industry have become
increasingly strict, the industry has been forced to treat their wastewaters
to remove dissolved metals. As mentioned previously, however, conventional
treatment methods such as lime precipitation result in the generation of
metals containing sludges. Since disposal of these sludges in landfills may
soon be banned under the amendments of RCRA, other nonsludge generating
methods of management will see increasing utilization. Data from the National
Survey of Waste Generators, which reflects 1981 practices, show that SIC 36
(electronics industry) ranks second among all industrial categories in offsite
use, reuse, recovery, or recycle (URRR) of hazardous waste. By contrast,
SIC 36 is not ranked in the top ten for onsite URRR. It is believed that
offsite URRR consists primarily of sending spent plating and etching solutions
back to the manufacturer of these solutions to be regenerated. Data in
Table 23 indicate that this type of practice has been common. Onsite recovery
of metals from rinsewaters has yet to achieve widespread use.
Some of the methods for onsite minimization of the quantity of hazardous
sludge include sodium borohydride reduction, ion exchange, eletrolytic
recovery, evaporation, reverse osmosis, and electrodialysis. These techniques
for recovering metals from wastewaters have probably become more common since
1981, and new methods are constantly being developed.
Since most spent organic solvents, when contamination is less than
5 percent, are still quite valuable, recovery has been a common method of
management. This is confirmed by the data in Table 24 which indicate that
40 million gallons of solvent waste were URRR by the SIC 36 in 1981. The
majority of this 40 million gallons (70 percent) was URRR offsite. In
contrast the chemical manufacturing industry (SIC 28) employed 87 percent
onsite URRR. One conclusion that may be drawn from this data is that solvents
used in the electronics industry require a high purity which is difficult to
achieve by standard solvent distillation practices. Consequently it is easier
to send these wastes offsite where the majority of the contaminants can be
removed, and the recovered solvent can be used in an application requiring
lower solvent purity.
In recovery of solvents by distillation, there is generation of a bottoms
product containing contaminants and up to 95 percent of the organic solvent.
Secondary recovery of the solution is many times possible through the use of
supplementary technologies such as steam distillation or thin film
evaporation. These methods significantly reduce waste product stream volume
(up to 90 percent of the solvent) and combined with positive photoresists and
aqueous based developers and strippers represent feasible and readily
implemented methods of hazardous waste management.
31
-------
TABLE 23. QUANTITIES OF METAL-CONTAINING WASTE URRR (GALLONS) [Versar, 1985]
to
10
Offsite URRR
2-Diglt
SIC Code
33
36
37
28
49
29
34
34
97
30
Generator
22,751,971
7,953,151
3,090,480
2,277,109
2,361,047
1,246,185
601,926
129,872
0
0
TSD
2,194,224
465,598
3,164,325
851,305
1,422
1,083,679
44,496
278,829
269,011
133,793
Total
24,946,195
8,418,749
6,254,805
3,128,414
2,362,469
2,329,864
646,422
408,701
269,011
133,793
2-Dlglt
SIC Code
37
29
33
28
39
50
30
32
34
31
Onslte URRR
Generator
463,384,736
29,401,566
3,305,980
929,989
4,357,800
0
0
0
155,247
180,545
TSD
202,215
52,751,679
35,267,641
23,164,213
0
2,591,306
994,365
235,232
63,429
0
Total
463,586,951
82,153,245
38,573,621
24,094,202
4,357,800
2,591,306
994,365
235,232
218,676
180,545
-------
TABLE 24. RECYCLING OF SOLVENT WASTES, LISTED BY SICa
SIC
0
1
7
10
14
16
17
20
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
42
47
Recycled offsite
1,260,842 (20)
10,691 (100)
593 (100)
620,232 (100)
0 (0)
9,542 (100)
96,225 (42)
58,642 (37)
385,388 (52)
26,739 (100)
199,085 (53)
605,906 (100)
1,019,037 (78)
1,248,469 (88)
51,677,963 (13)
173,644 (4)
2,742,552 (78)
80,274 (100)
103,335 (1)
1,472,571 (93)
7,386,188 (85)
11,879,873 (98)
27,283,111 (69)
8,161,110 (62)
959,990 (94)
1,727,729 (49)
40,709 (100)
4,066,556 (100)
0 (0)
Waste Volume (gals/yr)b
Recycled onsite
5,151,776 (80)
0 (0)
0 (0)
0 (0)
4,376,901 (100)
0 (0)
130,661 (58)
100,317 (63)
351,977 (48)
0 (0)
180,472 (47)
0 (0)
290,922 (22)
174,255 (12)
361,582,016 (87)
4,041,286 (96)
786,031 (22)
0 (0)
18,035,124 (99)
107,585 (7)
1,332,636 (15)
271,740 (2)
12,288,730 (31)
5,068,141 (38)
57,542 (6)
1,838,095 (51)
0 (0)
0 (0)
34,321 (100)
(continued)
33
Total recycled
6,412,618
10,691
593
620,232
4,376,901
9,542
226,886
158,959
737,365
26,739
379,557
605,906
1,309,959
1,422,725
413,259,979
4,214,930
3,528,583
80,274
18,138,459
1,580,156
8,718,824
12,151,613
39,571,841
13,229,251
1,017,532
3,565,824
40,709
4,066,556
34,321
-------
TABLE 24 (continued)
SIC
49
50
51
73
76
78
80
82
89
95
97
99
Total
Recycled offsite
A, 552, 807 (56)
16,272 (5)
373, 143 (100)
8,087 (0)
6,785 (100)
102,224 (54)
0 (0)
45,978 (100)
5,225 (100)
0 (0)
122,744 (60)
19,322 (34)
128,549,584 (23)
Waste Volume (gals/yr)b
Recycled onsite
3,560,738 (44)
295,541 (95)
0 (0)
3,723,766 (100)
0 (0)
86,478 (46)
3,328 (100)
0 (0)
0 (0)
12,523 (100)
82,029 (40)
36,813 (66)
424,001,745 (77)
Total recycled
8,113,545
311,813
373,143
3,731,853
6,785
188,702
3,328
45,978
5,225
12,523
204,773
56,135
552,551,329
aSource: - Versar, 1985.
Volumes in gals/yr. Numbers in parentheses indicate percentages.
34
-------
SECTION 3
FACILITY A CASE STUDY
FACILITY CHARACTERIZATION
Facility Description
Plant A was founded in 1981 and the Warwick, RI facility opened for
business in June 1985. The company is involved in the development of a
reclaimable product from metal plating waste baths and etchant dumps. Wastes
received at the plant are pretreated to adjust pH and/or remove cyanides, and
then converted to elemental metals or oxide sludges which are sold to smelting
operations in Europe for precious metal recovery. The facility is currently
operating under the precious metal recovery exemption of RCRA since sludge
product is sold for its precious metal content (Ni, Cu, and Au). The
facility's Part B TSD permit is currently being reviewed by Rhode Island DEM
and approval is anticipated shortly.
The treatment/recovery facility is located in a light industrial section
of the city of Warwick, RI. The recently constructed 30,000 sf facility
houses administrative offices, a full laboratory, tank truck and tote (300 gal
containers) unloading facilities, temporary storage (24 hr) for incoming
wastes (4-4,000 gal tanks), raw material storage (100,000 gal), reactors,
clarifiers, and solids handling facilities. Solids generated by the process
are recovered and dewatered on plate and frame filter presses. The dewatered
sludge is currently dried onsite although, during the site testing, offsite
drying was employed.
Waste Sources
Facility A processes concentrated dumps from the metal plating and
printed circuit board industries. These concentrated dumps include alkaline
etchants, acid plating baths, electroless plating cyanide baths, etc. Most of
these wastes fall into the following four categories which provide a logical
basis for segregation at Facility A:
• Acidic metals solutions;
• Alkaline, metals etchant solutions;
• Cyanides; and
• Chelated metals solutions.
35
-------
Average metals concentration in the process feedstock (incoming wastes) are
approximately 12- g/L (12,000 ppm).
Waste Management
The unit operations employed to detoxify the wastes and recover metals
include sodium hypochlorite oxidation of cyanides, pH adjustment, sodium
metabisulfite reduction, sodium borohydride reduction, sedimentation, rapid
sand filtration, dewatering (plate and frame filter press), and ion exchange
columns for effluent polishing. Figure 3 shows the process schematic for the
facility.
Waste Handling and Storage—
Incoming wastes or feedstock are transferred from 4,000 gallon tank
trucks or totes (300 gal capacity containers} in the receiving bay, which is a
fully enclosed multi-lined concrete-epoxy construction facility with
4,400 gallon capacity to contain spills. In order to minimize human error,
four special fittings are provided in the receiving bay Co handle each of the
four wastes described above. When the waste is received, samples are taken
for screening purposes to be compared to the waste anticipated from the
delivering facility. While the samples are being screened the tank truck load
is stored in one of four dedicated temporary storage tanks (less than
24 hours). If the screening results check, the waste is transferred to
short-term raw material storage tanks where it is segregated by the following
metals groupings as well as the four waste categories described earlier:
1. Cu, Ni, and other precious metals
2. Zn, Cd
3. Sn, Pb
4. Fe, Cr
If the screening results do not check properly with the contracted waste
specifications then a full set of analyses are performed and the waste is
classified prior to transferring the waste from temporary storage to raw
material storage.
Facility A has approximately 100,000 gallons of segregated raw storage
capacity. Six tanks provide 35,000 gallons storage for acid wastes, four
tanks provide 25,000 gallons storage for alkaline wastes, five tanks provide
25,000 gallons storage for cyanide wastes, and three tanks provide 15.UUO
gallons storage for chelated wastes. The storage tanks are fixed roof tanks
of concrete polyethylene lined construction. All piping is CPVC or
butt-welded polyethylene construction. All storage tanks are vented to a
building exhaust system equipped with a caustic scrubber.
36
-------
(-•
1
1
1 "
f 1
RA. 1ATIRIALS STORAGE
ACIDS CYANIDES
© ©
SODIUM
HYDROXIDE " »
.
ICKEl REMOVAL CYANIDE DESTRUCTION SODIUM HYDROXIDE
PRETREATMENT ~f PRETREATMENT "^ SODIUM HYPOCHLORIDE
• : !
^ 0.45 b FILTER •«•_«••
f+
1
1 1
NICKEL
SLUDGE HOLDING
TANKS (CLARIFIERS)
"
NICKEL
SLUDGE PLATE AND FRAME
FILTER PRESSES
SLUDGE , ©
OFFSITE SLDDCE
DRYING
r
PRODUCT DRUMMED
AND SHIPPED
KEY:
EMISSIONS — «. «
HASTE (T\
SAMPLING POINTS ^
AIR EMISSIONS m
SAMPLING ^
I
FILTRATE
<
BOROHYDRIDE
TREATMENT
HOLDING TANKS
©
SODIUM BOROHYDRIDE
TREATMENT
,®
BOROHYDRIDE SLDDCE
HOLDING TANKS
(CLARZFIERS)
1
SLUDGE PLATE AND FRAME
FILTER PRESSES
SLUDGE , ®
OFFSITE SLUDGE
DRYING
PRODUCT DRUMMED
AND SHIPPED
1
HYDROCHLORIC ACID * 1
SODIUM HETABISULFIDE BACKWASH 1
SODIUM BOROHYDRIDE 1
RAPID SAND FILTER J
1
CFfLUCNT 1
BACKWASH . ^^ 1
i
© • !
_ „ „_ __ i
1
CATION EXCHANGE ELUTRIATE
BEDS (2) | A
' g7) 1
1
ACID (HCI) ^DISCHARGE
ELUTRIATIOH | TO SEWER
I CONTROLLED 1
EMISSIONS 1
1
CAUSTIC J
SCRUBBER 1
SCRUBBER . .
UASTEWATER I f •
Figure3:. Sampling locations during Facility A Test.
37
-------
Batch Reactor Tanks--
Three batch reactors are used at Facility A to process wastes for
recovery of precious metals. One 15 cubic meter (3,963 gals) reactor is used
for cyanide waste treatment and an identical reactor is dedicated to metal
waste treatment. A larger 30 cubic meter (7,926 gallons) reactor is used for
treatment of both acidic and alkaline wastes. The reactors are of
polyethylene-lined concrete fixed roof tanks fitted with small manways,
agitators, and large probes for measuring pH and oxidation-reduction
potential. In addition, the reactors are equipped with chemical feed parts
for automatic metering of chemical reagent. All reactor emissions are
evacuated to the central exhaust system which is controlled by a caustic
scrubber.
The chelated waste will be handled separately from other "waste streams 'to
avoid recombination of the chelae ing compounds with other metal ions.
Chelated waste will be pretreated in approximately 4,000 gallon batches to
further reduce the chelated metal complex or tie-up the chelate. For example,
Facility A uses lime treatment to break EDTA complexes to form calcium EDTA
salts. Chelates based upon quadrols and citrates can be reduced with sodium
borohydride. The treatment technique applied is highly dependent on the metal
ion complexed, the chelating agents used, metal ion concentration, and pU of
the concentrate. Facility A initially conducts bench scale tests to determine
optimum treatment methods for each chelated waste stream and uses the
prescribed technique each time that it is received. Once pretreatment is
completed, the waste is further treated with sodium borohydride as discussed
below for acid/alkaline wastes.
The cyanide waste reactor generally serves as a pretreatment step to the
sodium borohydride step which is carried out in the acid/alkalai reactor.
Cyanide wastes are treated by alkaline chlorination in the cyanide reactor
prior to metal recovery. Batch sizes similar to the chelated wastes
(approximately 3,200 gallons) are pretreated to reduce CN concentrations to
less than 1 mg/L. Depending upon CN concentrations, batch reaction time
ranges from 3 to 12 hours. The reaction steps proceed in the following order:
1. Adjustment of pH to 11 with sodium hydroxide.
2. Sodium hypochlorite (NaOCl) addition; the reaction is controlled by
maintaining oxidation-reduction potential at +400 mv.
3. Cyanide levels checked periodically until CN level is less than
1 mg/L.
4. Ferrous sulfate (FeSO^) added to remove surplus chlorine (Cl^).
5. Pretreated waste is pumped to the acid/alkalai reactor for further
treatment.
Most process batches at Facility A are eventually transferred to the
acid/alkalai reactor for sodium borohydride treatment. Batch sizes in the
acid/alkalai reactor are approximately 6,340 gallons (24 m^). Reaction
times are generally 2 to 2-1/2 hours for normal acid/alkaline wastes and 3 to
3-1/2 for chelated wastes. The reaction steps begin with pH adjustment which
is usually accomplished in part by combining acid wastes with ^alkaline wastes
in the reactor. The treatment steps are:
38
-------
1. Adjustment of pH to 6 with sodium hydroxide or sulfuric acid.
2. Sodium metabisulfice addition (200-300 Ibs per 24 cubic meter batch)
for pre-reduction.
3. Sodium borohydride addition (10 percent solution, 100 L per batch)
with ORP and/or batch color monitored. When batch turns black, URP
is approximately -100 to -400 mv and reaction is complete.
During the above treatment significant quantities of hydrogen (H2) gas
may be liberated by foaming which takes place in the reactor. When the
reaction is complete the reactor contents are pumped to clarifiers which are
used as holding tanks prior to sludge dewatering.
Clarifiers (Sludge Tanks)—
Facility A employs three 30 cubic meter rectangular sedimentation tanks
for holding the reaction products from the reactor tanks while they are being
fed to the plate and frame filter press. The clarifiers were originally
designed to separate the reaction products (solids) from the aqueous
supernatants (clarifier overflow). However, since the original precipitation
reaction design was modified to borohydride reduction, the clarifiers are no
longer required because the filter presses can easily dewater the entire batch
volume. This is partly due to the large particle size of the agglomerated
sludge resulting from borohydride treatment.
Plate and Frame Filter Press—
Due to the plant modifications discussed above, the entire contents of
the sludge holding tanks (clarifiers) are pumped to one of two plate and frame
filter presses. Each press is capable of processing approximately 800 liters
of sludge per hour. Metal sludge produced on the filter press should be in
excess of 50 percent solids and 25 percent metal. From this point the sludge
was previously shipped offsite for drying to reduce the moisture content from
approximately 50 percent to 30 percent water. Since the completion of the
testing program, Facility A has installed onsite infrared sludge dryers.
Rapid Sand Filter—
The filtrate from the sodium borohydride plate and frame filter press is
fed to the rapid sand filter. A single media rapid sand filter, rated at
10 cubic m/nr, provides some additional effluent polishing in the event of a
sludge filter press failure and serves mainly to protect the cation exchange
columns.
Cartridge Filtration (Prefilter)—
A cartridge filter, rated at 10 m^/hr, is used for polishing the
discharge from the high pH (nickel pretreatment) plate and frame filter press
filtrate prior to the sodium borohydride treatment step. The filter employs a
design pore opening of 0.45 Mm.
-------
Cation Exchange Columns—
Two cation exchange columns in series serve as final polishing seeps for
the plant effluent prior to discharge to the City of Warwick, RI sewer
system. The cation exchange columns are periodically elutriated with HC1
which generates an acidic metals waste which is recycled to the acid/alkalai
reactors for treatment.
Waste Characterization/Process Monitoring
Waste screening analyses are performed in an onsite laboratory as
discussed earlier in the process description. Key process parameters which
are monitored near the end of each batch include:
• pH
• ORP
• Cyanide concentration
• Metals concentration (total and dissolved)
• Temperature
• Color (visual)
• Moisture content (percent by weight)
Periodic monitoring of all permitted discharge limits are also
conducted. In addition, the facility is equipped with cyanide alarms to warn
of airborne cyanide concentrations buildup within the facility.
PROCESS TESTING AND RESULTS
On December 11 and 12, 1985, GCA conducted field studies to evaluate
sodium borohydride waste treatment/reduction processes at Facility A. As tne
residuals (dried sludge) from the processes are sold for their precious metal
content, this technology significantly reduces wastes which would otherwise be
landfilled.
Test Deviations and Changes
Facility A plant process modifications required GCA to make alterations
to the originally proposed test program. Under the original test program,
samples were to be collected around unit process and from two process batches
which would then be combined to make a single sodium borohydride process
sludge sample. A process change (i.e., adding a sodium hydroxide
precipitation pretreatment step) created an additional process reactor and
sludge product to be sampled. In addition, this change dramatically increased
the length of time (number of batches) necessary to fill the sodium
borohydride sludge filter press. Thus, while the test plan originally called
for sampling all streams from a single process batch in one 8-hour shift, the
new process mode would have required nearly a week of sampling to obtain
single batch data.
40
-------
1C was noted chat because all of Che batches come from the same feedstock
tanks, sampling from different batches was presumed to be fairly
representative. Thus, it was agreed that Che investigators should spend an
additional day sampling Co collect as much daCa from a single batch as
possible. However, ic was also agreed that data from separate batches would
be acceptable Co Che program requirements. As a result, the Facility A
sampling program collected data from four process batches.
Results
In order to assess sodium borohydride as a viable waste treatment sludge
reduction alternative, its effectiveness in meeting effluent requirements and
obtaining low sludge volumes was evaluated. The parameters of interest are:
trace metals, TOC/TOX, cyanides, and hexavalent chromium. Each parameter was
examined for reduccion efficiency, sludge content, and regulatory compliance.
Other parameters of interest examined were air emissions, the effectiveness of
the cyanide destruct system and the metals content in the sodium hydroxide
sludge (Sample Point 5).
Trace Mecals—
Reduccion is defined as Che gaining of electrons by an atom, an ion, or
an elemenC thereby reducing its positive valence. The success of its metal
reduction is highly dependent on the mixing, residence time, and other process
conditions such as: pH, temperature, concentration, and reaction kinetics.
The purpose of the trace metals analysis is to evaluate sodium borohydride1s
effectiveness in the reduction of a mixed metal influent to a low volume, high
density sludge. The streams of interest are: the borohydride reactor
influent (Sampling Point 6), Che borohydride reacCor effluent (Sampling Point
7), and the borohydride sludge (Sampling Point 8).
1C was initially proposed that a mass balance would be developed across
Che whole borohydride reduccion process through the sludge filcer press and
its effluent. However, due Co apparent variations in batch compositions and
problems with the borohydride sludge press operation, it was necessary to use
the results form Batch 3 (85-12-1009) to assess sodium borohydride
effectiveness. The data from Batch 3 were particularly useful since effluent
samples were collected and filtered after treatment which in effect simulated
solids removal achieved in the sludge filter press.
Processing for Batch 3 cook place in the 24 cubic meter sodium
borohydride finishing reactor (acid/alkali) at ambient temperature and
atmospheric pressure. The total metals loading in Che batch reactor influent
was 7.25 kg per batch, of which over 83 percent was divalent copper. The
theoretical level of sodium borohydride (SBH) required for the total reduction
of all metals was 8 kg (58 licers of a stabilized aqueous solution of
1.2 percent SBH and 4.1 percenC causcic soda). The resulcs presenced in
Table 25 summarize plane operacions during testing. The actual sodium
borohydride solution usage was 9.8 kg (70 liters of 1.2 percent SBH). This
represents an accual/cheorecical SBH addicion ratio of 1.2, which falls well
within the range of 1.0 Co 1.5 reported in literature. Excess SBH is normally
required due Co nonoptimum reaction conditions and side reactions with other
species such as aldehydes, ketones, nitrates, peroxides, and persulfates.
41
-------
TABLE 25. SODIUM BOROHYDRIDE FINISHING REACTOR
PROCESS DATA
Treatment
pH ORPa
Commen t
10.5 -250
Start
AdJ 216 liters HCL 5.5 150 Lower pH
Add 200 Ib NaHS03 5.2
130
Stabilizer
(prereductant)
Add 60 liters NaBfy 8.0 -770 Reduction
Add 25 liters Flexon 7.8 -770 S02 suppressant
Add 10 liters
8.4 -830 Finish
42
-------
Sodium borohydride has a reducing capacity of 8 electrons/mole or an
equivalent weight of 4.75 g/molar electron and a standard electrochemical
potential of -1.25 volts [Purdue Research Foundation]. Table 26 compares SBH
reactor influent data with both filtered and nonfiltered effluent data for
8 selected metals. As expected, the nonfiltered effluent data demonstrate
little or no reduction due to the effluent solution becoming resaturated
during analysis for total metals. The filtered sample shows reduction with
efficiencies ranging from 16 to 99.8 percent. This wide range of reduction
efficiencies is likely a result of the concentration and chemical potential
(activity) of each of the metallic ions contained in the solute.
Table 27 summarizes the reduction efficiencies of each of the selected
metals in the filtered sample as a function of concentration and
electrochemical potential. Analysis of the results show that the
concentration of the metallic ion is often the determinant in reduction
efficiency. However, given equivalent concentrations, the ability of a
metallic ion to achieve equilibrium may be measured by its standard free
energy or chemical potential. An example of this behavior was exhibited by
lead and chrome which, under test conditions, have similar concentrations but
divergent electrochemical potentials. The resultant 40 percent drop in
reduction efficiency for trivalent chrome as compared to lead may be directly
attributed to the greater quantity of free energy (approximately six times)
required for chrome to achieve elemental form.
Analysis of filtered effluent showed that overall, of the 7.25 kg of
mixed metals, approximately 6.91 kg were reduced to elemental form. This
represents an overall reduction efficiency of 95.4 percent of total mixed
metals. The remainder of the metals influent loading (0.337 kg), of which
70 percent was calcium, was of sufficient quality that given efficient post
treatment, effluent limitations should be achieved.
A second objective in assessing sodium borohydride as a viable waste
treatment alterative is the ability to form a low volume, high density
sludge. An earlier study on hazardous sludge reduction (Centec Corporation)
reported that substitution of SBH treatment for lime treatment of mixed metal
wastewaters can result in a 68 percent sludge reduction. In addition, it has
been reported [PC FAB, May 84] that SBH reduction sludges typically contain
80 percent or more metals. These results compare favorably with the metals
content of hydroxide-lime sludge which generally contains less than 20 percent
metals. Since Facility A utilized both SBH and sodium hydroxide reduction, a
trace metals analysis was conducted to determine sludge loading
characteristics in each case.
The sludge samples collected from the nickel/cyanide sludge plate and
frame filter press (Sample Point 5) were analyzed for 17'trace metals. The
analytical results for these sludge samples (on a dry weight basis) are
summarized in Table 28. The feed to the filter press consisted of the entire
contents of the clarifier holding Batch 1 (85-12-1007) and Batch 2
(85-12-1008). These batches in turn consisted of the effluent from the
cyanide reactor (Sample Point 2) and the effluent from the nickel pretreatment
reactor (Sample Point 7) which included filtrate from the nickel/cyanide press.
43
-------
TABLE 26. SODIUM BOROHYDRIDE FINISHING REACTOR TRACE METALS
CONCENTRATIONS AND REMOVAL EFFICIENCIES
Element
Ag
Au
Cd
Cr
Cu
Ni
Pb
Zn
Reactor
influent (mg/L)
24.0
5.7
0.015
0.031
237.0
0.96
0.32
5.10
Reactor
effluent (mg/L)
un filtered8
6.2
4.76
0.01
0.03
207.0
0.902
0.31
4.76
Reactor
effluent (mg/L)
filtered b
0.06
0.15
0.01
0.026
0.47
0.422
0.14
0.79
Percent
of
removal
99.7
97.0
c
16.1
99.8
56.0
56.2
98.4
aNonfiltered sample.
^Filtered onsite at Plant A's Lab. In addition, a blank DI water sample was
filtered at the Plant's Lab onsite as a QC measure. Results for that sample
showed less than detection limits in all cases.
cUnable to obtain adequate precision.
44
-------
1ABLE 27. FACILITY A SLL'DGE CHAKACTiKlZATKi;. KisLLT
Dry weight EP Toxicity
concentration results
(percent) (mg/L)
Element
Ag
As
An
Ba
Ca
Cd
Cr
Cu
Fe
Mg
Ni
Pb
Rh
Se
Sn
Tl
Zn
Total
aSource
Federal
Ni/CN
sludge
0.019
0.004
0.134
0.001
0.205
0.008
0.029
11.000
0.720
0.042
19.400
0.130
0.865
0.003
0.305
2.840
35.705
SBH Ni/CN SBH
sludge sludge sludge
0.017 0.03 O.U6
0.017 0.04 0.05
0.328
0.001 0.224 0.163
0.089
0.001 0.589 0.016
0.003 0.294 0.032
5.250
0.049
0.003 0.0018 0.0022
0.293
0.046 4.6 0.03
0.050
0.001 0.04 0.04
0.015
0.361
6.624
EP Toxicity
standards3
(mg/L)
5.0
5.0
100.0
—
1.0
5.0
0.2
5.0
1.0
U.S. Environmental Protection Agency
Register V. 45 No. 98 98: 33122 May 14, 1980
-------
TABLE 28. REDUCTION EFFICIENCIES AS A FUNCTION OF
CONCENTRATION AND ELECTROCHEMICAL
POTENTIALS IN BATCH 3
Element
Cu
Ag
Au
Zn
Ni
Pb
Cr
Cd
Concentration
(fflg/L)
237.0
24.0
5.7
5.1
0.96
0.32
0.31
0.15
Electrochemical
potential
0.3402
0.7996
1.42
-0.7628
-0.23
-0.1263
-0.74
-0.4026
Reduction
efficiency (Z)
99.8
99.75
97.37
98.45
56.04
56.25
16.13
-
-------
Analysis of the sodium hydroxide (NaOH) and SBH sludges yielded total
metals dry weight fractions of 35.7 and 6.6 percent, respectively. Neither
sludge results supported Facility A's claim of 60-70 percent metals on a dry
basis. While Che SBH sludge result was significantly below performance
expectations, the exact cause of these results was not discernable. Possible
explanations include: 1) a possible process upset; 2) sampling error; or
3) analytical error. It seems most probable that a process upset was
responsible for these results, since blinding of the sludge press did occur on
the SBH press. Based on the results of other case studies on SBH reduction
conducted under this program, it is reasonable to assume that these results
are not representative, since typical sludge metals contents should be greater
than 70 percent.
Table 4 also presents Facility A sludge EP toxicity leachate results for
both the nickel/cyanide and sodium borohydride reactor sludges. The results
of the tests clearly show that for Facility A influent metals concentrations,
the sodium borohydride sludge is fairly stable in that its leachate
characteristics are below EP toxicity limits for all metals. However, note
that the waste is still classified as F006 hazardous waste.
An additional objective of this study was to demonstrate Plant A
treatment system compliance with final effluent limits. Once Che filtrate
leaves the SBH plate and frame filter press it is fed to a single media rapid
sand filter (Sample Point 10) to provide some additional effluent polishing.
Final polishing is performed in cation exchange columns (Sample Point 11)
prior to discharge to the city of Warwick sewer system. Samples were
collected at Sample Point 11 for final effluent from Batches 1 and 3.
Table 29 presents metals concentration results in comparison with local
pretreatment effluent limits. In both samples, the quality of the effluent
was inadequate to meet local effluent limits.
In an effort to remedy this problem, Facility A revised its waste
processing sequence in the following manner:
• Incoming noncyanide wastes have been adjusted to pH 7.5 with sodium
or magnesium hydroxide.
• Copper, zinc (refinery brass) and trace amounts of cadmium and
silver are reduced and pumped into filter press one.
• The filtrate from filter press one is transferred back to the
reaction tank and the pH is adjusted to 11.5 with NaOH, thus
precipitating any nonreduced metals as the hydroxide.
• Nickel and trace amounts of heavy metals are reduced and collected*
in filter press two.
• The filtrate from filter press two is transferred Co the SBH
finishing reactor and precious metals such as gold, platinum,
palladium, and chrome are reduced.
• The reduced metals precipitates are then removed through bag and
micron filters prior to final effluent polishing in Che rapid sand
filter.
46
-------
TABLE 29. PLANT A LOCAL PRETREATMENT
EFFLUENT COMPLIANCE DATA
Element
Ag
As
Cd
Cr
Cu
Ni
Pb
Se
Zn
Total metals
Batch 1
0.05b
0.62C
0.01
0.29C
0.146
0.767C
0.14
0.2
0.98C
concentrations
Batch 3
0.05b
0.29C
0.01
0.03
1.82C
0.861C
0.1
0.2
0.05
(mg/L)
Effluent
1 imi ta
O.U3
0.01
0.05
0.2
0.4
0.5
0.15
1.0
0.5
aCity of Warwick, RI effluent limits.
"Unable to obtain adequate precision.
cExceedence of effluent limits.
47
-------
In addition, Facility A has instituted the use of a QC holding tank
following the cation exchange columns to prevent column breakthrough. In this
manner Facility A is able to prevent any discharge to the sewer that might
exceed effluent limits. Since testing was completed with these revisions of
the process sequence, Facility A's effluent has been tested by the local sewer
district authority on several different occasions. Since implementation of
these changes, Facility A's effluent quality (based on local sewer district
authority sampling results) has improved considerably and is now consistently
meeting sewer authority guidelines.
Organic Indicator Results—
Total Organic Carbon (TOO and Total Organic Halide (TOX) samples were
extensively collected and analyzed for Batches 2, 3, and 4. Total organic
carbon results are summarized in Table 30 and total organic halide results are
presented in Table 31. As expected, both sample runs display little, if any,
reduction in organic concentrations after being processed through the SBH
finishing reactor. This phenomena may be due to the fact that the sludge
results do show organic constituents being concentrated in the sludge. Total
TOC/TOX concentration in the nickel/cyanide sludge and the SBH sludge were
2.6 and 4.0 percent, respectively. Thus, while the sludge results show some
concentration of organics, influent and effluent results showed little or no
removal.
Cyanides Results—
Total cyanide samples collected and analyzed for Batches 3 and 4 are
summarized in Table 32. The unforeseen presence of distillable organics in
the Facility A process streams may have shown a positive bias in the test
results, particularly in sample points 2 though 13. Results obtained for the
influent and effluent samples collected at the CN destruction reactor are
considered semi-quantitative due to marginal QC recoveries. These
semi-quantitative results showed a reduction of total cyanide from 2.25 to
less than 0.01 ppm cyanide. This represents a greater than 99.55 percent
destruction efficiently for the two streams tested. Analytical spike recovery
data for these cyanide results were only marginally acceptable. The cyanide
reactor effluent spike showed zero recovery as expected because excess
hypochlorite in the sample destroyed the CN spike.
Uexavalent Chromium--
Due to the complex nature of Facility A's process flow streams, the
investigators were unable to obtain any acceptable hexavalent chromium
results. The results obtained are presented in Table 33. The presence of
complex organics and strong reductants were apparently the cause of the poor
hexavalent chromium precision and accuracy results.
Process Emissions—
In addition to assessing wastewater effluent characteristics, the testing
program was designed to evaluate uncontrolled process air emissions. Table 34
summarizes the results of grab sample and integrated sample analysis of
process reactor exhaust ducts based on Drager tube analysis. The emission
results given in Table 34 show a continuous presence of hydrochloric acid and
48
-------
TABLE 30. SUMMARY OF TOTAL ORGANIC CARBON RESULTS (mg/L)
FACILITY A - WARWICK, RI
No.
1
2
3
4
5
12
6
£ 7
8
9
10
11
13
Sample point
Description (85-12-1007)
CN Reactor Influent
CN Reactor Effluent
NI Pretreatment Influent
NI Pretreatment Effluent
NI/CN Sludge
NI/CN Press Filtrate "
Borohydride Influent
Borohydride Effluent0
Borohydride Sludge
Borohydride Filtrate6
Sand Filter Effluent
Ion Exchange Resin Effluent
Ion Exchange Resin Elutriate
Hat>*»l« 9 DM ft- M It 1
Dflccn / t>a ten J
(85-12-1008) (85-12-1009)
729
25,100
468
511/528b
38,600
632
722
618
n - *. _ u /,
oaccn *4
(85-12-1010)
633
587
aSampled after Filter Press and 0.45 filter.
''Includes sludge from batches 2 and 3.
cln addition to effluent sample, a separate sample was collected and filtered onsite at the Plant lab.
As a QA measure, a DI water blank was also filtered by Plant A.
"Includes sludge from batches 1 and 2.
eSampled after sand filter and 0.45 filter
-------
TABLE 31. SUMMARY OF TOTAL ORGANIC HALIDE RESULTS (mg/L)
FACILITY A - WARWICK, RI
No.
1
2
3
4
5
12
6
7
8
9
10
11
13
Sample point
Description (85-12-1007)
CN Reactor Influent
CN Reactor Effluent
NI Pretreatment Influent
NI Pretreatment Effluent
NI/CN Sludge
NI/CN Press Filtrate3
Borohydride Influent
Borohydride Effluent0
Borohydride Sludge
Borohydride Filtrate6
Sand Filter Effluent
Ion Exchange Resin Effluent
Ion Exchange Resin Elutriate
4
D M k n 1* O D M * M L* 1 11 « * M iv A
Batch Z Batch J Batch 4
(85-12-1008) (85-12-1009) (85-12-1010)
34
6.5 13
1,000
23
17/13b
1,500
26
39
20
aSampled after Filter Press and 0.45 filter.
Includes sludge from batches 2 and 3.
cln addition to effluent sample, a separate sample was collected and filtered onsite at the Plant lab.
As a QA measure, a DI water blank was also filtered by Plant A.
Includes sludge from batches I and 2.
eSampled after sand filter and 0.45 filter
-------
TABLE 32. SUMMARY OF TOTAL CYANIDE RESULTS FACILITY3
Concentration (
-------
TABLE 34. SUMMARY OF DRAGER TUBE ANALYSIS RESULTS FOR UNCONTROLLED
PROCESS AIR EMISSIONS3
Parameter
Hydrogen Cyanide
Hydrogen
Sulfur Dioxide
Hydrogen Sulfide
Ammonia
Hydrochloric Acid
Gas Concentrations (ppm
Grab Sample Results
< 2 ppm
1.7 - 6.0d %
< 1 - 20 ppm
< 1 ppm
< 5 - 180 ppm
< 1 ppm
or % as noted)
Integrated
Sample Results
< 2 ppm
0.4 %
< 1 ppm
< 1 ppm
< 5 ppm
2 ppm
Threshold
limit value
(TLV)
short term
exposure
limitb
10 ppmc
—
5 ppm
15 ppm
35 ppm
5 ppmc
a Drager detector tubes are compound-specific for the parameter indicated.
Accuracy is estimated at ±5-20% of reading. Test conditions were as follows:
o Flowrate * 3,600 afpm
o Duct diameter 6 inches
o Duct area = 22.274 in2 or 0.196 ft2
o Volumetric flowrate at actual conditions = 0.196
ft2 X 3,600 afpm = 706.86 acfm.
bSource: ISBN 0.936712 - 61-9, 1985.
cTime weighted average value used in lieu of short term exposure limit.
d Five pump strokes were required (10 strokes standard) to reach saturation
concentration of 3%, thus extrapolated reading is 3.0% (10) = 6.0%
52
-------
'Hydrogen gas accompanied by occasional presence nf a-mnonia and sulfur
dioxide. Grab sample concentrations for ammonia and sulfur dioxide exceeded
adopted short term exposure limits (ACGIH, 1985) for these substances. One of
the hydrogen emissions grab sample results (6.0 percent) is significant since
this value is greater than the lower flammable limit for hydrogen
(4.0 percent). This is primarily due to hydrogen gas being evolved during SbH
treatment and is likely to be a function of the pH of the wastewater. This
problem may be eliminated through optimization of the treatment process and
should remain a design consideration for new applications.
ECONOMIC EVALUATION
As previously discussed, a sodium metabisulfite/borohydride was used at
Facility A to complex process solutions to metallic form. A primary obstacle
to the more widespread use of sodium borohydride has been its high cost.
Table 35 summarizes cost and performance of various reduction chemistries as a
function of pound of copper removed. Lime/ferrous sulfate is the least
expensive reducing agent, but will increase sludge generation by at least
68 percent relative to sodium bisulfite/borohydride [Centec Corporation). In
addition, lime/ferrous sulfate sludge is difficult to sell and refine due to
its low metals content (5 percent metals) and high gypsum content. An
inability to sell the sludge product would result in a RCRA permit violation
since Facility A currently operates under a precious metal recovery
exemption. Therefore, sodium borohydride was the reductant of choice since
its precipitant will yield finely divided metals which are easily recovered.
Chemical costs for SBH reduction typically range from $6.80 to $17.00 per
pound of copper removed depending on the actual to theoretical usage ratio
(1.0 to 5.0).
Initially, Facility A operated its sodium metabisulfite/borohydride
reduction reactor with plexon, an additional reducing agent. Plexon is a 12
weight percent dimethyl dithiocarbamate solution used to reduce nickel and to
lower sulfur dioxide emissions. However, a relatively high cost of $8 per
liter rendered the entire reduction process impractical at $19.80/lb copper
removed and also decreased sludge loading characteristics. By deleting plexon
from the reduction reaction, Facility A was able to decrease chemical costs
63 percent to approximately $7.30 per pound of copper reduced. This cost
figure compares favorably to the chemical cost of $6.00 per pound of copper
reduced, achieved in the Facility B case study (discussed in Section 4). The
Facility B reduction process operated at a 1.8 actual/theoretical sodium
borohydride usage ratio which resulted in a 99 percent reduction of the copper
influent loading.
Further process optimizations at Facility A have included the 2-stage
reduction process prior to the SBH finishing reactor and the installation of
onsite infrared drying ovens. These changes have reportedly resulted in a
sludge product that is 90 percent solids and 60 percent precious metals on a
dry weight basis. In addition, chemical costs have decreased to approximately
$6.00 per pound of metal reduced since only a small volume of sodium
borohydride solution is required.
53
-------
TABLE 35. ECONOMIC COHPAKISON: PRINTED CIRCUIT oOAAD WASTE*ATtR
TREATMENT
Treatment
FeS04 b
VenMet™Solutionc
Facility Ad
Facility Ae
Facility B
Ib chemical
Per Ib Cu reduced
9-44
5.7 - 14.2
21.7
13.9
2.0
Chemical costa
$/lb Cu reduced
0.9 - 4.4
6.8 - 17.0
19.8
7.27
3.5
Effluent
Quality (ppm)
2.7
1.0
0.47
-
0.26
aChemical Costs Source: Ventron Technical Brochure
$0.1/lb FeS04 • 7H20
$0.44/lb 40% Sodium dimethyIdithiocarbonate solution
$2.40/lb VenMet Solution
$0.25/lb Sodium Bisulfite
"Source: Ventron Technical Brochure
cSource: Ventron Technical Brochure, contains 12% by weight sodium
Borohydride and 402 NaOH, as well as 3.2 - 8.0 Ibs Sodium Bisulfite/lb
Cu reduced
^Measured during testing, contains VenMet solution, sodium bisulfite, and a
12% sodium dimethyldithiocarbonate solution (plexon)
eRevised system containing VenMet solution and sodium bisulfate
^Measured in separate case study, contains VenMet solution. Low organic
loadings in Facility B reactor influent stream made low chemical usage
possible.
54
-------
SECTION 4
FACILITY B CASE STUDY
FACILITY CHARACTERIZATION
Facility Description
Facility B is a manufacturer of printed circuit boards based in Santa
Ana, California. Ten years ago Facility B began operating on a job shop or
contract basis. The company currently employs 77 people, operating two
shifts, 5-1/2 days/week. The facility has now been in their present location
for 5 years and are planning an expansion of their operations. Printed
circuit board production is approximately a half-million square feet/year,
generating $7 million of gross sales.
Several years ago Facility B was discharging to the sewer as much as
40 Ib/day of untreated chelated and particulate copper. In April 1984,
legislation was introduced that would require printed circuit board
manufacturers to limit their effluent streams to 2.7 ppm copper. In response
to this pending legislation, Memtek Corporation in Woburn, Massachusetts was
hired by Facility B to perform a study to determine an appropriate waste
treatment system. As a result of this study, the facility installed a sodium
borohydride reduction and membrane ultrafiltration waste treatment system in
February 1983.
Facility B's wastewater discharge is permitted (Class I wastewater
discharge permit) and sampled on a quarterly basis by the Orange County
Sanitation District. If any of the effluent limitations are exceeded,
sampling is performed more frequently and corrective actions are taken.
Waste Sources
The three main methods of printed circuit board production are the
additive, subtractive, and semi-additive techniques. Additive techniques
involve the production of printed circuit boards through electroless plating
on unclad board materials. Subtractive involves the removal of large amounts
of copper foil from clad board material to create the desired circuit
pattern. The process used at Plant B is a hybrid between the two
aforementioned methods called semi-additive. Figure 4 illustrates the Plant
process and Table 36 presents some of the chemicals used in various steps of
the process. The waste stream of interest is metals containing wastewaters
from the rinses following the etching and plating operations as well as
production bath dumps. Consideration of the following general and specific
process areas can assist in evaluating metals sources within the production
process.
55
-------
KUJO MirMATlu*
1) CUT 10»IUII
2) OUUIK1I".
(IWFACE NtrUUTlM
1) t NCOlWlCAL I1NIM
1) KEQUMUU. IIUirilMC
I
CAIALTR Arrucxriw
CLICTUIXli PtATlne
i) cofrti run
2) fLUC HOLIl
))
rosiTivt MMZ nuwsru
i) siu laiwiNC «m
morouiin fwirrws
:j TDUOI UF
rrcxuc
eurue onotint ROIMT
nut moiouitn
SOLDIt HASK
i) nurtu roi
2) SOUtl HASK
FLATINC
or Tin
HTDIO saua
lUOHHt
rik*L PUCISSIHC
1) 1NSKCT10*
:> wAitUHC
3) CUTTING
.) TMTIHC
1) rtNAL
Figure 4. Plane B process flow diagram.
56
-------
TABLE 36. CHEMICALS USED IN PLANT B'S PROCESS
Process step
Chemicals used
Catalyst Application
Electroless Plating
Electroplating
Image Transfer
Etching
Hydro Solder Leveling Machine
Palladium Chloride (PdCl2)
Hydrochloric Acid (HC1)
Stannous Chloride
Copper Sulfate
Formaldehyde (CH20)
Sodium Hydroxide (NaOH)
Ammonium Persulfate (NH^,
used only occasionally to clean
electroless line
Copper Sulfate (CuStfy)
Sulfuric Acid (H2S04)
Caustic Soda, liquid and anhydrous
(NaOH) - used as a developer
Sodium Chloride (NaCl)
Sodium Chlorate (NaCl03>
Muriatic Acid (HC1)
Chloroe thane (C
57
-------
Drilling and Debarring—
During board preparation, the boards are sawed into blanks slightly
larger than is needed for the final product to allow for tabs and board
finishing. After mechanical or chemical cleaning the typical double-siaed or
raultilayered board is drilled by numerically controlled high speed spindle
drills. The resulting holes and board edges are deburred by rotating crushes
to remove any loose particulate matter or rough edges detrimental to
subsequent chemical processing.
Electroless Copper Deposition, Rinsing, and Neutralization—
After chemical cleaning and rinsing (to remove any dirt or surface oils),
the boards are catalyzed through the application of a thin layer of stannous
and palladium chloride. The stannous chloride layer is removed prior to
electroless plating by a mild fluoroboric acid solution (accelerator). This
removal exposes the palladium chloride ion which acts as a catalyst in the
subsequent electroless copper reaction. The electroless copper reaction
typically deposits a thin (25 to 85 micro-inch) layer of copper on the board
surface and in the drilled holes. This metallization provides electrical
contact between the surfaces and layers of the printed circuit board. After
electroless copper deposition, the boards are thoroughly rinsed and then
neutralized with a mild sulfuric acid solution.
Electrolytic Plating—
In order to ensure a uniform electrical conductivity, i.e., no breaks or
voids in the copper layer, a 1 to 2 mil deposit of electrolytic copper is
deposited on the thin electroless layer. The general processing procedure is
to activate the board surface with hydrochloric acid (to remove any surface
contaminants), plate, clean/rinse, and replate. Acid copper plating baths
contain sulfuric acid, copper sulfate pentahydrate, organic brighteners, and
50 to 70 ppm of chloride ions. Deposition takes place through the reduction
of cupric ions by electrical current which flows through the cell from anode
(phosphorized copper bars) to cathode (plating surface).
Positive Image Transfer—
Image transfer involves the production of a circuit pattern on a
metallized board surface with an ultraviolet light sensitive organic polymer.
In order to create a positive image, the photoresist is first applied directly
to the copper surface by a hot roll laminator or silk screening. Then a
stencil of the artwork is exposed to ultraviolet light while under vacuum to
produce the exact circuit pattern. Upon exposure, the photoresist surrounding
the polymerized circuit pattern becomes soluble. A caustic soda solution is
used to develop the photoresist and removes any nonpolymerized material. The
remaining nonsoluble photoresist, i.e., the circuit pattern, is now a chemical
inhibitor and acts as an etch resist.
Etching and Resist Strip—
Cupric chloride is used as an inexpensive final etch process for boards
without metallic etch resists. Its main constituents are cuprous chloride,
sodium chloride, sodium chlorate, muriatic acid, and water. The overall
reaction is:
-------
CUCL2 + CU J=i 2CUCL
This etching solution will remove all unwanted or excess copper from the
board, that is "ffot protected by the polymerized photoresist. Following the
etch process, the boards are rinsed (an important source of metallic
contamination), and then immersed in a high temperature, alkaline photoresist
stripping tank containing butyl cellosolve acetate as the active ingredient.
Solder Mask—
A solder mask or resist is a polymer coating which is applied to a
printed circuit board to prevent molten solder from adhering to preselected
areas. Solder resists act as a protective coating, preventing harmful
elements from degrading the circuits. In addition to physical protection,
solder masks also serve as an electrical insulator. The resist liquid is
applied through silk screen printing prior to curing, which fully crosslinks
the resist polymers to achieve proper end use characteristics.
GoId/Nickel Microplating—
The section of the printed circuit board that contacts the main assembly
is known as the gold edge connector. The edges are chamfered to allow easy
insertion and are designed and manufactured for maximum conductive and
corrosion resistance properties. Gold tends to form an intennetallic layer
with copper (changing its properties) and is too ductile for most
applications. Therefore, after a mild activation step, a 50 to 100 micro-inch
layer of nickel is plated over the copper to act as a hardening agent and
prevent the migration of copper molecules. Gold is then plated in a potassium
cyanide bath containing organic brighteners to provide the final protective
coating on the connector edges. Both nickel and gold baths are examples of
electrolytic plating and are followed by rinses. The gold, however, is
recovered directly in process due to its expense.
Hot Air Leveling—
The selective solder coating/hot air leveling process involves applying
an eutectic solder coating onto the copper areas not covered by the solder
mask. Prior to application, the gold edges are masked with tape Co prevent
solder adhesion, the board is precleaned and then thermally conditioned/
activated by a water soluble flux. As the board exits from the solder, it
passes through two heated, horizontal air knives, producing a quality,
selected deposit. After final rinsing and solvent cleaning using
1,1,1-trichloroethane (to remove any residual flux), the board goes to final
inspection.
Waste Management
As discussed above, the main sources of metallic contamination to the
wastewater stream emanate from the rinses following plating or etching
operations. Copper contamination is confined to the rinses following copper
chloride etching, electroless and electrolytic plating, and the activation
baths on the electrolytic and microplate lines. Nickel contaminants are
introduced solely from the rinse following the electrolytic nickel plating
operation on the microplate line. While lead contamination is introduced into
Che waste stream in the rinsing and cleaning operation following hot air
leveling.
-------
The purpose of this case study is to evaluate metals containing hazardous
sludge (RCRA code F006) minimization technologies. The process of interest is
the semicontinuous, sodium borohydride ultrafiltration system. The primary
operation is the precipitation of heavy metals through the use of a strong.
reducing agent (sodium borohydride) followed by liquids-solids separation
through membrane ultrafiltration and sludge filtration. A brief description
of the two technologies is presented below.
Sodium Borohydride Reduction—
Sodium borohydride (SBH) is a strong reducing agent and provides a simple
and efficient method of metal precipitation and recovery. SBH is able to
reduce metal contaminants to their elemental form which results in a low
volume, high metal content sludge. In addition, the use of SBH promotes good
settling characteristics which minimize the need for a flocculant.
An understanding of the chemistry associated with the use of SBH is
helpful. The basic reduction reaction involves the donation of
8 electrons/molecule of SBH to an electron deficient metal cation. The
following half-reaction occurs when SBH is added to an aqueous effluent:
NaBH4 + 2H20 ?=* NaB02 + 8H* + 8e- (1)
If this reaction takes place in the presence of metal cations, reduction
occurs according to the following reaction:
8M* + 8e T— 8M° (2)
If there are no inorganic or organic reducibles, hydrolysis takes place:
NaBH. + 2H,0 =J NaBO + 4H_ (3)
42 2
Combining Equations 1 through 3 yields the overall reaction:
NaBH + 2H20 + BMX ;=? NaB02 + 8M° + 8HX (4)
where: M = metal (valence +1), and
X = anion (chloride, carbonate, etc.).
-------
This reaction is dependent on the following process conditions and
operational parameters: pH, temperature, metal concentrations, the kinetics
of competing reactions, agitation, residence time and the method of
liquid-solid separation. The pH should be in the slightly alkaline range.
Though reaction temperatures vary from one application to the other, it should
be noted that reduction is usually rapid at ambient temperature.
Ultrafiltration—
Ultrafiltration (UF) is the second aspect of Plant B's system to be
discussed. The UF process discriminates on the basis of molecular size, shape
and flexibility. Suspended solids and large molecule colloidal solids
(0.002 to 10 um) can be filtered in this process. In conventional filtration,
flow is perpendicular to the surface of the filter. In Ultrafiltration, flow
is in the direction parallel to the surface of the filter. A pressure drop of
35 psi is employed. A summary of the Facility B's Ultrafiltration system
design parameters is shown in Table 37.
A description of the process used by Plant B in the reduction and
precipitation of incoming complex wastes represented in Figure 5. Process
water from the plant (streams 1 and 2) is collected in a reservoir, combined
with filtrate from the filter press (stream 6), and subsequently pumped into a
chemical reaction tank for pretreatment. A level controller in the reservoir
activates a sump pump that initiates the transfer of influent upon reaching
the required volumes. Since the transfer rates are highly contingent upon a
variable wastewater flow rate, all subsequent operations have been designed to
perform in a semicontinuous or batch mode.
The wastes entering the chemical reaction tank (stream 3) are
automatically adjusted for pH through the use of sodium hydroxide or sulfuric
acid. Since many heavy metals are only soluble under acidic conditions, to
facilitate the reaction, pH is maintained in the alkaline range (7 to 11). An
ORP (oxidation reduction potential) controller automatically meters a dilute
(12 percent) sodium borohydride solution to ensure complete reaction and
optimize chemical consumption. The resultant reaction products flow by
gravity into the concentration tank.
The concentration tank acts as a repository for the solids generated in
the sodium borohydride reaction. Concentration is achieved by forcing the
permeate through the 0.1 micron pores of the Ultrafiltration membrane. The
suspended solids are rejected and returned to the concentration tank while the
permeate is discharged to the sewer (stream 5). Multiple passes are employed
to improve overall removal efficiency while a level controller prevents the
pump from cavitating. On a regular basis (twice/day), the concentration tank
is drained and its contents dewatered to a 40 to 60 percent solids in a filter
press. The resultant filtrate is returned to the sump (stream 6), while th'e
filter cake is disposed offsite (stream 7).
To prevent membrane blockage or damage, chemical cleaning by a sodium
hypochlorite solution is performed daily. The cleaning solution is circulated
throughout the Ultrafiltration unit for approximately 20 minutes, dissolving
any metals that have became trapped in the membrane. In the case of critical
blockages, the system design permits isolation and easy replacement of any of
the seven Ultrafiltration modules.
61
-------
TABLE 37. PLANT B ULTRAFILTRATION SYSTEM
SPECIFICATIONS*
Parameters
Design
Solids content of influent (recycle)
Pressure drop
Waste throughput
Filter area (total)
Cycle time
Pore size
Flux
Tube diameter
Number of tubes
1-2Z
35 psi
2 gpu
15 ft2
batch
0.1 mm
200 gal/ft2/day
I ia.
10
a?er module, 6 to 10 modules/system.
Source: Memtek Ultrafiltration Systems
62
-------
SUMP rum
RCCIRCULAI 1011 SIRE All
(COnCEIIII(AIE)
tICIIIHC
WASIES (2)
(I)
PLAIIMG
WAS US
MEIIDKAME PERMIAIE
MEIERS
DISCHARGE 10
SEWLR (SUMP)
RLACIIUII COIILEIIIRAIIOII
IAIIK IAIIK
I IIIRAIE (/b)
DLIER
PRESS
SLUDGE
(7o)
orrsiiE
DISPOSAL
Figure 5.' Process schematic showing Plant B plating/etching waste treatment system.
-------
Sludge Filter Press—•
The mechanical sludge dewacering device in use ac Facility B is a Delta
Unifilter low-pressure filter press. Pressure filters of this type dewater
sludge by pressurizing it and forcing the permeate out through a membrane.
The Facility B filter press has an operational area of 7.5 sq. ft. and a line
pressure of 65 to 75 psf. Sludges at Facility B are usually dewacered to 20
to 40 percent solids. The final solids concentration depends on the length of
time the sludge remains in contact with the filter and the operating pressures
applied to the sludge.
PROCESS TESTING AND ANALYTICAL RESULTS
Process Testing
On January 7, 1986, GCA personnel performed sampling of the Facility tt
waste treatment system beginning at 9:15 a.m. As stated in the QA Plan,
samples were taken on an hourly basis or as available. Table 38 summarizes
the sampling times for each stream and indicates deviations from the hourly
sampling schedule in several instances. The pH of streams 3, 4, and 5 were
measured using GCA field instrumentation.
Flow measurement readings were also monitored from plant instrumentation
during the initial stages of testing. However, it was soon obvious that the
flow metering equipment at Plant B was malfunctioning and other means for
estimating flowrate were employed.
One source of flowrate estimates was obtained by contacting the Orange
County Sewer Authority for recent data on Facility B wastewater flowrates.
This source indicated an average flowrate of 22,000 gpd was measured during
quarterly monitoring. A second estimate of actual flowrates was developed
from the throughput of the SBH/ultrafiltration wastewater feed pump
(1-1/2 hp). The feed pump, which is manufactured by Gould Inc., is rated for
27 to 35 gpm with a 4-inch suction and a 3-inch discharge. However the pump
operated on an intermittent basis (approximately 75 percent of the time), thus
the effective flowrate was approximately 1,215 to 1,575 gph for 16 hours or
19,440 to 25,200 gpd. The two methods for estimating flowrate provided a good
check against each other. Based on this information, the measured value of
22,000 gpd was used.
Analytical Results
The objective of the sampling program was to evaluate the effectiveness
of the waste reduction technology utilized by Facility B. The effectiveness
was measured in terms of volume reduction of hazardous waste streams and
minimization of other releases to the environment. The waste stream
parameters analyzed under the sampling and analysis program included total
metals (copper, nickel, lead, and zinc), EP toxicity metals (sludge filter
cake only), total organic carbon (TOC) and total organic halides (TOX). Each
parameter is evaluated in a comprehensive mass balance which focuses on three
main streams of interest. The main streams of interest consist of the
combined influent into the reaction tank (stream 3), the permeate effluent
from the ultrafiltration unit (stream 5), and che sludge filter cake from tne
filter press (stream 7a).
64
-------
TABLE 38. PROCESS OBSERVATIONS AND MEASUREMENTS
Reaction tank
Time pHa ORP Comments
9:15 14.42 -440 Stream I not flowing enough to collect a
sample.
9:45
9:55
10:10
10:20
10:45
10:55
11:10
11:20
11:45
12:10
12:20
12:30
12:45
1:05
1:20
1:45
2:00
2:35
2:45
3:00
3:10
14.51
14.51
14.48
14.47
13.67
8.60
8.58
8.53
11.83
11.95
11.67
14.34
14.42
14.24
14.23
8.91
14.48
14.51
14.52
14.05
8.92
-570
-565
-835
-863
-838
+050
-052
+090
-461
-578
-277
-450
-459
-257
-339
+058
-286
-319
-297
-252
+054
Membrane cleaning began at 10:00. Used
75 gallons of NaOCl.
Membrane cleaning ends.
Reaction tank contents are blue.
Reaction tank contents are black.
Reaction tank contents are green/brown.
Influent to sump stopped at 11:30
(lunch break).
apH readings from Plant metering equipment are apparently high, readings
greater than 14.0 are not valid.
65
-------
The data summarized in Table 39 show influenc and effluent stream
concentrationa for metals of interest in chis study. Influent copper
concentrations during testing (786 ppm) significantly exceeded normal levels
indicated by the facility (150 to 200 ppm). This apparent anomaly may have
been due to process abnormalities during testing, although the plant personnel
did not mention any unusual occurrences. During GCA's sampling, the influent
waste stream was observed to turn a dark shade of blue for a short period of
time, indicating a temporary increase in metallic copper content. Although
this fluctuation does not affect the daily mass balance, drawing any annual
estimates from the concentration data could be misleading. A second
observation drawn from the raw data would be the 78 percent (dry weight) of
copper in the filter cake, which compares favorably with the vendor supplied
data of 80 to 95 percent (Linsey and Hackman, 1985). Both total organic
carbon and total organic halide were virtually unaffected by the reduction
process with only 9.75 and 0.57 percent losses to the sludge stream.
Trace Metal Results—
The objective of the trace metal analysis was to evaluate both the
efficiency of reaction and the removal efficiencies observed for the sodium
borohydride/ultrafiltration treatment system at Plant B. The first evaluation
utilized a process mass balance approach to determine actual and theoretical
reagent requirements and calculate the effectiveness of the sodium borohydride
reagent in reducing the metals of interest. The second evaluation involved an
assessment of influent and effluent concentrations and a comparison of these
with local and Federal effluent limitations to determine process viability.
In developing a mass balance for metals contained in the entering and
exiting wastestreams, it was necessary to make the following assumptions:
1. Flowrate was constant at 22,000 gpd;
2. Wastewater flow to Stream 7 was small compared to Streams 3 and 5;
and
3. The influent flovrate (Stream 3) was approximately equal to the
effluent flowrate (Stream 5) at 22,000 gpd.
While assumptions 1 and 3 above violate continuity, it follows that if in fact
the flow of Stream 7 is small compared to Streams 3 and 5, then these
assumptions hold. These assumptions were necessary due to difficulties
encountered during testing with Plant B's wastewater flow metering equipment.
Metals concentration data were used in conjunction with waste stream
flowrate information to develop the mass balance results presented in
Table 40. Based on these data the sodium borohydride treatment system at
Plant B showed high removal/recovery of copper, zinc and lead while showing-
reduced effectiveness for nickel. During testing, the total metals loading to
the SBH reactor and ultrafiltration system was approximately 145 Ib/day of
which over 99 percent was divalent copper (Cu*2). The theoretical
requirement for the total reduction of all metals present in the SBH reactor
66
-------
TABLE 39. SUMMARY OF TEST DATA FOR SODIUM BOROHYDRIDE TREATMENT AT PLANT B
Stream 3 Stream 5 Stream 7
Influent waste Effluent wastewater Sludge3
Parameter (mg/L) (mg/L) (ug/g)
Total organic carbon 40.0
Total organic halide 1.76
Total trace metals
Cu 786.0
Ni 0.055
Pb 0.57
Zn 3.86
EP toxic metals
Ar
Ba
Cd
Cr
Pb
Hg
Se
Ag
36.1 184.8
1.75
1.49 780,000
0.03 54.7
0. 10 300
0.028 1,430
0.03
0.522
0.002
0.003
1.8
0.0002
0.04
0.50
aResults given on a dry weight basis for sludge.
67
-------
TABLE 40. MASS BALANCE OF TRACE METALS RESULTS FOR
SODIUM BOROHYDRIDE TREATMENT AT PLANT B
Total
metals
analyte
Scream 3
Influenc waste
Ub/day)
Scream 5
Effluent wastewacer
(Ib/day)
Scream 7
Sludge
(Ib/day)
Percent
recovery (%)
Cu 144.1524 0.2b04 143.892 99.82
Ni 0.0101 0.0055 0.0046 45.54
Pb 0.1045 0.0183 0.0862 82.49
Zn 0.7079 0.0051 0.7028 99.28
68
-------
influent scream was estimated to be 175.5 Ibs of sodium borohydride
(15 gallons of an alkaline solution containing 12 percent by weight of i>UH).
The actual quantity of solution consumed during the reduction reaction was
148.6 Ibs (12.7 gallons of solution). This consumption rate represents an
actual to theoretical SBH solution addition ratio of 0.84. Previous case
studies (PC FAB, February, May 1984) have shown that SBH requirements are more
typically 1 to 1.5 times the theoretical requirement. It is speculated that
SBH reduction operated below stoichiometric requirements due to an absence of
competing reactants (e.g., aldehydes) and relatively high metals concentration
found in the reactor influent.
Metals concentrations in the effluent stream were used to determine the
effectiveness of the SBH reduction system in both meeting effluent guidelines
and minimizing releases to the environment. Table 41 describes effluent
performance characteristics in terms of reduction efficiencies and effluent
compliance. Analysis of these characteristics show that copper was reduced
most efficiently at 99.82 percent, while nickel removal was the least
efficient at 45.54 percent. The wide disparity in removal efficiencies seems
to be mainly a function of concentration (higher concentrations are removed
more efficiently), but the chemical potentials (quantity of free energy
required for an ionic species to obtain equilibrium) may also have been a
factor. Approximately 144.7 Ibs of total metals were reduced to elemental
form by the SBH ultrafiltration system, representing a total reaction
efficiency of 99.8 percent. Overall the quality of the effluent produced by
the SBH reaction system was quite good. Metals currently discharged to the
sewer are now meeting stringent local and Federal EPA pretreatment standards.
Previously Facility B was unable to meet County standards using a batch
filtration system which frequently failed, unintentionally discharging
precipitated copper (Circuits Manufacturing, September 1984). However, since
the installation of the SBH ultrafiltration system, Facility B's effluent has
been consistently below discharge requirements based upon sampling and
analysis by the Orange County Sewer Authority.
An additional criteria in the assessment of the SBH/ultrafiltration
treatment system is the characterization of the sludge filter cake. As
Facility B currently ships the sludge product offsite for land disposal, the
E.P. Toxicity leachate characteristics of the SBH sludge have been evaluated.
Analysis of the raw data shows that the sludge product contains greater than
78 percent (dry weight) elemental copper. This concentration in conjunction
with 1,430 ug/g of reduced zinc combine to form a product called refinery
brass. This intermetallic product can be easily recovered by a smelter, thus
eliminating the need for land disposal and limiting any liabilities thereof.*
However, if the sludge product must be landfilled, the results in Table 42
show that leachate resulting from the sludge dry filter cake is within Federal
EPA guidelines. However, it is noted that this resultant sludge would still
be classified RCRA waste code F006 under current Federal regulations.
Total Organic Carbon/Total Organic Halide—
The objective of the total organic carbon (TOO and total organic halide
(TOX) analysis was to determine total organic loadings within the SBH reactor
system. As mentioned previously SBH is an extremely efficient reductant and
-------
TABLE 41. PLANT B EFFLUENT PERFORMANCE CHARACTERISTICS
Effluent5
Metal*
Cu
Ni
Pb
Zn
Concentration
(mg/L)
1.42
0.030
0.10
0.028
Loading
(Ibs/day)
0.2604
0.0055
0.0183
0.0051
Percent
Recovery
99.82
45.54
82.49
99.28
county-
standards
(Ibs/day)
0.50
0.70
0.10
0.70
reaerai.-
1 imitations
(mg/L)
3.72
3.51
0.67
2.64
*Measured as total trace metals method 3050.
^Permeate from Memtek ultrafiltration unit discharged to sewer.
cOrange County Sanitation District.
^Daily maximum (mg/L) for electroplating point source effluent limitations
U.S. EPA, Federal Register U.77, No. 169:38477, August 31, 1982.
70
-------
TABLE 42. EP TOXICITY LEACHATE RESULTS FOR PLANT B
SODIUM BOROHYDRIDE SLUDGE
Element
Concentration
(mg/L)
EPA standards3
(mg/L)
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
0.03
0.522
0.002
0.003
1.8
0.0002
0.04
0.56
5.0
100.0
1.0
5.0
5.0
0.2
1.0
5.0
aU.S. Environmental Protection Agency,
Federal Register. Vol. 45, No. 98,
98:33122, May 14, 1980.
71
-------
will reduce organic, as well as inorganic species. Often the presence of
nonmecallic compounds such as aldehydes, kecones, nitrates, peroxides, and
persulfaces will reduce reactor efficiency and consume up to twice the
theoretical quantity of SBH solution required. However, the relatively low
concentrations of organics in the SBH reactor influent (presented in T?ble 43J
showed little reduction of nonmetallic species. For example, TUC reduction
was only 9.75 percent while TOX reduction was 0.57 percent. It is believed
that low concentrations of these organic species contributed to the overall
success of the SBH reactor system with respect to metals as discussed earlier.
ECONOMIC AND ENVIRONMENTAL EVALUATION
Economic Evaluation
One of the objectives of this study was to evaluate the economics of the
waste minimization technology. For this case study the economics of the
treatment system tested (SBH) are presented along vicb lime-ferrous sulfate
conventional technology which would be used in its place.
Central to any discussion or comparison between sodium borohydride and
lime-ferrous sulfate is the actual to theoretical chemical usage ratio. In
sodium borohydride applications, the high unit cost of sodium borohydride
solution ($2.7/lb) versus ferrous sulfate (30.11/lb) necessitates the careful
control of chemical usage. Parameters of importance are: presence of
nonpriority reducibles, pH adjustment, good mixing and settling conditions,
adequate reaction time, and liquid/solid separation. These site-specific
factors combined with effluent limitations and total treatment and disposal
costs, can significantly affect the economics of employing sodium borohydride
treatment technology.
The SBH application at Facility B is fairly uncharacteristic in that they
utilize a cupric chloride instead of an ammonical etenant. This is
significant in that cupric chloride contains very few complexants which would
interfere with the SBH reduction reaction, lowering reaction efficiency and
driving up treatment chemical costs.
Complexing and chelating agent applications in the electronic components
industry are typically for the suspension of metals in plating or etching
solutions. Major sources of complexing agents are alkaline (ammonical
chloride) and ammonium persulfate etchants. Borohydride may react with these
other compounds (i.e., ammonia) in the wastewaters, thus reducing its
availability for metal ions. Therefore, most electronics components users
find that a large excess of borohydride is frequently required to ensure rapid
and complete metals reduction.
A detailed cost analysis for both sodium borohydride and lime/ferrous
sulfate (LFS) technologies is presented in Table 44. The CCA test data
clearly established that SBH treatment is superior to LFS treatment in this
application, when capital costs are held constant. Chemical costs for the SBH
72
-------
TABLE 43. ORGANIC LOADING RESULTS FOR
FACILITY B SBH REACTOR SYSTEM
Stream ID Description TOC (ppm) TOX (ppn)
3 SBH reactor system influent 40.0 I.756
6a SBH reactor system effluent 36.1 1.746
7a SBH filter cake 184.8
7J
-------
TABLE 44. PLANT B ANNUAL TREATMENT AND DISPOSAL COSTS
FOR SODIUM BOROHYDRIDE AND LIME FERROUS
SULFATE PRECIPITATION TECHNOLOGIES
Basis
Unit cose
(3)
Sodium borohydride
treatment
system coat
(*)
Lime ferrous
treatment
system cose
(S)
Chemical costs
SBH solution
Sodium hydroxide
Ferrous sulfate
Hydrated lime
2.7/lb
0.32/gal
O.ll/lb
50.O/ton
100,298
12,500
35,888
495
Total chemical cost
Disposal costs
Sludge disposal
Annual costs
Total annual cost
Cost/lb metal reduced
200/ton
112,798
I3,278b
126,076
3.5
36,383
205,5bOc
241,943
6.7
a35 percent total solids
b78 percent metal in the solids
c 5 percent metal in the solids
74
-------
treatment system were three times greater than LFS, however the more
significant sludge disposal costs for SBH reduction are shown to be
93.5 percent less. Aa a result of these factors, SBH treatment was able to
reduce overall operating expenses by 48 percent (in comparison to LFS
treatment), wh*Tle decreasing sludge production at the same time. In addition,
Facility B will soon be practicing sodium borohydride sludge reclamation
onsite. This will not only further reduce operating expenses, but also
potentially lower liabilities associated with hazardous waste land disposal.
Environmental Evaluation
Results indicate that the use of SBH is an effective means for metallic
waste precipitation and solid waste management. As stated previously, SBH
application is very site specific and the presence of oxidizing agents such as
complexants can increase chemical demand by as much as SO to 100 percent.
However, its cost-effective performance in achieving discharge limits and
reducing hazardous waste at Facility B makes it a practical alternative to
comparative waste treatment technologies*
75
-------
SECTION 5
FACILITY C CASE STUDY
FACILITY CHARACTERIZATION
Facility Description
This facility manufactures electronic computing equipment including
logic, memory and semiconductor devices, multilayer ceramics, circuit
packaging, intermediate processors and printers. Approximately 11,000 persons
are employed at this particular location.
Waste Sources
Two of the major processes in which hazardous waste streams are generated
are the manufacture of semiconductors and the manufacture of printed circuit
boards. As mentioned in the previous section, the manufacture of these two
products can involve the use of organic solvents both for the cleaning of
surfaces and the developing and stripping of photosensitive resists. The
photoresists are' used to form either a positive or a negative image of the
circuit pattern on the substrate chip or circuit board. After application of
the photoresist to the substrate material, a mask of the circuit pattern is
placed over the board or chip and Che surface is exposed to light. Since in
this case a negative photoresist material is used, the resist polymerizes upon
exposure to light, while the resist that is covered by the photomask does not.
Following this exposure to light, developer solvent is used to remove the
resist material which has not been stabilized. The developer solvent used at
this facility is methyl chloroform (1,1,1-trichloroethane). Subsequent to
developing the resist, the exposed areas of the substrate material are etched
and/or metal plated. Once this has been done, the resist has served its
purpose and it can be "stripped" from the surface. Either acids or organic
solvents may be used for photoresist stripping. At this facility methylene
chloride is used to strip photoresist from electronic panels. However, the
spent solvent from this operation is handled separately and will not be
addressed further in this report.
71>
-------
Waste Characteristics and Quantities—
Waste solvent streams will vary in composition both according to whether
they were used for stripping or developing, and whether they were used in the
manufacture of circuit boards or semiconductors. Several different organic
solvents are used ac this facility,including:
Methylene Chloride - Resist stripping of Electronic Panels
Methyl Chloroform - Resist Developing of Electronic Panels and
Substrate Chips
Freon - Surface Cleaning and Developing of Substrate Chips
Perchloroethylene - Surface Cleaning of Electronic Panels
The major difference in the waste solvents from resist stripping and
resist developing is thac resist stripping solvents will contain polymerized
resist while resist developing solvents will contain unpolymerized resist.
Unpolymerized resists may polymerize if they are heated to a certain
temperature, and therefore waste containing these materials may have to be
heated differently than wastes containing already polymerized resists. In
both of the waste streams, resists are present as dissolved solids at maximum
concentrations of 1 percent by weight. The exact concentration of dissolved
solids in the solvent will depend on the volume of work processed in the
solvent. Some days the solvent will have close to 1 percent dissolved solids
and other days the concentration will be closer to zero.
After developing or stripping, the work piece is generally rinsed in
water to remove the residual solvent. This results in a waste solvent stream
contaminated with water. Gravity settling is employed to separate the solvent
and water fractions directly after the developing or stripping operation , but
some residual amount of water may remain in the solvent fraction. The water
fraction is sent to wastewater treatment.
Most of the solvent waste streams are kept segregated to facilitate
recovery, but the developing process for substrate chips involves the use of
both methyl chloroform and Freon. Consequently, the spent solvent stream from
this process apparently contains a mixture of both of the solvents, 90 percent
being Freon and 10 percent being methyl chloroform.
Another solvent that is used at the facility is perchloroethylene. It is
used for precleaning the surface of electronic panels to remove dirt, oil or
grease which may have been deposited during previous manufacturing operation.
The spent solvent from this cleaning operation is handled separately and will
not be addressed further in this report.
77
-------
Waste Management
The purpose of this study is to evaluate onsite methods of recovering
and/or recycling hazardous wastes. At Facility C the primary
recovery/recycling operation are the recovery and reuse of solvents by
distillation or evaporation. Several types of equipment are used to recover
spent solvents at this facility. Box distillation units are used to recover
methylene chloride and perchloroethylene, flash evaporators are used to
recover methyl chloroform and a distillation column is used to recover Freon.
The operation of these pieces of equipment is described below.
Flash Evaporation of Methyl Chloroform-
Two flash evaporation units are used to recover spent methyl chloroform
from several different resist developers. The spent solvent from each of
these developing areas is first treated to remove water and then pumped to a
waste solvent collection tank. From this waste collection tank, the solvent
is then pumped to the flash evaporation units, where Che contaminants are
removed, and the recovered solvent is pumped to a clean solvent storage tank.
Then, virgin methyl chloroform is added to the recovered methyl chloroform (to
replenish corrosion inhibitors), and returned to the developers. A schematic
of this system is shown in Figure 6.
Figure 7 shows the major components of the flash evaporation units and
Table 45 indicates normal operating parameters. This type of a unit is used,
instead of a conventional still, because of the presence of unpolymerized
photoresist in the methyl chloroform. As mentioned above, the photoresist
will polymerize when subjected to high temperatures such as those required to
boil MCF at atmospheric pressure. If the resist polymerizes onto the heating
coils of the still, the operation of the unit would be adversely affected. In
flash evaporation, the solvent is preheated to approximately 100°F and then
enters a "flash" chamber where a vacuum of 20.5" Hg causes the preheated
liquid methyl chloroform to vaporize. At atmospheric pressure, the MCF must
be heated to 165"F for boiling to occur.
In the flash chamber, a certain fraction of the MCF vaporizes, and the
other fraction, containing the contaminants, remains in liquid form. The
vapor passes through a condenser and is recovered at an average rate of
600 gallons per hour. The fraction that does not vaporize collects in the
bottom of the chamber and is recirculated through the heat exchanger at a rate
of 490 gallons per minute. A certain amount of this liquid is bled off every
ten minutes and pumped to the still bottom storage tank.
The flash evaporators are fed from a 15,000 gallon spent solvent
collection tank, and recovered solvent is collected in an adjacent
15,000 gallon tank. When the level of spent solvent reaches 12 or 13 thousand
gallons, and the level of recovered solvent is down to 2 or 3 thousand
gallons, the evaporators are turned on and operated until these quantities are
reversed. This amounts to recovering approximately 10,000 gallons, which at a
recovery rate of 600 gallons per hour for each unit, requires the operation of
both units for approximately 8 hours per day.
78
-------
Figure 6. Schematic of methyl chloroform recovery system.
-------
I
Oc
c
9
AIR EHISS10HS VENT
TO CARBON ADSORPTION
UNIT
STFAH
Figure 7. Schematic of flash evaporator.
-------
TABLE 45. METHYL CHLOROFORM FLASH EVAPORATION SYSTEM OPERATING INFORMATION
NORMAL DISTILLATION UNIT OPERATION
Operating Parameters
1. Solvent
Recovery Rate
Boiling Temperature
Recycle Rate
Distillate Temperature
Separator Pressure
2. Chilled Water
Flow
Inlet Temperature
Outlet Temperature
3. Steam
Flow
Pressure at Still Supply
4. Hot Water
Inlet Temperature
Outlet Temperature
Recycle Rate
5. Electrical Requirements
Normal Conditions
600 gph
99°F (clean)
113°F (14.52 non-volatiles)
490 gpm
70°F
20.5" Hg Vacuum
46 gpm
47°F (normal)
778F (normal)
800 Ibs/hr
30 psig
140°F (max.)
133°F
400 gpm
30 hp
81
-------
TABLE 50. ESTIMATED COST OF METHYL CHLOROFORM RECOVERY
Cost Item
Capital Cost
- Equipment
- Engineering
Other
TOTAL CAPITAL
Annual O&M
Electricity
Steam
- Cooling Water
- Labor
Maintenance
Residue Disposal
TOTAL O&M
Annual Costs
Annualized Capital
Annual O&M
Solvent Cost
TOTAL COST
Ons ite recovery
Quantity Cost8
2 320,000
32,000
32,000
384,000
30 kv 8,400
800 Ibs/hr 4,300
9,600 gal/hr 13,400
2,800 hrs/yr 42,000
38,400
129,000 gallons 0
106,500
68,000
106,500
129,000 gallons 329.800
504,300
Offsite recovery
Quantity Cost3
— — — _
—
— - —
—
_ _
—
—
—
—
3,619,000 gallons (904,700)
(904,700)
—
—
3,619.000 gallons 10,911,300
10,006,600
al986 dollars.
-------
TABLE 51. ESTIMATED COST OF RECOVERY OF FREON/METHYL CHLOROFORM
SO
•o
Cost Item
Capital Cost
Equipment
Engineering
Other
TOTAL CAPITAL
Annual O&M
Electricity
Steam
Cooling Water
Labor
Maintenance
- Residue Disposal
TOTAL O&M
Annual Costs
Annual ized Capital
Annual O&M
Solvent Cost
TOTAL COST
Onsite recovery
Quantity Cost8
1 130,000
13,000
13,000
156,000
1.5 kw 600
470 Ibs/hr 3,800
4,320 gal/hr 9,100
2,800 hrs/yr 42,000
15,600
119,500 gallons 0
71,100
12,600
71,100
23,900 gallons 190,900
274,600
Offsite recovery
Quantity Cost8
0 0
0
295,800 gallons (74,000)
(74,000)
(74,000)
176,300 gallons 1,408,000
1,334,000
a!986 dollars.
-------
Methyl Chloroform—
The estimated costs associated with onsite recovery of methyl chloroform
by flash evaporation are displayed on che left side of Table 50, and those
associated with offsite recovery are displayed on the right-hand side of the
same"Table. The onsite recovery costs are based on the use of an APV
Paraflash evaporator capable of handling a 600 gallon per hour feed ra.te. The
FOB cost of each of the two units is 160,000 dollars and this includes a plate
heat exchanger, a vapor liquid cyclone separator, shell and tube main
condenser, shell and tube vent condenser, hot water set for indirect stream
heating, a set of pumps and instrumentation for automatic operation
[APV Crepaco, 1986]. The equipment would also be fully preassembled on a skid
with manual valves and piping. The capital cost estimate also includes
engineering and "other" costs because at this facility modifications are
generally made to equipment to meet site-specific conditions.
O&M costs are based on the operation of each of the units, for 8 hours
each day 350 days per year. Labor costs were estimated only for operation of
the units. It was assumed that one person would be assigned to monitor
operation of the two units, 8 hours each day 350 days per year.
In looking at the costs, it is evident that the major cost is associated
with the purchase of virgin solvent. Since the purchase cost of virgin methyl
chloroform is 4.50 dollars per gallon versus approximately 5 cents per gallon
(of recovered solvent) to operate the flash evaporator, it certainly makes
economic sense to recover the solvent onsite. When the solvent is sent
offsite for recovery it can be bought back at two-thirds the price of virgin
solvent, but this is still a cost of 3 dollars per gallon.
The annual savings resulting from onsite recovery is greater than
9 million dollars. One of the major reasons for the tremendous savings is
that the amount of spent solvent generated is so large. At a smaller
facility, savings would not be quite so impressive.
Freon/Methyl Chloroform—
Costs associated with the recovery of Freon are presented in Table 51.
The equipment cost is based on the use of a 1,200 gallon per day APV Batch
distillation system. The 130,000 dollar FOB price includes a 28 foot
distillation column with metal mesh packing, a 1,450 gallon batch pot, a U
bundle reboiler, U bundle condenser, shell and tube vent condenser, bottom and
top product pumps, bottom and top product shell and tube coolers, and
instrumentation for automatic operation [APV Crepaco, 1986]. The unit would
be preassembled and include all valves and piping.
O&M costs for this system are based on operating this unit 24 hours per
day, 350 days per year. This unit is operated continuously as long as spent
solvent is available for input. Labor costs are those for operating the
unit. Since the unit is equipped with instrumentation to allow for automatic
operation, it is not necessary to have someone monitor operation 24 hours per
day. Instead, an estimate of 8 hours per day was used as a maximum amount of
operating labor required.
93
-------
Onsite recovery of Freon results in a cost savings of approximately
1 million dollars per year compared to sending the waste solvent offsite for
recovery. Cost savings per gallon of spent Freon recovered are even higher
than in the recovery of methyl chloroform because of the extremely high cost
of virgin Freon. The total annual savings are less only because less spent
Freon is generated.
ENVIRONMENTAL EVALUATION
The purpose of this study is to present case studies of methods of waste
management that are alternatives to land disposal. The intent was to show the
reduction in the quantity of land disposed waste achieved through use of the
alternative technology. In this particular case study, the spent solvents
have been recovered on site, in some degree or another, for over ten years.
In addition, even if they were not recovered on site, they could easily be
sent offsrte for recovery. Land disposal of spent halogenated solvents of
almost 99 percent purity, particularly in the quantities that this facility
generates, is not an economically intelligent practice. Consequently,
economic and not environmental factors are the driving force behind onsite
solvent recovery.
In any event, certain residues and emissions are generated by onsite
distillation and evaporation. The primary residues generated are still
bottoms. These still bottoms, (see tables) contain at least 90 percent
solvent, and therefore they are sent offsite for further recovery. The
quantities of still bottoms that were sent offsite during the years 1981
through 1984 are presented in Table 52. These quantities are only 5 percent
of what would be sent offsite if onsite recovery were not practiced* One
environmental benefit of onsite recovery is that the quantity of solvent
requiring offsite transport is much less. Therefore, the chances of an
accident occuring in which spent solvent is spilled into the environment are
reduced. Another benefit is that loading and unloading the solvent from tank
trucks is greatly minimized therefore reducing the possibilities of air
emissions and spills.
The other potential environmental impact of onsite solvent recovery is
emission of volatile solvents to the atmosphere. The primary source of these
emissions would most likely be the vacuum pump associated with the flash
evaporator. Unfortunately, it was not possible to measure these emissions.
As shown in Figure 7, however, these emissions are vented to a carbon
adsorption unit whose removal efficiency would generally range from 85 to
95 percent. Since the solvent recovery operations are all indoors, it is
possible to vent any other fugitive emissions from pumps, valves and other
fixtures to carbon adsorption units also. Consequently, the majority of the
air releases from solvent recovery are captured. After the capacity of the
carbon in these units is spent, the absorbed solvents are desorbed by steam
stripping, the water/solvent mixture is decanted, and the solvent fraction is
reused after recovery by distillation or evaporation.
94
-------
TABLE 52. QUANTITIES OF STILL BOTTOMS GENERATED FOR OFFSITE RECOVERY
Yearly quantity generated
Waste type 1981 1982 1983 1984
Methyl Chloroform
gals 150,000 166,000 90,000 129,000
IDS 1,400,000 1,550,000 840,000 1,200,000
Methyl Chlorofonn/Freon
gals. — — 77,640 119,460
Ibs. — — 880,600 1,355,000
-------
SECTION 6
FACILITY D CASE STUDY
FACILITY CHARACTERIZATION
Facility Description
Facility D Manufactures mobile communications equipment in an operation
consisting of 260 employees. Process operations consist of a small metal
forming shop, prepaint and painting lines, electroplating, and electroless
plating of printed circuit board components. An onsite wastewater treatment
includes cyanide destruction, hexavalent chromium reduction, and acid/alkaline
neutralization. Organic solvents involved in photoresist developing and
stripping are recovered in-process through distillation. The waste stream of
interest is a spent developing solution, consisting of 1,1,1-trichloroethane
and non-stabilized Dupont Riston photoresist.
Waste Sources
Dry film photoresists such as Dupont Riston are accepted by the industry
as the most reliable technology for producing printed circuit panels at high
yields. The typical printed circuit panel prior to imaging consists of a one
ounce per square foot of copper foil clad on a fiberglass-epoxy substrate. A
complete list of Printed Circuit Board (PCB) manufacturing unit operations is
presented in Table 53, while the discussion below highlights those operations
relevant to Facility D.
Drilling and Deburring—
After mechanical or chemical cleaning the typical double-sided or
multilayered board is drilled by numerically controlled, high speed spindle
drills. The resulting holes and board edges are then deburred by rotating
brushes to remove any loose particulate matter or rough edges detrimental to
subsequent chemical processing operations.
Electroless Copper Deposition, Rinsing and Neutralization—•
The first chemical process is the deposition of a 25-85 microinch layer
on the surface and in the drilled holes of the panel. This thin layer
provides electrical contact from surface to surface and layer to layer. The
reaction does not require electrical current and, therefore, is dependent
mainly on three factors: chemical activity, mechanical agitation, and
temperature. After electroless copper deposition, the board is thoroughly
rinsed, neutralized with a mild acid, and dried to ensure that the surface
will be receptive to the photoresist.
96
-------
TABLE 53. PLANT D PRINTED CIRCUIT BOARD PROCESSING
1. Copper Clad Board
2. Board Cleaning
3. Drilling
4. Debarring
5. Electroless Copper Pretreacments
6. Electroless Copper Deposition
7. Hot Roll Lamination
8. Image Transfer
9. Developing
10. Electroplating - Copper
11. Electroplating - 60/40 Solder
12. Immersion Tin
13. Resist Stripping
14. Aramonical Etching
15. Electroplating - Tabs
16. Finished Board
97
-------
Hot Roll Resist Lamination—
Dry film photoresist is an ultraviolet light sensitive organic polymer
applied directly to the copper foil surface by a hot roll laminator. All
surfaces must be absolutely clean because any foreign particles laminated
under the resist will cause electrical shorting if it falls across a circuit.
To prevent resist overhang, edges should be trimmed flush with the panel.
Image Transfer—
In the photo imaging operation, a stencil of the artwork is exposed to
ultraviolet light under vacuum. This maintains good contact between the photo
tool and the resist surface, preventing blurred images. All desired-features
such as circuit traces, hole pads, and connector tabs remain under the shaded
area of the stencil and, therefore, unexposed. The dry film covering the
undesired areas, such as excess copper cladding, is exposed and polymerized.
Since polymerized photoresist acts as an electroplating inhibitor, only the
areas remaining unexposed will receive the required metallization in the
subsequent plating operations.
Resist Developing—
Following exposure, the printed circuit panels are introduced to a high
pressure spray of organic solvent. The solvent, 1,1,1-trichloroethane,
dissolves the unexposed photoresist and reveals the bare copper underneath.
Spent TCE containing unstabilized resist is automatically gravity fed from the
developing solution tank to a Dupont-Riston solvent recovery still.
Electrolytic Plating--
At Facility D, the panels are racked and immersed in a MacDermid
acid-copper electroplating bath until the drilled holes, circuit traces, etc.,
acquire the sufficient plating thickness. The required 1-2 mils of copper
plating insures a uniform electrical conductivity throughout the panel.
Following copper plating, the panels are then immersed in a 60/40 Sn/Pb alloy
electroplating bath in which they receive 30-50 microinches of plate that acts
as an etch resist.
Resist Strip—
To remove polymerized photoresist, the boards are rinsed, dried and then
run through a conveyorized stripper. The active ingredient is methylene
chloride/methanol which is applied to the board surface by spray nozzles that
penetrate into the narrow channels between circuits to remove resist.
Incomplete removal of resist will cause problems in subsequent processing,
particularly the aomonical etchant step.
Developer Waste Management
The purpose of this case study is to evaluate developer solvent waste .
minimization technologies, particularly the state-of-the-art application at
Facility D. The process depicted in Figure 9 represents a semi-continuous,
2-step, distillation recovery system. Until recently, all still bottoms from
the primary distillation unit; a Dupont Riston SRS-120 solvent recovery still,
were drummed and shipped offsite for reclamation. In October 19b5, Facility D
purchased a Zerpa Recyclene RX-35 solvent recovery system to reclaim still
bottoms onsite. A brief description of the two technologies is presented
below.
98
-------
SPENT SOLVENT
97% TCE
SPENT TCE
MAKE-UP
99% SOLVENT
22.5 GPD
DEVELOPER
PROCESS
SEPARATOR
SOLVENT
99%
GRAVITY
SEPARATOR
CONDENSER
SRS-
120'
STILL
85 GAL
CHARGE
SEPARATOR
RAFFINATE
.39% TCE
(NOMINAL)
OLD SYSTEM
STILL BOTTOMS
98% SOLVENT
STILL
OVERHEAD
99% SOLVENT
RECYCLENE
STILL
CONDENSER
NEW SYSTEM
STILL
BOTTOMS
7.5 % SOLVENT
Figure 9. Plant D 2-stage solvent recovery system.
99
-------
Dupont Riston SRS-120—
The SRS-120 is a liquid phase recovery system whose operation is
consistent with conventional batch distillation technology. The apparatus
consists of a 100 gallon boiling chamber, an overhead condenser, and a
3 gallon water separation system. The still operates at atmospheric pressure
and a boiler temperature of 165*F, which is the boiling point of
1,1,1-trichloroethane. The major recovery criteria is overhead solvent
specification and contaminant content in the product. Prior to operation, the
still must be filled with an 85 gallon charge consisting of overhead from che
recyclene unit and any required make-up fluid. Feed from the developer
solution tank is gravity fed to the still on a semiconcinuous basis. The feed
varies in solids content depending on the type of board being processed, but
it typically ranges from 1-3 percent photoresist. The heat of vaporization is
supplied by a closed coil circulating system using low pressure steam
(2-15 psi and 250°F) to provide the heat input. Due to the wide range in
component boiling points encountered in this type of application, this system
is economically feasible for feed streams with low solids concent of 5 percent
or less [Solvent Recovery in the United States 1980-1990, The Pace Company].
The low operating temperature (165*F) insures that the photoresist will not
polymerize and foul heat transfer surfaces, decreasing still efficiency.
The second operating phase is the condensation of the solvent vapor in a
simple shell and tube heat exchanger, using ethylene glycol as the cooling
medium. Vapor is removed from the still as fast as it is formed without
appreciable condensation or reflux. The overhead product which is virtually
solids-free is gravity fed into a 3-gallon decanter to separate any water
introduced in the spent feed from the product solvent. TCE from the bottom of
the separator is gravity fed to a ten gallon sump where it is temporarily
stored. A level control on the sump returns this product to the developing
fluid holding tank which displaces a similar quantity into the developing
tank, completing the closed loop system.
Since little true fractionation occurs, increasing recovery will result
in degradation of the overhead product. Therefore, after the completion of
each recovery cycle (typically 2 days), the vessel is opened and the remaining
sludge and solvent mixture is drained through ports in the vessel bottom. The
solvent mixture, which may contain up to 8 percent solids, is then pumped into
barrels awaiting transfer to the Recyclene unit.
RX-35 Recyclene Still—
The Recyclene RX-35 solvent recovery system is a batch distillation
apparatus analogous to the Dupont Riston SRS-120. The system (Figure 10)
consists of a 30-gallon capacity, silicone oil immersion heated stainless
steel boiler, a non-contact water-cooled condenser, and a 10 gallon temporary
storage tank. The boiler is equipped with a vinyl liner inside a Teflon bag.
The Teflon bag provides temperature resistance and the vinyl bag collects
solid residue, eliminating boiler clean-out and minimizing sludge generation
after distillation. Two thermostats control the temperature of the boiler and
the vapor, automatically shutting down the boiler when all the solvent has
evaporated.
10U
-------
SPENT SOLVENT
HEAT
EXCHANGER
COOLING
WATER
LOOP
SINK
\/
CONDENSER
BOILER
CLEAN
SOLVENT
REUSE
Figure 10. Process flowsheet - recyclene distillation solvent recovery.3
Source: Dietz J.D. an3 Cherniak C.M., 1984.
101
-------
Prior to operation, a 20-25 gallon charge is transferred into the boiler
from the barrels containing the Dupont Riston SRS-120 still bottoms. The
boiler which operates at 1 aera is heated to 370°F through use of an
electrically heated silicone oil jacket. Cleaned solvent vapors (99.5 percent
pure) rise through a water cooled condenser (85 gph) and are collected in the
10 gallon temporary storage tank. Typical batch node operation time requires
90 minutes. The unit of Plant D is charged 4 to 6 times daily. Following
complete evaporation, the bottom product is a dry solid consisting of
1,1,1-TCE, waste photoresist, and residual contaminants. The residual
contaminants are believed to be trace amounts of electroless copper solution
and corrosion inhibitors. Since TCE in the presence of water can liberate
hydrochloric acid and react violently with aluminum, inhibitors are necessary
to prevent corrosive reactions. Facility D is also recycling 42 still bottom
drums which have been stored onsite in anticipation of acquiring the recyclene
unit.
The truly unique feature of Recyclene RX-35 is the bag liner system which
keeps the heat transfer surface dry and clean, consequently making it easier
to operate the system. [Hazardous Materials and Waste Management, Nov./Dec.,
1984] Otherwise, contamination would result in reduced efficiency, increased
energy requirements, and decreased distillation rate efficiency. It also
concentrates as much of the waste as possible, thereby drastically reducing
waste volume.
A few technical limitations of the Recyclene unit should be noted. The
maximum operating temperature is 39°F, so that recovery of solvents with
higher boiling points would not be practical. Addition of a chiller is often
necessary to condense compounds with depressed boiling points.
PROCESS TESTING AND ANALYTICAL RESULTS
Process Testing
On January 23, 1986, CCA conducted a field study to evaluate waste
minimization operations at Facility D. Sampling was conducted over the course
of a normal day's operation when both the Dupont distillation and Zerpa
solvent recovery units were in operation (see Table 54 for recyclene still
process information). Seven separate sampling locations were utilized to
provide a comprehensive process evaluation and mass balance. Those locations
are: la) the initial Riston still charge, Ib) the spent developer solvent
prior to entry into the Riston still, 2) clean solvent exiting the water
separator, 3) the contents of the water separator, 4) the still bottoms
product from the Riston distillation unit, 5) recovered solvent from the
recyclene unit, and 6) the final bottoms product from the Recyclene unit.
Upon arrival at Facility D several discrepancies were noted between
actual operation and the original process description in the QA Sampling Plan
(these deviations have been corrected in the case study process description):
(1) still bottoms from the Dupont Riston SRS-120 were not pumped continuously
to the Recyclene RX-35 solvent recovery still, but were pumped when needed
from 55 gallon storage drums, (2) Dupont Riston still was not filled
continuously from the developing solution tank, but was filled with recovered
102
-------
TABLE 5A. FACILITY D PROCESS INFORMATION
RX-35 RECYCLENE SOLVENT STILL
Parameter*
Capacity (gph)
Thru put (gph)
Temperature (°F)
Pressure (ATM)
Overhead (Z)
Purity
Yield (Z)
Design3
35
10-35
390
1
99.5
99+
Operation during testing
25
12
370
1
99.9
99.8
aSource: Blodgett, W.A., 1985.
103
-------
TCE from storage drums, (3) no virgin TCE was used in Che process. Other
discrepancies during testing were a two hour delay in the sampling of the
Dupont Riston still overhead due to a faulty valve assembly and a 15 gallon
process spill.
Due to these slightly different operating conditions, allowances were
made and sampling/testing proceeded in the following manner. The Recyclene
still bottoms were analyzed for extractable and volatile organics, and EP
Toxicity metals since it is expected that this product will be land filled.
The other six streams were sampled for volatile organic compounds, extractable
organics, metals, and total solids. The spent solvent influent and Riston
still overhead were sampled at 5 (instead of 8) hourly intervals over the day
to provide a representative composite. The SRS-120 still bottoms was grab
sample, at the beginning of the sampling effort while the recyclene hoccorns,
overhead and water separator contents were collected at the end. In addition,
the principal investigator -elected to grab sample the initial Dupont Riston
SRS-120 still charge. This charge consisted of 85 gallons of Recyclene still
overhead collected from previous Recyclene batch still operations. The
initial Riston charge was then analyzed for volatiles, extractables, and
solids to fully characterize the system. Finally, the total metals analyses
proposed, were not conducted in order to reduce program analytical costs.*
During the previous day's production, 1500 (12"xl8"), 2-sided boards were
developed through the spray application of 1,1,1-trichloroethane. Roughly
50 percent (2,250 ft^) of the photoresist was dissolved during this
operation and then accumulated in the 85 gallon capacity of the Dupont Kiston
solvent recovery still. On the day of testing, the still bottoms were drained
from the SRS-120 and pumped into two 55-gallon drums. The feed to the
Recyclene RX-35 still consisted of 25 gallons of contaminated solvent
transferred from one of these drums.
Analytical Results
The various process flow streams are primarily either solvent or
dissolved photoresist (with the possibility of trace metals). Thus, the
composition of the streams can be determined through both a volatile organic
and total solids analysis. The results of these analyses are summarized in
Tables 55 and 56, respectively. Note that a large percentage (6.7-11.0) of
the total solvent mixture is composed of carbon tetrachloride. The presence
of this solvent is unexplainable since it is not normally found in solvent
waste streams typical to printed circuit board manufacturing. In fact, the
relatively high concentrations of carbon tetrachloride in the feed and product
streams came as quite a surprise to both the investigators and plant
personnel. While it was determined that Facility D had not used carbon
tetrachloride for some time, it is possible that it was introduced into the.
system from old solvents stored onsite in contaminated containers and not as a
breakdown product from the recyclene still.
*In accordance with revised proposal to EPA Project Monitor Harry Freeman
dated 28 February 1986.
104
-------
TABLE 55. SUMMARY OF ANALYTICAL RESULTS FOR VOLATILE ORGANIC COMPOUNDS
Analytical remit* (w/wZ)
Paraaiitcr
I.I, l-frichloroutliano
Other lolvcnt (total)
Metliylene chloride
Acetone
1, l-Dicliloroethenc
1 . 2-UicMoroct liane
2-Butanone
Carbon let raclilor ide
Vinyl acetate
2-lleManone
Tetracliloroetliene
••Slrcun la
Kiaton
still
initial
charge
92.0
12.14
0.5
—
0.64
--
~
II. 0
—
~
—
Slrc.ua Ib
Hilton
• till
coiit inuoui
feed
100.0
IS. 09
0.52
O.'i9
0.64
0.89
3.6
B.l
0.46
O.U
0.22
Strcooi 2
Rliton
it ill
diatillute
100.0
9. 83
0.48
0.17
0.18
0.20
I.J
6.7
0.08
0.08
0.08
StreOB ]
Water
•epiratar
discharge
0.19
0.20
0.08
0.01
0.01
0.01
0.02
0.01
-
-
-
Stream 4
Rilton
• till
bultoiai
92.0
10.0}
0.28
0.19
0. II
0.11
1.2
8.4
0.11
0.11
0.11
Struiim 5
Kecyclene
• till
distillate
92.0
12.05
0.2l'
0.22
0.66
0.12
1.1
9.8
0.12
0.12
0.12
Si re. in 6
Recyclenc
• till
buttona
I.S
1.01
0.01
0.01
0.98
0.06
O.OI
0.79
0.01
0.01
0.07
-------
TABLE 56. RESULTS OF SOLIDS ANALYSIS
Waste scream Description Concentration (mg/kg)
la
Ib
2
3
4
5
Initial Riston Charge
Riston Still Feed
Riston Still Distillate
Water Separator Discharge
Riston Still Bottoms
Recyclene Still Distillate
460
1,200.0
1.7
-
23,000.0
6.4
106
-------
For the purpose of this case study, 1,1,1-trichloroethane and carbon
tetrachloride were combined into the general category of solvent. Other
components such as 1,1-dichloroethene or 2-butanone (MEK) were either
1,1,1-trichloroethane breakdown products, buffering agents, or corrosion
inhibitors. The final assumption necessary for a complete mass balance is
that all components less than 0.12 percent (below detection limits) are
equivalent to 0. The only exceptions are 1,1-dichloroethane,
1,2-dichloroethane, and tetrachloroechane in the recyclene still bottoms.
Dupont Riston Still Characterization—
The Dupont Riston still characterization (see Table 57) consisted of
sampling; the initial Riston still charge (Stream la), the Riston still feed
(Stream Ib), the distillate (Stream 2), and the water separator discharge
(Stream 3). The initial Riston charge (IRC) which was described previously,
was analyzed to contain 98.81 percent solvent, 1.14 percent "other" volatile
components, and 0.05 percent total solids. The solids concentration in the
IRC was greater than expected, due to a resaturation of the solvent by
polymerized photoresist which had collected on the sides of the still. Solids
were continually added during the distillation process by the feed stream
which contained approximately 99 percent volatiles and 0.12 total solids.
Distillation of the IRC and feed Stream resulted in a clear overhead which
contained 96.62 percent solvent, 3.37 percent other volatiles, and less than
0.0002 percent solids. The water separator discharge which was grab sampled
at the end of the day was found to contain virtually no solids, 0.46 percent
solvent, and the remainder was other components, primarily water.
TABLE 57. DUPONT RISTON SRS-120 SYSTEM CHARACTERIZATION
Concentration (wt Z)
Parameter
Solvent
Solid
Other
IRC
(Stream la)
98.81
0.05
1.14
Influent
(Stream Ib)
92.89
0.12
6.99
Distillate
(Stream 2)
96.62
0.0002
3.37
Water separator
discharge (Stream 3)
0.46
-
99.54a
aC6nsists of 99+ percent water
Recyclene Still Characterization-
While the sampling and analytical assessment dealt with the entire two
stage solvent distillation system described earlier, the primary focus of this
case study is the performance of the Recyclene batch still. Thus, streams of
primary interest include Stream 4 (Riston Bottoms/Recyclene Feed), Stream 5
(Recyclene Distillate) and Stream 6 (Recyclene Bottoms). Table 58 details the
loading distributions calculated from the solvent mass balance for the
107
-------
parameters of interest. The distillation of the contaminated solvent resulted
in a clear overhead containing 92.8 percent solvent, 7.1 percent other
volatile components, and less than 0.001 percent of total solids. The low
accumulation of non-volatiles (solids) in the distillate resulted in a total
overhead purity of 99.99* percent and a volatile component yield of
99.78 percent. In comparison, manufacturer's specifications for the kecyclene
RX-35 solvent still were 99.5+ percent and 95+ percent, respectively.
Approximately 2.5 percent of the initial solvent charge was recovered as
bottoms product, with only 8.3 percent (4.36 Ibs) of the residual weight
classified as solvents. This represents a 97.5 percent decrease in waste
volume generation and a significant (99.8 percent) reduction in non-fugitive
emission related solvent losses.
TABLE 58. RECYCLENE STILL MASS BALANCE
Loading (lb/batch)a
Parameter
Solvent
Solid
Other
Influent
(Stream 4)
1853.7
47.0
142.9
Distillate
(Stream 5)
1849.45
0.01
142.79
Bottoms
(Stream 6)
4.25
46.99
0.11
•
aBased on 25-gallon charge.
Process Residuals—
Since it is expected that the residual bottoms product will be disposed
of through land disposal, GCA investigated the manufacturer's claim that in
some cases the RX-35 will convert hazardous residue to nonhazardous residue.
[Hazardous Materials & Waste Management, Nov./Dec., 1984] This goal was
accomplished through an EP Toxicity Metals analysis, an organic extraccables
analysis and a volatile organic analysis. Table 59, which compares Plant D's
EP Toxicity Metals results to Federal guidelines, clearly shows that the
bottoms product is well within current compliance standards and fairly low on
metallic contaminants. Organic extractables results for Plant D were all
below detection limits (see Table 60).These low concentrations (less than
0.0003 weight percent) resulted in these compounds not being included in the
process flow stream characterization. Therefore, the bottoms product is
assumed to contain no or very little priority pollutants. However, the
volatile organic analysis, previously presented in Table 57, show the
recyclene bottoms product (Stream 6) contains 7.5 percent by weight of
1,1,1-trichloroethane. This concentration classifies the recovery process
residue as a F002 (trichloroethane recovery still bottom) toxic hazardous
waste.
108
-------
TABLE 59. PLANT D E.P.TOXICITY METALS RESULTS FOR
RECYCLENE STILL BOTTOMS
Element
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Concentration
(mg/L)
0.03
0.106
0.002
0.003
0.03
0.0009
0.04
0.01
EPA standards
(mg/L)a
5.0
100.0
1.0
5.0
5.0
0.2
1.0
5.0
aU.S. EPA, Federal Register, V.45, No. 98:33122,
May 14, 1980.
109
-------
TABLE 60. FACILITY D SEMI-VOLATILE ANALYSIS DETECTION LIMITS
Sample I.D.
Description
Dececcion Limit (mg/kg)
Stream la
Stream Ib
Stream 2
Stream 3
Stream 4
Stream 5
Stream 6
Initial Riston
Charge
Riston Still
Feed
Riston Still
Distillate
Water Separator
Discharge
Riston Still
Bottoms
Recyclene Still
Distillate
Recyclene Still
Bottoms
0.76
0.75
0.75
0.20
15
0.76
500.0
110
-------
ECONOMIC AND ENVIRONMENTAL EVALUATION
Economic Evaluations
One of the goals of this program was to evaluate the economic
practicability of the Recyclene RX-35 batch still at Plant D. Based on the
recyclene still mass balance (Table 58), the quantity of waste generated prior
to the installation of the RX-35 unit was roughly 10,625 gallons per year.
Operation of the 2-stage solvent recovery system resulted in the recycling of
10,602 gallons per year of solvent not lost through fugitive emissions. This
figure represents a 97.5 percent reduction in waste volume and a 99.8 percent
recovery of solvent in the overhead. While these results are very
encouraging, the economic feasibility of the recyclene process will ultimately
determine the extent to which it is applied.
The capital costs for the RX-35 include the installed purchase price for
the basic unit ($25,850.00) and a start-up service fee ($300.00).
Differential energy consumption was calculated on Che basis of 47 Kwh per
batch at $0.06 per Kwh. In addition to electricity, operating expenses at
Plant D include labor (one manhour per batch) and liner consumption (1.5
batches per nylon liner) for a total cose of $6,478 per year. Finally, 6400
Ibs of residual solids were estimated to be disposed of throu'gh landfill ing,
at a'cost of $200.00 per ton. Thus, the total first year cost for
implementation and operation of a Model RX-35 with auto-fill at Plant D was
found to be $34,473.
Table 61 lists the annual cost savings and waste reduction calculated for
Plant D. As already stated previously, over 10,600 gallons of solvent were
estimated to be recycled in the first year of operation. This represents a
disposal savings, at $0.35 per gallon of solvent, of $3,71(1 per year.
However, more substantial is a savings of $47,709 per year in virgin solvent
purchases (at $4.50 per gallon). When the two savings totals are summed, the
aggregate annual savings is $51,428. This figure represents a net first year
savings of $16,955 (includes total capital cost) and an estimated investment
payback period, after considering credit for reclaimed solvent and reductions
in waste transportation and disposal costs, of 7.3 months.
Environmental Evaluation
While the Recyclene RX-35 solvent recovery still does significantly
reduce the volume of hazardous waste generated, it does not eliminate it
completely. In the advent of a total ban on the land disposal of hazardous
solvent wastes, alternate methods of disposal, such as solidification or
incineration, would have to be investigated.
Ill
-------
TABLE 61. ANNUAL COST SAVINGS AND PAYBACK FOR RECYCLENE RX-35 AT PLANT D
Cost item
Contaminated Solvent
Recyclene Bottoms
Differential Solvent
Purchase
Differential Energy
Consumption
Replacement Liners
Teflon
Nylon
Additional Labor
TOTAL COST
ANNUAL COST SAVINGS
Number of Cost per
units (per yr) unit ($)
10,625 gal 0.35
3.2 tons 200
10,602 4.50
20,092 kwh 0.06
52 bags 45.15
155 bags 6.50
208 hrs 15.00
(18C year)
RECYCLENE RX-35 PURCHASE AND INSTALLATION COST
PAYBACK PERIOD
Cost prior Cost after
to installation ($) installation (i)
3,719
640
47,709
1,205
2,348
1,010
3,120
51,428 8,323
43,105
26,150
7.3 mo.
-------
SECTION 7
FACILITY E CASE STUDY
FACILITY CHARACTERIZATION
A description of the general operating characteristics of Facility E is
provided below. This is followed by a brief summary of waste types generated
from specific processing areas and their respective management methods.
Finally, this section concludes with a detailed discussion of the activated
carbon treatment method which is utilized by Facility E to maintain the
electroplating baths.
Facility Description
Facility E began operations in January 1982 as a manufacturer of
customized, fine-line multilayer printed circuit boards. Facility E utilizes
a subtractive process to produce boards with up to 22 layers, which are then
shipped to other facilities for assembly operations. The plant employs
600 people and has an annual production volume of 600,000 ft* of finished
boards. Production volume is expected to double within the next few years
when the facility completes planned construction of an additional plating
line. Facility E currently operates 5 or 6 days/week, 24 hours/day. Its
onsite wastewater treatment plant operates 7 days/week.
Facility E initiated an ambitious waste minimization program in
mid-1984. Since that time, production has roughly doubled, but liquid
discharge to the wastewater treatment plant has remained constant and
wastewater sludge generation has dropped roughly 30 percent. Waste
minimization efforts continue to center around in-process modifications to use
nonhazardous or reclaimable solutions, to reduce water consumption and bach
dump frequency, and to optimize wastewater treatment operations. These
programs and a description of Facility E's production and waste treatment
processes are described below.
Facility E treats all process rinse waters and spills in the onsite
wastewater treatment plant. Aqueous process baths are treated in-line, in Che
treatment plant, or are temporarily stored in a tank farm to be reclaimed
offsite. The only solvent used in significant quantity in the plant is
1,1,1-trichloroethane (TCE), which is used as a presolder mask cleaning
agent. This is recovered onsite in a still equipped with provisions for
secondary recovery of solvent from drummed still bottoms. Solid hazardous
113
-------
wastes include wastewater sludge, ICE still bottoms, spent activated carbon
solids, and potassium hydroxide resist stripper sludge. Other filtered
solids, filter paper, and waste board materials (15 to 25 percent of
production) are nonhazardous and, therefore, disposed in a sanitary landtill.
Through the use of nonhazardous and reclaimable process solutions, Facility E
has significantly reduced the quantity of waste which would otherwise have to
be disposed of as hazardous material. A summary of the primary waste streams
is provided in Table 62. More detailed description of their origin,
disposition, and constituents of concern is provided below under the process
area in which they are generated.
Waste Sources
The production facility houses administration offices, raw material
storage, a wastewater treatment operation with a wet/dry laboratory, and
production lines. Shipping/receiving and a tank farm (21 tanks) are located
on the northeastern end of the production building. The onsite wastewater
treatment plant processes 280,000 to 300,000 gpd of complexed and noncomplexed
rinsewater and process baths in both flowthrough and batch systems. The
manufacturing facility includes separate processing areas for inner and
outer-layer operations. The process is described below with reagent usage and
waste generation discussed for each processing area.
Board Cutting/Inspection—
Facility E uses two-sided, copper foil clad, epoxy/glass cloth boards as
the base material for its printed circuit boards. Reject boards are disposed
in a sanitary landfill, along with any unreclaimable boards which are
defective due to improper processing. Facility E does not currently have
plans to investigate recovery options for these boards.
Inner Layer Chemical Clean—
Boards are chemically cleaned in two processing lines. The first line
employs sulfuric acid tfl^SO^) and Metex E-250 (50 percent potassium^
Yhydroxide; KOH), and the second uses hydrogen peroxide/sulfuric acid/baths
with copper sulfate (CuSO^) and a stabilizer (sodium salts and phosphoric
acid). These solutions and rinsewaters are sent to the waste treatment plant.
Inner Layer Image—
Boards are spray cleaned with 10 percent ^SO^, mechanically
scrubbed, and air dried. Resist (Dynachem Film Laminar TR, containing
methylmethacrylate)>is applied in a dry film laminator using rollers and then
exposed to UV light. The cleaning solution is discharged to the waste
treatment plant.
Inner Layer Develop, Etch, and Strip—
Resist is developed by dipping in three tanks containing 1 percent
potassium carbonate (1(2003).' This solution is discharged continuously to
waste treatment. Other process baths are filtered in-line using spiral-wound,
Sethco particulate filters to minimize bath dump frequency. The nonhazardous
solids and filter material are neutralized and disposed in a sanitary
landfill. Together with filtered solids from other processing areas,
2,200 gallons of this waste is disposed annually. All liquid discharges to
114
-------
TABLE 62. FACILITY B PROCESS WASTE SUMMARY
Chemical name
Ammonium Chloride
Nitric Acid
Sulfuric Acid
Metal llydroxyde Sludge
(F006)
Potassium Cyanide
Solution
1,1, 1-t rlcliloroetliane
Still Bottoma
Pol an slum Hydroxide
Sludge/Filter Paper
Potassium Carbonate
Developing Solution
Scrap Boards,
Edge Cuttings
Pb/Sn Solder Plating
Bath
Uae in process
Etchant
Rack stripping in
electrolesa and
electroplate
operations.
Epoxy smear removal
in electrolesa
plating area.
Uaatewater treatment
plant aludge.
Cold stripping
(closed process).
Solder mask cleaning
agent.
Inner layer resist
stripping solution
sludge collected on
paper filtera.
Inner layer and
outer layer.
In all process areaa.
Pattern plating.
Annual quantity
gal/yr (Ib/yr) Diaposition
208,000 Reclaimed and
supplied by CP
Chemical in Sumter.
8,100 Reclaimed by SCA
(CSX) in HJ.
24.000 Reclaimed by
City Servicea.
(1,040.000) Landfilled aa a
hazardous waate.
Recovered offaite.
220 Landfilled as a
hazardoua waate.
8.800 Landfilled.
Wastewater treatment.
(129,000 s.f.) Sanitary landfill.
220 Recovered.
Comments
Hazardous waste. Replaced CuLI.
Facility E planta to use Mercer
process (2-stage solvent
extraction, electrovent ing Cn) lor
acid reclamation in the futurp.
Contains III Cu with Nll^Otl. INI.
NII4CI. and (NII4)zCOj.
Will treat onsite beginning in
January. Contains Pb from solder
plating.
Excemption aa recyclable material.
9)1 pure in spent solution
(981 in virgin). Plsnning to
discontinue use in future.
Contains Cu. Ca, Na. and airvlit
photoresist. Small amounta of fl> .
Sn, Ag, Hn and high pH. No In.
TCE ia recovered onsite in a cloicil
loop, 2-stage distillation process
(901 recovery).
High pll. Consults primarily and resist
(acrylic polymer). Solidified
with lime.
Resist is too hydrolyiable to
settle by gravity.
Cu, Sn, Pb on boards. Looking
for reclaimer.
Send to vendor as hazardous
waste and reclaimed.
(continued)
-------
TABLE 62 (continued)
Chemical name
KUII Desist Stripping
Solution Sludge
Fluor inert /Filters
Sodium Hydroxide
Filtered Solids and
Filter Paper
Hydrochloric Acid/
Tliiourea
Na2SO4 Hicro-Etch
Use in process
Outer layer resist
stripper.
Fuse pre-cleaii.
Fuse pre-clean.
Primarily in
elcctroless and
pattern plate.
Fune prc-clean.
F.lpctrolcss and
micro-plate.
Annual quantity
gal/yr (Ib/yr) Disposition
4,300 Landfitled as a
hazardous waste.
Recovered offsile.
12,870 Wastewater treatment.
2,200 Landfill.
I), 730 Wastewater treataent.
37,860 Wastewater treatment.
Comment «
Resist 14 settled by gravity and
consolidated in an inclined
hydro-sieve filter. The si mine is
dewatered to 151 solids in a
holding bsg prior to drumming.
Hazardous due to Cu (IOX CuSU4),
Pb and pH.
Recovered by supplier.
prior to landfilling.
Electroleis elch 920O gal tjnk)
is dumped once per day. Cu
Monoe t linnn I an I ne
Stripping Solution
rotassiun Carbonate
Developer
I'alturn plating and
outer layer strip.
Solder mask.
16,730
Variable.
Waate treatment.
Waste treatment.
concentration ramies froa |O,()i>i) lu
50,000 ppm. Micro-plate elch is
dumped in snail quantities, I clrum
every month.
Used for in-line stripping to
correct image errors.
-------
waste treatment resulc from displacement of used solution in the tanks by
addition of make-up, which is added automatically to maintain necessary bath
characteristics (e.g., pH). This general arrangement is used in most other
discharged process baths. Counterflow rinse water is used to generate the
developer make-up solution. This conservation effort alone has resulted in a
20 percent reduction (5 gpm) in water requirement in this area.
juranpnium chloride (NfyCl) is currently being used as the board
etchant. This solution contains 12 oz. of Cu/gallon, as well as ammonium
hydroxide and ammonium carbonate. Etchant is used at a rate of roughly 2 gpm
and is reclaimed offsite by MacDermid, the raw material supplier. Together
with spent outer-layer etchant, over 2 million pounds of this waste are
generated annually. Facility E is currently exploring options for onsite
recovery (Mercer Process).
A ,2.5 percent 'solution of lpo^sjiuraJiyd^oxj.d^(KOH)JJis used as the inner
layer stripper. Spent solution is continuously fed to two in-line gravity
paper filters equipped with automatic paper advance. Approximately, 160 drums
of solid waste is generated annually consisting of 95 percent filter paper',
and only,5 percent~acrylic polymer sludge;. It is neutralized and solidified
with lime^prior tolandfilling as a hazardous wastej AC this juncture,
Facility E has not identified an alternative technology to effect separation
of Che highly hydrolizable resist. They are investigating the use of
fine-mesh, reusable filters.
Inner Layer Surface Treatment—
___ Inner layer boards undergo surface preparation prior to lamination. A
bronze oxide/potassium hypochloride solution is used to generate a rough
copper oxide layer which prevents peel ing.when the-board is laminated.
Cleaning solutions contain KOH, H2SO^, NaOH, and JlaClC^i A small
quantity of copper is stripped off during surface treatment so the Line is
equipped with a counterflow recovery rinse. Boards are air dried prior to
lamination. No chemical drying agents are used in this facility.
Lamination, Drill, and Deburr—
The only waste generated in lamination consists of fines (epoxy, acrylic,
some copper) which are filtered out of a recirculating water wash which keeps
the laminator clean. These fines are generated in small quantities and dumped
in a sanitary landfill.
Electroless Copper Platingr-
\Acid/alkali soluTions-are used for cleaning, rinsing, conditioning, arrd
activating the board Surface for palladium catalyst deposition. These
solutions contain Na2C(>3, H2S04> NaOH, NaF, KMnC<4 (residue oxidizer),
HC1 (activator), SnCl2. hydrazine/H2SC-4 (accelerator), PdCl2, and
organic activators such as ethanolamine. Many of these solutions are
proprietary mixtures supplied by MacDermid. The residue oxidizer (KMnC^) is
dumped to a waste treatment complexed solution system in small batches
(ISO gallons) where it is used to help break complexes.
117
-------
Tnicfo'-eTcB is used to prepare Che surface for catalyst^
application.- This"700 gallon tank has copper concentrations of 1 to 5 percent'
and is dumped once eachTday to the wastewater treatment-plant.-_ Facility £ is
currently experimenting with electrolytic recovery of copper from this bath,
but has not yet identified a viable method.
Ac^trvator and'Tccelerator solutions^ used for cataj.yst__application are
reclaimed offsite~by the vendor. Citric acid and H2&04 washesJ which
precede Che electroless copper placing tanks,' are discharged to wastewater
treatment. [Electroless baths We proprietary solutions containing copper
salts, formaldehyde, small quantities of CN (4 ppm), organic chelators
(e.g., EDTA), and NaOH. These electroless baths are reclaimed offsitte by
MacDermid. tountercurrent rinses and in-process filtration are used to reducfe
water consumption and extend bath life, respectively.; ' ~ "~"
Outer Layer Image Transfer—
Boards are spray cleaned with recirculated 10 percent fy50^*
mechanically scrubbed, and air dried prior to image transfer. Dupont Riston
3620 (contains methacrylatesT~rs applied'in"a dry film laminator and developed
through exposure"to UV radiations ~Sul"furic acid solutions^are 'discharged to
wastewater treatment.
Outer, Layer Developing— t . . ^
/Potassium carbonate (^(X^) vin a^ 6.5 percentjsolution is used as the
outer layer developing agent. This solution is continuously-discharged to
waste treatment as it is displaced by make-up fluid which is added to maintain
pH. This line is equipped with a KOH strip tankjwhich is used to reclaim
boards with image errors. The outer layer developing process is. currently
being upgraded by Facility E to miiiimizej water consumption in similar fashion^
-to-its inner_layer-counterpart (e.g.,^reusejrinse water in developing fluid
make-up).
Pattern Plating— . ,
Boards are pattern plated with ^ei'ght'acicPcopper, and ,one aqueous lead/tin
plating baths in a 48 tank plating line. ""The"line begins with~~_a^(nitric^acid_i
(HN03) rack strip tank.} Spent acid isf.hazardous> due to its lead contenjf
from solder plating". This is combined with HN(>3 rack strip 'from the
electroless line and reclaimed (8110 gpy^ offsitey These tanks are filtered
continuously to reduce"dumping frequency (twice per year). Filtered solids
are neutralized and disposed. <•
After the racks are stripped.) boards are loaded and then undergo rinsing,
cleaning with phosphate solutiojiSjd^PO^ Electroclean PC2000), and more
rinsing before-being plated.,j Acid copper baths contain^CuS04, suIfuric
acid, an organic brighter,_and chlorides with copper concentrations of
12 oz/gallon.l the solder__plating bath contains IHB03,__BF3,_ Pb(_BF_4)2i->
Sn(BF4)2, and organic acids.] The general processing procedure is to
activate the board surface (HC1), plate, clean/rinse, and replate.
Cleaning baths are continuously filtered and discharged to^wastewater
treatment by make-up fluid displacement. They are sent out for reclamation
when copper levels reach 1 to 1-1/2" Ib/gallon~as determined by in-process
monitoring. All rinses are countercurrerit flow and are discharged to waste
treatment.
118
-------
Copper and solder electroplating bacFs are ^treated with activated carbon
once every 3 months and every month, respectively. These electroplating
.never have to be dumped with this arrangement under normal processing
'conditions. The^activated carbon treatment process is described in detail in
the next section".
Outer^Layer Strip and Etch—
'Ammonium chloride (NH^Cl^'is also used as the outer layer etchant and
is reclaimed in similar fashion to the inner layer etchant. Hydrochloric acid
in a 10 percent solution is used as a post-etch solder activator and.cleaning
solution and is discharged to wastewater treatment. Potassium hydroxide (KOU)7
in a 5 to 20 percent solution is used as the resist stripper. It is
continuously recovered through gravity separation of the resist from the
solution in a Hydro-Sieve inclined cascade filter without requiring any
chemical addition. Filter sludge is collected, dewatered by gravity to
15 percent solids, solidified with lime, and disposed at a rate of 4,3uO gpy
in a hazardous waste landfill. The resist sludge is high in Cu (10 percent
CuS04>, Pb, and pH. Potassium hydroxide (KOH) solution is discharged to
waste treatment as it is displaced by make-up fluid.
Fuse-Preclean—
Solder is fused in a Fluorinert vapor blanket in a completely.closed
system. Fluorinert is a proprietary, long aliphatic carbon chain containing
fluorine^ The system is equipped with filters which are reelaimed?by the
chemical supplier. The vapor blanket is followed by a spray cleaner
containing 10 percent NaOH and a finishing solution spray containing
10 percent HC1 with thiourea. These spent solutions are discharged to
wastewater treatment at a rate of 45 gpd and 55 gpd, respectively.
Microplate--
Facility E plates nickel and gold/on board tabs in a microplate line.
Tabs are micro-etched with a sodium persulphate solution" (contains C.uS(>4 and]
M2S0^» which is discharged to wastewater treatment in small quantities
(one drum each month). Tabs are then nickel-plated using a NiSCfy J>ath.
Sulfuric acid^CH^SO^'and-. NiC03~are added for pH adjustment, (boric acid
^HBC^pis added as a buffer for nickel salts, and organic "surfates/aldehydes
are added as a stress reducer/brightner. ./Gold is plated.in.a gold cyanide~
bath containingGold plating and rinse solutions" are
,recovered in an adjacent line.^ The recovery process uses a stripping solution
""containing Technistrip Au II (aromatic hydrocarbons^) andlKCN^and lionic\
^exchange .equipment. Recovered .rinse water is returned to the process and
,spent ionic resin is reclaimed offsite. In addition, a gold recovery line has
also been installed adjacent to the wastewater treatment plant for stripping
gold off tabs of reject boards.
Solder Mask-
Solder masking is accomplished by first cleaning with
,1,1,1-trichloroethane (TCE)} followed by dry film application and aqueous
solution developing. TCE is recovered in a closed-loop still, which operates
with a recirculation flow of 1 gpm for a 6 to 8 hr/day to yield greater than
95 percent recovery/7 The system includes solvent storage tanks for used
product and for virgin make-up which deliver solvent into the still and
119
-------
process tank, respectively. The sci.ll operates automatically when ic receives
a charge from Che spent TCE tank. Heat is supplied co Che still througn a
Teflon heat exchanger and noncontact cooling water is used in Che Teflon
condenser. Still bottoms are dumped roughly once each month into a drum
enclosed in a heating jacket for second-stage recovery of solvent. Although
the liquid content of the waste could be reduced further, Facility £ fills and
disposes one drum of these bottoms every 90 days in order to comply with
hazardous waste drum storage regulations. Approximately one drum of make-up
TCE is added to the system each month.
Electroplating Bath Waste Management
Fundamental to the success of any modern printed circuit board is the
certainty that electroplated deposits will withstand the forces and stresses
that the board will encounter. For example, thermal changes during soldering
and power-up/power-down place stresses on the electroplated layers which can
cause cracks or failures. These failures are often the result of organic
contamination from addition agent breakdown products. Multilayer boards of
the type manufactured by Facility E are regulated by MIL-STD-55110 which
prohibits such failures in the finished product. To prevent the loss of
military certification, printed circuit board manufacturers lacking a bath
regeneration system, would typically be forced to either discharge the spent
plating bath in wastewater treatment, or send it offsite for reclamation.
i
The purpose of this case study is to evaluate the(extension ofj
electroplating .bath lifetimes,(and subsequent waste reduction) by activated]
ccarbon removal of.organic.brightner breakdown products^ The acid copper baths
were selected for study since recovery of this solution results in the most
significant amount of waste minimization.
Prior to the discussion of the regeneration technology, it is useful to
discuss the composition, function, and limitations of organic brightner
systems. Organic addition agents are a blend of leveling, carrying, and
ductilizing compounds. The carrier component, a high molecular weight
carbon-oxygen compound, acts as a plating inhibitor to prevent overplating and
burning in high current density areas. .The leveling agents are often amine
compounds or heterocyclic sulfur .compoundsjrequired to eliminate small-hole
wall imperfections due to drilling. The brightening (grain refining)
compounds are usually complex reaction products of sulfur and nitrogen;
'compounds used to improve the overall appearance of the deposit, tensile
strength, and ductility. Good tensile strength (usually greater than
40,000 psi) and elongation (greater than 10 percent) are required for the
deposit to withstand thermal cycling and thermal stress tests.
Organic jiddition'agents} which ensure optimal physical deposit
properties, must be stable under conditions of high agitation, current
density, and solution temperatures. Unstable brightner additives frequently
break down and may become incorporated into the copper deposit adversely
affecting the deposit properties. Large volumes of production work on a
continual basis will significantly shorten the lifespan of—Che-addition agents^
and, consequently, jthe placing solutions.' Therefore7>at Facility E, it is"
„necessary^tp__treat VhTeJereTtrolytic-copper-batb^with/activated carbon"7
"approximatel-y-every'-3~ months. >
120
-------
A secondary source of organic contamination in acid copper bachs is the
breakdown of photoresist during the pattern plating operation. As previously
mentioned, photoresists are light sensitive, organic thermoplastic polymers
which harden (polymerize) upon exposure co ultraviolet light. Incomplete '
exposure, developing, or rinsing can insult in a defective, nonstable resist
which can then break down and dissolve into the plating bath. While this
problem is less prevalent than organic brightner breakdown, it is still
detrimental to overall plating quality. This residue may also be removed
through activated carbon treatment.
At Facility E, jttivated carbon treatment is performed in a batch mode
for acid copper, solder, and nickel microplating baths in three separate '
systems. These systems consist of a holding tank, mixing tank, and MEFIAG
paper-assisted filtration unit. For acid copper treatment, 2,400 gallons of
contaminated solution is pumped into a 3,000 gallon mixing tank. Hydrogen
peroxide is added to oxidize volatile organic species and the temperature of
the bath is maintained at 120 to 130°F for 1 hour. Powdered activated carbon
(80 pounds) is added and the contents are mixed for 3 to 4 hours, allowing
sufficient time for adsorption of the organic breakdown products.
The solution is prefiltered by diatomaceous earth, followed by
recirculation through a paper-lined MEFIAG filter to remove the suspended
activated carbon. The filter solids and paper are removed as needed when a
predetermined pressure drop across the filter is reached. When the bulk of
the activated carbon has been removed (generally after three passes of the
solution through the filter), the filter is precoated with approximately
0.67 cu ft (5 gal) of diatomaceous earth. The partially treated solution is
further recirculated through the filter until a particulate test indicates
sufficient solids removal (no residue detected on visual examination of
laboratory filter paper), ^otal spent solids from plating bath purification
is approximately 3.5 cu ft per batch (about 1-1/2 drums every 3 monthsj/which
is disposed in a sanitary land fill. ^ "A schematic of the activated carbon'
treatment process is1 presented in Figure 11.
- -3 ^ \ )*»
PROCESS TESTING AND ANALYTICAL RESULTS
Process Testing
GCA tested the plating bath carbon reclamation technology at Facility E
during the week of February 10, 1986. Testing of the electrolytic recovery
system described in the QA Project Plan was not conducted for two reasons:
1) Facility E had inadvertently dumped the static rinse batch which was
critical to the sampling; and 2) delays at Facility E would make testing of
the new system difficult under the time frame for this program. Sampling
activities conducted at the site were limited to the bath reclamation system
as discussed below.
Samples for the activated carbon treatment system evaluation were taken
from the 3,000 gallon agitated treatment tank and the 30 gallon MEFIAG
activated carbon filter. A brief discussion of how each waste stream was
sampled is presented below:
121
-------
3.000 GAL
HOLDING
TANK
o
•ACTIVATED
CARBON
ADDITION
N>
ACID COPPER
PLATING
BATH TANK
3,000 GAL
AGITATED
TREATMENT
TANK
\
30 GAL CAPACITY
HEFIAG ACTIVATED
CARBON FILTER
-*—FILTER PAPER
AND SUPPORT
PLATE
SPENT ACTIVATED
CARBON/FILTER
PAPER
Figure 11. Facility E activated carbon treatment system.
-------
• Contaminated Copper Solution - Samples of Che contaminated copper
solution were collected from the top of the continuously stirred
treatment tank at approximately 11:00 p.m. The contaminated copper
solution sample, obtained from the tank by use of a plastic ladle,
was taken by one of tfte plant employees. Duplicate samples of
volatile organics, extractable organics, metals and TOC , were taken
from the tank at this time. These samples were held in the sample
cooler until the sampling activities were completed the next day.
• Spent Activated Carbon - Spent carbon samples from the filter press
were field composited for metals, extractable organics, and TUC over
the duration of the treatment operation. Volatile organic samples
were taken at discrete sample times. The composite samples were
volume composited, based on the approximate percent recovery of
carbon found in each filter run. The first sample composite
contained diatomaceous earth which is used as a prefilter. Carbon
was scraped off the filter paper and into the bottle. Samples were
taken for volatile organics, EP TOX, and metals. Two samples of
diatomaceous earth were also taken for metal analyses.
t Clean Copper Solution - The treated-filtered copper solution was
taken from a valve on the feed line to the clean storage tank.
Samples for extractable organics, metals, TOC and volatile organics
were collected at the end of the 18-hour run.
• At the completion of sampling, all samples were placed in coolers
with ice and vermiculite and shipped to the analytical laboratory by
Federal Express.
A complete summary of the measurements, parameters, and observations
recorded during the sampling period are listed in Table 63. In addition to
the sampling matrix already outlined in the Quality Assurance Project Plan, a
cyclic voltaic stripping (CVS) analysis was performed by the SEL-KEX division
of the OMI Corporation. Since SEL-REX is the manufacturer of the 70/30 acid
copper bath used by Facility E, they routinely conduct this analysis as a
customer service to determine brightner concentration.
Analytical Results
As previously stated, the test plan was designed to characterize the
plating solution and determine the effectiveness of activated carbon for
plating bath regeneration. The sampling parameters examined were trace
metals, total organic carbon, semi-volatiles, volatiles, and CVS. A
discussion of the results for each analytical parameter is presented below.
Trace Metals—
The objective of the trace metals analysis was to determine, through
atomic absorption analysis, the concentration of metallic ions at each
sampling location. The resultant data allows both a mass balance and spent
activated carbon characterization to be determined. Table 64 presents a raw
data summary and describes the metallic ion loading for the contaminated and
filtered plating solution, as well as the activated carbon residue. The
123
-------
TABLE 63. SUMMARY CF PROCESS OBSERVATIONS ANTJ SAMPLING AT FACILITY E
Time Process observation measurement
11:50 p.m. Sampled spent acid copper plating bath.
12:00 p.m. Began heating bath via steam.
00:52 a.m. Add 4 gallon 50% (H202).
01:00 a.m. Reached process temperature of 138°F.
03:30 a.m. Add 88 Ibs powdered activated carbon.
08:00 a.m. Load diatomacious earth onto filters
from holding tank sludge.
08:05 a.m. Pause for electrical repair.
08:25 a.m. Filtration begins with pressure at 20 psig.
08:35 a.m. Checked return which was gray/black in
appearance.
09:30 a.m. Checked pressure - 20 psig.
09:35 a.m. Break down filter, drainage emptied to
holding tank.
09:40 a.m. Sampled first filter run. Composited 502
of sample volume from top filler paper.
Sampled VOAs.
10:08 a.m. Start second filter run.
10:12 a.m. Checked pressure - 20 psig.
11:30 a.m. Break down filter.
11:35 a.m. Sampled second filter run. Composited 30Z
of sample volume from top filter paper.
Sampled VOAs.
12:00 a.m. Began third filter run.
01:15 p.m. Break down filter.
(continued)
124
-------
TABLE 63 (continued)
Time Process observation measurement
01:20 p.m. Sampled third filter run. Completed
sample composed from cop two filter papers.
Sampled VOAs from third and fourth filter
papers.
01:45 p.m. Began fourth filter run.
02:45 p.m. Checked pressure - 20 psig.
02:50 p.m. Sampled virgin diatomaceous earth.
03:25 p.m. Break down filter. Unable to sample since
all but carbon fines had been removed.
03:50 p.m. Slurry diatomaceous earth.
04:00 p.m. Coat filters with diatomaceous earth for
final run.
05:50 p.m. Sampled clean bath. Completed testing.
125
-------
TABLE 64. FACILITY E TRACE METAL SUMMARY AND MASS BALANCE
Parameter
description
Concentration (mg/L)a
Copper
Lead
Tin
Loading (Ibs/batch)
ACT-L
Contaminated
solution
21,500
1.1
6.9
Sample ID and
ACT- 2
Spent
activated
carbon
107,000
99
420
description
ACT-3DIA
Virgin
diatomaceous
earth
95.6
57
710
ACT-3M
Filtered
solution
21,400
0.66
8.3
Copper
Lead
Tin
471.33
0.024
0.151
9.42
0.009
0.037
461.91
0.015
0.114
aACT-2 and ACT-3DIA concentrations are ug/g.
126
-------
initial copper loading of the contaminated plating solution was 471 Ibs or
12 oz/gallon of copper sulface (CuS04.5H20). The filtered/carbon treated
plating solution contained 461.9 Ibs of copper or approximately 11.8 oz/gallon
of copper sulface. This represents a loss of copper to the spent activated
carbon of only 9.4 Ibs or 2 percent of the initial metallic copper charge.
Since the recommended range for electrolytic grade acid copper sulfate
solutions operating in the 0 to 40 amp/ft^ range is 9.3 to 13.4 ozjgallon,
the metallic copper loss was not detrimental to placing specifications.
In addition to metallic copper, the contaminated plating solution
contained small amounts of other trace metals such as tin and lead. These
metals are not recommended for optimum plating performance and represent a
source of inorganic impurities. Foreign anions such as tin and lead are
incrementally introduced into a plating solution throughout the solution's
operational lifetime. Common sources of metallic ion contamination include
leaching of parts, tanks and racks, or drag-in from previous plating
operations. If allowed to accumulate, these inorganic impurities will
exacerbate placing quality in Che following manner:
• increased resistance to flow of currenc;
• decreased bright range;
• increased tendency to burn;
• rough and pitted deposits; and
• reduced covering power.
While the^primary purpose of Che activated carbon filtration process at
Facility E is to remove organic molecules, Che spenc activated carbon data in
Table 64 shows that inorganic impurities~are adsorbed as,well.' /This
co-adsorption of inorganic contaminants has Che net effect of reducing total
lead and cin loadings in Che filtered soluCion (37.5 and 24.5 percenC,
respectively), as well as extending bach life and improving plating
performance. Therefore, it can be concluded that while activated carbon
treatment does remove a small quantity of divalent copper (approximately
2 percent of the bath content), the co-adsorpcion of inorganic impurities such
as tin and lead, is beneficial. r \
3 *K.\
The final objective of the trace metals analysis was to determine the
suitability of Che spenc activaced carbon residue as a landfilled hazardous
waste. Spenc activated carbon, which is currently classified as a hazardous
waste, is generated at the rate of 0.67 cu ft per batch. However, the results
of the EP Toxicity leachate test shown in Table 65 demonstrate that the spent
activated carbon contains metals concentrations which are within Federal
guidelines. This low toxic metals content, combined wich Che relatively
nonhazardous nature of the organics (chiocarbamoyl-chio-alkane sulfonate
class), many render the residue suitable for delisting.
127
-------
TABLE 65. EP TOXIC LEACHATE SUMMARY
Concentration EPA standard8
Element (mg/L) (rag/L)
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
0.03
0.042
0.002
0.065
0.17
0.0004
0.04
0.12
5.0
100.0
1.0
5.0
5.0
0.2
1.0
5.0
aU.S. EPA, Federal Register, Vol. 45,
No. 98: 33122. Hay 14, 1980.
Total Organic Carbon—
The objective of the Total Organic Carbon (TOG) analysis was Co determine
the overall organic carbon removal efficiency by the periodic oxidation and
activated carbon filtration system at Facility E. However, as previously
stated, brightner compounds are in the class of thio-carbamoyl-thio-alkane
sulfonates and are complex reaction products of sulfur groups (thiols) and
nitrogen compounds (amines). As such, these reaction products are usually
oxidized by the addition of hydrogen peroxide and volatilized during the
subsequent elevation of bath temperature from ambient to 120 to 130°F. The
apparently low removal efficiensies (13 percent) for total organic carbon
shown in Table 66 are somewhat suprising. Possible explanations for these low
removals include the presence of activated carbon residuals in the filtered
solution (ACT-1-3) or difficulties in analyzing the sample matrix. These
results do not necessarily indicate that the brightner system was not
preferentially adsorbed as shown below. For example, the high molecular
weight polynuclear aromatics of the type present in the carrier component of
the brightner system will readily adsorb to powdered carbon (EPA-6UO/8-80-023).
Conversely, low molecular weight carboxylic acid derivatives such as methyl
formate, which do not effect plating quality, will not be easily adsorbed,
especially in an acidic environment.
Volatiles and Semivolatiles—
The volatiles test results presented in Table 66 show little, if any,
adsorption of low molecular weight carboxylic acid derivatives such as methyl
formate, methyl acetate. On the other hand, sulfur dioxide, which is thought
to be a by-product of the brightner system, was completely removed. It is
more likely that the sulfur dioxide was volatilized during treatment than
adsorbed by the activated carbon. It must be remembered, however, that the
adsorption of organic substances from mixed solution is a complex phenomenon.
This can manifest itself in preferential adsorption of one substance over
others, nonadsorption if a substance is only weakly adsorbed, or Che
displacement of a weakly adsorbed substance by a strongly adsorbed substance.
128
-------
TABLE 66. FACILITY E TOTAL ORGANIC CARBON AND VOLATILE ANALYSIS
Samples ID and description
ACT-l-L ACT-1-2 ACT-1-3
Parameter
description
Contaminated
Contaminated solution-
solution duplicate
Average
Filtered removala
solution (percent)
Total organic carbon (mg/L)
Volatiles (mg/L)
257.9
241.4
218.4
ACT-1 - ACT-1-3
aZ removal
x 100
ACT-1
where: ACT-1 * (ACT-l-i + ACT-l-2)/2.
bDetection limit for GC/FID analysis.
CNC * Not appropriate for calculation.
13
Sulfur dioxide
Methyl formate
Methyl acetate
Acetone
Unknown
Unknown
2.1
1.9
0.43
—
0.22
0.12
5 x 10"6
1.9
0.52
—
0.17
0.10
<5 x 10'6
2.3
< 0.65
<0.08
—
~
99.998
NCC
NC
NC
—
—
129
-------
The semivolatiles test results, including nontargec compound analyses,
indicated that all compounds that could be identified were below priority
pollutant detection limits of 0.1 mg/L, except for phenols which were below
the detection limit of 5.0 mg/L. It may be concluded then, that the spent
plating bath and CTie spent activated carbon residue did not contain any
priority pollutants, possibly making the residue suitable for del is ting.
Cyclic Voltaic Stripping--
Cyclic voltaic stripping is an electrochemical analysis recently
developed by Haatt, Ogden, and Tench (Plating and Surface Finishing,
September 1979) for the determination of brightner concentrations in acid
copper baths. The method is one in which the potential of a rotating platinum
disc electrode is cycled at a constant rate. Copper is alternately deposited
on the electrode and stripped off by anodic dissolution. The resultant
current density is plotted against the electrode potential to determine
brightner concentration (Plating and Surface Finishing, December 1985). The
analysis is able to determine brightner concentrations with 5 to 10 percent
variance and a sensitivity of 0.2 mg/L on total brightner concentrations of
approximately 5 mg/L.
Previously, with standard analytical techniques such as spectrometry or
chromatography, it was difficult to control the concentration level due to the
interference of other bath components. However, with the CVS analysis, the
brightner concentration removal value could be easily determined and would be
in direct proportion to the decomposition product removal rate. On May 23,
1986, a CVS analysis was performed by the SEL-REX division of the OHI
Corporation (see Appendix). The brightner concentration was determined before
and after activated carbon treatment (ACT-1-1 and ACT-3-1). Prior to
activated carbon treatment the brightner concentration in the plating solution
was 6.42 raL/L. After carbon adsorption the total brightner concentration was
3.40 mL/L, representing a 47 percent adsorption of brightner and byproducts.
ECONOMIC AND ENVIRONMENTAL EVALUATION
Economic Evaluation
Evaluation criteria for the processing of contaminated electrolytic
plating baths for recovery and reuse include compliance with environmental
regulations and overall economics. Regulatory justification is based on RCKA
cradle-to-grave hazardous waste disposal responsibilities which include
ultimate liability for the mismanagement of hazardous waste. Economic
justification for the use of spent plating bath reclamation technology is
related to the increasing costs of raw materials and regulatory compliance
(waste treatment and disposal). A detailed current (i.e., 198b) cost estimate
and economic evaluation is presented in Table 67.
Capital Costs—
Capital costs for the treatment system are based on a Baker Brothers
Model 3020 Y activated carbon filtration unit with slurry tank (see
Appendix B). The stainless steel unit consists of 22 filter pads with a total
available filtration area of 40.5 ft2. Nominal capacity is 3,700 gpm,
although at open pumping capacity, throughput is increased to 4,800 gpm. The
130
-------
TABLE 67. ECONOMIC EVALUATION OF FACILITY E'S FILTER TREATMENT SYSTEM
Cose 1C em
Unit cost ($)
Cost C)
Capital Costs
(1) Model 3020Y*
Filter Treitment System
Miscellaneous
TOTAL CAPITAL
Annual OlM
Mefias, Filter Papers8
Electricity6
Maintenance
Labor
Powdered Activated Carbonc
50S Hydrogen Peroxide6
TOTAL O&M
Annual Costs
Annualired Capital'
Annual O&M
Annual Spent Carbon Disposal8
TOTAL COSTS
Annual Credit
Hazardous 'Waste Disposal6
Recovered Plating Solution*
TOTAL CRZDIT
TOTAL NIT CRZDIT (annual basis)
8,356
102 of purchase price
166/250
0.05/KHH
10Z of Total Capital
15/hr
TO.-96/lb
0.56/lb
0.177
UO/drum
(Copper) 10,000/bath
(Tin/Lead) 15,000/bath
8,356
S36
9.192
1.227
25
919
3,360
1.859
583
7,973
1.627/yr
7.973/yr
10,153/yr
iO,000/yr
15.000/vr
67,420
57.267
•Baker Brothers Technical Bulletin.
^Department of Energy, Zner^y Information Administration. National Average,
December 1986.
'McKesson Chemical Technical Brochure.
dAnnual costs derived by using a capital factor:
i(l-i)
CRT
where: i • interest rate and n • life in tne investment. A CR5 of 0.177
was used to prepare cose estimates in this docuaent. This
corresponds to an annual interest rate of 12 percent and an
equipment life of 10 years.
'As quoted by Clean Harbors Inc.
fOKI SSL-RZX Technical Brochure.
131
-------
COPPR-:_AD BOARD
. *INSr.ATtt
UNSrwATER
RXNSEUATB
AQUEOUS DEVEUriNC SOLUTION
UNSEUATEt
UNSEWATER*
sroit HATING SOLUTION
UNStVATSX*
AOUIOUS murraie SOLUTION
UNSSWATIR*
SPBiT ETOUN6 SOLUTION
UNSZUATE*
rucxr XAOCI
V:TR corrs
AN8 IIN/LEAB
I
S?E.T
OLCfllDI
SFSiT S7&IF7IHS SOLUTION
SOAU
•Dcnotu mfti ict<»« of tnutiit ce Chi* CM* ftudv.
Figure 3. Workflow diagram for Facility F.
132
-------
equipment is delivered preassembled and FOB cost for one unit is $6,356. The
capital cose estimate includes a contingency charge of 10 percent for cms ite
equipment modifications and related costs.
Operation and Maintenance Costs—
O&M costs are based on the operation of the activated carbon filtration
unit for 8 hours per treatment, 28 treatments per year. This includes four
activated carbon filtrations per year (one every 3 months) on each of the
,7-four, 2,400 gallon acid copper baths, and 12 (once/month) filtrations on the
1,200 gallon 60/40 tin-lead placing bath. Electricity costs are based on the
operation of one 440V, 3 hp. TEFC motor required to recirculate contaminated
solution throughout the treatment system. Labor costs were estimated only for
the operation of the unit and consist of 8 labor hours per treatment, at
28 treatments/year. Treatment chemical costs consist of 88 Ibs of powdered
activated carbon and 4 gallons (47.4 Ibs) of 50 percent reagent grade hydrogen
peroxide per acid copper treatment. Each 60/40 tin-lead bath treatment was
estimated to consume approximately half of these quantities due to
differential bath volumes. MEFIAG filter papers are used at a rate of
22 paper-lined filters/pass, three passes/treatment. Replacement costs are
$166/case of 250 filter papers. In addition to these operational costs, an
annual maintenance charge of 10 percent of the total capital has been included.
Total Annual Costs-
Total annual costs for the implementation of the activated carbon
filtration system in use at Facility E were approximately $10,153 and consist
of total capital, operation and maintenance, and spent activated carbon
disposal. The total capital cost was amortized over 10 years at 12 percent
interest. Annual spent activated carbon disposal costs (28 activated carbon
treatments/year) are based on telephone conversations with several hazardous
waste disposal companies. However, analytical test results seem to indicate
that if the spent activated carbon is noncorrosive in nature, it may be
suitable for delisting. Del is ting would further minimize the quantity of
hazardous waste generated at Facility E and decrease annual treatment costs by
an additional 5 percent.
Total Annual Cost Savings—
Total annual savings for the implementation of the activated carbon
filtration system were approximately $67,000 and consist of raw material
purchase and hazardous waste disposal costs. The raw material purchase
savings consist of 10,800 gallons of recovered plating solution at a cost of
$10,000/acid copper bath and $15,000/solder bath. Recovery volumes are based
on the assumption that prior to one full year of operation, organic impurity
concentrations are noncritical. After this point (based on a high volume work
flow), deposits will become burnt and powdery in nature and the bath will have
to be replaced. Disposal costs are related to current (1986) hazardous waste
facility pricing and represent a significant quantity of the total savings
(18.4 percent). As evidenced from Table 67, approximately $57,267 of net
savings were realized annually by utilizing the activated carbon filtration
system. This represents a payback period of under 3 months for this
application.
133
-------
Environmental Evaluation
The reduction in quantity of hazardous waste that could possibly be land
disposed is significant. Although most of the metals are reclaimed by either
the manufacturer of the plating bath or a commercial hazardoA waste treatment
facility, significant quantities of hazardous waste treatment sludge (FOUfo)
are produced and are currently landfilled. Activated carbon treatment,
therefore, is a cost-effective and environmentally sound technology for
reducing the quantity of hazardous waste generated by electrolytic plating
baths.
134
-------
SECTION 8
FACILITY F CASE STUDY
FACILITY CHARACTERIZATION
Facility Description
Facility F manufactures primarily double-sided single-layer printed
circuit boards using the subtractive method. The company is a job shop
employing approximately 300 people and producing an average of 40,000 square
feet/month of boards. Figure 12 details the process operations employed at
Facility F.
Waste Sources
Approximately 75,000 gallons/day of metals contaminated rinsewaters are
generated at Facility F. The rinsewaters which are of most concern, due to
contamination with dissolved metals such as copper and lead, are shown (marked
with an asterisk) in Figure 12. These rinsewaters generally contain very
dilute concentrations of the process bath constituents. Standard bath
constituents and their approximate concentrations (in the concentrated baths)
are shown in Table 68. Another wastewater source is the rinses from
photoresist developing and stripping operations. The photoresist used at this
facility is developed using an aqueous solution containing sodium carbonate
and butyl carbitol. Following the electroplating step, the light-exposed
resist is stripped using a different aqueous solution containing glycol ethers
and low and high molecular weight alcohols. Because both of these solutions
are primarily water, they are combined with other wastewaters and then
discharged to the sewer.
In addition to rinsewaters and resist developing and stripping solutions,
there are several process baths that are sent offsite after use to be
regenerated or disposed of. These include spent copper etching solutions,
spent solder stripping solutions, solutions used to strip metal from plating
racks, and spent solvent used to strip epoxy inks.
The copper etchant solution is an aqueous, alkaline solution containing
12 percent ammonium chloride. After a period of use, it will accumulate up to
15 percent dissolved copper. At this point, it will have lost its
effectiveness and must be replaced with fresh etchant. The spent etchant is
placed in a 4,500-gallon storage tank where it awaits pickup and regeneration
by the manufacturer. Approximately 70,000 gallons are sent offsite each year.
135
-------
TABLE 68. COMPOSITION OF PROCESS BATHS
Process bath
Major
constituents
Concentration (g/L)
Acid copper
a,b
Solder bath
(602 Tin - 402 Lead)a«b
Copper ecchanec
Tin/lead stripc
Copper microetch6
Electroless copper**
Sulfuric acid 52.5-135
Copper sulfate 160-300
Copper (Cu*2) 40-75
PC gleam (brightening agent)
Chloride ion 20-80 ppm
Stannous tin 56.2
Lead 26.2
Fluoboric acid 100.0
Boric acid 26.2
Peptone 5.2
Ammonium Chloride 122
Hydrogen Peroxide 102
Fluoric Acid 202
Sulfuric Acid
Hydrogen Peroxide
Copper 1.5-2.4 g/L
Formaldehyde 1.5-3.0
NaOH 6-8
chelacing agents
aEPA-600/2-83-033.
bBaths which contaminate rinses that were tested in this study.
cFrom plant-supplied Material Safety Data Sheets.
Product literature from Shipley Company, Inc.
136
-------
The solder stripping solution is also regenerated ofisice by the
manufacturer. It is composed of 20 percent fluoric acid and 10 percent
hydrogen peroxide, and will have accumulated high concentrations of tin and
lead after use. Approximately 30, 55-gallon drums are generated per month.
The only waste stream which is actually disposed of, as opposed to being
regenerated, is the spent rack stripping solution. These racks hold the
boards as they are immersed in the copper and the tin/lead plating solutions,
and consequently they also are plated with metal. Periodically, the racks are
placed in a solution composed of 50 to 70 percent nitric acid to remove the
plated metal from the stainless steel rack. The spent solution, containing
dissolved metals, is then picked up for offsite disposal as a hazardous
waste. Approximately 8,000 gallons of spent solution are generated annually.
Finally, a waste stream is generated by stripping epoxy inks from circuit
boards using methylene chloride. These inks are applied prior to gold plating
the tabs by silk screening through a mask with the image of the circuit
pattern on it. If a mistake is made due to misalignment of the mask, the
epoxy ink with methylene chloride is removed prior to curing. The spent
methylene chloride is sent offsite to be reclaimed.
Waste Management
Background—
The offsite management of several waste streams was mentioned above. Of
concern to this study, however, are onsite methods of reducing the quantities
of waste that would otherwise be managed offsite. At this facility, the major
process of this type is electrolytic recovery of metals from rinsewaters.
Electrolytic recovery has been practiced on rinses following several plating
baths for a little more than 1-year. The primary purpose of the electrolytic
reactors is to reduce the concentration of metals in rinsewaters which are
released to the wastewater sump. Prior to recovering metals from these
rinsewaters, a simple two-stage rinse system was used. This resulted in the
release of up to 3,000 ppm of copper and lead to the sump which would have
necessitated some type of end-of-pipe treatment system in order to comply with
increasingly strict pretreatment standards. Instead of installing an
end-of-pipe treatment system, however, a decision was made to try to attain
compliance by removing the contaminants at the source. Not only would this be
a much less expensive alternative, but it would also eliminate the generation
of large quantities of hazardous sludge that are associated with most
conventional treatment systems.
The location of the electrolytic recovery units has been changed several
times in order to achieve the greatest recovery of metals. For example,
originally there was one unit used to recover copper from the rinse following
electroless copper plating. The amount of copper recovered, however, was low
and so the unit was moved to the copper electroplating rinse where there are
higher concentrations of dissolved copper with potential for recovery. There
are now four individual copper recovery units associated with this process,
and there are three units associated with the solder (tin/lead) electroplating
process.
137
-------
Currently, there are no longer recovery units associated with the copper
etching rinses, the third major source of metal-containing rinsewater. This
is because the electrolytic units which are currently in use do not have the
capability to recover copper from etching solutions. Electrolytic recovery
from copper etching solutions is difficult because the purpose of the etching
solution is to remove plated copper, and so once the copper is plated onto the
cathode of the electrolytic cell, it is quickly etched back into the
solution. Facility F is currently investigating the use of more powerful
units to recover copper from this waste stream.
Electrolytic Recovery System—
The installation of electrolytic recovery units required converting the
primary rinse tank into a static dragout tank, as shown in Figure 13, and
leaving the second rinse tank as a flowing rinse. The contents of the dragout
tank are continuously circulated through the electrolytic reactor(s) and back
into the dragout tank. As the solution passes through the reactor a snail
amount of metal is plated onto the cathode. Since plating solution,
containing dissolved metals, is continuously input to the dragout tank, the
removal of metal by the electrolytic reactor is only able to maintain a
certain concentration of metals in this solution. The concentration is
maintained, however, at a low enough level so that drag-in of metals to the
secondary rinse is minimal. The secondary rinse solution can then be released
to the wastewater sump containing only a low concentration of metals.
The electrolytic reactors used at this facility are simple, compact
units. They consist of a vastewater sump, a pump, and the anode and cathode,
contained within a rectangular box with dimensions of approximately
22 in. x 10 in. x 22 in. The anode is cylindrical and is encircled by a
stainless steel cathode with a diameter of 8 inches and a height of 6 inches
[Agtnet Equipment Corp.]. The anode material used for copper plating solutions
is titanium. For tin/lead plating solutions, however, the anode material is
columbium. The columbium anode is required for the tin/lead rinse because the
fluoroboric acid in these solutions was found to be extremely corrosive to
titanium. Other pertinent characteristics of the electrolytic reactors tested
are presented in Table 69. These units are operated at constant voltage, and
so the amperage will vary according to the conductivity of the solution. The
higher the concentration of electrolyte in the dragout tank, the higher will
be the corresponding current.
TABLE 69. ELECTROLYTIC REACTOR CHARACTERISTICS
Cylindrical anode within
Design parameter cylindrical cathode
Cathode area 1 ft2
Reactor volume 1.3
Maximum flowrate 16.3 gallons/minute
Amperage 0.5 to 20 amps
Power 110 VAC
Cathode material Stainless steel
Anode material Columbium or titanium
138
-------
DIRECTION OF BOARD
TRAVEL
ELECTROPLATING
BATH
CATHODE
ELECTROLYTIC REACTOR
SAMPLING LOCATIONS
750 GALLON
DRAGOUT BATH
750 GALLON
"SECONDARY
RINSE"
RINSEWATER DISCHARGE
TO WASTEWATER SUMP
(25-50 PPM DISSOLVED METALS)
2 GPM RINSEWATER
Figure 13. Electrolytic recovery units employed at Facility F.
-------
The electrolytic units are operated 24 hrs/day, 7 days/week, except for
period maintenance requirements. Maintenance includes replacing parts,
especially related to the pump mechanism, and removing the metal foil that has
been plated onto the cathode. Foil removal is usually necessary once a week
for each of the units. The time required to clean each unit is about a
1/2-hour. Copper recovery per week has averaged about 10 pounds, and lead
recovery has averaged about 5 pounds/week.
PROCESS TESTING AND ANALYTICAL RESULTS
Process Testing
Sampling of rinsewater and process streams associated with copper and
tin/lead electroplating was conducted on February 18-19, 1986. There were
three sampling locations associated with each of the electroplating
processes. These were the plating bach itself, the dragout bath and the
secondary rinse. The sampling activities were conducted over a 24-hour period
starting and ending at approximately 9:00 a.m. The most important parameter
to define was the concentration of dissolved metals (copper, lead and tin) in
the dragout bach. Since Che dragout bath is circulated through the
electrolytic reactor 24 hours/day, samples were taken every 4 hours over the
24-hour period. The printed circuit board plating line is only operated for
16 hours out of the 24-hour period (between 8:00 a.m. and midnight) and,
therefore, samples of the secondary rinse bath were taken every four hours
during this period.
Samples were also collected for analysis of total organic carbon (TOC).
These samples were collected to provide a general indication of the fate of
organic compounds when subjected to the electrolytic reactor. Because lesser
importance was attached to these types of compounds, sample collection was
less frequent.
Operation on the day of testing deviated from normal due to several
factors. The first of these was that the dragout tank following copper
electroplating had been emptied the previous day in order to fix the weir. It
was then refilled with fresh water. Consequently, the concentration of metals
in the tank was not as high as ic normally would be. In addition to this, one
of the electrolytic units stopped operating during the middle of the testing
because of a broken pump impeller. The effect of this would be to reduce
overall metal recovery from the bath 25 percent (since there are four units
altogether), and thus result in higher concentrations of metal in the dragout
bath and the secondary rinse tank.
Finally, che cathodes of one tin/lead and one copper recovery unit were
weighed at the beginning and at the end of the 24-hour sampling period in .
order to determine the quantity of metal that was recovered from solution.
For the copper recovery unit, che weight increase was 0.30 pounds. For the
tin/lead unit there was no difference in the weight of the cathode at the
beginning and the end of the 24-hour period, indicating that there were
problems with the unit.
140
-------
Analytical Results
Copper Electroplating-
Table 70 presents the measured concentrations of copper, tin, lead, and
total organic carbon in the copper electroplating, dragout and rins^ baths.
The concentrations of these constituents, particularly copper were measured to
determine the performance of the electrolytic reactors with regard to removing
metals from the dragout bath, and as a result reducing the concentration of
metals in the secondary rinse bath. Several indicators of reactor performance
are discussed below.
Secondary Rinse Copper Concentrations—The concentration of copper or
other constituent in this stream is important to know because this is the
stream that is actually released to the sewer (after mixing with other
wastewater streams). As listed in Table 70, the concentration of copper
ranged from 70 to 90 mg/L over the 24-hour period with the highest
concentration being at 8:00 a.m. The contribution of 90 mg/L of copper at
2 gpm to the final plant effluent can be estimated assuming the total plant
effluent is 75,000 gallons/day. This calculation is shown below:
90.4 mg/L x 2 gal/min x 960 min/day / 75,000 gal/day ' 2.31 mg/L
The maximum allowable daily discharge of copper is 4.5 mg/L [40 CFR 413].
Therefore, the other sources of copper must be kept below 2.2 mg/L for these
discharge limits to be met. Since there are several other sources of copper,
particularly that from the rinse following copper etching, these limits may be
difficult to achieve without increasing the number or power of the
electrolytic reactors.
Copper Recovery Rate—The rate of recovery of copper by the electrolytic
reactors can be determined in several ways. One of these ways is to monitor
the concentration of copper in the dragout tank over time. A. plot of this
relationship is shown in Figure 14. This plot shows that the copper
concentration increases in an approximately linear fashion between 1:00 p.m.
and midnight at a rate of 21.5 mg/L/hour. This increase in copper
concentration occurs despite the removal of copper by the electrolytic
reactors, indicating that the rate of input of copper due to dragout from the
plating bath is greater than the rate of removal achieved by the electrolytic
reactors.
From midnight until 9:00 a.m., when there is no dragout of copper from
the plating bath, the copper concentration decreases at a rate of
approximately 5 mg/L/hour. Knowing that the size of the dragout tank is
1,000 gallons, the mass rate of removal of copper is calculated to be
equivalent to almost. 19 grams/hour.
The rate at which copper is removed from the dragout tank was also
estimated by weighing the cathode of one of the electrolytic reactors both at
the beginning and the end of the 24-hour testing period. This weight
measurement showed a 0.30 Ib increase over the 24-hour period indicating that
an average of 5.7 grams/hour of copper were removed from solution and plated
onto the cathode. Since there are four electrolytic units used on the copper
141
-------
TABLE 70. CONSTITUENT CONCENTRATIONS FOR COPPER ELECTROPLATING PROCESS (mg/L)
IN
Nl
Plating bath
Time Cu Pb Sna TOC
9:00 am 27,000 3.2 17 873.1
1:00 pm
5:00 pm
9:00 pm
1 : 00 am
5:00 am
8:00 am 888.6
9:00 am
Dragout
Cu Pb Sna
326 0.45 3
416 0.54 2
498 0.38 5
530 0.64 3
508 0.21 3
490 0.45 3
tank Second rinse
TOC Cu Pb Sna
11.42
.2 14.05 70.5 0.26 4.0
.8 16.90 77.6 0.08 4.3
.1 83.6 0.16 5.3
.2
.7
23.13 90.4 0.35 5.0
.8
aTin data considered invalid, see Section 6.
-------
600 -I
40O -
30O -
a
E
200 -
100 -
O
(BAM)
X - CURRENT IN ELECTROLYTIC
REACTOR
A - COPPER
O - TIN
D - LEAD
MIDNIGHT (plallng stops)
n
n
- a
-e
CO
Q.
ill
tc.
oc
3
u
- 2
10
20
I
26
TIME (hrs)
Figure 14. Copper concentration in dragout finse vs. time.
-------
clragout tank, the total amount of copper removed would be approximately
22.7 grams/hour. During testing, however, one of the units broke down and did
not function for the full 24-hour period. Therefore, the quantity of copper
removed would be slightly less than expected. Nonetheless, this method and
the graphical method of estimating copper removal rates yield similar results.
Finally, one can determine the efficiency of the reactors by comparing
the actual amounts of copper recovered to the theoretical maximum amount that
could be removed. Faraday's Law states that the amount of material that can
be produced electrochemically is proportional to the amount of charge in
coulombs [Snoeyink, V. L. and D. Jenkins, 19801. A coulomb is equivalent to
the amount of charge transferred when 1-ampere of current flows for 1-second.
Using this fact and knowing that the average current over the 24-hour period
was 5 amperes (see Figure 14), one can calculate the theoretical amount of
copper recovered in 1-hour.
31.7 g Cu x 5 amperes x 3,600 sec/hr 96,500 coulombs - 5.9 g/hr
equivalent equivalent
This is the amount of copper recovered per reactor, so with four reactors the
amount that theoretically could be recovered is 23.7 grams/hr. Then, the
current efficiency of the reactors can be determined by dividing the estimated
"actual" copper removal rate into this theoretical rate. Depending on which
estimate of actual removal rate is used (18.9 or 22.7) the current efficiency
is calculated to be between 80 and 90 percent. As indicated in Table 71,
however, the removal efficiency of the electrolytic reactors, based on a
copper input rate of 100.3 grams/hour is between 18 and 22 percent.
TABLE 71. COPPER RECOVERY DATA
Copper input to dragout tank* 100.3 grams/hour
Copper removal rate 18.9 - 22.7 grams/hour
Theoretical recovery rate 23.7 grams/hour
Current efficiency 80 - 96 percent
Removal efficiency 18 - 22 percent
Calculated by assuming that total copper input is
equal to the rate of increase of copper concentration
between 8:00 a.m. and midnight, plus the rate of
decrease of copper concentration between midnight and
8:00 a.m.
Tin/Lead Electroplating—
Table 72 presents the measured concentrations of tin, lead and copper in
the tin/lead electroplating dragout and rinse baths. Due to the complexity of
the sample matrix, the accuracy of the tin analyses on all baths and the
analyses of lead in the plating bath were very poor. As a result, it was not
posible to use these data to make any definitive conclusions. In addition, as
is discussed below, the data on the tin/lead baths in general does not show a
144
-------
TABLE 72. CONSTITUENT CONCENTRATIONS FOR TIN/LEAD PLATING PROCESS (mg/L)
Plating bath Dragout tank Second rinse
Time Cu Pb Sna Cu Pb Sna Cu Pb Sna
9:15 am 2.91 5,000* 2,400
1:15 am
5:15 am
9:15 am
1:15 am
5:15 am
4.88
4.76
4.28
3.36
4.22
2,400
2,500
2,100
2,500
2,300
2,200 1.57 45
2,100 0.43 56
380 0.38 64
410
1,500
38
11
13
8:15 am 3.02 5,700* 2,700 0.28 22 9.0
9:15 am 4.23 1,900 1,700
aData considered invalid, see Section 6.
145
-------
clear relationship between netal addition and removal as did :ne copper aata.
This may be in pare due co the analytical difficulties, but it may also
indicate that the electrolytic reactors were noc functioning properly.
Nonetheless, it was possible to make some general conclusions. These are
discussed below.
Plating Bath—The average, measured concentrations of tin and lead in the
tin/lead plating bath are, respectively, 2,550 and 5,350 mg/L. These
concentrations are almost an order of magnitude lower than the concentration
of copper in the copper electroplating bath. Consequently, the amount of
metal which will be dragged out of the bath should also be lower.
Dragout Bath—The measured concentrations of lead in the dragout bath
range from 1,900 to 2,500 mg/L. These concentrations are at most 85 percent
less than the concentration of the metal in the plating bath itself, and more
commonly the concentration is only 50 percent less or close to equivalent.
This indicates that the lead is only being removed to a very small degree by
the electrolytic reactors.
Figure 15 shows a very erratic curve for metal concentration vs. time in
contrast to the copper plating case; this data does not show a clear
relationship between input of metal to the dragout bath and removal by the
electrolytic reactor. Instead, both the tin and lead concentration appear to
drop during the period when it would be expected that the input of metal would
exceed the removal by electrolytic recovery. Then, during the midnight to
8:00 a.m. period, when there is no input of metal to the dragout bath, the tin
concentration rises from less than 500 mg/L to greater than 1,500 mg/L. As
mentioned above, the difficulties in analyzing these samples may be the cause
of these unexplainable results.
Two other indicators of poor recovery of lead and tin from this dragout
bath are:
• The low amperage of the electrolytic reactors; and
• The unmeasurable amount of metal plated onto the cathode of one of
the reactors.
Firstly, the current indicated by the ammeter on the electrolytic units
remained below 1-ampere during the entire test period. With a current of
1-ampere, the maximum removal of tin and lead would be, respectively, 1.1 and
1.9 grams/hour/unit (from Faraday's Law). To increase the rate of recovery,
the amperage and/or the number or electrolytic recovery cells could be
increased. Increasing the amperage, however, may increase the generation of
gases (such as oxygen and fluorine) at the anode and result in the plated
metal being of poorer quality. This may then result in the etching of plated
metal back into solution.
The other method for quantifying the removal of metals by the
electrolytic reactor was to weigh the cathode of one of the units at the
beginning and the end of the 24-hour period. In doing this, it turned out
that the weight of the cathode did not change, indicating that no metal was
removed from solution by this electrolytic reactor.
146
-------
2500 -
2000 -
~ 1500 -
S
1000 -
500 -
O - TIN
D - LEAD
A — COPPER
MIDNIGHT (plating stops)
A
A
A
A
A
A
i I l I I
0 5 10 15 20 25
(8AM) TIME (hrs)
Figure 15. Tin/lead concentrations in dragout bath vs. time.
-------
Second Rinse—The concentrations of netals in the secondary rinse are
also erratic and do not clearly correspond to Che concentrations in the
dragout bach. This may be due Co analytical difficulties with these samples.
Nonetheless, averaging the four analyses for each metal results in a lead
concentration of 47 mg/L. Assuming a total plane efficient of 75,000 gal/day,
the concentration of lead in the final effluent due Co this 47 mg/L would be
1.2 mg/L. Maximum daily allowable levels of lead are 0.6 mg/L. In order Co
achieve this level, Che concentration of lead in the secondary rinse would
have to be lowered from 47 to 23 mg/L.
ECONOMIC AND ENVIRONMENTAL EVALUATION
This facility installed electrolytic recovery units in order Co reduce
Che amount of metals in its final effluent. It appears, however, that this
technology is noc one that can be used by itself Co achieve effluent discharge
limits, particularly for lead. Its advantage, then, is to remove some of che
metals from the rinsewaters at the source, thereby lessening the amount of
end-of-pipe treatment chaC must be done. For example, if a precipitation
system is used to Creat che total plane effluent, the amount of metal
hydroxide sludge that will be generated by this system can be reduced by
removing some of the metals upstream using electrolytic recovery. Reducing
the quantity of sludge that is generated will be beneficial in both economic
and environmental terms.
Economic Evaluation
Tables 73 and 74 present cost estimates for electrolytic recovery of
metals from dragout baths following copper and tin/lead plating. Separate
estimates are presented for Che copper recovery and the tin/lead recovery
systems because the equipment cost for the two systems is different, and also
because maintenance requirements for the tin/lead system are expected to be
higher than for che copper system. Both of these cost differences are due to
the fact chat Che tin/lead bath is extremely corrosive, therefore
corrosion-resistant columbium anodes are required, and more frequent
maintenance is required to replace corroded parts. The basis for other
elements of these cost estimates is discussed below.
Capital Cost--
As mentioned above, the purchase cost for one unit to recover copper from
an acid copper rinse is less Chan che cost of one unit Co recover lead from a
Cin/lead fluoborate bath. The difference of 1,000 dollars, 3,500 vs.
4,500 dollars, is due Co che use of a columbium vs. a titanium anode.
Otherwise, the two units are identical. The cost quoted here is the cost for
which these units (Agmet Model 5200) were purchased in 1985.
The other element included in the capital cosC is for miscellaneous items
associated wich Che installation of che unit. These include piping and any
changes that have to be made to the rinse baths in order to install the
units. Miscellaneous coses have been set at 10 percent of che equipment cose.
148
-------
iLECTORYLTIC COPPER RECOVERY
Basis
Unit cost ($)
Cost ($)
Capital costs
• 4 recovery units
with Titanium anode*
• Miscellaneous costs such
as installation/piping
Total capital
Operation & maintenance
• Electricity
- for electrolysis
(3 volts, 10 amps/unit)
~ for pumps
(1/8 HP/unit)
• Maintenance
• Labor (250 hrs/yr)
Total 0 & M
Annual costs
• Annualized capital
(10S over lOyrs)
• 0 & M
Total annual cost
Annual savings
• Recovered copper (10 Ibs/wk)
• Sludge disposal
2.3 tons at 20Z solids or
23 tons at 2Z solids
• Waste treatment chemicals
(0.4 cons Ca(OH)2
Total annual savings
3,500a
102
$0.05/kwhrb
10?
il5/hr
0.1627
S0.22/lbc
4200/ton
*50/tond
45
139
2.505
5.474
7,980
114
455-4550
20
589-4684
aAgmet Equipment Corporation.
^Department of Energy, Energy Information Administration. National Average.
December 1986.
cPrice quoted to Facility F.
^Chemical Marketing Reporter. May 5, 1986.
149
-------
TABLE 7i. ELECTKOLYTIC TIN/LEAD RINSE KECCVERY
Basis
Unit cost ($)
Cose ($)
Capital costs
• 3 recovery units A,500
with columbium anode
• Miscellaneous costs 102
Total capital
Operation & maintenance
• Electricity $0.05/kwhx
- for electrolysis
for pumps
• Maintenance 20Z
• Labor (250 hrs) $15/hr
Total 0 & M
Annual costs
• Annualized capital 0.1627
(102 over 10 yrs)
• 0 & M
Total annual cost
Annual savings
• Sludge disposal $2QO/ton
0.9 tons at 20Z solids
or 9 tons at 2% solids
• Waste treatment chemicals $50/ton
(0.2 tons Ca(OH)2
Recovered tin/lead (5 Ibs/wk) $0.10/lba
Total savings
13,500
2,416
185-1,850
10
26
221-1,886
aPrice quoted to Facility F.
150
-------
Operation and Maintenance Costs—
The first of Che coses listed under this heading is electricity costs.
Electricity is required for the 1/8-horsepover pump contained within each
electrolytic reactor and also for generating the electric current necessary to
plate the metal onto the cathode. The electricity use for copper recovery is
based on operation at 3 volts and 10 amperes, and for tin/lead the recovery
voltage is 6 and the amperage is 5.
Maintenance costs for the tin/lead units are assumed to be 20 percent of
capital costs vs. 10 percent of capital costs for the copper units. As
mentioned above, this is due to the anticipated higher frequency of parts
replacement resulting from the highly corrosive nature of the fluoroboric acid
in the tin/lead plating solution. Personnel at this facility have indicated
that proper maintenance of the units is extremely important in order to
achieve maximum recovery.
Labor associated with these units is primarily for removing the plated
metal foil from the cathode of each of the units. This must be done
approximately once per week for each unit. Labor is also required for fixing
units that have broken down. Approximately 5 hours/week for the copper units
and the tin/lead units is assumed to be required.
Annual Costs—
Annualized capital cost was estimated using an interest rate of
10 percent over a 10-year period. Assuming a rate of recovery of 10 Ibs of
copper/week, the total annual costs/pound of copper recovered would be
approximately 15-dollars. For tin/lead recovery of 5 Ibs/week, the cost would
be almost 30 dollars/pound.
Annual Savings—
The use of electrolytic recovery units to remove metals from rinsewaters
at the source of generation will lower the amount of metals that must be
removed in an end-of-pipe treatment system. The "savings" that are presented
at the bottom of the table are those that would be accrued if electrolytic
reactors were used upstream of an end-of-pipe lime precipitation system. When
the electrolytic reactors are used, less metals reach the precipitation
system, and so less lime is required and less hydroxide sludge is produced.
The savings are based on recovery of 10 Ibs/week of copper and 5 Ibs/week of a
1:2 tin/lead mixture. The quantity of sludge not generated as a result of
recovering these metals would vary in volume depending on whether it was
thickened and dewatered. A range of cost values, one based on 20 percent
solids and the other based on 2 percent solids, is presented. A sludge of
20 percent solids is ten times less voluminous than one of 2 percent solids
and, therefore, the cost for disposing it would be correspondingly lower.
However, some type of equipment for dewatering the sludge, most likely a plate
and frame filter press, would be required to achieve 20 percent solids.
Therefore, the decreased sludge disposal costs would be offset by increased
equipment costs.
Tables 73 and 74 indicate that the annual costs associated with
electrolytic recovery exceed the annual savings. The savings, however, are
based on sludge disposal costs of 200-dollars/ton. With upcoming land
-------
disposal restrictions on certain m^tal-bearvng hydroxid-e sludges, however,
their disposal will most likely become much more expensive. Therefore, in tne
near future, Che cost savings may become much greater.
Environmental Evaluation
The environmental benefit of electrolytic recovery is that the quantity
of metal hydroxide sludge (RCRA code F006) that is generated by an end-of-pipe
treatment system is minimized. The removal of 10 Ibs/week of copper and
5 Ibs/week of tin and lead from dragout rinse baths reduces by 32 tons/year
the quantity of sludge (at 2 percent solids) that would otherwise be generated
by precipitation. Instead of being converted to hydroxide sludges, the
copper, tin and lead are plated onto the cathode in a metallic form so that
they can be reused.
152
-------
SECTION 9
QA SUMMARY
INTRODUCTION
Quality Assurance/Quality Control (QA/QC) procedures followed in this
program were based upon routine laboratory and field practice and the Quality
Assurance Project Plans prepared for this program in December, 1985. This
Quality Assurance section will summarize areas where changes in laboratory
and/or field procedures were made, and will address EPA comments on the
Project Plan made in memoranda dated February 28, 1986. To facilitate review
of pertinent QC data, this section will follow the outline of the QA Plan.
For a detailed description of QA/QC data or procedures for each facility refer
to either the Draft Report and/or the QA Plan for each facility.
PROJECT ORGANIZATION AND RESPONSIBILITY
During the course of this program, several major organizational changes
were made. In the analytical laboratory, Dr. Peter Lieberman replaced
Ms. Mary Kozik as Inorganic Section Head, and Ms. Joan Schlosstein replaced
Ms. Andrea Cutter as Analytical QC coordinator. In the field measurements
department, Mr. Howard Schiff replaced Mr. Graziano as Field QC coordinator.
PRECISION, ACCURACY, COMPLETENESS, REPRESENTATIVES AND COMPARABILITY
Analytical precision was estimated through the analysis of replicate
sample aliquots. Analytical accuracy was determined through the analysis of
EPA Environmental Monitoring and Support Laboratory (EMSL) Quality Control
Samples and the analyses of matrix spiked sample aliquots. Results of these
analyses broken down by facility, are presented in Tables 75 through 80 and
are discussed below.
Completeness, defined as the percentage of all measurements whose results
are judged valid, was determined to have been between 0 and 100 percent.
Wherever possible, reference methods and standard sampling procedures were
used as stated in the QA Plan to ensure comparability with other
representative measurements made by Alliance or another organization.
Facility A
Quality control procedures for trace metals, total organic carbon, and
total organic halide determination included the preparation and analysis of a
laboratory method blank, for which final results were corrected, a laboratory
153
-------
TABLE 75. FACILITY A QUALITY ASSURANCE SUMMARY
Precision
Accuracy
Compleceness
Parameter
QA
QA Com- objective Com- QA Corn-
objective pliance (2 pliance objective pliance
(Z RPD)a (2) recovery) (%) (%)b U)
Trace Metalsc
Trace Metalsd
Total Organic0
<30
<50
<30
100
100
100
70-130
50-150
70-130
100
100
100
95
95
95
100
100
100
Carbon
Total Organicd £50 100 50-150 100 95
Carbon
Total Organic6 £30 100 70-130 25 95
Halide
100
100
Total Cyanides
Total Cyanides
Hexavalent Chrome
Hexavalent Chrome
<30
I50
<30
<50
f
f
100
g
70-130
50-150
70-130
50-150
18
100
9
100
95
95
95
95
0
0
0
0
aRPD •= Relative Percent Difference
''Percentage of all measurements whose results are judged valid
cLiquid waste matrix
dSolid waste matrix
eAccuracy and precision analyses for solid wastes matrix were not performed
^Interfering substance (excess chlorine) rendered precision analysis
results invalid
8Accuracy and precision analyses for solid wastes matrix were not performed
154
-------
TABLE 76. FACILITY B QUALITY ASSURANCE SUMMARY
Precision
Parameter
Copper
Nickel
Lead
Zinc
Total Organic
Carbon
Total Organic
Carbon
Total Organic
Halide
Total Organic
Halide
QA
objective
-------
TABLE 77. FACILITY L (JlAim JiSSCKANCE. SL'M>iARY
Precision
Accuracy
Completeness
Parameter
QA
QA Com- objective Com- QA Corn-
objective pliance (Z pliance objective pUance
(Z RPD)a (2) recovery) tt) (2)b (Z)
Volatile Organics £40
Totalc Solids £50
Extractable Organics £75
100
100
100
50-160 100
-
10-150 100
95
-
95
1UO
-
100
8RFD * Relative Percent Difference
^Percentage of all measurements whose results are judged valid
cPrecision and accuracy goals for total solids were not set in the
Facility C Quality Assurance Plan
TABLE 78. FACILITY D QUALITY ASSURANCE SUMMARY
Precision
Accuracy
Completeness
Parameter
QA
QA Com- objective Com- QA Corn-
objective pliance (Z pliance objective pliance
(Z RPD)a (Z) recovery) (Z) (Z)b (Z)
Total Solids £50
Volatile Organics <40
Volatile Organics £75
Extractable Organics £75
100
100
100
100
__-c
50-160
50-160
10-150
100
100
100
95
95
95
100
100
100
aRPD •• Relative Percent Difference
^Percentage of all measurements whose results are judged valid
cPrecision and accuracy goals were not set in Facility D QA Plan
dGC/FID Analysis
eGC/MS Analysis (ERT)
156
-------
TABLE 79. FACILITY E DUALITY ASSl'KANCE Sl'kNnRY
Precision
Accuracy
Completeness
Parameter
QA
QA Com- objeccive Com- QA Corn-
objective pliance (2 pliance objective pliance
(2 RPD)a (2) recovery) (2) (2)D (2)
Tin
Lead
Copper
Total Organic
Carbon
Volatile Organics
Extractable Organics
_30
<30
^30
<_30
<50
^
50
50
100
100
100
— c
75-125
75-125
75-125
75-125
50-160
50
100
100
50
100
95
95
95
95
95
100
100
10U
100
100
aRPD •• Relative Percent Difference
^Percentage of all measurements whose results are judged valid
cMatrix interference rendered analyses inconclusive
TABLE 80. FACILITY F QUALITY ASSURANCE SUMMARY
Parameter
Precision
Accuracy
Completeness
QA
QA Com- objective Com- QA Corn-
objective pliance (2 pliance objective pliance
(2 RPD)a (2) recovery) (2) (2)b (I)
Copper
Tin
Lead
Total Organic
Carbon
<20
<20
<20
<20
100
0
100
100
80-120
80-120
80-120
80-120
100
0
90
100
95
95
95
95
100
0
91
100
aRPD •= Relative Percent Difference
^Percentage of all measurements whose results are judged valid
157
-------
control sample, duplicate sample aliquocs a.\d matrix spikes o: Juplicace
aliquocs. Laboratory control samples were obtained from U.S. EPA
Environmental Monitoring and Support Laboraotry-Cincinnati and prepared as
directed to check instrument calibration. Results which are presented in
Table 75 indicate that precision (<_30 relative percent difference), accuracy
(70-130 percent recovery), and completeness (95 percent valid) goals were met
for the liquid wastes matrix determinations with the exception of total
organic halides which had only 25 percent accuracy. Precision (^50 relative
percent difference), accuracy (50-150 percent recovery) and completeness
(95 percent valid) goals were met only for trace metals and total organic
halides for the solid wastes matrix determination. Precision and accuracy
goals for total organic halides in the solid wastes matrix was not performed.
Quality control procedures for the determination of total cyanides and
hexavalent chromium included the preparation and analysis of a laboratory
method blank by which final results were corrected, a laboratory control
sample, duplicate sample aliquots and matrix spikes. However, due to the
presence of interfering substances such as excess chlorine froa Che cyanide
oxidation process and distillable organics precision and accuracy goals for
total cyanide and hexavalent chrome were inconclusive or met in only a few
cases. Completeness was judged to be 0 for these analyses.
Facility B
Quality control procedures for trace metals, total organic carbon, and
total organic halide determination included the preparation and analysis of a
laboratory method blank, by which final results were corrected, a laboratory
control sample, a duplicate sample aliquot and a matrix spike of duplicate
sample aliquots. Laboratory control samples were obtained from U.S. EPA
Environmental monitoring and Support Laboratory, Cincinnati, and were prepared
as directed. Results are presented in Table 76. Precision goals for total
organic carbon and total organic halide determinations (both liquid and solid)
were met. However, the success in meeting the trace metals precision goals
cannot be determined since duplicate analyses as opposed to triplicate
analyses were performed. All accuracy and completeness goals were met, except
for the total organic halides accuracy analyses on the solids which was not
performed in a deviation from the QA plan.
Facility C
Quality control procedures for solids, extractable and volatile organics
determination included the analysis of duplicate aliquots of sample and matrix
spiked samples by GC/FID. Samples were directly injected so no method blank
was prepared. A field bias blank, collected with the samples was analyzed and
found to contain less than 0.1 percent Freon TF and 1,1,1-trichloroethane. •
Laboratory control samples were not available. Precision, accuracy, and
completeness goals were met as indicated in Table 77 except accuracy goals for
total solids which were not set in the Facility C QA project plan.
158
-------
Facility 3
Quality control procedures for volatile organics determinations by GC/F1D
included the preparation and analysis of a laboratory method blank, by which
final results were corrected, a field bias blank, duplicate sample injections
and matrix spikes of duplicate sample aliquots. A field bias blank was
collected along with the samples to measure possible contamination from
handling and storage. Quality control procedures for volatile and extractable
organic compounds determinations by GC/MS at CRT Analytical Laboratory
included the analysis of a method blank, and surrogate and matrix spikes from
duplicate sample aliquots. Percent recovery of matrix spiked compounds was
calculated as a measure of analytical accuracy. Results of these analyses,
presented in Table 78, indicate that precision and accuracy goals were met and
completeness, defined as the percentage of all measurement whose results are
judged valid, was determined to be 100 percent.
Quality control procedures for solids determination included the
preparation and analysis of a laboratory method blank, by which final results
were corrected, and duplicate sample aliquots. Laboratory control samples and
matrix spiked samples were not available for analysis because of the nature of
the sample matrix, therefore accuracy cannot be determined. Precision goals,
set at <50 relative percent difference, were met. Precision of analysis
conducted on water matrices was not determined.
Facility E
Standard Quality control procedures were implemented whenever possible
for program analysis including analysis of a laboratory method blank, an LCS,
duplicate sample aliquots, and a matrix spike of duplicate sample aliquots.
Completeness for all analyses was 100 percent. Trace metals precision results
(<30 RPD) as indicated in Table 79 were not met for tin and lead, while
accuracy goals (75-125 percent recovery) were not met for tin. Accuracy goals
for total organic carbon were not met while precision and completeness were.
Quality control procedures for volatile organics indicate that both precision
(£50 RPD) and accuracy (50 to 160 percent recovery) goals were met. However,
matrix interferences during sample extraction and subsequent sample dilutions
reduced spike concentrations on extractable organics to below detection limits
(5 g/L).
Facility F
Quality control procedures for trace metals determination included the
analysis of a method blank, by which final results were corrected, a
laboratory control sample (LCS), duplicate sample aliquots and matrix spikes
of duplicate sample aliquots. LCS's were provided by U.S. EPA Environmental
Monitoring and Support Laboratory, Cincinnati. Due to the complexity of the
sample matrix, precision and accuracy goals for trace metals were difficult to
meet as indicated in Table 80. Completeness was determined to be 0 percent
for tin, 91 percent for lead and 100 percent for copper. All tin results were
considered invalid. Lead results on the plating bath were also considered
159
-------
invalid. Precision goals for total organic carbon, set at ^JU rela:ive
percent difference, were met. Matrix spiked samples were prepared oy spiking
a duplicate sample aliquot with known concentrations of the compounds of
interest. Results, provided in Table 80, indicate that accuracy goals (80 to
120 percent recovery) were met. Completeness for total organic carbon was
100 percent.
SAMPLING PROCEDURES
The sampling procedures outlined in Section 4 of the QA Plans were
followed with minor deviations. These sampling procedure deviations have been
presented in detail in the individual draft final reports. However, to
preserve clarity in the summary report only major sampling procecure changes
for each facility will be addressed.
Facility A
Volatile organic analysis were eliminated from the program as they
were noncritical parameters in the metals reduction evaluation and
added substantially to the program cost.
Since it was difficult to sample one batch completely due to length
of time necessary to fill the SBH filter press, CCA collected as
much data from a single batch (85-12-1009) as possible. In
telephone conversations with the EPA Project Officer, it was agreed
that data from separate batches would be acceptable.
Facility B
A flow meter malfunction necessitated the use of flowrate
estimates. The estimates were obtained by contacting the Orange
County Sewer Authority for recent data on Facility B wastewater
flowrates, thus verifying this data by calculating the throughput of
the SBH/ultrafiltration wastewater feed pump.
Facility C
Samples for TOX (total organic halogen) were not collected. The
high corrosivity of some of the samples may have adversely affected
the analytical instruments.
Facility D
The total metals analyses proposed were not conducted in order to
reduce program analytical costs in accordance with the revised
proposal to EPA Project Monitor Harry Freeman dated 28 February 1986.
Facility E
The electrolytic recovery system was not tested for two reasons:
1) Facility E had inadvertently dumped the static rinse batch which
GCA had planned to sample; and 2) delays at Facility E made testing
of the new system difficult under the time frame tor this program.
160
-------
Facility F
• Sampling of process and waste screams associated with the box
distillation process were eliminated. This was done at the request
of the EPA project officer to cut costs.
SAMPLE CUSTODY
Sample custody procedures described in Section 5 of the QA Plans were
followed during the sampling program.
CALIBRATION PROCEDURES AND FREQUENCY
Calibration procedures described in Section 6 of the QA Plans were
followed during the sampling program.
ANALYTICAL PROCEDURES
Analytical procedures described in Section 7 of the QA Plans were
followed during the sampling program.
DATA REDUCTION, VALIDATION, AND REPORTING
Data reduction, validation, and reporting procedures described in
Section 8 of the QA Plans were followed during this program.
INTERNAL QUALITY CONTROL CHECKS
Internal QC procedures described in Section 9 of the QA Plans were
followed during this program.
PREVENTIVE MAINTENANCE
Preventive maintenance procedures described in Section 11 of the QA Plans
were followed during this program.
ASSESSMENT OF PRECISION, ACCURACY AND COMPLETENESS
Analytical precision was reported in terms of relative percent difference
using the following equation:
RPD » Xl " X2 x 100
where: RPD * relative percent difference
X. * larger individual measurement
smaller individual me
average of X. and X.
X. * smaller individual measurement
161
-------
Accuracy assessments were based on inn results of analyses o£ cPA
Standard Reference Materials and of matrix spiked samples and reported in
terms of percent recovery which was calculated as shown below:
Percent Recovery * 100/Mea3ured Value \
\ True Value /
The following formula was used to estimate completeness:
C * Percent completeness
V » Number of measurement judged valid
T * Total number of measurements
CORRECTIVE ACTION
There were no Corrective Action Request forms initiated in regard to this
program.
QUALITY ASSURANCE REPORTS
All pertinent quality control data and activities have been summarized in
this Final Report.
162
-------
SECTION 10
CONCLUSIONS AND RECOMMENDATIONS
ELECTRONIC INDUSTRY WASTE MANAGEMENT
In the manufacture of printed circuit boards and semiconductors, major
waste streams of concern are spent organic solvents (RCRA codes F001-F005) and
metals-containing wastes and wastewater sludges (RCRA code F006-F009).
Organic Solvent Wastes
Organic solvents are used for wafer/board cleaning and for the developing
and stripping of photoresist materials used in the image transfer and/or
circuit fabrication processes. The electronics component industry ranks high
relative to other industries in the generation of solvent waste.
Semiconductor manufacturers are ranked 12th and electronics component
manufacturers not elsewhere classified (which includes the manufacture of
printed circuit boards) are ranked 19th. As companies continue switching to
photoesist materials with an aqueous or semiaqueous base as opposed to an
organic solvent base, quantities of organic hazardous waste generated by this
industry should decrease. However, many companies will continue to employ the
solvent-based process due to the high capital costs associated with
conversion. For these companies, onsite waste reduction will become an
important means to reduce waste treatment costs and future liabilities. Thus,
onsite solvent still bottoms recovery will see increasing prevalence as land
disposal costs and offsite processing costs continue to rise.
Since most spent organic solvents are still quite valuable, recovery has
been a common method of management. Solvents used in the electronics industry
require a high purity which is difficult to achieve by standard solvent
distillation practices. Consequently, it is easier to send these wastes
offsite where the majority of the contaminants can be removed, and the
recovered solvent can be used in an application requiring lower solvent purity.
Recovery of solvents by distillation results in the generation of a
bottoms-product containing contaminants and up to 95 percent of the organic1
solvent. Secondary recovery of the solvents is often possible through the use
of supplementary technologies such as steam distillation or thin film
evaporation. These methods significantly reduce waste product stream volume
and represent feasible and readily implemented methods of hazardous waste
management.
163
-------
Metals-Containing Wastes
Metals are essential to all electronic components due to their conductive
and resistive properties. The most common forms of application are
electroless and electrolytic plating, in which an adherent metallic coating is
deposited on an electrode (the part being plated) to produce a surface with
properties or dimensions different from those of the basic metal. These
metals are introduced into the waste stream through either the disposal of
concentrated plating baths or running rinses directly following the
electroplating process. A second major source of metallic contaminants is the
chemical etch step utilized as part of the electroplating preclean operations
or in the removal of excess surface metal. Etching rinses will contain
relatively high concentrations of metals along with dilute levels of etching
solution. Conventional waste treatment for metals containing waste includes
chemical precipitation, clarification, and dewatering, which results in the
landfill ing of hazardous sludges (RCRA code F006 through F009).
As effluent discharge limits for the electronics industry have become
increasingly strict, the industry has been forced to treat their wastewaters
to remove dissolved metals. However, conventional treatment methods such as
lime precipitation results in the generation of large quantities of metal
containing sludges. Since disposal of these sludges in landfills may soon be
banned under the amendments to RCRA, other nonsludge generating methods of
management will see increasing utilization.
Offsite use, reuse, recovery or recycle (URRR) consist primarily of
sending spend plating and etching solutions back to the manufacturer of these
solutions to be regenerated. Onsite recovery processes, however, such as the
electrolytic recovery of metals form rinsewaters, has yet to achieve
widespread use. Methods for onsite reduction of the quantity of hazardous
metals-containing sludge include sodium borohydride reduction, ion exchange,
electrolytic recovery, evaporation, reverse osmosis, and electrodialysis.
These techniques for recovering metals from wastewaters have become more
common since 1981 and new methods are constantly being developed.
CASE STUDY FINDINGS
The findings of the six waste minimization case studies tested under this
program are presented in Table 81, which includes data collected by the
facilities and verified by sampling and laboratory results. These results
indicate that a variety of technologies exist to minimize metals-containing
and solvent wastes produced by the printed circuit board and semiconductor
industries. The technologies discussed range from simple changes in treatment
system reagents with nominal capital costs to large onsite solvent reclamation
facilities with significantly higher capital costs.
Four of the case studies investigated under this program focussed on
technologies to reduce metal-plating rinsewater sludges. Two of the case
studies, evaluating the use of sodium borohydride reduction as a substitute
for lime/ferrous sulfate precipitation, found that the technology was a viable
substitute in one case and appeared to be marginally acceptable in another.
The case study on carbon adsorption removal of harmful organic contminants
from plating bath wastes found that this technology significantly reduced both
disposal costs and waste volume. The case study of electrolytic recovery
indicated that this technology is highly waste stream specif i.e. An acid
164
-------
copper electroplating rinse is an ideal waste stream for electrolytic
recovery. However, other metal-bearing rinses, such as those from solaer
(tin/lead) plating, or etching are not appropriate for use of electrolytic
recovery. Electrolytic recovery units are, however, generally inexpensive to
purchase and can be used in many cases to supplement an end-of-pipe treatment
process.
Two of the case studies presented in this program involved the recovery
of spent halogenated solvents using batch distillation units. Both of these
case studies indicate that onsite solvent recovery is successful from a •
technical and an economic standpoint. In both cases, over 95 percent of the
waste solvent was recovered and reused onsite. Solvent recovery appears to be
a technology that can be applied to a number of printed circuit board
manufacturing facilities.
The results of this project indicate that waste reduction can be achieved
through the use of an appropriate technology, and it can be achieved with
significant reductions in cost. The case studies also indicate that Che
success of waste reduction is in many cases waste stream specific. The
technologies will not necessarily be successful in all cases. A slight
variation between one waste stream and another may make waste reduction either
technically or economically impractical. Therefore, successful waste
reduction is dependent on a thorough knowledge of waste quantities and
characteristics.
RECOMMENDATIONS
As the case studies presented in this document indicate, cost-effective
application of waste reduction technologies is dependent on site specific
factors such as waste volume, waste characteristics, and availability of
existing onsite facilities and technical expertise. The latter is
particularly lacking in small businesses which often do not possess specially
trained personnel that are able to devote the time required to investigate
waste treatment options. Due to this factor and economies of scale, these
businesses currently land dispose a disproportionately high percentage of
their wastes whereas large quantity generators are more apt to employ waste
minimization and recycling practices. Thus, the land disposal restrictions
and consequently dissemination of waste reduction information, will have a
more significant impact on smaller waste generating firms. This is
particularly true now that the small quantity generator exclusion limit has
been lowered.
Industries will also be impacted to varying extents based on the type of
wastes they generate and the effective dates for promulgation of the land
disposal restrictions for these wastes. Solvent wastes, with total organic
content of one percent or more, are the first waste types to be banned from
land disposal, effective November 8, 1986. Industries which currently land
dispose large quantities of these wastes include a wide range of small volume
generators including metal finishers, electronic component and equipment
manufacturers, and dry cleaners. In addition, these industries consume
relatively large quantities of halogenated solvents. Since these wastes tend
to be generated in smaller volumes, are more restricted in terms of available
disposal options, and are more expensive to purchase relative to their
non-halogenated counterparts, they are particularly well suited to the
application of waste minimization and recycling technologies. Thus, future
165
-------
EPA information dissemination should focus on substitutes and recovery and
treatment alternatives for halogenated solvents. In particular, performance
data are lacking for high solids processing units, like the Recyclene
distillation unit used at Facility 0, and disposal options for the resulting
residuals.
Residual disposal costs will represent an increasingly important factor
in overall system cost-effectiveness as the scale of the operation increases.
This is particularly true for chlorinated or metals containing residuals with
moderate organic contents. These are expensive to incinerate and are not
amenable to conventional stabilization/encapsulation techniques. Additional
guidance on optimal treatment process selection and research on alternative
residual disposal methods is required to assist generators of these wastes.
Although large generators are likely to be impacted less severely than small
generators as a result of the land disposal ban, it must be recognized that
they are responsible for the majority of waste generation and disposal. Thus,
research which is oriented towards the management of large quantity generator
wastes will result in the greatest overall reduction in waste disposal costs
and its associated environmental hazard.
Candidate technologies which appear promising but for which performance
data are currently limitted include chemical fixation, encapsulation, use as a
fuel substitute in aggregate kilns and blase furnaces, and dechlorination
techniques for halogenated solvents. Non-halogenated organics are more
ammenable to conventional thermal destruction techniques. Similarly, other
waste types (e.g., corrosives, metal bearing sludges) are also ammemable to
conventional disposal techniques (e.g., neutralization,
stabilization/encapsulation) and thus will not be subject to as high an
increase in disposal costs as can be expected for halogenated organics.
However, the large volume of these wastes justifies further research and
information dissemination to assist industry in complying with the land
disposal ban in the most cost-effective manner. In particular, additional
performance data are required for membrane and other metal recovery
technologies that can withstand corrosive environments such as that found in
many pickling, etching and plating baths.
In summary, EPA activities to date in this and other programs have
focussed on the identification of and dissemination of information on waste
reduction and treatment technologies. This effort has served to inform
industry of current cost-effective practices and to identify wastes for which
currently available data are lacking. Future efforts should target specific
wastes which create the most significant disposal problems in terms of overall
cost to industry and severity of impact on specific industries. In addition,
research should focus on those technologies which are most likely to result in
cost-effective compliance with the land disposal ban regulations.
166
-------
TABLE 81. SUMMARY OF FINDINGS OF WASTE REDUCTION CASE STUDIES
Fac i 1 ity name
fac i I i t y A
Facility B
i— *
a>
"• Facility C
Fac i 1 ity D
Faci lity E
Facility F
Technology
Sod tun boroliydride reduction
Sodium borohydride reduction
Solvent batch distillation
2-Stanc solvent distillation
Carbon adsorption
plating both reclamation
Agmet electrolytic
recovery unit
Waste reduction
Metals sludge
Metals sludge
Hethylene chloride
Metlil chloroform
Freon
1,1, 1 ,-Tr ichloroe thane
Resist developer
still bottoms
Plating bath wastes
(metals sludge)
Metals sludge
Annual Waste
reduction
achieved
a
962 tons
6,152,000 gal
10,625 Ral
10,600 gal
32 tons
Capital costs
(i)
Nominal
Nominal
709 .'.00
26.150
9,200
30,350
Projected
annual cost
savings
(i)
__b
115,870
16,000.000
43.105
57,267
(10.685)b
"Not quantifiable, but a significant waste reduction was realized.
b«ot ilemnnstrated during testing.
c( ) indicates negative value.
-------
REFERENCES
Agmet Equipment Corporation. Product Literature on Met-tronic 5200.
Cranston, Rhod£ Island
Aldrich, Roberts. Handel, Centec Corporation. Hazardous Sludge Reduction.
The 70th AES Annual Technical Conference (June 1983)
Proceedings, Indianapolis, IN
APV Crepaco, Inc. Price Quotation 0D-9585-RG. March 24, 1986.
Baker Brothers Corporation. Product Literature on Model 3020 SY Carbon
Filtration Unit.
Baron-Blakeslee, Inc. Telecon with M. Arienti, GCA Technology Division, Inc.
March 28, 1986.
Blodgett, W.A. Assessment of Solvent Distillation Equipment. NEESA
20..3-013 Naval Energy and Environmental Support Service.
December 1985.
Card, D. Flat Year Forecast for Independent PCB Markers. Electronic
Business, 9/1/85, pp. 112-132.
Chemical Marketing Reporter. May 5, 1986.
Chillingworth, et al. Volume IV of Industrial Waste Management
Alternatives and Their Associated Technologies/Processes.
GCA-TR-80-80-G. Section 5. pp. 40-48.
Canoby, J.F. Circuit Board Facility Emphasizes Water Recovery and Reuse.
Plating and Surface Finishing. April 1984. pp. 38-40.
Conway, R.A., and R.D. Ross. Handbook of Industrial Waste Disposal.
1980 Edition. Van Nos Reinhold. pp. 213-215.
Department of Energy, Energy Information Administration. National Average.
December 1986.
Dietz, J.D. and Cherniak, C.M., University of Central Florida, Orlando.
Evaluation of New and Emerging Technologies in the Metal Finishing
Industry. U.S. EPA Water Engineering Research Laboratory, Cincinnatti,
Ohio. December 1984.
Electronic Business. A Semi Tough Scene for the Chip Kings. March 1, 1985*
pp. 138-140.
Electronic Business. -Top Semi Companies: A Changing of the Guard.
January 1, 1986 pp. 78-80.
Engineering Science. Supplemental Report on the Technical Assessment of
Treatment Alternatives for Waste Solvents. Contract No. 68-03-3149.
Final Report for EPA Office of Solid Waste. 1984.
168
-------
REFERENCE (continued)
Environmental Regulations and Technology. The Electroplating Industry.
EPA/625/10-85/001.
EPA-600/8-80-023. Carbon Adsorption Isotherms for Toxic Organics. April 1980.
EPA-600/2-83-033. Industrial Process Profiles for Environmental Use:
Chapter 36. The Electronic Component Manufacturing Industry. U.S. EPA
April 1983.
EPA-625/5-85/016. Environmental Pollution Control Alternatives:
Reducing Water Pollution Control Costs in the Electroplating Industry.
September 1985.
Flick, E.W. Industrial Solvents Handbook. Third Edition, Noyes Data
Corporation. Park Ridge, N.J. 1985.
Heleba, S.F. EPA Effluent Compliance and Sludge Control. PC FAB May, 1984
pp. 58-61.
Lindsay, Hackman. Morton Thiokol Inc. Sodium Borohydride Reduces Hazardous
Waste. Purdue Research Foundation, West Lafayette, IN 47701. 1985.
Lopez, N. Chelated Copper Extricated by Membrane Filtration System, Effluent
Cut to Less than 0.2 ppm. Chemical Processing. October 1984. pp. 94-95.
Lopez, N. Fair Treatment. Circuits Manufacturing. September 1984.
pp. 116-118.
Neraec, M.M. Zerpa Offers On-Site Solution for Solvents. Hazardous Materials
and Waste Management, Nov./Dec. 1984.
Pace Inc., Solvent Recovery in the United States 1980-1990. Houston, TX.
Prepared for Harding Lawson Associates, January 1983.
PEDCo Environmental Inc. Industrial Process Profiles for Environmental Use:
Chapter 36. The Electronic Component Manufacturing Industry.
EPA-600-2-83-033. U.S. EPA. April 1983.
Peters, M.S. and K.D. Timmerhaus. Plant Design and Economics for
Chemical Engineers. McGraw-Hill Third Edition. 1980.
Plating and Surface Finishing. Cyclic Voltammetric Determination of
Brightner Concentration in Acid Copper Sulfate Plating Baths
December 1985.
Recyclene Products Inc. Technical Product Bulletin, 405 Eccles Avenue
So. San Francisco, CA 94080. 1985.
Removal and Recovery of Heavy and Precious Metal with Sodium Borohydride.
Metal R&R Newsletter. Issue No. 4. Morton Thiokol, Inc. 1985.
Shipley Company Inc. Product Literature on Cuposit CP-78. Newton, MA.
169
-------
REFERENCE (continued)
Snoeyink. C.L. and D. Jenkins. Water Chemistry. John Wiley and Sons, Inc.
1980.
Ulman, J.A. Control of Heavy Metal Discharge in the Printed Circuit Industry
with Sodium Borohydride. 1984 AES SUR/FIN Annual Technical Conference
and Exhibit.
Versar, Inc. National Profiles Report for Recycling—A Preliminary Assessment.
Contract No. 68-01-7053. Draft Report for EPA Waste Treatment Branch.
1985.
WAPORA Inc. Assessment of Industrial Hazardous Waste Practices - Electronic
Components Manufacturing Industry, EPA SW-140C, January 1977.
Weast, R.C., Editor. CRC Handbook of Chemistry and Physics. 58th Edition.
CRC Press, Inc., 2255 Palm Beach Lakes Blvd., West Palm Beach, Florida.
Wing, R.E. Complexed and Chelated Copper-containing Rinsewaters. Plating
and Surface Finishing. July 1986. pp. 20-22.
Wopschall, R.N. High-Density Yield PWB Imaging with Dry Film Photoresist.
Circuits Manufacturing, July 1984, pp. 36-40.
Yeshe, P. Low-Volume, Wet-Scrap Processing. Chemical Engineering Progress
September 1984, pp. 33-36.
170
------- |