RECYCLING NICKEL ELECTROPLATING RINSE WATERS BY LOW TEMPERATURE
EVAPORATION AND REVERSE OSMOSIS
BY
Timothy C. Lindsey
Hazardous Waste Research and Information Center
Champaign, Illinois 61820
Contract Number
CR-815829
Project Officer
Paul M. Randall
USEPA Pollution Prevention Research Branch
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
NOTICE
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency (EPA) under Cooperative Agreement No. 815829. This
document has been subjected to the Agency's peer and administrative reviews, and it ^has been
approved for publication as an EPA document. This approval does not necessarily signify that
the contents reflect the views and policies of the EPA. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products frequently
carry with them the increased generation of materials that, if improperly dealt with, can threaten
both public health and the environment. The U.S. Environmental Protection Agency is charged
by Congress with protecting the nation's land, air, and water resources. Under a mandate of
national environmental laws, the agency strives to formulate and implement actions leading to
a compatible balance between human activities and the ability of natural systems to support and
nurture life. These laws direct the EPA to perform research to define our environmental
problems, measure the impacts,a nd search for solutions.
The Risk Reduction Laboratory is responsible for planning, implementing, and managing
research, development, and demonstration programs to provide an authoritative, defensible
engineering basis in support of the policies, programs, and regulations, of the EPA with respect
to drinking water, wastewater, -pesticides, toxic substances, solid and hazardous wastes, and
Superfund-related activities. This publication is one of the products of that research and
provides a vital communication link between the researcher and the use community.
This document presents the results of experiments conducted to evaluate two technologies
currently available to irunirnize environmental problems associated with rinse waters generated
during electroplating operations. The objective was to concentrate nickel to the same level as
the electroplating bath so that the nickel could be reused. The high cost of nickel and of treating
and disposing of electroplating wastes is an incentive to develop these recovery technologies to
reduce wastes at the source.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory'
-------
ABSTRACT
Electroplating operations generate rinse water wastes that are classified as hazardous due
to the presence of heavy metals. Disposal of the rinse water waste without prior treatment is
cost prohibitive and environmentally unacceptable. End-of-pipe treatments that utilize chemical
precipitation and destruction of the plating chemicals waste valuable materials that could be
recycled. This project examined two alternative technologies for reducing the waste generated
and recovering heavy metals such that they can be recycled in the electroplating process.
Specifically, low temperature evaporation and reverse osmosis systems were evaluated on a pilot
scale to process rinse water collected from a nickel electroplating operation. i
The low temperature evaporation system exhibited consistent and predictable results
throughout the tests. It was capable of concentrating the rinse water feed solution to contain
nickel levels well above the eight percent required for replacement into the plating bath. The
cleaned rinse water was high quality and essentially metal free and could be either reused for
rinse water or discharged to a POTW. Disadvantages associated with the evaporation system
include its relatively high capital costs and high energy requirements.
The reverse osmosis system exhibited superior productivity at the beginning of the tests
while productivity dropped off dramatically after about 60 percent of the feed solution had been
processed. The decline commenced when nickel levels in the feed solution reached about 4,000
to 5,000 mg/L. The decline continued until the productivity of the reverse osmosis equipment
was reduced to almost nothing. At this point the reverse osmosis system had concentrated the
feed solution to nickel concentrations of 12,560 to 18,200 mg/L which are well below the eight
percent nickel concentration required for the plating bath. The cleaned rinse water produced by
the reverse osmosis equipment was directly related to the quality of the feed solution pumped
into the unit. The quality of the cleaned rinse water from the reverse osmosis test was
acceptable for reuse as rinse water but unacceptable for discharge to the POTW. The reverse
osmosis system offers relatively low capital cost and energy requirements when compared to an
evaporation system. However, its application is limited by the extent to which it concentrates
solutions and the quality of the cleaned solution it produces.
Both systems offer advantages under specific operating conditions. The low temperature
evaporation system appears to be best suited to processing solutions with relatively high nickel
concentrations. The reverse osmosis system is best adapted to conditions where the feed solution
is of relatively low (less than 4,000 to 5,000 mg/L) nickel concentration. In electroplating
operations where relatively dilute rinse water solutions must be concentrated to levels acceptable
for replacement in the plating bath, a combination of the two technologies might provide the best
IV
-------
process alternative. The reverse osmosis system could be used to initially concentrate the feed
solution followed by low temperature evaporation processing to concentrate the solution to levels
acceptable for replacement in the plating bath. ,t
~y' , • '•/?*•
This report was submitted in partial fulfillment of Contract No. CR-815829 under the
sponsorship of the U.S. Environmental Protection Agency. This report covers a period from
November, 1991 to August, 1992 and work was completed as of December, 1992.
-------
TABLE OF CONTENTS
Foreword • • •, m
Abstract , iv
List of Figures v&
List of Tables
Acknowledgements
IX
.x
1. Introduction . • • • 1
Industry background ! 1
Graham Plating .......'..... 2
Low temperature evaporation - 4
Reverse osmosis ;..... 5
2. Materials and Methods |- .... 9
Low temperature evaporation testing 9
Reverse osmosis testing 9
Chemical analysis , 10
3. Experimental Procedure ... 12
Field sample collection . . „ 12
Low temperature evaporation procedure 12
Reverse osmosis procedure 13
4. Results and Discussion 14
Equipment productivity !..... 14
Nickel concentrations -17
Total organic carbon concentrations '...,, 21
Electrical conductivity 25
Parameter relationships 28
5. Economic Analysis . ;.-.•..* 35
6. Conclusions and Recommendations . 43
System efficiency !.... 43
Recommendations 45
References 47
Appendix 49
Quality Assurance 49
VI
-------
FIGURES
Number
1 Processing Line 2 at new Graham Plating facility 3
2 Basic flow diagram for single effect evaporator 6
3 Components of tubular, spiral-wound, and hollow fiber membranes . 8
4 Percent of drum volume processed versus time; low temperature
evaporation tests 15
5 Percent of drum volume processed versus time; reverse osmosis tests , . . 15
6 Permeate flux rates versus time; reverse osmosis tests 16
7 Permeate flux rate versus percent of drum volume processed; reverse
osmosis tests , 17
8 Concentrate nickel concentration versus percent of drum volume processed;
low temperature evaporation tests 18
9 Concentrate nickel concentration versus percent of drum volume processed;
reverse osmosis tests . .19
10 Permeate nickel concentration versus concentrate nickel concentration;
reverse osmosis tests 21
11 Concentrate TOC concentration versus percent of drum processed; low
temperature evaporation tests 22
12 Concentrate TOC concentration versus percent of drum volume processed;
reverse osmosis test 23
13 Permeate TOC concentration versus percent of drum volume processed; reverse
osmosis tests 24
VI!
-------
Number ;
14 Concentrate electrical conductivity versus percent of drum volume processed;
low temperature evaporation tests . . . ......:... 25
15 Concentrate electrical conductivity versus percent of drum volume processed;
reverse osmosis tests : ... 27
16 Permeate electrical conductivity versus percent of drum volume processed; ,;
reverse osmosis tests : ... 27
17 Permeate flux rate versus concentrate nickel concentration; reverse \
osmosis tests > ... 29
18 Permeate flux rates versus permeate nickel concentration; reverse ;
osmosis tests 29
19 Permeate flux rates versus concentrate electrical conductivity;
reverse osmosis tests 30
20 Permeate flux rates versus permeate electrical conductivity; reverse
osmosis tests • • • • 30
21 Concentrate nickel concentration versus electrical conductivity; low :
temperature evaporation tests 31
22 Concentrate nickel concentration versus electrical conductivity;
reverse osmosis tests j 32
23 Permeate nickel concentration versus permeate electrical conductivity;
reverse osmosis tests , | 32
i
24 Concentrate TOC concentration versus electrical conductivity; low
temperature evaporation tests ...» 34
25 Concentrate TOC concentration versus electrical conductivity; reverse
osmosis tests 34
VIII
-------
TABLES
Number
1 Comparison of nickel concentrations in concentrate, distillate and permeate nickel
concentrations 18
2 Average nickel concentrations in distillate and permeate - 20
3 Comparison of concentrate, distillate, and permeate TOC
concentrations 22
4 Average TOC concentrations in distillate and permeate 24
5 Comparison of concentrate, distillate, and permeate
electrical conductivities 26
6 Average electrical conductivity values in distillate and
permeate 28
7 Assumptions for economic calculations 35
8 Low temperature evaporation cash flow summary 37
9 Economic summary - low temperature evaporation option : 38
10 Reverse osmosis cash flow summary 39
11 Economic summary - reverse osmosis option 40
12 Combined system cash flow summary 41
13 Economic summary - combined technology option 42
IX
-------
ACKNOWLEDC
i
This project was performed jointly by the US EPA and the Hazardous Waste Research
and Information Center, Champaign, Illinois. The cooperative efforts of Graham Plating,
Chicago, Illinois; licon, Inc. Pensacola, Florida; and Osmonics, Minnetonka, Minnesota;
facilitated successful completion of this project. A list of individuals who made significant
contributions to the management and execution of this project is provided below. '
Mr. Paul Randall USEPA ;
Mr. Tim Lindsey HWRIC
Mr. Clayton Graham Graham Plating
Mr. Ken MajewsM Graham Plating J
Mr. John Campbell licon, Inc.
Mr. Curt Weitnauer Osmonics
-------
SECTION 1
INTRODUCTION
This is the final report for one of five projects constituting the Illinois/EPA WRITE
(Waste Reduction Innovative Technology Evaluation) Program. The project was a joint effort
of Graham Plating, Chicago, Illinois, an electroplating firm; the Hazardous Waste Research and
Information Center (HWRIC) which is a division of the Illinois Department of Energy and
Natural Resources, Champaign, Illinois; and the Pollution Prevention Research Branch of the
U.S. Environmental Protection Agency's Risk Reduction Research Lab, Office of Research and
Development, Cincinnati, Ohio. Assistance and direction was also provided by Licon, Inc. of
Pensacola, Florida and Osmonics of Minnetonka, Minnesota.
The purpose of this project was to evaluate, compare, and document the effectiveness of
two technologies for recovery and reuse of water and plating bath chemicals associated with
electroplating rinse waters. The recovery technologies examined in this project include 1) low
temperature evaporation and 2) reverse osmosis. These treatment systems were evaluated based
on their effectiveness with respect to the following considerations:
O Recovery efficiency and purity of the treated rinse water
O Recovery efficiency and purity of the rinse water concentrate
O Anticipated reduction in water use and chemical utilization due to recycling
O Anticipated reduction in waste volume associated with installation of the treatment
systems, and
O Economic analysis of installation and operation of the treatment and reuse systems
INDUSTRY BACKGROUND
During electroplating operations, metal parts are immersed in a bath containing dissolved
plating metals and chemicals. Typical metals and alloys used for plating include cadmium,
copper, iron, lead, nickel, gold, silver, platinum, brass and bronze. When an electrical potential
is applied, an electrochemical cell is created and the dissolved metal deposits on the part surface.
Parts are then conveyed to a series of rinse tanks to remove residual plating solution. Rinsing
prevents spotting, uneven metal deposition and cross-contamination of the plating chemicals with
later operations. This source of rinse water contamination is called dragout. When the dragout
in the rinse tanks reaches high enough concentrations, rinsing is no longer effective and the rinse
water must be replaced (Hunt, 1988). Some electroplating operations deal with this problem by
replacing the entire rinse water volume on an as needed basis while others replace the rinse
-------
water with a continuous inflow of fresh water. In addition to the dissolved metals utilized in the
plating solutions, some baths contain organic compounds that modify the growth of the metal
deposit to produce bright, semi-bright or satin finished surfaces.
Electroplating rinse water wastes are classified as hazardous due to the presence of heavy
metals and, in some cases, cyanide in these solutions. Direct discharge of rinse waters can
pollute natural resources, inhibit or destroy biological activity and sewage treatment processes,
and corrode sewer lines and structures. Fortunately, treatment technologies exist which can
minimize the harmful effects of rinse waters to human health and the environment. These
processes render inactive or remove the hazardous components from the rinse waters.
Electroplating rinse waters can be treated by either end-of-pipe or in-plant recovery
techniques. End-of-pipe treatments rely on chemical reactions such as pH adjustment to
precipitate metal and other plating chemicals. These methods provide an effective means to
remove the metal and cyanide species from the rinse water thus enabling reuse or discharge of
the water. However, in most plant treatment systems, waste streams from the various plating
lines are combined prior to treatment. Consequently, the resulting sludge must either be
disposed of or treated with a high temperature metals recovery system to recover the metals.
These options tend to be wasteful and expensive to implement. In-line treatments that make it
possible to recover and return lost plating chemicals to the electroplating bath do exist. These
recovery techniques are superior to the end-of-pipe treatments which merely reduce the waiste
volume. Additionally, in-line techniques require less processing than end-of-pipe recovery
techniques. Examples of effective in-line recovery techniques include; low temperature
evaporation, ion exchange, electrowinning, electrodialysis and reverse osmosis. Low-
temperature evaporation and reverse osmosis techniques were selected for examination in (Ms
project because officials from the participating company, Graham Plating, believed that they
offered the most promise for their facility.
GRAHAM PLATING
Graham Plating is a large "job-shop" which has been located for many years on the north
west side of the city of Chicago. A new modern building has recently been completed in
Arlington Heights, Illinois and Graham Plating plans to relocate the plating operations to the
new facility. Incorporated into the design of this new structure and in the new plating line
layouts are special features which promote waste reduction. Examples of the waste reduction
features are large underground rinse water collection tanks which provide the means to segregate
the rinse waters by principal metal component and to store these waters for later processing and
reuse.
The plating lines at the new Graham Plating facility will be like those at most plating
operations in that they contain multiple bays for the different phases of the plating process.
Objects to be plated are processed through a number of cleaning, pretreatment, and rinsing steps
before the final metal plate is deposited. Part of the pretreatment may include plating with
several other metals before the final desired finishing metal is applied. Figure 1 depicts plating
-------
CO
CO
CO
8
CO
01
CO
OI
CO
04
v>
o
a>
So
o
o
a
o
E
of
RO
$ $ .M Qj o
CO
v>
o
CO
Figure 1.
Processing Line 2 at new Graham Plating facility
3
-------
line 2 at the new Graham Plating facility which will be used for processing a variety of metals.
Once the final metal has been plated, the object will be rinsed using both spray and
counter flow rinses. Water from the rinses will be collected in tanks for subsequent treatment
through a vacuum evaporation unit. This unit is capable of processing 7,200 gallons of rinse
water per day. If additional processing capacity is required, the low temperature evaporation
unit may be supplemented with a reverse osmosis unit to expand the ability of the facility to
process rinse water. The unique design of the new Graham Plating facility is such that these
rinse waters can be collected in large underground storage containers allowing :complete
segregation of the rinse waters by their major metal component. The rinse water can then be
collected in storage tanks and treated in batches. :
LOW TEMPERATURE EVAPORATION I
Rinse water from the plating operation will be pumped into a low temperature
evaporation unit where it will be treated such that the water can be reused for rinsing operations
in the plating lines and the concentrate solution can be replaced in the plating bath. This type
of operation has been successfully used in some plating operations because it allows the recovery
of not only the water but also the plating bath chemicals. The economics associated with
purchase of the equipment are usually favorable and utilization of the equipment can facilitate
compliance with environmental regulations (Making Pollution Prevention Pay. 1983),
i • •
Evaporation units can be classified as either atmospheric or vacuum. In iboth cases the
water is heated to produce vapor which is later condensed resulting in distilled water. Impurities
in the untreated water remain as a concentrated slurry or solution of chemicals, in mis case, the
latter. In vacuum evaporation, the process takes place at pressures lower than atmospheric
pressure. The pressure reduction lowers the boiling point of the water which permits the use
of lower temperatures to create the steam and finally, the distilled water (Kusfaner and Knshner,
1981). Commercial systems may be constructed of a combination of glass, titanium, fiberglass,,
PVC, and stainless steel. Although the capital and operating costs of fhese systems are
expensive when compared to other waste treatment options, their corrosion resistance makes
them ideally suited to use with electroplating wastes.
In addition to facilitating recovery of the plating rinse waters, the lower boiling
temperatures of the vacuum type unit enable recovery of the plating chemicals which might
decompose during a standard evaporation process (Electroplating Engineering Handbook. 1984).
This is accomplished by concentrating the metal salts and organic compounds present in the rinse
water to a point that is acceptable for replacement in the plating bath. At the new Graham
plating facility the recovery process will begin with the collection and storage of the rinse waters
from both spray and counter flow rinses. Rinse water from nickel plating will collect in the
designated storage tank until it is filled. Processing will begin by pumping water into the
vacuum evaporator condenser. Here the water is heated under low pressure to produce steam
which is condensed. At this point, inorganic contaminants will have been removed from the
rinse water but organic additives are likely to remain since they tend to volatilize with the steam.
To remove the organics, the water can be passed through a carbon filter. The recovered water,
-------
which will then be suitable for reuse in the plating line rinses, will continue through a closed
loop system to a reservoir for storage until it is needed. The inorganic salts and the non-volatile
organic compounds will remain in the condenser as a concentrated solution or slurry. Prior to
replacement of the concentrate into the plating bath:, the concentrate will be removed from the
condenser for characterization and adjustment by the laboratory. A schematic drawing of
materials flow through a vacuum evaporator is provided in Figure 2.
Characterization of the concentrate is perhaps the most important step in the chemical
recycle process. Changes in plating quality will result if the integrity of the plating bath is not
maintained. To do this one must be certain of the composition of everything that is being added
to the bath including the recycled rinse water chemicals. It is almost certain that the concentrate
will be somewhat different from the original plating bath solutions. The anticipated inorganic
components will be metallic nickel, nickel sulfote, nickel chloride, nickel sulfamafe, boric acid,
and hydrochloric acid. But other inorganic components, such as metal from the object being
plated, may also be present (Metal Finishing. 1987).
The amount of organic compounds in the concentrate and plating solutions is important.
Since these organic compounds are critical in obtaining the desired finishes, knowledge of their
concentration in the concentrate solution is critical. Volatile organic compounds can be removed
by a carbon filter but there will be non-volatile organics present in the concentrate as well.
Knowing the concentration of all of the inorganic and organic constituents of the concentrate
is vital in determining if the solution can be recycled in the plating bath and what supplements
(if anjr) need to be added to the concentrate before recycling. Additionally, chemical
characterization of the concentrate would help determine if contaminants were concentrated
during rinse water processing.
REVERSE OSMOSIS
Reverse osmosis is a proven technology for concentrating dilute solutions. In a closed-
loop electroplating operation, reverse osmosis can be used to recycle purified water to the rinse
tanks and return plating chemicals to the bath. Reverse osmosis systems can be used in
combination with other treatment systems, such as low temperature evaporation and ion
exchange, or it can by used alone. The higher metal ion concentrations of the processed rinse
water produced with reverse osmosis systems can improve the efficiency of the other treatment
systems. When reverse osmosis is used alone, it can purify rinse water for rinsing and reclaim
lost metals and plating chemicals for reuse (Rozelle, 1973).
Reverse osmosis is a pressure-driven membrane separation process in which a feed
stream under pressure (200-800 psig) is separated into a purified "permeate" stream and a
"concentrate" stream by selective permeation of solution through a semi-permeable membrane.
The pressure required to force the permeate through the membrane is dictated by the osmotic
pressure of the feed stream (Rousseau, 1987). Three important parameters describe the
performance of the reverse osmosis process: recovery, flux, and rejection. Recovery is defined
as the percentage of the feed that is converted to permeate. Flux is the rate at which the
permeate passes through the membrane per unit of membrane surface area. Rejection is the
-------
Cooling
Water-<—
Outlet
Cooling
G§ Water—*{_
In
Steam or
Water
Supply
Hot
Water
Return^
_._.
| © bbtniita Vapor ©
-•- i <" -^
-> 1
.1
<}._._.. i
| Bayonet ^xchanger Evaporator
*
Feed
s
C
f
4
1
t
t •
oncentra
•
/•
9
i 1
Concentrate
Legend
Feed/Concentrate Q
Steam or Hot Water £fc
Dlstmale <*\
*- Cooling Water
Feed
o
Figure 2.
Basic flow diagram for single effect evaporator.
-------
ability of the membrane to restrict the passage of dissolved salts into the permeate, and is related
to particular salt species (Cushnie, 1985).
One of the most significant operating; problems ctwnmon to all membrane types is gradual
reduction in performance associated with plugging by suspended solids. Another potential
problem is caused by the precipitation of dissolved solids in the feed solution as it is
concentrated in the reverse osmosis unit. With proper pretreatment (suspended solids filtration,
oxidation, pH adjustment, and so forth), performance can be trouble-free and the equipment
easily maintained.
Three major types of membrane modules are available for commercial use. These
include; tubular, spiral-wound, and hollow fiber (Figure 3). Tubular membranes are not
susceptible to plugging by suspended solids and can be operated at high pressures. However,
they are expensive to operate and require large amounts of space. Spiral-wound and hollow-
fiber modules cost considerably less to operate. The hollow-fiber modules require less space
to operate while the spiral-wound modules are less susceptible to plugging by suspended solids.
All three types of membranes can be constructed from a variety of materials, such as: aromatic
polyamide, cellulose acetate, and polyether/amide.
Reverse osmosis systems offer advantages over evaporation and ion exchange systems
for treating rinse waters in that they are cheaper and require less energy for operation.
However, the application of reverse osmosis technology is limited due to the modest degree of
metal salt concentration it can achieve compared to other technologies. Additionally, the
permeate purity resulting from treating concentrated feed streams may not be acceptable for
reuse as rinse water.
Currently, no plans exist to install a reverse osmosis system at the Graham Plating
facility. However, if the low temperature evaporation system is not capable of processing the
quantities of rinse water they will be generating at their new facility, Graham Plating will
consider supplementing the evaporation system with reverse osmosis. If the decision is made
to install a reverse osmosis unit at the new Graham plating facility, results obtained from tests
conducted during this project will be utilized to determine the type, size, and operating
parameters of the system.
-------
"HOOKXASS SUCTORTTVSE
BOOTIEM.
Tubular
Spiral-wound
COMCEKITCATE
oomouto* noc
Hollow Rber
Figure 3. Components of tubular, spiral-wound, and hollow fiber membranes.
8
-------
SECTION 2
MATERIALS AND METHODS
LOW ITEMPERATURE EVAPORATION TESTING
The Low Temperature Evaporation unit utilized in this study was manufactured by Licon,
Inc. of Pensacola, Florida. This unit is a model C-3, single effect, pilot scale evaporator
specially designed for conducting pilot scale tests on a variety of feed solutions. The full scale
unit to be installed at the new Graham Plating facility was also manufactured by Licon, Inc.;
however, it is a double effect type evaporator. Double effect evaporators differ from single
effect evaporators in that heat generated in a first evaporation step is conserved and used in a
second step. The primary advantage of utilizing a double effect system is that only half of the
energy is required to process solutions as is required with a single effect evaporator. The quality
of the concentrate and distillate produced by the two types of evaporators is virtually the same.
Consequently, with the exception of energy usage, the pilot scale unit provided an accurate
representation of performance that might be expected from a full scale system.
The Low Temperature Evaporation testing was conducted at HWRIC's pilot laboratory
using the pilot scale unit described above. A brief discussion of the low temperature evaporation
process that takes place in this unit is provided below and shown on Figure 2. Feed solution
is pumped from a reservoir (in this case, 2-55 gallon sample drums were used) to a 6 gallon
concentrate tank on the unit where it is stored for processing. Float limit switches maintain the
concentrate tank at a relatively constant level. Solution is pumped from the concentrate tank into
a bayonet exchanger evaporator cell which is equipped with heat exchangers that provide the
heat necessary to separate water vapor from the feed solution. The bayonet exchanger
evaporator cell maintains a vacuum of approximately 23 to 25 inches which enables the solution
to boil at a temperature of approximately 150 to 160° F. The vapor rises through a separator
cell until it collects' on a water cooled condenser which transforms the water vapor into liquid
water. The liquid water is then collected in a 6 gallon capacity distillate tank and
ultimately discharged to a drum or holding tanlc where it can be recycled for use as rinse water
or any number of other purposes.
REVERSE OSMOSIS TESTING
The reverse osmosis unit used in this project was manufactured by Osmonics of
Minnetonka, Minnesota. The unit is a model PES/OSMO-19T-80SSXXC Reverse Osmosis
machine for process evaluation. It is capable of operating at pressures ranging from 100 to 800
-------
psi and can accommodate one membrane cartridge 2-1/2 inches in diameter x 39 inches long.
The feed inlet to this unit is equipped with a prefilter which was supplied with a 5 micron
cartridge filter for the duration of this project. Prefiltering the solution helps prolong the life
of the membrane by removing dirt and other abrasives that might damage the membrane. This
unit is capable of utilizing a variety of membrane cartridges. For the purposes of this project,
an Osmonics model number 192T-MSO5 thin film composite membrane cartridge was used to
process the rinse water solution. This is a spiral wound type membrane cartridge and is
constructed of a polyamide film cast over a polysulfone backing. The membrane cartridge
contains approximately 19 square feet of membrane and has a molecular weight cutoff of 100 -
150. This membrane was selected because Osmonics representatives indicated that it has been
used extensively for processing nickel rinse waters with good results.
Two 55 gallon drums of Graham Plating rinse water solution were processed individually
through the reverse osmosis system at HWRIC's pilot lab facility. The solution was pumped
through a 5 micron filter prior to reverse osmosis processing. The filter had to be changed once
during the prefiltering stage due to buildup of a waxy film on the filter cartridge which restricted
flow. Based on preliminary electrical conductivity analysis performed on the rinse water (10230
umhos for Drum C and 6110 umhos for Drum D), it was assumed that the solution had an
osmotic pressure of approximately 10 psi. Reverse osmosis systems can be effective on
solutions with osmotic potentials as high as 200 psi. Therefore, it was estimated that the
equipment could concentrate each dram of the solution approximately 20 fold to about 2 3/4
gallons per dram. Based on these calculations, it was estimated that primary pressures of 250
to 380 psi would have to be maintained in the system during solution processing. The pressure
was increased periodically as the volume of feed solution decreased and became more
concentrated. Temperatures were maintained at 74 to 80 degrees fahrenheit through use of a
heat exchanger which is mounted on the unit. The concentrate flow rate was maintained at
approximately 3 gallons per minute for the duration of the testing while the permeate rate started
at approxunalely 15 to 25 gallons per hour and decreased steadily as the solution became more
concentrated until a final rate of less than 3 gallons per hour was attained.
CHEMICAL ANALYSIS
In an effort to determine the effectiveness of both the low temperature evaporation and
reverse osmosis systems with respect to treating the rinse water, samples were collected during
the tests and analyzed for key chemical constituents. Concentrate and distillate samples
comprised the bulk of samples collected during the low temperature evaporation testing while
concentrate and permeate samples were the predominant sample types collected in the reverse
osmosis testing. Analyses conducted on the samples included electrical conductivity, pH, total
organic carbon, and nickel.
As the testing progressed for both technologies, samples were analyzed for pH and
electrical conductivity immediately after collection in the pilot lab. A Beckman model number
32 pH meter was used to measure the sample pH. Using method number 9040 from SW846
(EPA 1986). An Orion model number 140 Conductivity/Salinity meter was used to| determine
10
-------
electrical conductivity using method number 9050 from SW846 (EPA 1986). Both instruments
were operated according to procedures provided in the instruction manuals.
; •
.Analysis for total organic carbon (TOC) was conducted to provide an indication of the
fate of ithe organic constituents present in the rinse water. The TOC analysis was conducted on
a Rosemount Analytical Dohrmann model DC- 190 TOC Analyzer according to method 9060
from SW846 (EPA 1986). Samples were diluted according to strength to ensure the instrument
was working within the instrument working range.
Nickel analyses were performed to determine the efficiency of both the low temperature
evaporator and reverse osmosis systems with respect to removing this metal from the rinse water
and concentrating it for potential recycling. Samples were digested according to method 3010
and analyzed according to method number 7520 from EPA SW846 (EPA 1986). A Varian
Spectra-10 Atomic Absorption Spectrophotometer was utilized to perform the analyses. Most
analyses were run at 352.4 nm, the wavelength most appropriate for the high concentrations hi
many of the samples. Instrument settings were hi accordance with the manufacturer's AAS
operating/procedures manual. Because of the large range hi concentrations in these samples,
alternative wavelengths were also used. Samples were diluted as necessary to ensure the
concentration was within a range that would provide optimum accuracy.
11
-------
SECTION 3
EXPERIMENTAL PROCEDURE
FIELD SAMPLE COLLECTION
Since Graham Plating has not yet relocated to their new facility in Arlington Heights,
Illinois, field samples of the rinse water had to be collected from their Chicago facility I Rinsing
of nickel plated parts is accomplished at the Chicago facility by manually dipping baskets of
plated parts into two separate (primary and secondary) rinse tanks which contain about 150
gallons of rinse water each. Four - 55 gallon drums of electroplating rinse water (henceforward
referred to as Drums A,B,C, and D) were collected from a secondary nickel rinse water tank
at the Graham Plating facility in Chicago and shipped to the HWRIC pilot lab facility by a DOT
licensed hazardous materials transporter. The samples were collected by pumping the solution
intermittently into the drums over a 2 day period. Upon their arrival at the pilot lab, 2
subsamples were collected from each drum and placed in 250 Ml. nalgene bottles. These
samples were analyzed to determine baseline nickel concentrations present hi the drums. Results
of these analyses are summarized below:
, O Drum A 4,820 to 4,940 mg/L
O DramB 2,680 to 2,750 mg/L
O DrumC 2,740 to 2,770 mg/L
O Drum D 1,580 to 1,590 mg/L
Based on these data, calculations were performed to estimate the volume reduction
necessary to produce a concentrate solution of approximately eight percent nickel (the target
nickel level maintained in the Graham Plating nickel bath). In addition to the bulk samples
collected at the Graham Plating facility, a sample of the nickel plating bath solution was
collected and analyzed to verify that the bath contained eight percent nickel.
LOW TEMPERATURE EVAPORATION PROCEDURE
Drums A and B were processed separately through the low temperature evaporation unit.
The unit processed the solution at a rate of approximately 3 gallons per hour. A total of 14
operating hours were required to process drum A (approximately 42 gallons) while 16 hours
were required to process drum B (52 gallons). Numerous operating parameters associated with
various pump and system pressures and temperatures were monitored throughout the' testing to
ensure that optimum conditions were maintained. '
12 i
-------
Processing continued until the rinse water volume was reduced to a point where its
viscosity limited the ability of the unit's feed pump to take in the solution. At this point,
virtually all of the remaining unprocessed solution was contained in the unit's plumbing
(approximately 1 gallon). Samples of concentrate and'distillate were collected at one hour
intervals hi 250 Ml. nalgene bottles and immediately tested for pH and electrical conductivity.
The samples were then diluted with 2% nitric acid and stored in a 4 degrees Celsius cold storage
room until the other chemical analyses could be performed. Nickel analysis was performed on
all samples collected during these tests. Additionally, total organic carbon (TOC) was
determined on samples collected at the beginning of the testing, and approximately every 4 hours
thereafter until concentrations above 3,500 mg/L of TOC were attained. Beyond that point,
TOC levels increased dramatically as the volume of the feed solution was reduced. Therefore,
TOC measurements were performed on samples collected on an hourly basis for the duration of
the testing.
REVERE OSMOSIS PROCEDURE
In order to determine baseline flux characteristics of the membrane used in this test, clean
tap water was processed through the reverse osmosis unit for a 30 minute period prior to and
after testing with the rinse water solution. These "clean water flux" measurements were
determined utilizing gauges mounted directly on the reverse osmosis machine. Following
determination of the clean water flux measurements, Drums C and D were processed separately
through the reverse osmosis unit. Flux measurements were collected at 15 minute intervals
during the tests. Results of these tests indicated that, at the beginning of the testing, the reverse
osmosis system processed solution at rates of 15 to 25 gallons per hour for drums C and D,
respectively. However, as the testing progressed and the feed solution became more
concentrated, the flux rate associated with both drums fell off to about 3 gallons per hour.
A primary operating pressure of approxiinately 250 to 300 psi was maintained for the
Drum C test. The Drum D test was conducted at a primary operating pressure of approximately
350 to 380 psi. The higher pressures implemented in the Drum D test were used to determine
the impacts of increased operating pressure on the equipment's productivity and the quality of
the solutions processed. Overall, the two drums (approximately 52 gallons each) were processed
in 6.25 and 5 hours, respectively. However, the solution could only be concentrated down to
a final volume of about 6 gallons (as opposed to the 1 gallon volume reached with the low
temperature evaporation system) before the flux rate plunged to the 3 gallon per hour level.
Samples of concentrate and permeate were collected at 15 minute intervals during these
tests and placed hi 250 Ml. nalgene bottles. These samples were immediately analyzed for pH
and electrical conductivity. The samples were then diluted with 2% nitric acid and placed in a
4 degrees Celsius storage room until analysis for the other chemical parameters could be
accomplished. Nickel analysis was performed on all samples collected during the reverse
osmosis testing. Additionally, total organic carbon was determined in samples collected at the
beginning of testing and approximately every 75 minutes during the duration of the reverse
osmosis testing.
13
-------
SECTION 4
RESULTS AND DISCUSSION
Results of the pilot scale testing and laboratory analysis were utilized to determine the
effectiveness of the two technologies tested in this project with respect to 1) the success in
concentrating the nickel levels for reuse in the electroplating process, 2) cleaning the
electroplating rinse waters for possible reuse or discharge to the POTW, and 3) the relative
productivity and economic feasibility associated with installing these systems in an electroplating
facility. The nickel concentrations of the various solutions monitored were determined
throughout this project to determine its fate with respect to the processes studied. Total organic
carbon concentrations were also monitored during the course of the project to provide an
indication of the fate of the organic additives that accumulate in the rinse water concentrate.
Electrical conductivity was monitored to determine if relationships exist between this parameter
and the other parameters tested. Electrical Conductivity is a relatively quick and inexpensive
parameter to monitor during actual operating conditions. Therefore, identification of
relationships between electrical conductivity and the other parameters could prove to be
important when low temperature evaporation and reverse osmosis technologies are actually
implemented in the electroplating process. Additionally, pH was monitored in all samples
collected in this project. Distillate and permeate samples tended to exhibit pH.levels from 0.5
to as much as 2 units higher than the concentrate pH measurements. Some fluctuations occurred
with respect to solution pH as the tests progressed although the pH variations did not play an
integral part in detenruning quality of either distillate, permeate, or concentrate. Therefore, the
pH data are not presented in this report. :
EQUIPMENT PRODUCTIVITY
The low temperature evaporator processed the solutions at a relatively constant rate of
3.4 gallons per hour for drum A and 2.8 gallons per hour for drum B. As shown in figure 4,
the productivity rate varied little during the course of these tests. The evaporator processed the
solution down to a final volume of approximately 1 gallon which corresponds to la volume
reduction of nearly 98 percent.
The reverse osmosis unit processed the solutions at a relatively constant |rate until
approximately 60 percent of the drum volume was treated. Productivity beyond this point began
to decrease steadily until about 80 percent of the solution was treated. Beyond the 80 percent
level, the productivity decreased dramatically (see Figure 5). The reverse osmosis equipment
14
-------
processed! the solution from Drum D faster than the Drum C solution. This occurred due to the
higher operating pressures that were utilized in the Drum D test.
•a
0>
n
en
ID
O
o
O
E
E
Q
100 -I
20-
H Drum A
* DrumB
Figure 4,,
cm
•a
c>
LI
s
o.
E:
st
E
Z
a
Elapsed Time (Hours)
Percent of drum volume processed versus time; low temperature
evaporation tests.
100 n
80-
60-
Elapsed Time (Hours)
Figure 5. Percent of drum volume processed versus time; reverse osmosis tests.
15
-------
Figure 6 shows how the permeate flux rate changed over time in the two reverse osmosis
tests. Due to the higher operating pressures used in the Drum D test (350 to 380 psi), the unit
produced permeate at an initial rate of about 20 gallons per hour which corresponds to a flux
rate of about 0.0031 gallons/hour/sq.ft. of membrane/psi. This rate was substantially higher
than the initial productivity observed in the Drum C test (operating pressures of 250 to 300 psi)
which started at about 13 gallons per hour (0.0027 gallons/hour/sq.ft. of membrane/psi).
However, the Drum D flux rates decreased more rapidly over time than those observed in the
Drum C test such that the Drum C flux rates were actually higher after about 2 hours of
operation. These data indicate that increases in operating pressure will result in improved initial
productivity. However, the increased pressure also causes the membrane to foul at a faster rate
resulting in a more rapid decline in productivity.
^ 0.004-1
OT
Q.
"5
«3
t*
JQ
E
o
S
u.
s
CD
et
CC
X
ul
0.003-
0.002-
0.001-
O.OOO
a DrumC
RA2 •= 0.971
• DrumD
RA2 = 0.982
Figure 6.
Elapsed Time (Hours)
Permeate flux rate versus time; reverse osmosis tests.
Figure 7 shows how the permeate flux rate changed as the relative volume of the two
drums was processed through the reverse osmosis equipment. The flux rate for both tests
dropped initially as the first 10% of the drum volume was processed. The flux rates then
leveled off until approximately 60% of the volume had been processed. Beyond this point, the
flux rates dropped off dramatically for both of the drums tested. It is noteworthy that in this
comparison the flux rates associated with the Drum D test were consistently higher than the
Drum C flux rates. Consequently, it took about 1.5 hours longer to process Drum C than Drum
D. These differences in flux rates can be attributed to the higher operating pressure utilized in
the Drum D test. |
The Drum C solution was processed to a final volume of about 6 gallons (an 88 % volume
reduction) while the Drum D solution was processed to a final volume of 3 gallons'. (a 94%
16
-------
volume reduction). Reverse osmosis processing was terminated at these endpoints because of
the drastically reduced production rates (approximately 1 to 3 gallons per hour). Clean water
flux measurements taken before and after treatment of each of the two drums indicated that the
membranes recovered fully after the tests even though no attempt was made to clean the
membranes.
a DrumC
R*2 «= O.S90
* DfumD
R*2 = O.SS4
0.000
Figure 7.
% of Drum Volume Processed
Permeate flux rate versus percent of drum volume processed;
reverse osmosis tests.
NICKEL CONCENTRATIONS
In the low temperature evaporation tests, nickel concentrations started at 4140 mg/L
(Drum A) and 2540 mg/L (Drum B) respectively, (Table 1) and increased at a steady rate until
concentrations of approximately 25,000 to 30,000 mg/L were reached. As shown in Figure 8,
this level corresponds to a point where approximately 80 to 85 % of fee rinse water volume had
been processed. Beyond this point, nickel concentrations increased dramatically to
concentrations of 179,000 mg/L (Drum A) and 128,000 mg/L (Drum B), respectively. In the
Drum A test, the concentrate solidified shortly after the final concentrate sample was collected.
A sample of this solid concentrate was collected from the evaporator's plumbing and was found
to contain a concentration of 18.6% of nickel. Table 1 shows the nickel concentrations of the
concentrate solutions at the beginning and end of both the low temperature evaporation and
reverse osmosis tests.
-------
TABLE 1. COMPARISON OF NICKEL CONCENTRATIONS IN CONCENTRATE, DISTILLATE, AND PERMEATE
NICKEL CONCENTRATIONS
Low Temp. Evap. Reverse Osmosis
Product Drum A Drum B DrumC Dnnn D
Concentrations at beginning of test (mg/L):
Concentrate
Distillate
Permeate
Concentrations at end of test
Concentrate
Distillate
4,140 2,540
2.5 2.2
-
179,000 128,000
1 0.3
2,580 1,425 •
- -
44.5 14.5
12,560 18,200
— — •
Permeate
210
790
Ratio of «tig*i11«fi'- permeate to concentrate:
Distillate 0.0296
Permeate —
0.01%
1.49%
1.5456
o
o
o
c
o
O
z
o
o
.o
c
o
o
Figure 8.
Drum A
DrumB
-I
20
1 - • - T
40 60 80
of Drum Volume Processed
100
Concentrate nickel concentration versus percent of drum volume
processed; low temperature evaporation tests.
18
-------
Nickel levels used in the reverse osmosis tests started at 2580 mg/L and 1425 mg/L
respectively (Table 1). Figure 9 depicts how nickel concentrations in these samples changed as
the solutions were processed. Nickel concentrations increased steadily until about 60% of the
rinse water volume was processed. At this point, nickel concentrations were about 4,000 to
5,000 mg/L, in the two drums. Beyond this point, nickel concentrations increased more rapidly
until final concentrations of 12,560 mg/L (Drum C) and 18,200 mg/L (Drum D) were reached.
The final concentrations attained in the reverse osmosis tests correspond to about 7 to 14% of
the nickel concentrations achieved with the low temperature evaporation tests.
20000 -t
E
c:
o
c:
a>
o
c:
o
O
10000-
1C
ID
•a
c
o
o
13 DrumC
* DmmD
Figure 9.
% of Drum Volume Processed .
Concentrate nickel concentration versus percent of drum volume
processed; reverse osmosis tests.
The increased nickel concentrations observed in concentrate samples from both the low
temperature evaporation and reverse osmosis tests were relatively consistent with the volume
reductions that were accomplished in these tests. This phenomenon, along with the data
presented above, suggests, that the equipment effectively concentrated die nickel.
Distillate nickel concentrations determined in the low temperature evaporation tests
ranged from 2.5 to 0.1 mg/L during the course of testing. Distillate samples collectecLfrom the
Drum A test averaged 0.71 mg/L while Drum B samples averaged 0.37 mg/L. Table 2 provides
a summary of the average nickel concentrations in the dfcrilfyte samples. The ratios of distillate
to concentrate nickel concentration averaged 0.01% (Drum B) to 0.02% (Drum A) of the
concentrate concentrations during the course of the testing. The nickel concentrations of the
distillate did not appear to be affected by the rising nickel concentrations in the concentrate as
the test progressed. Due to the fact that nickel salts are not volatile, distillate samples collected
at the beginning of the tests contained nickel concentrations similar to those collected at the end
19
-------
of the test. The levels of nickel La the rinse water are low enough hi the distillate such that it
could be both reused as rinse water or discharged to the POTW without additional treatment.
TABLE 2. AVERAGE NICKEL CONCENTRATIONS IN DISTILLATE AND PERMEATE
Distillate Ni Concentration Permeate Ni Concentration
Mean (mg/L) STD (mg/L) Mean (mg/L) STD (rag/L)
A(n=13) 0.71 0.63
B (n=16) 037 0.52
C (n=22)
D (n=17)
- - ;
- - •
89.55 49.22
13438 202.19
The low temperature evaporation test data presented hi Tables 1 and 2 do not include several
samples which contained elevated nickel concentrations as a consequence of the evaporator
malfunctioning. On several occasions, the evaporator feed solution overflowed and contaminated
the distillate. Representatives from licon have indicated that this phenomenon tends to occur
in the pilot-scale evaporator due to its sensitivity to operating conditions and the necessity to
calibrate all operating parameters manually. Full-scale production evaporators are not as
susceptible to overflowing because the operating parameters are more fully automated. Based
on this information, it was decided not to include the samples contaminated by the bojl over hi
the Table 1 and 2 analysis. j
Nickel concentrations in permeate samples (Table 2) averaged 89.55 mg/L hi the Drum
C test and 134.38 mg/L in the Drum D test. The difference hi these tests can be attributed
primarily to the higher pressures associated with the Drum D test which caused more of the
nickel salts to permeate through file membrane. The ratio of permeate to concentrate nickel
concentration averaged IA9% (Drum Q and 1.54% (Drum D), respectively, during the course
of the testing. Figure 10 shows the relationship between concentrate and permeate nickel
concentrations during the reverse osmosis tests. This graph indicates that the relationship
between concentrate and permeate nickel concentrations is relatively linear until the concentrate
reaches a level of about 12,000 mg/L. Beyond this point, additional increases In concentrate
nickel result in dramatically higher nickel concentrations in the permeate. The nickel levels hi
the permeate samples indicate that initial treatment with reverse osmosis would not produce
water of adequate quality for discharge to a POTW. However, the processed water would be
of sufficient quality for raise as rinse water. Further, the quality of this solution could be
further improved through additional reverse osmosis processing to remove additional nickel.
20
-------
800
H DrumC
RA2 = 0.990
* DfumD
RA2 « 0.974
10000
20000
Concentrate NI Concentration (mg/L)
Permeate nickel concentration versus concentrate nickel
concentration; reverse osmosis tests.
Figure 10.
TOTAL ORGANIC CARBON CONCENTRATIONS
Total organic carbon (TOC) concentrations in the Graham Plating nickel plating baths
are maintained at around 14,000 mg/L. TOC concentrations in rinse water samples collected
during Hhe low temperature evaporation tests started at 990 mg/L (Drum A) and 550 mg/L
(Drum B), respectively, TOC levels in the concentrate increased slowly until a concentradon
of 4,000 to 5,000 mg/L was reached. These levels were reached when approxhriately 80 to 9055
of the drum volume had been processed. Beyond this point, TOC concentrations increased
dramatically until final concentrations of 26,000 mg/L A to 25,000 mg/L (Drum B) were
attained. These concentration increases were relatively consistent wim the volume reduction that
was achieved. Figure 11 shows how levels of TOC increased in the concentrate solution as the
rinse water was processed through the evaporator. Table 3 includes a summary of the TOC
levels at the start and finish of these tests.
21
-------
en
E
c
0
o
o
O
O
O
30000 -i
20000-
10000
Q Drum A
* DrumB
2O
40
60
80
100
% of Drum Volume Processed
Figure 11. Concentrate TOC concentration versus percent of drum processed;
low temperature evaporation tests.
TABLE 3. COMPARISON OF CONCENTRATE, DISTILLATE, AND PERMEATE TOC CONCENTRATIONS
Low Temp. Evap. Reverie Osmosis
Drum A Drum B Drum C Drum JD
Concentrations at Beginning of Test (mg/L) :
Concentrate
Distillate
Permeate
990
2.6
-
550 590
63
9.1
340
' : -
: 1-9
Concentrations at End of Test (mg/L)
Concentrate
Distillate
Permeate
26.000
63
-
25,000 2,800
1 -
- 16
3,500
i.
12
Figure 12 depicts changes in concentrate TOC levels that occurred as the volume of rinse
water was processed through the reverse osmosis equipment. This information is summarized
further in Table 3. As shown, concentrate TOC levels started at 590 mg/L (Drum C) and 340
22
-------
mg/L (Drum D), respectively. TOC levels increased steadily until concentrations of
approximately 1,000 mg/L were reached. This level was reached when about 60% of the drum
volume was processed. Beyond this point, TOC levels increased more rapidly until final
concentrations of 2,800 mg/L (drum Q and 3,500 mg/L (Drum D) were reached.
3000 n
of Drum Volume Processed
Figure 12.
Concentrate TOC concentration versus percent of drum volume
processed; reverse osmosis test.
TIae increased TOC concentrations observed in the concentrate samples from both the low
temperature evaporation and reverse osmosis tests were faMy consistent with the volume
reductions that were accomplished in these tests. This phenomenon suggests that the organic
compounds present in the rinse water were effectively concentrated by the equipment and, in
general, were not allowed to pass through the processes to the distillate and permeate solutions.
While this information illustrates the fete of the quantity of organic compounds, it does not take
into account changes in the quality of these compounds which may have occurred as a result of
the processes. Evaluation of me mdivib^orgardcco^^
would be costly and time consuming. While this mformation would be useful, it is beyond the
scope of this study.
'• ••
Permeate TOC levels were closely related to TOC levels present hi the concentrate
(Figure 12). Figure 13 provides information regarding changes in permeate TOC concentrations
as the volumes of rinse water were processed through the reverse osmosis system. Comparison
of this information to Figure 12 suggests that permeate TOC levels are closely related to TOC
levels piesent in the concentrate. Permeate TOC concentrations increased relatively slowly at
first but increased dramatically after the first 70 percent of the solution volume had been
processed.
23
-------
3000 -\
en
E
c
o
o
t=
o
o
O
O
o
"o
o
0.
2000 -
1000
a DrurnC
• DairnD
0 20 40 60
•% «f Drum Volume Processed
Figure 13. Permeate TOC concentration versus percent of drum volume
processed; reverse osmosis tests,, ;
Table 4 provides a summary of TOC levels in distillate and permeate solutions. As
shown, the distillate samples averaged 3.5 (Drum A) and 3.03 (Drum B) mg/L, respectively.
The permeate samples contained substantially more TOC, averaging 19.46 (Drum C) and 21.9S
mg/L (Drum D), respectively. The permeate samples contained 5 to 7 times as much TOC as
the distillate samples. TOC levels observed in the distillate samples averaged only '0.13 to
0.3196 as much as the concentrate TOC levels. TOC was present in the permeate at levels
which averaged 1.24 to 1.48% of fee corresponding concentrate levels. This data suggest that
the organic constituents (brighteaers, etc.) associated with the rinse water are able to penetrate
the reverse osmosis membrane bat are not distilled in the low temperature evaporation process.
Table 4. AwerageTOC CooeeataSoos m Distillate and Penneate ' '
Drum
DoSIateTOC Concenintiaa
STD (mg/L)
Permeate TOC Co
Mean (mg/L)
STDOag/L)
A (n=3)
3.50
3J03
2.48
2.55
19.46
21.98
15.48
32.91
24-
-------
ELECTRICAL' CONDUCTIVITY
In the low temperature evaporation tests, electrical ^conductivity values in concentrate
samples started at 16,600 (Drum A) and 10,150 (Drum B) umhos, respectively. As shown in
Figure 14, conductivities increased steadily as the first 80 percent of the drum volume was
processed to about 40,000 umhos (Drum B) and 65,000 umhos (Drum A). As the final few
gallons of rinse water were processed, electrical conductivities increased dramatically to 98,700
umhos (Drum C) and 108,800 umhos (Drum B). Table 5 summarizes changes in concentrate
ECs as the tests progressed.
•£• 120000
o
Drum A
DrumB
0'
0 20 40 60 80 100
% of Drum Volume Processed
Figure 14. Concentrate electrical conductivity versus percent of drum volume
processed; low temperature evaporation tests.
The increases in concentrate electrical conductivity values observed in the reverse
osmosis tests were fairly consistent with the volume reductions that were achieved in these tests.
This phenomenon combined with the data presented above suggests that the equipment was
effective at keeping most of the salts from crossing the membrane into fee permeate solution.
Electrical conductivity values in concentrate samples collected from the first 80 percent of the
low temperature evaporation tests were also reasonably consistent with the volume reduction that
occurred. However, electrical conductivity values associated with concentrate samples collected
from the final 20 percent of the low temperature evaporation tests are well below what might
be expected based on the reduction of feed solution volume that occurred. These data suggest
that at conductivities above about 60,000 umhos, salts begin precipitating out of solution and,
consequently, are not measured in the electrical conductivity analysis.
25
-------
T«ble5. Comparison of Concentrate, Distillate. «nd Permeate Electric«l Conduarrjaea
Low Temp. Evap. . Rwene Orao.is
Dtum A Dnim B DrumC Dnim D
Electrical Conductivities at Beginning of Test (umhos)
Concentrate 16,600 10,150 10,300 6,270
Distillate 35'8 30-5 ~ ; ~ .
500 242
Permeate _^______ — :
Electrical Conductivities at End of Test (umhos) \
Concentn,te 98,700 108,800 37,300 47,900
Distillate 27-3 2I'6
3,720 9.190
Permeate '• —— — ' '
As shown in Figure 15 and Table 5, concentrate electrical conductivity values in the
reverse osmosis tests started at 10,300 and 6,270 umhos for Drums C and D, respectively. As
the test progressed, electrical conductivity increased steadily until approximately 60 percent of
the drum volume had been processed. At this point, Drum D displayed an electrical
conductivity of about 10,000 umhos while Drum C's electrical conductivity was about 18,000
umhos. As the next 25 to 35 percent of the solution was processed, electrical conductivity
increased dramatically until final conductivities of 37,000 umhos (Drum Q and 48,300 umhos
(Drum D) were attained. The higher conductivity reached during the Drum D test was most
likely a consequence of the higher pressures utilized in this test. The higher pressures enabled
the reverse osmosis equipment to force more permeate across the membrane and concentrate the
feed solution further.
Changes in electrical conductivity values associated with permeate samples collected in
the reverse osmosis tests were- very similar to those which occurred in the concentrate samples.
Figure 16 and Table 5 indicate that permeate electrical conductivities started at 500 umhos
(Drum C) and 242 umhos (Drum D), respectively. Permeate electrical conductivities increased
steadily until about 70 percent of the drum volume was processed. Approximately 70 percent
of the test volume had been processed at (bis point. Permeate conductivity i increased
dramatically as the next 15 to 25 percent of the rinse water volume was processed .until final
conductivities of 3720 umhos (Drum C) and 9190 umhos (Drum D) were reached. The Mgher
conductivities associated with the Drum D test may again be attributed to the higher pressores
utilized in this test. The higher pressures forced more salts to pass through the membrane into
the permeate. '••"•„
26
-------
E
3
50000
4000O
~ 30000
JO
o
ej
o
o
o
Ul
20000
10000
s DiumC
* DoimD
20 40 60 80
of Drum Volume Processed
too
Figure 15.
Concentrate electrical conductivity versus percent of drum volume
processed; reverse osmosis tests.
3
•a
c
o
O
ct
o
III
o
E
10OOO
8000
6000
4000
2000
e DrumC
* OmmC
60 80
% of Drum Volume Processed
TOO
Figure 16.
Permeate electrical conductivity versus percent of drum volume
processed; reverse osmosis tests.
27
-------
Table 6 provides a summary of electrical conductivities present in permeate and distillate
samples. As shown, the low temperature evaporation tests produced distillate samples that
averaged 22.94 (Drum A) and 25.24 (Drum B) umhos, respectively. The reverse osmosis tests
produced permeate samples which averaged 1331 (Drum C) and 1794 (Drum D)| umhos.
Distillate samples collected during the low temperature evaporation tests exhibited electrical
conductivities which averaged only 0.09 to 0,15 percent as high as the concentrate
conductivities. Permeate samples from the reverse osmosis tests exhibited electrical
conductivities which averaged 5.97 to 6.45 percent as high as the conductivities displayed in the
concentrate samples suggesting that more salts can pass through the reverse osmosis membrane
than the evaporation process.
Table 6. Average Electrical Conductivity Values in Distillate and Permeate ;
Distillate EC Permeate EC
Mean (un&os) ' STD (umhos) Mean (umhos) STD (umhos)
A (n=13)
B (n=16)
C
-------
c
«s
l_
J3
E
c
or
W
X
a
a
a
tr
X
3
0.004 n
0.003-
0.002-
0.001 -
0.000
Q OrumC
RA2 = 0.964
10000
20000
Figure 17.
c
c
a
E
e
5
O"
to
Q
C
a
tr.
x
u!
Concentrate HI Concentration (mg/L)
Permeate flux rate versus concentrate nickel concentration;
reverse osmosis tests.
0.004-1
0.003-
0.002-
0.001
0.000
DrutnC
R*2 « 0.9SS
DfumD
R*2 = 0.917
Figure 18.
Permeate Nickel Concentration (mg/L)
Permeate flux rate versus permeate nickel concentration;
reverse osmosis tests.
29
-------
c
a
Ja
E
ti-
er
a
O
o
tr
x
ul
Figure 19.
tn
a.
o
c
O
cr
CO
z
*
a
a
CC
X
uZ
0.004
0.003
0.002
0.001
13 "DrumC
R*2 «= 0.951
o.ooo-i « E * 1 • « • ' ' l
0 10000, 20000 30000 40000 50000
Electrical Conductivity of Concentrate (umhos)
Permeate flux rate versus concentrate electrical conductivity;
reverse osmosis tests.
0.004 n
0.003-
0.002-
0.001
0.000
DrumC
RA2 « 0.963
* Drum D
R*2 «* 0.957
20QO
4000
eooo
8000
10000
Figure 20.
Electrical Conductivity of Permeate {umhos)
Permeate flux rate versus permeate electrical conductivity;
reverse osmosis tests.
30
-------
Electrical conductivity also proved to be a good indicator of the nickel concentrations
present in the concentrate solutions produced by both the low temperature evaporation and
reverse osmosis tests. Figure 21 shows that an exponential relationship existed between
concentrate electrical conductivity and nickel concentrations. This relationship appears to be
linear until concentrate electrical conductivities reach about 70,000 umhos. Then nickel
concentrations increased exponentially as increases in electrical conductivity occurred. Figure
22 shows the relationship between the electrical conductivity of concentrate samples and their
respective nickel concentrations for the reverse osmosis tests. As shown, this relationship
appears to be very linear. However, it should be noted that the highest conductivities observed
in these tests were in the 40,000 to 50,000 umhos range which is well below the point (70,000
umhos) where the concentrate electrical conductivities began increasing exponentially. Figure
23 shows the relationship between electrical conductivities associated with permeate samples and
their respective nickel concentrations. This relationship appears to be similar to the one defined
in Figure 22 for the concentrate samples. Permeate nickel concentrations increased linearly as
the electrical conductivity levels increased.
200000 -i
ro
E
©
o
c
o
o
o
32
o
c,
93
O
er
o
O
1000OO-
B Drum A
R*2 = 0.920
* DrumB
R*2 = 0.939
20000 40000 60000 80OOO
1 ' 1
1OOOOO 120000
Electrical Conductivity, of Concentrate {umhos)
Figure 21. Concentrate nickel concentration versus electrical conductivity;
low temperature evaporation tests.
31
-------
20000
c
o
e
o
o
c
o
O
o
.!£
U
O
U
c
o
O
10000
EI DrumC
R*2 « 0.992
» DrumD
R*2 •= 0.993
1
10000
1 « 1 1 J « I
20000 30000 40000 50000
Figure 22.
O)
E
c
o
o
c
o
o
o
_o
Z
o
g
o
a.
Figure 23.
Electrical Conductivity of Concentrate (umhos)
Concentrate nickel concentration versus electrical conductivity;
reverse osmosis tests.
8001 '
600
400
200
DrumC
R*2 «= 0.995
Drum D
R*2 = 0.990
2000
6000
8000
10000
Electrical Conductivity of Permeate (umhos)
Permeate nickel concentration versus permeate electrical
conductivity; reverse osmosis tests.
32
-------
Electrical conductivity data was also well correlated with total organic carbon
concentrations in both the low temperature evaporation and reverse osmosis tests. Figure 24
shows the relationship between electrical conductivity of concentrate samples and corresponding
total organic carbon levels. This relationship appears to be linear through electrical conductivity
levels of approximately 50,000 umhos. At higher electrical conductivity levels, total organic
carbon levels increased exponentially. Figure 25 shows a linear relationship between concentrate
electrical conductivity and total organic carbon results in the reverse osmosis tests. It should
be noted that the highest electrical conductivity values attained in the reverse osmosis tests were
below the 50,000 umhos level where nickel concentrations began increasing exponentially in the
low temperature evaporation tests.
The relationships described above will be very useful hi developing operating standards
for the low temperature evaporation and reverse osmosis systems. It is fortunate that electrical
conductivity data are well correlated with the other parameters because conductivity
measurements are relatively quick, simple, and inexpensive to obtain. Further, electrical
conductivity instruments can be incorporated into operating systems to help provide process
control.
33
-------
30000 -i
CO
c
o
o
o
o
O
€>
"o
O
o
20000-
10000 -
H Drum A
R*2 •= 0.971
* DfumB
R*2 «= 0.873
20000 40000 60000 80000
T « 1
100000 120000
Electrical Conductivity of Concentrate (umhos)
Figure 24. Concentrate TOC concentration versus electrical conductivity;
low temperature evaporation tests.
O
O
c
o
O
O
O
o
o
c
o
O
4000 n
3000-
2000-
10OO-
B DrumC
R*2 K 0.99S
» DiumO
RA2 * 0.999
40000
Figure 25.
0 10000 20000 30000 40000 50000
Electrical Conductivity of Concentrate (umhos}
Concentrate TOC concentration versus electrical conductivity;
reverse osmosis tests.
34
-------
SECTIONS
ECONOMIC ANALYSIS
The costs and benefits associated with installing low temperature evaporation and/or
reverse osmosis systems at the new Graham Plating facility were analyzed to determine the
economic feasibility of these technologies. Assumptions regarding inflation rale, discount rate,
federal tax rate, depreciation schedule, project life and various operating expenses were entered
into a IJOTUS spreadsheet (General Electric, 1987) which calculates a number of economic
indices. The assumptions utilized in these calculations and their sources are presented in Table
7. Projected costs of future liabilities associated with hazardous waste generated from the
Graham Plating facility were not included in this assessment due to the lack of an accurate
means to assess these liabilities. Therefore, the economic indices presented below are most
likely a conservative estimate of the monetary benefits that might be expected due to
implementation of these technologies.
TABLET. ASSUMPTIONS FOR ECONOMIC CALCULATIONS
Item
Inflation Rate
Discount Rate
Federal T«x Rate
Depreciation Schedule
Project I jfe
Power Costs
Thermal Unh Costs
Labor Costs
Salvage Value
Sludge Disposal Costs
Water Costs
Reclaimed Ni Value
Factor
4%
7.7256
3456
7 years
10 years
$.10/ldlowatt
$.40/100,000 BTU
$15/bour
1056 of Capital Cost
$7.27/gaHon
$3/1,000 gallons
$3.75/pound
Source
Consumer Price Index
10 year treasury bffl rate +0-5%
General Electric, (1987)
Genera! Electric, (1987)
Osmonics and Licon
Commonwealth Edison
Commonwealth Edison
Graham Plating
Osmonics and Licon
Graham Hating
Arlington Heights Public Works
Stutz Metal Finishing Products
In addition to the assumptions presented in Table 7, it was further assumed that all of the rinse
water treated with the systems would be produced from a nickel electroplating line. Under
actual plant conditions, rinse water from other electroplating lines would be processed with the
equipment. However, since nickel rinse water was the only feed solution tested in this project
35
-------
it was not appropriate to make assumptions regarding equipment performance with respect to
other metals. It was also assumed that all equipment evaluated in the economic assessments
would operate 24 hours per day, 5 days per week at only 80% of its capacity (5,760 gallons per
day). This estimate takes into account down time associated with maintenance activities and
production fluctuations. -
These assumptions along with inforEnation collected in the low temperature evaporation
and reverse osmosis tests were used to prepare the economic analysis on three alternative: process
systems that could be implemented at th© new Graham Plating facility. The alternatives
examined include: 1) a low temperature evaporator system, 2) a reverse osmosis system, and
3) a combined system which would utilize feoth technologies. Detailed cost/benefit assessments
of these alternatives are provided below. ;
The low temperature evaporation economic assessment was based on the anticipated
utilization of the Licon evaporator purchased for use at the new Graham Plating facility. This
unit is capable of processing up to 7,200 gallons of electroplating rinse water per day. Table
8 provides a cash flow summary associated with implementation of the evaporator system. As
shown, this unit requires a significant capital investment of $140,000. Additionally, the
evaporator requires considerable energy input (5,000 BTU per gallon of solution processed)
which would result in energy costs of just snider $30,000 per year (based on 1992 costs) or $20
per 1,000 gallons. Replacement of pump seals, miscellaneous repairs and equipment monitoring
activities would cost about $3,060 per year. Replacement of the various pumps utilized on this
equipment would take place twice during the project life at the projected costs displayed in Table
8- !
Considerable savings could be expected through utilization of the evaporation system to
reclaim nickel salts from the rinse water.. It was estimated that over 12 thousand pounds of
nickel could be salvaged from the nickel zinse water in a year's time through utilization of the
evaporation system. Based on 1992 dollars, these salts would have an approximate value of
$36,660 per year as compared to me cost ©Fbuying an equivalent amount of new plating crnps.
Additional savings would be realized throusgh reduction in water utilization ($4492 per year) and
sludge generation and disposal ($19,200 per year). :
Based on the parameters described above, the LOTUS spreadsheet was utilized to
calculate the economic indices presented MI Table 9. As shown, the economics associated with
installing a Low Temperature Evaporation system at the Graham Plating-facility are relatively
favorable. It is estimated mat the payback period associated with this technology would require
about 6.9 years. An implied rate of retom of about 10.6% would be realized while the net
present value would be $54,017.
36
-------
TABLE 8: LOW TEMPERATURE EVAPORATION CASH FLOW SUMMARY
[Dollar Amounts Before Taxes and Depreciation]
AR
n 2 345 6 7 8 9 10
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
EXPENDITURES
son Double Effect Evaporator
PERATING EXPENDITURES
nergy Consumption
[7.5 billion BTU/year)
quipment Maintenance*
Replace Concentrate Pump
Replace Rea'rculaion Pump
Replace Distillate Pump
Replace Pump Seals
Miscellaneous Repairs*
Monitor Equipment*
FOTAL EXPENDITURES
SAVINGS
Reclaimed Nickel
9776 fosJyear @ $3.75 per Ib.
Water Consumption
1497000 ga^year @ $3/1000 ga.
Waste Disposal
2640 gaJyr. sludge @ $7.27/ga.
Equipment Salvage
TOTAL SAVINGS
CASH FLOW '
140000
29952 31150 32396 33692 35040 36441 37899 39415 40991 42631
450 526
787 921 !
787 921
1000 1040 1082 1125 1170 1217 1265 1316 1369 1423
500 520 541 562 585 608 633 658 684 712
1560 1622 1687 1755 1825 1898 1974 2053 2135 2220
173012 34332 35706 39159 38619 40164 41771 45810 45179 46986
36660 38126 39651 41238 42887 44602 46387 48242 50172 52179
4492 4672 4859 5053 5255 5465 5684 5911 6148 6394
19200 19968 20767 21597 22461 23360 24294 25266 26277 27328
• 14000
60352 62766 65277 67888 70603 73427 76365 79419 82596 99900
-112660 28433 29570 28729 31983 33263 34593 33608 37416 52913
Prices include labor
37
-------
TABLE 9. ECONOMIC SUMMARY - LOW TEMPERATURE EVAPORATION OPTION ;
Capital Invested ' $140,000
Payback Period 6.9 years
Net Present Value (NPV) ! $54,017 |
Implied Rate of Return (IRR) 10.6%
The economic assessment for the reverse osmosis system was based on the utilization of
an Osmonics unit equipped with 12,4 inch diameter membrane cartridges and a ten horse power
motor. Based on results of the reverse osmosis tests and manufacturers recommendations, this
unit should be capable of processing rinse water volumes of about seven to eight thousand
gallons per day. This estimate takes into account the low processing rates that can be expected
as the concentrate solution is condensed to the maximum concentrations attainable with this
technology. A cash flow summary explaining the costs and benefits associated with installing
and operating this unit is provided in Table 10.
Capital costs required to install this unit would be approximately $50,000. The
membrane cartridges for this unit would have to be replaced every other year at a cost! of $800
each (1992 dollars). The reverse osmosis unit would require about $3,724 annually (1992
dollars) in electrical power costs or $2.50 per 1,000 gallons. Other expenses required to operate
this unit would include costs associated with prefilters, cleaning the membranes, and monitoring
the equipment which would cost approximately $4608 per year (1992 dollars).
Based on the results of t&e reverse osmosis tests conducted in this project, the unit could
only concentrate the nickel solution to a concentration of about 1.2 to 1.8 percent nickeL This
concentration is well below the eight percent nickel concentration normally utilized in the
Graham Plating operation's plating baths. Therefore, the concentrate solution produced through
the reverse osmosis process coold only be placed in the plating baths to replace water losses.
It is estimated that about 300 gallons of concentrate containing approximately 1.5% nickel would
be produced daily from the reverse osmosis process. This volume would be well beyond that
which could be used to replace water in the plating baths. For the purposes of economic
calculations, it was assumed that 3/4 of this concentrate (about 58,656 gallons per year) could
be used to replace plating bath, water losses. This figure probably greatly exceeds that which
could reasonably be used in the plating operation but it is used in this economic assessment to
provide a "best case" scenario. The remaining 1/4 of the concentrate solution (about 19,552
gallons per year) which could not be utilized to replace plating bath water losses would have to
be treated further by precipitation and shipped off site to a facility that conducts metal reclaiming
operations at a net cost of just over $45,000 per year.
Substantial savings would be realized as a result of installing a reverse osmosis system
at the Graham Plating facility. Approximately $27,000 (1992 dollars) per year in savings would
be realized through reclamation of nickel salts from the processed rinse water (as compared to
having to purchase fresh salts). An additional $4492 (1992 dollars) per year could be antidpated
38
-------
TABLE 10: REVERSE OSMOSIS CASH FLOW SUMMARY
[Dollar Amounts Before Taxes and Depreciation]
•&•
\
'EAR
JAPITAL EXPENDITURES
Dsmonics 12 membrane system
i
DPERATING EXPENDITURES
Energy Consumption
(37,240 kw/year)
Equipment Maintenance *
Replace Membranes
Replace pre-filters
Clean Membranes
Monitor Equipment
Replace Pump
Waste Disposal**
364 drums/year @ $1 1 0/drum.
4- $5000 shipping
TOTAL EXPENDITURES
SAVINGS
Reclaimed Nickel
7223 IbsJyear @ $3.75 per Ib.
Water Consumption
1497000 gaJyear @ $3/1 000 ga.
Waste Disposal
2640 gaJyr. sludge @ $7.27/ga.
Equipment Salvage
TOTAL SAVINGS
CASH FLOW
1
1993
50000
3724
2028
1020
1560
45040
103372
27084
4492
19200
50776
-52596
2 3
1994 1995
3873 4028
9984
2109 2193
1061 1103
1622 1687
46842 48715
65491 57727
28167 29294
4672 4859
19968 20767
52807 54919
-12684 -2808
4 5
1996 1997
4189 4357
10799
2281 2372
1147 1193
1755 1825
50664 52690
70835 62438
30466 31684
5053 5255
21597 22461
57116 59401
-13719 -3037
6 7
1998 1999
4531 4712
11680
2467 2566
1241 1291
1898 1974
6083
54798 56990
82698 67533
32952 34270
5465 5684
23360 24294
61777 64248
-20922 -3285
8
2000
4901
12633
2669
1342
2053
59270
82867
35641
5911
25266
66818
-16049
9
2001
5097
2775
1396
2135
61640
73043
37066
6148
26277
69490
-3553
10
2002
5300
13664
2886
1452
2220
64106
89629
38549
6394
27328
5000
77270
-12359
* Prices include labor
** Assumes 3/4 of concentrated nickel solution can be used in plating
! while other 1/4 must be disposed of bath makeup
: 39
-------
through reduction in water utilization. Waste disposal costs associated with electroplating sludge
riddance would be reduced by about $19,200 per year.
The economic indices (payback period, net present value, and implied rate of return)
associated with the reverse osmosis system (Table 11) show that for this operation, the reverse
osmosis system is not an economically viable option. The primary factor contributing to the
unfavorable economic conditions associated with this system can be traced directly to the costs
required for disposal of the excess nickel concentrate. If all of this solution could be recycled
into the electroplating process, the economics associated with the reverse osmosis option would
be much more favorable. It is possible tfeat other companies could recycle all of the concentrate
solution from the reverse osmosis process. An economic assessment performed on the feasibility
of a reverse osmosis system installed at such a facility would probably suggest that the reverse
osmosis technology is acceptable and may be the;most beneficial of all rinse water recycling
options. ! '•
TABLE 11. ECONOMIC SUMMARY T REVERSE OSMOSIS OPTION i
Capital Invested ! $50,000
Payback Period . ! Never
Net Present Value (NPV) (578^17)
Implied Rate of Return (TERR) : (9.8%) !
The economic assessment for the combined technologies option was based on the
utilization of an Osmonics reverse osmosis unit equipped with six membrane cartridges and a
Licon double effect evaporator. Based on results from the reverse osmosis tests and
manufacturers recommendations., the reverse osmosis unit would be capable of processing 5,500
to 6,000 gallons per day. Although this unit would be equipped with only six membranes, the
individual membranes would be more productive than the membranes placed in the unit specified
in the reverse osmosis system. The increased productivity could be attributed to the fact that
the reverse osmosis system used in the combined technologies option would only be utilized on
feed solution containing less tnaa 4,000 mg/L of nickeL At these concentrations, the membreines
would not foul as rapidly resulting in higher productivity and less down time for membrane
cleaning. The evaporator specified for the combined technology option would be a licon double
effect evaporator capable of processing about 2,000 gallons of rinse water per day. , . '
\ - *
The reverse osmosis system would be used to concentrate about 80 percent of the rinse
water volume from an initial concentration of about 784 mg/L nickel to about 4,000 mg/L
nickel. The solution concentrated by fhe reverse osmosis system would then be transferred to
the low temperature evaporation system. The evaporator would concentrate the remaining 20
percent of the rinse water solution to a concentration of about eight percent nickel which could
subsequently be replaced in the plating bath. Using me equipment within its optimum operating
ranges would augment the ability of the systems to process the rinse water with maximum
efficiency while supplying the electroplating operation with high quality concentrate, distillate,
and permeate solutions for reuse. Since the equipment would always be functioning within
40
-------
TABLE 12: COMBINED SYSTEM CASH FLOW SUMMARY
[Dollar Amounts Before Taxes and Depreciationl
EAR
lAPITAL EXPENDITURES
Ucon Double Effect Evaporator
Osmonlcs 6 Membrane System
DPERATING EXPENDITURES
_OW TEMPERATURE EVAPORATION SYSTEM
Energy Consumption
(1 .5 billion BTU/yeai)
Hquipment Maintenance*
Replace Concentrate Pump
Replace Recirculation Pump
Replace Distillate Pump
Replace Pump Seals
Miscellaneous Repairs
REVERSE OSMOSIS SYSTEM .
Energy Consumption
[37.240 few/year]
Equipment Maintenance*
Replace Membranes
Replace pre-filters
Clean Membranes
t 23,4 6 « 78 0 10
1893 i,1»94 1095 ":...499ft 1997 1998 1999 2000 2001 - 2002
85000
30000
5990 6230 8478 S738 7007 7288 7579 7882 8198 &S28
334 461
675 790
675 790
800 832 865 900 936 973 1012 1053 1095 113S
400 416 433 450 468 487 506 526 647 569
3724 3873 4028 4189 4357 4531 4712 4901 5097 5300
4992 5399 5840 6316 6S32
806 838 872 907 943 981 1020 1061 1103 1147
460 478 498 517 538 560 582 605 630 655
4867
BOTH SYSTEMS
Monitor Equipment
TOTAL EXPENDITURES
SAVINGS
Reclaimed Nickel
9651 tbsJyear @ $3.75 per Ib.
Water Consumption
1497000 gaJyear @ $3/1000 ga.
Waste Disposal
;19200 gaJyr. sludge @ $1/ga.
Equipment Salvage
TOTAL SAVINGS
CASHFLOW
•Pricesinclude labor
1560 1622 1687 1755 1825 1898 1974 2053 2135 2220
128740 19282 14861 22599 16074 27423 17385 26437 18804 26388
36191 37639 39144 40710 42338 44032 45793 47625 49530 51511
4492 4872 4859 5053 5255 5465 5684 5911 6148 63S4
19200 19968 20767 21597 22461 23360 24294 25266 26277 2732S
11500
59883 62278 64769 67360 70055 72857 75771 78802 81954 96732
-68857 42996.72 4990856 44781.71 53980.78 45433.46 58385.61 52364.87 63149.88 70343.98
41
-------
optimum concentration ranges, smaller reverse osmosis and low temperature evaporation units
could be implemented than if the individual units were used alone.
Table 12 provides a cash flow summary for the combined technologies option. As
shown, a capital investment of $115,000 would be required to install the two systems described
above into the new Graham Plating facility. Energy costs to operate the two systems would
require about $9714 (1992 dollars) per year. Costs for replacing pre-filters and j cleaning
membranes associated with the reverse osmosis unit would be about $1,266 per year.
Membranes for the reverse osmosis unit would have to be replaced every other year at a cost
of about $4,800 (1992 dollars). Pumps for the reverse osmosis and evaporator units would have
to be replaced at various times during the projected life of the equipment at the approximate
costs projected on Table 12. Additionally, about $2,760 (1992 dollars) would have to|be spent
on a yearly basis to replace pump seals, perform miscellaneous repairs and monitor equipment
performance. . , '
The economic indices calculated for the combined system are presented in Table 13. As
shown, this option offers a net present value of $177,057, an implied rate of return of 27.6
percent, and would payback the capital investment in about 2.8 years. Based on the assumptions
and conditions used in this assessment, the combined option appears to offer the best economic
alternative for processing the electroplating rinse water. The superior indices produced by this
alternative may be attributed to the fact that the two technologies proposed in this option would
be utilized under conditions for which the equipment is best suited. The reverse osmosis
equipment is more efficient at processing the dilute (less than 4,000 mg/L) solution while the
evaporator is more efficient at processing rinse water containing nickel in concentrations in
excess of 4,000 mg/L. i
TABLE 13. ECONOMIC SUMMARY - COMBINED TECHNOLOGY OPTION
Capital Invested S11S,000
Payback Period • 2.8 years
Net Present Value (NPV) $177,057 '.
Implied Rate of Return QRK) 27.6%
42
-------
SECTION 6
CONCLUSIONS AND RECOMMENDATIONS
SYSTEM:
The low temperature evaporation system exhibited consistent productivity throughout the
tests This performance feature was unfailing regardless of the chemical concentrations of the
feed solution provided to the system. In addition to the steady production rate of the low
temperature evaporation system, it was capable of concentrating the rinse water feed solution
to nickel levels as high as 13 to 18 percent which are well above the 8 percent required for
return to the plating bath. The low temperature evaporation system concentrated the organic
constituents of the rinse water to total organic carbon levels of 25,000 to 26,000 mg/L. The
concentration rate of the organic components paralleled the nickel concentration rate suggesting
that few organic materials were lost to volatization. Characterization and quantitation of the
individual organic constituents was not performed in this project and should be a consideration
in future investigations. Prior to reuse of the concentrate in a plating bath, these organics may
need to be removed to prevent buildup and adverse affects on plating quality.
The quality of the cleaned rinse water is of great importance to electroplating shops,
particularly if it is to be reused in the electroplating process. Distillate produced by the low
temperature evaporation system was very low in both nickel concentration (average 0.31 to 0.71
mg/L) amd TOC concentration (average 3.04 to 3.50 mg/L). This water is of good enough
quality that it could be both reused within the electroplating facility or discharged to a POTW.
Disadvantages of the low temperature evaporation system include its relatively high
($140,000) capital cost and high energy requirements C$20 per 1,000 gallons processed). The
implied rate of return of 10.6 percent determined in the economic assessment for the low
temperaiure evaporation system suggests that it is a marginal investment opportunity by today's
standards. However, these estimates do not take into account the future liabilities that would
be Hiinimized as the result of reducing the quantity and type of hazardous waste discharges from
the facility.
The reverse osmosis system exhibited superior productivity at the beginning of the tests
and productivity dropped off dramatically after about 60 percent of the feed solution had been
processed. At this point, feed solution nickel concentrations were approximately 4,000 to 5,000
mg/L. Beyond these levels, the productivity of the reverse osmosis equipment decreased
dramatically as solids began to precipitate and foul the membrane. The reverse osmosis system
43
-------
was capable of concentrating the feed solution to nickel concentrations of 12,560 to 18,200
mg/L. These concentrations are we3i below the 8 percent nickel concentration required for the
plating bath. Some of this solution could be used to replace water losses in the electroplating
process. However, it is likely t&at the reverse osmosis system would produce too much
concentrated rinse water containing 1.2 to 1.8 percent nickel. This material would have to be
further processed by using an alternative technology such as low temperature evaporation or
shipped to a facility that could extract the nickel for utilization in other industrial processes.
The reverse osmosis system concentrated the organic constituents present in the rinse
water to levels of 2,800 to 3,500 mg/L. These concentrations suggest that the organic bath
constituents are concentrated by the reverse osmosis equipment at rates that parallel the nickel
concentration rates. No information was obtained in these tests regarding the quality and nature
of the organic bath constituents. However, the low operating temperatures utilized by the
reverse osmosis equipment (74 to 80 degrees F) should have prevented degradation of the
organic molecules.
The quality of the cleaned rinse water produced by the reverse osmosis equipment was
directly related to the quality of the feed solution pumped into the unit Permeate produced, by
the reverse osmosis system averaged 89 to 134 mg/L nickel. These concentrations are
acceptable for reuse as rinse water. However, this solution would not be acceptable for
discharge to POTWs. It should be noted that the permeate quality could be significantly
improved if the reverse osmosis system was utilized only on rinse water feed solution with lower
(4,000 to 5,000 mg/L or less) nickel concentrations. TOC concentrations averaged 19.46 to
21.98 mg/L in the permeate solution suggesting that some of the organic compounds are able
to permeate through the membrane. It is not known if some of the compounds have a greater
tendency to permeate the membrane than others and this possibility should be investigated in
future studies. Advantages of tihe leverse osmosis system include its relatively high production
rates with respect to low concentration (less than 4,000 mg/L) feed solutions. Additionally, it
requires lower capital investment (about $50,000) than a comparably sized low temperature
evaporation system. Energy costs to operate a reverse osmosis system would require only about
$2.50 per 1,000 gallons processed. .
Disadvantages associated with a reverse osmosis system include its inability to
concentrate the feed solution to levels beyond the 12,560 to 18,200 mg/L levels revealed in this
study. This factor alone would prevent utilization of a stand alone reverse osmosis system at
the Graham Plating fecfliiy due to the ampracticalities associated with utilization of the
concentrate produced by the system. Another disadvantage associated with the reverse osmosis
system is the lower quality permeate produced by the system. This solution would most likely
have to be reused within the plant and could not be discharged to the POTW.
Both the low temperature evaporation and reverse osmosis systems appear to offer
advantages under specific operating conditions. The low temperature evaporation system appears
to be best adapted to processing solutions with relatively high nickel concentrations. It can
process these solutions such that concentrate solution comprised of eight percent or more nickel
is produced along with a very high quality distillate solution. The reverse osmosis system is best
44
-------
adapted to conditions where the feed solution is of relatively low nickel concentration. It can
process the low concentration feed solution with relatively high efficiency to a level of about
4,000 to 5,000 mg/L. At this point the solution could be transferred to a low temperature
evaporator or other acceptable process for further concentration. Utilizing the equipment within
its optimum operating ranges would augment the ability of the systems to process the rinse water
with maximum efficiency while supplying the electroplating operation with high quality
concentrate, distillate, and permeate solutions for reuse. This relationship is consistent with the
combined technology economic assessment provided above in which the two systems would be
utilized in tandem. The implied rate of return of 27.6 percent associated with this assessment
is the most favorable of the economic scenarios examined due to the projected utilization of the
equipment under its most favorable conditions.
Electrical conductivity measurements taken during operation of both the low temperature
evaporation and reverse osmosis systems could be of great value during actual plant operating
conditions. The electrical conductivity data obtained in this project was well correlated with
nickel concentration, TOC concentration, and membrane flux characteristics. Accurate
assumptions regarding concentrate, permeate, and distillate quality could be based on electrical
conductivity measurements taken throughout the work day. Further, the equipment could be
automated to accumulate and discharge the various solutions based on in process electrical
conductivity measurements that could activate pumps, valves, and/or switches when preset levels
are attained.
RECOMMENDATIONS
This study analyzes the performance of the two technologies with respect to processing
rinse water from a nickel electroplating process. Additional tests utilizing rinse water from other
electroplating lines using other metals should be performed to determine these technologies*
usefulness with respect to processing the entire spectrum of rinse water streams mat would be
produced at a full-scale electroplating operation. Detailed analysis of all organic and inorganic
rinse water components (organic brighteners, sulfate, chloride, etc.) would be useful in future
studies to determine the effects of low temperature evaporation and reverse osmosis processing
on the relative quality and quantity of these constituents.
When Graham Plating implements the low temperature evaporation system in their new
facility, on site testing should be performed to allow comparison of this full-scale system with
the pilot-scale tests performed in this study. Detailed study of the performance of the
concentrated rinse water that is returned to the plating bath should be performed. Efforts should
be made to determine the effects of the recycled plating chemicals on plating quality and bath
longevity. The information obtained from this study is based on very short term observations.
Therefore, information should also be obtained regarding the long term performance of the full-
scale system at the Graham Plating facility.
Evaluations of other technology options such as ion exchange and electrowinning should
be performed in studies similar to those described in this project to provide a comparison with
respect to all alternatives currently available. This project has provided a good foundation for
45
-------
evaluating options for processing electroplating rinse waters. Execution of the additional studies
recommended above would provide an accurate assessment of all of the available alternatives
and their respective potential regarding the processing of electroplating rinse water.
46
-------
REFERENCES
Cushnie, George C., 1985. Electroplating Wastewater Pollution Control Technology. Park
]Ridge, New Jersey: Noyes Publications
Electroplating Engineering Handbook. 1984. Fourth edition. L. J. Dumey, editor. New York,
NY: Van Nostrand Reinhold.
General Electric. 1987. Financial Analysis for Waste Management Alternatives. Fairfield,
Connecticut.
Hunt, Gary E. Waste Reduction in the Metal Finishing Industry. JAPCA. VoL 38, No. 5, pp.
672-680. May, 1988.
Kushner, J.B., and A.S. Kushner. 1981. Water and Waste Control for the Plating Shop.
Second edition. Cincinnati, OH: Gardner Publication, Inc.
Making Pollution Prevention Pav in the Electroplating and Metal "Finishing Industries. 1983.
Summary of a Workshop held April 13 and 14, 1983. B. Paitington, preparer.
Charlotte and Raleigh, NC: Water Resources Institute of the University of North
Carolina (UNQ, North Carolina Department of Natural Resources and Community
Development, American Electroplaters Society, UNC-Chadotte Urban Institute Waste
information and Education (WISE) Program, and North Carolina State University
Industrial Extension Service.
Mrtal Finishing! 55th Guidebook - Directory. 1987. P.H. Langdon, M. Murphy, and SX.
Congdon, editors. Hackensack, NJ: Metals and Plastics Publications, Inc.
• Miller, Gary D. and Jackie Peden. Quality Assurance Proi«* Plan for Evaluation of the
Effectiveness of Low Temperature Evaporation for Chemical TtecycHng/Rense of
Electroplating Rinse Waters. Cincinnati, Ohio: U.S. Environmental Protection Agency.
April, 1991.
Rousseau, Ronald. 1987. Handbook of Separation Process Technology. New York, NY: John
Wiley and Sons.
47
-------
Rozelle, L., C. Kopp, and K. Cobian. jNew Membrane!
Metal Finishing Effluents. EPA-660/2-73-033.
December, 1973.
for Reverse Osmosis
Washington, D.C.:
U.S.
Standard Methods for the Examination of Water and Wastewater. 1989. Seventeenth edition.
L.S. Clesceri, A.E. Greenberg, and "SLR. Trusseil, editors. Washington, D.C.: American
Public Health Association, American Water Works Association, and Water Pollution
Control Federation.
United Nations Environment Program (UNEP).
Finishing industry: A Techmcal Guide.
Office (LEO).
1989. Environmental Aspects of the Metal
Paris, France: Industry and Environment
U.S. EPA, Office of Solid Waste and Emergency Response.
Solid Waste - Physical/Chemical
1986. SW-
Methods for
3rd edition. Washington, D..C.
48
-------
APPENDIX
QUALITY ASSURANCE
The quality assurance project plan submitted for this project (Miller, 1991) was written
to validate the evaluation of a full-scale low temperature evaporation system at the new Graham
Plating facility. However, Graham Plating has not been able to relocate their electroplating
operations to their new facility. Therefore, evaluation of the full-scale low temperature
evaporation process was not possible. An alternative study plan was developed in cooperation
with Graham Plating and the U.S. Environmental Protection Agency project officer to evaluate
both low temperature evaporation and reverse osmosis technologies on a pilot scale. The focus
of this project was modified to evaluate these technologies with respect to their capabilities for
processing electroplating rinse water produced from a nickel electroplating operation.
Specifically, the tests were structured to examine the fate of nickel, total organic carbon, and
electrical conductivity as the solutions were processed and concentrated through the tow
temperature evaporation and reverse osmosis systems.
Although significant modifications were made with respect to the project location and
scope., great efforts were made to ensure that the quality assurance objectives established for tins
project were not compromised. The specifications outlined in the quality assurance project plan
were followed with respectto sampling procedures, analytkalprocedures, instrument caHbzalion,
internal quality control checks, performance audits, and data reduction, calculation, vaBdatipn
and reporting. Compliance with these quality assurance objectives has resulted in high quality
data validating the success of the two test methods. This simulation will have broad applicability
to numerous electroplating firms.
NICE3EL ANALYSIS CALIBRATION /-
A five point calibration was used for all analyses performed by this technique, in
accordance with the manufacturer's operating/procedures manual. The instrument software
automatically processes the standards information and prepares a standard curve consistent with
the data. Standard curves were manually plotted and checked on a recurring basis to ensure the
accuracy of the computer algorithm, and to verify that the software was operating properly. An
operating range was selected to ensure that the standard carve still had sufficient slope on the
high end to facilitate unambiguous differentiation between sample concentrations of interest.
Calibration curves were compared between analytical runs to verify proper standards preparation
and consistent instrument performance.
49
-------
TOTAL ORGANIC CARBON ANALYSIS CALIBRATION >
The Dohnnann instrument was calibrated with a single standard of potassium hydrogen
phthalate in accordance with the manufacturer's operating/procedures manual. The calibration
standard, usually 1000 mg-C/L, was chosen to be higher in concentration than the highest
expected sample. Calibration was verified at approximately every 10th sample by rerunning the
standard. j
MEASURES OF ACCURACY AND PRECISION
Results of precision and accuracy determinations for nickel and TOC are included in the
accompanying table.
NICKEL ANALYSIS
The wastewaters from the nickel plating process contained initially high levels (4,000
mg/L) of nickel and these levels were increased significantly in the concentration processes. By
contrast, nickel levels in the treated water were much lower, ranging from 0.1 to about 100
mg/L. As a consequence, samples were run under several different wavelength conditions on
the AAS, trying to match the analytical range to the sample concentrations to minimize the
dilution factors. This resulted in a number of different check standards, one set for each
concentration range. . Check standards were monitored at approximately every 10 sample
analyses, and were compared for consistency within the run and between runs. Drift in excess
of 10% in check standard concentrations during the course of a run, which was occasionally
observed, was used to reject data sets or parts thereof. A chart of check standard responses
through a period in July, 1992, when the instrument was calibrated between 10 and 50 mg/L,
is attached. As a matter of course, samples that were above or below the range of standards
were reran at a different dilution or wavelength.
On average, one duplicate and one spike analysis was run with each 10 samples analyzed.
Duplicates were invariably within 5% relative difference throughout the project. Differences
as large as 10 55 would not have compromised the experimental objectives of the project because
of the distinct differences in treated water versus concentrates. So duplicate results strongly
supported the appropriateness of the nickel data for evaluating project results.
Spike results also were satisfactory for meeting project needs. Most spike recoveries fell
between 90-100%, with a few values in the 80% range. In a few cases, spike values fell
significantly below 80 %, suggesting errors in preparation or spike levels inappropriate to sample
concentrations. These spikes were prepared over and rerun to verify accurate performance of
the instrument.
Calibration standards were also routinely inserted into sample sets to serve as checks on
instrument performance during the course of a run. One standard was inserted for each 10
samples analyzed; the concentration of the standard was varied so that in our standard 30 sample
runs, three different standards were routinely checked. Instrument responses to these-standards
50
-------
were used in conjunction with check standard data to indicate drift or other problems with
instrument performance.
Numerous samples were rerun during the course of the several months of analytical
support provided this project. Reruns were performed to ensure comparability of results over
time and to check on results that were out of range or suspect due to problems with check
standards or spikes. These reruns regularly confirmed the validity of the original
determinations, often to within 1% relative difference.
Specific conductivity was measured at the point of sample collection and was used as a
control parameter in the experimental procedure. These data were used to guide analytical
decisions on concentration range, dilution, and spike levels. They were also used as a final
validation of the nickel data, since the relationship between specific conductance and nickel
concentration was shown to be consistent within a sample set. In a few instances, deviations
from this relationship were used to identify samples for reanalysis. Such reanalysis often
revealed problems in the initial concentration measurement and justified adjusting reported
concentration values.
TOTAL ORGANIC CARBON
Duplicate samples and reruns (on subsequent dates) were used to verify accuracy of TOC
results. Only selected samples were analyzed for TOC, to give indications of trends in organic
carbott concentration through an experimental run. The range of concentrations was large, 1 to
26,000 mg/L, and to some degree, paralleled the nickel results. The duplicates and reruns
strongly supported the reprodudbffity of the instrumental measurement, and calibration checks
suggested the results were accurate. Comparisons to changes in nickel and conductivity data
also served as an independent check on TOC values.
A duplicate and spike was prepared for the set of five samples run for chloride and
sulfate determination. Percent differences between duplicates was 1% or less and spike
recoveries ranged from 99 to 104% for these analyses.
REFERENCE STANDARDS
The laboratory did not participate in any external perfbrmance evaluation (PE) for nickel
analysis by AAS during the course of this project. The laboratory is involved in a quarterly PE
exercise with the USGS, but the low metals concentrations in these PE samples argues for their
analysis by ICP/MS. We have consistently performed well in these PE exercises, with nickel
analysis always yielding good to very good results. The significance of these PE exercises to
the current project is that they provide an independent check on the nickel standard quality.
A Spex certified standard was used for nickel analysis in this project (a copy of the
Certificate of Analysis for the nickel standard is attached). Check standards were prepared,
whenever feasible, from a separate stock, and consistently resulted in measured concentrations
at or near the anticipated concentration.
51
-------
No reference standards were used for TOC or inorganic ion determinations. Standards
for these determinations were prepared from the highest quality reagent chemical available on
an analytical balance with a calibration history traceable to NIST. Previous performance in the
USGS PE program has shown that we have performed well on sulfate and chloride analysis by
ion chromatography.
DETECTION OMITS
Instrument detection limits were never an issue in this study. Concentrations were quite
high in the untreated samples and concentrates so that 1 mg/L as a bottom limit offered
treatment efficiencies of greater than 99.9%. Some measurements were made in the! 0.1 to 1
mg/L range, well within the capabilities of the instrument. Most were made hi the 10 to 50
mg/L range. Performance at the lower end, the operationally defined project detection limit,
was verified by evaluating the quality of the standard curve and the results of sample spikes.
This lower limit of 0.1 mg/L proved more than adequate hi addressing all of the samples.
Some treated samples had TOC levels below 10 mg/L. These levels are near the
effective lower limit of TOC determination for our instrumental system. However, the exact
value of the TOC at this level was less critical to the project than just knowing concentrations
were in this range. Consequently, no formal determination of lower detection limit was
performed. Duplicate determinations at 16 mg/L showed consistent results. Evaluation of 5 and
10 mg/L standards hi this low range yielded quantitation at 93 % of nominal. The treated water
also provided a very clean matrix for these measurements, contributing to our confidence in the
results.
Instrument detection limits for inorganic ions were not a factor hi this study as the few
samples analyzed had concentrations of analytes at or above 10,000 mg/L. ;
METHOD BLANKS
i
A distilled water blank was used in each calibration of the AAS. A reagent blank,
containing a concentration of HNQj (the diluent for all diluted samples, and the stabilizer for
undiluted samples) similar to all of the samples, was analyzed, on average, with every 10
samples. Because of the fairly high levels of nickel in nearly all of the samples, the reagent
blank results, usually between 0 and 0.05 mg/L in the higher concentration measurement ranges
(10-50 mg/L), were never considered significant Blanks in the lower concentration
measurement range (0.1-1.0 mg/L) were indistinguishable from zero.
TOC and ion chromatographic analyses employed a distilled water blank, and consistently
gave results indistinguishable from zero. ;
DATA REDUCTION
I
Data reduction was limited to adjusting direct concentration measurements ftxrai the
instruments for dilution factors, and to providing % difference and % recovery values for
52
-------
duplicate and spikes, respectively. Calibration on the instrumental systems allows for direct
readout of concentrations on all samples analyzed. In some cases, dilution factors can also be
built into the instrument program. However, we rarely used this latter feature, preferring
instead to maintain close contact with the actual instrument response for the sample. Dilution
corrections were made manually (spreadsheet) after examination of the instrument result^ All
data transcriptions to the spreadsheet were checked and rechecked, sometime by two different
individuals. The spreadsheet was also used for calculation of % difference and % recovery
information.
53
-------
it:
a1
H E
CO O
CO C
lo
05 Ot
o>
o o
SO
a «
if
wD O c
•O
CO
a.
4^
to
CO L
3?*
CO
TJ
o
"5
a
a
•i o
o
oil £
E
CO
I
5
i!
II
o
M- JD %
O CO Q
CD
E «
wo 9
o
CN;
o
o
in
CO
a.
•o
o
o
CO
CD
O5
O O
o o
»- °l °*
CO IO CO
i- CM CM
CS
=£ a.
Q. 3
89 T3
n
S
3 m 3
•DC "O
O 00 O 00
OS O> T- «* oooooooooo«-«-*-«-»-'«—«-«-»— ^oi
30 CO CO 00 OO CO C3)CnO5O5O)O5C5O)O3CJ3C3>O5C5O3O3C'3O3O5CJ5CJ5O>
O5
iyCSCO^-lO®r^C»Cp»^^T^»^T^»^»-r^r^^e^CSWC^tNeNC>lC>J
-------
o
o
o
o
o
°0
00
co . to in
T— CD CM in CD
o
O5
in
CM
CM
?^
• •
=3
o
0
o
o
.
O CD T- i- CM V V CO CM /
en CM os es CM 03
o
o
CO
os u?
T- en
o
in
o «-
<— CM n CD in co
CM CM CM CM «- r-
cn cj o> o> p p
*— i— r— «— CS CM
CM CM CM CM CM CM
cn o> o> os oj en
co en o o *-• CM
CM CM co in 10 in
f^ H H H H H
55
-------
-------
c
«
O
0|
2
15
o
o
eg «-
r V
O
o
03
Q.
o
O
t*- CO
^ CO
O
O
in «-
es V
O
5
CO
«-
CO .
^ V
03 V
O
in
oo
a>
(35
§
CO
~*i
*- O
O
O
co
O
O
«- cs
O3»-O>»-O3COO3COO3
-------
o
c
m
I
O
Ot^
£E
2 g
TJ ^
*w o
ro c
15
ll
§ g
<-£
t/> o
II
"m
Il
o
§§
«- CM
"fck./ «—
en «-
to
O 0.0
O O O
o o o
m*1^*-7
"? CO •* CM
CO «- CO *-
«-• +* *»
o "5 o o
a. 3
co T>
. - -
3 O 3 3
T3 03 T3 tJ
O
eo^q
ID C5
CD
_ ts
3 O
21
co -o
C3
o o o o o o
oo 03 ^j- r^ co in
to ia j*» r^ «3 r
O
O
CO
tO V— V—
x o
CM n
4-1 «J «v «^
O3 CO C9 TO •_
_« _O _O _O 3
"5. "5. "5. "5. "?
3 3 3 3 CM
•D -0-T3 T3 CM
«^
CO
^J
"5.
3
T3
cDcooco
ooooooooo
c
3
CD M-
CM CM
CD CO
CM CM
o 3 s-^-p-p-^-p-p
~?~? excMin"?"?oju>oiooscM«ncMCMtoio
OCO £CM»-O)CneSI»-COCMCMCMCMCMCMeMCM
to
to
o •'.';•
cncncn
eMCMCM
oscnco
CMCMCM
CO CO tO
CM CM CM
m en
CM «-
OS
CO
in
O
in co i~* co
in in in in
o o o o
CMCM OCMCMCMCM
cjjcn ccncncncn
^SOOOOOOCOCOCOtOCOCOCOCOCOCOCOCOCOCOCO^'COCOO
OCMCMeMe^WeMeMCMCMCMCMeMCMCMcAjCMeMeMeMCMCMCMeMCMCM
ccncncncncQencoocncncncncncncncncncncncncocncncno^
«— CMCOOS
-------
eMotcMeywcviestc^cjtcjiopcpco
66666666666666666
cctccctrccccccocccccocccccir
-------
OOOOOOO
o o o o
(OOOOOOOO .
GOlOCMCDC^COO^CNlO^
&jci«-^co'clinin|^
t-T-CM»-CSjir-CM«
o
feg§°-^«2:iB2^22S«SSSSS^SSg«cococS|§g|||g|
b|§§S§§SSSSSS3SSc?c?Sc?c?c?Sc?c?Sc?c?Sc?c5c?c?c?Sc?c5
SC^^^^C^C^^CMC^C^^^^W^^W^^CJJ^CSCSCMS^W^WWCSW^
«*U3CDr^COO3O»-
COCOCOCOCOCO'*''*
f>. |CO
be
-------
u
"
a »
5
aj
5(
o
CO
en
c
w
w <>
s g
20
tg
a.
to to
o
r~
o
*^
to
"5.
ID
o
o
-®
-------
------- |